Anda di halaman 1dari 497

De Gruyter Graduate

Leclerc, Gauvin ∙ Functional Materials

Also of Interest
Polymer Surface Characterization
Luigia Sabbatini (Ed.), 2014
ISBN 978-3-11-027508-7, e-ISBN 978-3-11-028811-7

Organic and Hybrid Solar Cells – An Introduction

Lukas Schmidt-Mende, Jonas Weickert, 2015
ISBN 978-3-11-028318-1, e-ISBN 978-3-11-028320-4

An Introduction to Surfactants
Tharwat F. Tadros, 2014
ISBN 978-3-11-031212-6, e-ISBN 978-3-11-031213-3

Seema Agarwal, Andreas Greiner (Editor-in-Chief)
ISSN 1618-7229

Journal of the Mechanical Behavior of Materials

Aifantis, Elias C. (Editor-in-Chief)
ISSN 2191-0243

For Energy, Sustainable Development and
Biomedical Sciences

Edited by
Mario Leclerc and Robert Gauvin
Prof. Mario Leclerc Dr. Robert Gauvin
Department of Chemistry Scientific Liaison Officer
Université Laval Québec Center for Functional Materials (CQMF)
Quebec, Canada Quebec, Canada
Email: Email:

ISBN 978-3-11-030781-8
e-ISBN (PDF) 978-3-11-030782-5
e-ISBN (EPUB) 978-3-11-038819-0

Library of Congress Cataloging-in-Publication Data

A CIP catalog record for this book has been applied for at the Library of Congress.

Bibliographic information published by the Deutsche Nationalbibliothek

The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie;
detailed bibliographic data are available on the Internet at

© 2014 Walter de Gruyter GmbH, Berlin/Boston

Cover image: PASCAL GOETGHELUCK/SPL/Agentur Focus
Typesetting: PTP-Berlin, Protago TEX-Produktion GmbH,
Printing and binding: Hubert & Co. GmbH & Co. KG, Göttingen
♾ Printed on acid-free paper
Printed in Germany
When asked to write a foreword for this new book on functional materials, I initially
hesitated to comply with the request for a few simple reasons. First of all, new journals
and books on functional materials are sprouting these days like cabbage in the spring.
This is even more the case for original research papers printed in the usual disci-
plinary journals in chemistry, physics, medicine, and engineering. Consequently, it is
almost impossible to critically keep track and differentiate important studies from the
less significant ones. Moreover, not every original contribution contains completely
novel work, as many of them are instead additions to already well-known facts or well-
accepted lines of thought. In addition, the word “functional” implies that the material
serves a specific and defined purpose in a unique manner; but many contributors to
the field forget or overlook the fact that the material under consideration or subject
under study is just one of many components in a device and its function only evolves
through interplay with all the other components in the device. Note that even the sim-
plest of all rechargeable batteries as an example of a device contains at least a dozen
different functional materials, which all need to comply with the overall performance
of the battery. Thus, compatibility of materials is an important issue frequently over-
looked by eager researchers.
These critical remarks should not be misunderstood. We do not argue against aca-
demically motivated research in materials science. On the contrary, we think that high
quality work aiming at improved, optimized, and – to some extent – novel functional
materials is a necessity. We would merely like to express a warning against exagger-
ated promises and hopes that have no basis in the world of practice. After all, research
on functional materials must envisage practical solutions and is never justified as a
purely academic exercise. The clash between academic ideas and the reality of the
industry and the markets is nowhere harsher than in this field of science.
Despite these critical remarks, it is highly appreciated that this book is not simply
a review of the vast body of literature of the recent years, as it holds the focus upon
various aspects of application. Moreover, it selects only a few topics in favor of a solid
and thorough treatment of the relevant aspects. It is particularly pleasing that poly-
mer colloids receive much attention. Polymer colloids have seen an impressive revival
as a subject of academic research. This is a bit surprising at first sight, given the sit-
uation that emulsion and suspension polymerization processes have served as major
industrial processes in the mass production of polymers, to supply markets with ma-
terials for the coatings and adhesives industry, as textile and leather modifiers and as
constituent in cosmetics, just to mention a few. However, the widespread interest in
nanotechnology has guided the researchers to develop methods which not simply aim
at the synthesis of new molecules, but rather target the production of defined objects.
These nanosized objects are not only composed of different molecules, each type of
molecule contributing different functions to the object, but they also exhibit a defined
VI | Foreword

shape and size distribution, have a defined and controllable surface structure, and a
structured internal composition.
Similar cases are characteristic of the synthesis of mesoporous materials, which
have a long history as absorbents and filtering materials in industry, but have also
found new attention, with the desire to embed more functions into these materials
than was previously possible. This includes better control of pore structure and pore
size distribution. Here again, the target of the synthesis is the object, and not the indi-
vidual molecule which would only later be subject to further processing. Similarly,
nanoscale pigments have been available through industrial processes for decades.
Nevertheless, academic interest centers on the development of novel pigments with
particular size distribution, surface structure, and function, as well as specific elec-
tromagnetic properties. The definition of viable conditions for mass production by in-
dustrially acceptable processes is a key to the success of such research.
All of these aspects are treated in this book. In addition, materials for energy
storage (batteries and capacitors) as well as fuel cells are brought to the readers’s
attention. Last but not least, materials aiming towards application in the vast field
of biomedicine are discussed, described, and critically evaluated. Biocompatibility,
safety aspects, and the essential approval by the authorities are important aspects not
to be overlooked in this area, requiring extensive knowledge in both materials science
and medicine.
In summary, this book comes at a good time, when a large body of academic liter-
ature has been accumulated and is waiting for a critical inspection in the light of the
real demands of application. This book will provide valuable information to a critical
readership and should receive wide-spread recognition from the scientific community.

September 2014 Professor Gerhard Wegner

Emeritus scientific member
Max-Planck Institute for Polymer Research
Mainz, Germany
Energy, health, and environment are certainly at the top of the list of priorities in the
challenges facing society worldwide for the next 50 years. Functional materials are
providing new solutions and opportunities to ensure sustainable energy and environ-
ment, as well as improved medical care for the future. By its nature, the field of func-
tional materials is highly interdisciplinary, comprising basic sciences such as physics,
chemistry, and biology, as well as applied sciences and engineering. In this regard,
this book brings together the expertise of multiple experts and gives an overview of
emerging trends in this field of research. It allows the reader to associate a multitude
of functional materials with their respective application, in a way that has not pre-
viously been done in existing books covering smart or functional materials. As one
can imagine, this interdependence between synthesis, structure, properties, and per-
formance of functional materials requires applied and theoretical chemistry, physics,
biology, and engineering tools in order to ensure the development of novel and effi-
cient technologies (Fig. 1).





Fig. 1. The field of functional materials is by definition interdisciplinary, involving scientists from
various backgrounds in designing and synthesizing new materials with structure and properties
tailored in order to achieve optimal performance.

This book is addressed to experts as well as graduate students interested in the most
recent developments in the field of functional materials. It includes 15 chapters which
are organized in three sections focusing on (1) the synthesis and applications of func-
tional materials, (2) novel materials for energy applications, and (3) new trends in
functional materials for sustainable development and biomedical applications. The
chapters have been written by recognized experts in their respective fields and cover
topics ranging from fundamental concepts in synthesis to applied research projects.
Recent examples of how functional materials are improving many aspects of technol-
ogy currently available to us on a daily basis are also discussed. The authors are all
VIII | Preface

members of the Quebec Center for Functional Materials (CQMF), a research center lo-
cated in Quebec, Canada, and are internationally known for their work in the devel-
opment of innovative and state-of-the-art functional materials.
In summary, this book provides an overview of the recent advances made in syn-
thesis, characterization, and computer modeling, as well as various applications in-
volving functional materials, and should be of great interest for the scientific commu-
nity working in this area of research. The authors are grateful to Mrs. Patricia Basque
(CQMF), and Mr. Philippe Dufour (NanoULaval) for their collaboration, and to Mrs.
Karin Sora (De Gruyter), Mrs. Julia Lauterbach (De Gruyter), and Mrs. Kathleen Prüfer
(De Gruyter) for their assistance in editing this book. We also acknowledge financial
support from the FRQNT-Strategic Networks Funding Program, which provides the op-
portunity to bring scientists and engineers together in order to develop novel and in-
novative technologies.

Québec, August 2014 Mario Leclerc

Robert Gauvin
Foreword | V

Preface | VII

Contributing authors | XVII

About the editors | XXIII

R. Gauvin
1 Introduction | 1

Part I: Functional materials: Synthesis and applications

A. Al Shboul, F. Pierre, and J. P. Claverie

2 A primer on polymer colloids: structure, synthesis and
colloidal stability | 9
2.1 Introduction | 9
2.2 Polymer colloids inside out | 10
2.2.1 How many polymer chains per particle? | 10
2.2.2 How many particles? | 10
2.2.3 Are the chains immobile within the nanoparticle? | 12
2.2.4 Morphology of polymeric nanoparticles | 13
2.3 Preparation of polymer nanoparticles | 17
2.3.1 Emulsion polymerization | 18
2.3.2 Miniemulsion polymerization | 22
2.3.3 Microemulsion polymerization | 24
2.3.4 Self-assembly in selective solvents | 25
2.4 Colloidal stabilization | 26
2.4.1 Electrostatic stabilization | 26
2.4.2 Steric stabilization | 30
2.4.3 Depletion stabilization | 31
2.4.4 Future directions | 33

S. Rondeau-Gagné and J.-F. Morin

3 Synthesis, functionalization and properties of fullerenes and
graphene materials | 37
3.1 Introduction | 37
3.2 Fullerenes | 37
X | Contents

3.2.1 General considerations | 38

3.2.2 Synthesis and purification of fullerenes | 39
3.2.3 Chemical and physical properties of C60 | 40
3.2.4 Chemical functionalization of C60 | 42
3.2.5 Applications | 45
3.3 Graphene | 47
3.3.1 Production of graphene | 49
3.3.2 Graphene in energy conversion devices | 52

J. Florek, R. Guillet-Nicolas, and F. Kleitz

4 Ordered mesoporous silica: synthesis and applications | 61
4.1 Introduction | 61
4.2 Ordered mesoporous silica (OMS) | 62
4.2.1 Principle of synthesis | 63
4.2.2 Mesostructure diversity and tailoring | 69
4.3 Functionalization of ordered mesoporous silica | 78
4.4 Morphology control | 80
4.5 Selected applications of functionalized ordered
mesoporous silica | 82
4.5.1 Functionalized MSNs as controlled drug delivery platforms | 83
4.5.2 Functionalized mesoporous materials for extraction
chromatography (EXC) applications | 88
4.5.3 Mesoporous organic-inorganic hybrid membranes for
water desalination | 91

A. Ritcey
5 Nanoparticles: Properties and applications | 101
5.1 Introduction | 101
5.2 Synthetic methods | 101
5.2.1 Particle nucleation and growth | 102
5.2.2 Synthesis in inverse micelles | 104
5.3 Particle aggregation and stabilization of colloidal suspensions | 105
5.4 Colloidal quantum dots | 107
5.5 Metal nanoparticles | 110
5.6 Metal oxide nanoparticles | 112
5.6.1 Titanium dioxide | 112
5.6.2 Iron oxide | 113
5.6.3 Silica | 115
5.7 Polymeric nanoparticles | 115
5.8 Advanced architectures and hybrid systems | 117
Contents | XI

N. Allard and M. Leclerc

6 Conjugated polymers for organic electronics | 121
6.1 Introduction | 121
6.2 Processable conjugated polymers | 122
6.3 Applications in renewable energy | 126
6.3.1 Organic solar cells | 126
6.3.2 Conjugated polymers for organic solar cells | 128
6.4 Applications in micro-electronics | 130
6.4.1 Field-effect transistors | 130
6.4.2 Conjugated polymers for field-effect transistors | 132
6.5 Applications in lighting | 133
6.5.1 Light-emitting diodes | 133
6.5.2 Conjugated polymers for light-emitting diodes | 135
6.6 Summary | 136

A. Soldera
7 Theoretical tools for designing microscopic to macroscopic properties
of functional materials | 139
7.1 Methods | 140
7.1.1 The link between microscopic and macroscopic scales | 140
7.1.2 Ab initio methods | 142
7.1.3 Bridging the gap between ab initio and atomistic levels | 146
7.1.4 Atomistic simulation | 147
7.1.5 Bridging the gap between atomistic and mesoscale levels | 151
7.2 Examples | 151
7.2.1 Quantum studies | 152
7.2.2 Atomistic simulation | 156
7.3 Summary | 164

Part II: Development of new materials for energy applications

S. B. Schougaard and D. Bélanger

8 Electrochemical energy storage systems | 171
8.1 Introduction | 171
8.2 Metrics and performance evaluation | 171
8.3 Models and theory of electrochemical charge storage | 173
8.3.1 Battery operation – a Faradaic process | 174
8.3.2 Electrochemical capacitor operation – a non-Faradaic process | 175
8.4 Electrolytes | 178
8.5 Electrode materials | 180
8.5.1 Electrochemical capacitors | 180
XII | Contents

8.5.2 Hybrid electrochemical capacitors | 181

8.5.3 Lithium battery electrode materials | 183
8.5.4 Negative (anode) electrode materials | 184
8.5.5 The positive (cathode) electrode | 185
8.5.6 Electrode production | 186
8.6 Summary | 186

D. Rochefort
9 Functional ionic liquids electrolytes in lithium-ion batteries | 189
9.1 Introduction | 189
9.1.1 Historical overview | 190
9.1.2 What are ionic liquids? | 191
9.1.3 Key properties as electrolytes | 192
9.2 Ionic liquids as Li and Lithium-ion battery electrolytes | 193
9.3 Functional ionic liquid electrolytes | 194
9.3.1 Overview of functional ionic liquids | 195
9.3.2 Solid electrolyte interphase | 196
9.3.3 Transport of lithium ions | 197
9.3.4 Electroactive ionic liquids as redox shuttles | 198
9.3.5 Perspectives | 202

C. de Bonis, A. D’Epifanio, B. Mecheri, S. Licoccia, and A. C. Tavares

10 Solid polymer proton conducting electrolytes for fuel cells | 207
10.1 Introduction | 207
10.2 Proton exchange membranes | 209
10.2.1 Nafion® | 210
10.2.2 Alternative sulfonated ionomers and membranes | 213
10.3 Characterization of solid polymer electrolytes | 218
10.3.1 Proton conductivity | 218
10.3.2 States of water and water mobility | 222
10.4 Summary | 233

P. Bénard, A.-M. Beaulieu, D. Durette, and R. Chahine

11 Supercritical adsorption of hydrogen on microporous adsorbents | 241
11.1 Introduction | 241
11.2 Fundamentals of supercritical adsorption | 242
11.3 Supercritical adsorption isotherms | 246
11.3.1 Virial expansion of the excess density in terms of pressure | 246
11.3.2 Basic analytic models of the adsorption isotherm | 252
11.3.3 Self-consistent approaches | 256
11.4 The thermodynamics of adsorption | 257
11.4.1 Properties of surface potential | 259
Contents | XIII

11.5 Microporous adsorbents for hydrogen storage | 261

11.5.1 Activated carbons | 261
11.5.2 Single wall nanotubes | 262
11.5.3 Metal organic frameworks | 263

Part III: New trends in sustainable development and biomedical


D. Mantovani, L. Levesque, G. Sabbatier, M. Leroy, D. G. Seifu, R. Tolouei,

V. Montaño, M. Cloutier, I. Bilem, C. Loy, M. Byad, C. Paternoster, C. A. Hoesli,
B. Drouin, G. Laroche
12 Advanced materials for biomedical applications | 277
12.1 Introduction | 277
12.2 History of biomaterials | 278
12.3 Basics in material science for biomaterial applications | 280
12.3.1 Biomaterial properties | 280
12.3.2 Biometals | 280
12.3.3 Bioceramics | 281
12.3.4 Biosynthetic polymers | 282
12.3.5 Natural polymers | 284
12.4 Biomedical applications | 286
12.4.1 Cardiovascular system | 286
12.4.2 Musculoskeletal system | 291
12.4.3 Visceral organs | 300
12.4.4 Nervous system and sensory organs | 304
12.4.5 Esthetic applications | 310
12.4.6 Skin | 312
12.5 Future trends | 319
12.5.1 Tissue engineering basic concepts | 319
12.5.2 Scaffolds | 319
12.5.3 Surface modification | 323
12.5.4 Stem cells | 323
12.5.5 Bioreactors | 324
12.5.6 Computational models | 324
12.6 Summary | 326

M.-A. Fortin
13 Nanoparticles for magnetic resonance imaging (MRI) applications
in medicine | 333
13.1 The basics of MRI in medicine | 337
13.2 Relaxivity: the performance of MRI contrast agents | 339
XIV | Contents

13.3 Synthesis and characterization of magnetic nanoparticles | 340

13.3.1 Synthesis of magnetic nanocrystals | 340
13.3.2 Nanoparticle coatings for MRI applications | 344
13.3.3 Physicochemical characterization | 346
13.4 Physical properties of magnetic nanoparticles | 347
13.5 MR relaxation properties of magnetic nanoparticles | 352
13.5.1 Relaxivity of paramagnetic CAs | 353
13.5.2 Relaxivity of superparamagnetic CAs | 356
13.5.3 Relaxometric performance of MRI CAs at clinical magnetic field
strengths | 358
13.6 Biological performance of magnetic nanoparticles for MRI | 358
13.6.1 In vivo barriers | 360
13.6.2 Impact of nanoparticle size and surface on colloidal stability and
blood retention | 361
13.6.3 Directing nanoparticles in vivo | 362
13.6.4 Toxicity | 363
13.7 Summary | 364

J. Greener
14 Microfluidics for synthesis and biological functional materials:
from device fabrication to applications | 375
14.1 Introduction | 375
14.2 A practical introduction to microfluidic reactors for
material synthesis | 376
14.2.1 Microfluidic reactor geometries | 376
14.2.2 Device fabrication materials | 377
14.2.3 Fabrication of polymer-based planar microreactors
and components | 380
14.3 Manipulating and measuring precursor reagent streams
in microchannels | 383
14.3.1 High surface area to volume ratios in microchannels | 383
14.3.2 Rapid heat transfer | 384
14.3.3 Control of concentrations | 384
14.3.4 Controlling “time on chip” | 386
14.3.5 Control of hydrodynamics and mass transfer | 386
14.3.6 Characterization in microchannels | 389
14.4 Microfluidics for polymer microparticles | 391
14.4.1 Manipulating the shaping of liquid precursors | 392
14.4.2 Effect of the channel wall | 392
14.4.3 Emulsification of precursor droplets | 393
14.4.4 Channel geometries to achieve emulsified droplets | 393
14.4.5 Multiple emulsions | 395
Contents | XV

14.4.6 Forming linear threads and two-dimensional interfaces | 395

14.4.7 Converting liquid precursors into solid micro-materials | 397
14.4.8 Scale up: a circuit analysis of microfluidic flow in a highly
parallelized microreactor | 397
14.5 Microfluidics for synthesis of functional nanoparticles | 400
14.5.1 Microfluidics for highly controlled nanoparticle synthesis | 401
14.6 Biomaterials | 402
14.6.1 Tissue engineering and membranes | 403
14.6.2 Microenvironments for encapsulated cells | 404
14.6.3 Biofilms | 406
14.6.4 Microdevices utilizing functional biomaterials | 407
14.7 Summary | 410

T. Lefèvre, F. Byette, I. Marcotte, and M. Auger

15 Protein- and peptide-based materials: a source of inspiration
for innovation | 415
15.1 Introduction | 415
15.2 Basics of proteins, peptides and polypeptides | 417
15.2.1 Polypeptides are sequences of amino acids | 417
15.2.2 Polypeptides can adopt various conformations | 418
15.2.3 Polypeptides possess various levels of structural organization | 419
15.3 Functional materials from fibrous proteins | 420
15.3.1 Resilin & abductin | 421
15.3.2 Byssus (mussel anchoring threads) | 422
15.3.3 Silk | 425
15.4 Functional materials from globular proteins | 429
15.4.1 Natural proteins | 429
15.4.2 Artificial proteins | 430
15.5 Functional materials from synthetic peptides | 432
15.6 Summary | 435

B. Riedl, V. Vardanyan, W. N. Nkeuwa, A. Kaboorani, V. Landry, B. Poaty, M. Vlad,

and C. Sow
16 Nanocomposite coatings | 443
16.1 Introduction | 443
16.2 Coating formulations | 446
16.2.1 Chemical components | 446
16.2.2 Mixing techniques | 447
16.2.3 Application and curing | 449
16.3 Nanoparticle additives | 449
16.4 Coating characterization | 454
16.4.1 Mechanical properties | 454
XVI | Contents

16.4.2 Optical properties | 456

16.4.3 X-ray imaging and particle aggregation | 459
16.4.4 Weathering and artificial aging | 459
16.5 Bio-based coatings | 460
16.6 Future developments | 462
16.7 Summary | 463

Index | 465
Contributing authors
Nicolas Allard Ibrahim Bilem
Department of Chemistry Department of Mining, Metallurgical
Université Laval and Materials Engineering
Quebec, QC, Canada Université Laval
e-mail: and
Chapter 6 Centre de recherche du CHU de Québec
Québec, QC, Canada
Ahmad Al Shboul e-mail:
Department of Chemistry Chapter 12
Université du Québec à Montréal (UQAM)
Montréal, QC, Canada Michael Byad
e-mail: Department of Mining, Metallurgical
Chapter 2 and Materials Engineering
Université Laval
Michèle Auger and
Department of Chemistry Centre de recherche du CHU de Québec
Université Laval Québec, QC, Canada
Quebec, QC, Canada e-mail:
e-mail: Chapter 12
Chapter 15
Frédéric Byette
Ann-Marie Beaulieu Department of Chemistry
Research Institute on Hydrogen Université du Québec à Montréal
Université du Québec à Trois-Rivières (UQTR) and
Trois-Rivières, QC, Canada Université de Montréal
e-mail: Montréal, QC, Canada
Chapter 11 e-mail:
Chapter 15
Daniel Bélanger
Department of Chemistry Richard Chahine
Université du Québec à Montréal (UQAM) Department of Electrical Engineering and
Montréal, QC, Canada Computer Engineering
e-mail: Université du Québec à Trois-Rivières
Chapter 8 Trois-Rivières, QC, Canada
Pierre Bénard Chapter 11
Research Institute on Hydrogen
Université du Québec à Trois-Rivières (UQTR) Jerome P. Claverie
Trois-Rivières, QC, Canada Department of Chemistry
e-mail: Université du Québec à Montréal (UQAM)
Chapter 11 Montréal, QC, Canada
Chapter 2
XVIII | Contributing authors

Maxime Cloutier Marc-André Fortin

Department of Mining, Metallurgical Department of Mining, Metallurgical,
and Materials Engineering and Materials Engineering
Université Laval Université Laval and
and Centre de recherche du CHU de Québec
Centre de recherche du CHU de Québec Québec, QC, Canada
Québec, QC, Canada e-mail:
e-mail: Chapter 13
Chapter 12
Robert Gauvin
Catia de Bonis Quebec Center for Functional Materials
Department of Chemical Science and Québec, QC, Canada
Technology & NAST Center e-mail:
University of Rome Tor Vergata Chapter 1
Rome, Italy
e-mail: Jesse Greener
Chapter 10 Department of Chemistry
Université Laval
Alessandra D’Epifanio Québec, QC, Canada
Department of Chemical Science and e-mail:
Technology & NAST Center Chapter 14
University of Rome Tor Vergata
Rome, Italy Rémy Guillet-Nicolas
e-mail: Department of Chemistry and Centre de
Chapter 10 Recherche sur les Matériaux Avancés (CERMA)
Université Laval
Bernard Drouin Québec, QC, Canada
College François Xavier Garneau e-mail:
e-mail: Chapter 4
Chapter 12
Corinne A. Hoesli
David Durette Department of Mining, Metallurgical and
Department of Chemistry, Biochemistry Materials Engineering
and Physics Université Laval
Université du Québec à Trois-Rivières Centre de recherche du CHU de Québec
Trois-Rivières, QC, Canada and
e-mail: McGill University
Chapter 11 Québec, QC, Canada
Justyna Florek Chapter 12
Department of Chemistry and Centre de
Recherche sur les Matériaux Avancés (CERMA) Alireza Kaboorani
Université Laval Department of Wood and Forest Sciences
Québec, QC, Canada Université Laval
e-mail: Québec, QC, Canada
Chapter 4 e-mail:
Chapter 16
Contributing authors | XIX

Freddy Kleitz Lucie Levesque

Department of Chemistry and Centre de Department of Mining, Metallurgical and
Recherche sur les Matériaux Avancés (CERMA) Materials Engineering
Université Laval Université Laval
Québec, QC, Canada and
e-mail: Centre de recherche du CHU de Québec
Chapter 4 Québec, QC, Canada
Véronic Landry Chapter 12
Department of Wood and Forest Sciences
Université Laval Silvia Licoccia
Québec, QC, Canada Department of Chemical Science and
e-mail: Technology & NAST Center
Chapter 16 University of Rome Tor Vergata
Rome, Italy
Gaétan Laroche e-mail:
Department of Mining, Metallurgical Chapter 10
and Materials Engineering
Université Laval Caroline Loy
and Department of Mining, Metallurgical
Centre de recherche du CHU de Québec and Materials Engineering
Québec, QC, Canada Université Laval
e-mail: and
Chapter 12 Centre de recherche du CHU de Québec
Québec, QC, Canada
Mario Leclerc e-mail:
Department of Chemistry Chapter 12
Université Laval
Québec, QC, Canada Diego Mantovani
e-mail: Department of Mining, Metallurgical and
Chapter 6 Materials Engineering
Université Laval
Thierry Lefèvre and
Department of Chemistry Centre de recherche du CHU de Québec
Université Laval Québec, QC, Canada
Québec, QC, Canada e-mail:
e-mail: Chapter 12
Chapter 15
Isabelle Marcotte
Marie Leroy Department of Chemistry
Department of Mining, Metallurgical Université du Québec à Montréal
and Materials Engineering Montréal, QC, Canada
Université Laval e-mail:
and Chapter 15
Centre de recherche du CHU de Québec
Québec, QC, Canada
Chapter 12
XX | Contributing authors

Barbara Mecheri Bouddah Poaty

Department of Chemical Science Department of Wood and Forest Sciences
and Technology & NAST Center Université Laval
University of Rome Tor Vergata Québec, QC, Canada
Rome, Italy e-mail:
e-mail: Chapter 16
Chapter 10
Bernard Riedl
Vanessa Montaño-Machado Department of Wood and Forest Sciences
Department of Mining, Metallurgical Université Laval
and Materials Engineering Québec, QC, Canada
Université Laval e-mail:
and Chapter 16
Centre de recherche du CHU de Québec
Québec, QC, Canada Anna Ritcey
e-mail: Department of Chemistry
Chapter 12 Université Laval
Québec, QC, Canada
Jean-François Morin e-mail:
Department of Chemistry Chapter 5
Université Laval
Québec, QC, Canada Dominic Rochefort
e-mail: Department of Chemistry
Chapter 3 Université de Montréal
Montréal, QC, Canada
William Nguegang Nkeuwa e-mail:
Department of Wood and Forest Sciences Chapter 9
Université Laval
Québec, QC, Canada Simon Rondeau-Gagné
e-mail: Department of Chemistry
Chapter 16 Université Laval
Québec, QC, Canada
Carlo Paternoster e-mail:
Department of Mining, Metallurgical Chapter 3
and Materials Engineering
Université Laval Gad Sabbatier
and Department of Mining, Metallurgical and
Centre de recherche du CHU de Québec Materials Engineering
Québec, QC, Canada Université Laval
e-mail: and
Chapter 12 Centre de recherche du CHU de Québec
Québec, QC, Canada
Florian Pierre e-mail:
Department of Chemistry Chapter 12
Université du Québec à Montréal (UQAM)
Montréal, QC, Canada
Chapter 2
Contributing authors | XXI

Steen B. Schougaard Ana C. Tavares

Department of Chemistry Institut National de la Recherche Scientifique
Université du Québec à Montréal (UQAM) Énergie, Matériaux et Télécommunications
Montréal, QC, Canada (INRS-EMT)
e-mail: Varennes, QC, Canada
Chapter 8 e-mail:
Chapter 10
Dawit G. Seifu
Department of Mining, Metallurgical Ranna Tolouei
and Materials Engineering Department of Mining, Metallurgical
Université Laval and Materials Engineering
and Université Laval
Centre de recherche du CHU de Québec and
Québec, QC, Canada Centre de recherche du CHU de Québec
e-mail: Québec, QC, Canada
Chapter 12 e-mail:
Chapter 12
Armand Soldera
Department of Chemistry Vahe Vardanyan
Université de Sherbrooke Department of Wood and Forest Sciences
Sherbrooke, QC, Canada Université Laval
e-mail: Québec, QC, Canada
Chapter 7 e-mail: vahe.vardanyan.1@
Chapter 16
Caroline Sow
Department of Wood and Forest Sciences Mirela Vlad
Université Laval Department of Wood and Forest Sciences
Québec, QC, Canada Université Laval
e-mail: Québec, QC, Canada
Chapter 16 e-mail:
Chapter 16
About the editors
Mario Leclerc was awarded a Ph.D. in chemistry from Univer-
sité Laval, Quebec City, Canada, in 1987, under the guidance of
Prof. R.E. Prud’homme. After a short post-doctoral stay at INRS-
Energie et Matériaux near Montréal with professor L.H. Dao, he
joined the Max-Planck-Institute for Polymer Research, in Mainz,
Germany, as a post-doctoral fellow in the research group of
Prof. Dr. G. Wegner. In 1989, he accepted a position of professor
at the department of chemistry of Université de Montréal. He
returned to Université Laval in 1998 where he has held since 2001 the Canada Research
Chair for Electroactive and Photoactive Polymers. Prof. Leclerc has co-authored about
250 papers published in leading scientific journals which have been cited more than
16 000 times. According to Science Citation Index, he has an h-index of 64. His cur-
rent research activities include the synthesis and characterization of new oligomers
and polymers for applications in micro-electronics, energy, sensors, and genomics.

Robert Gauvin holds a Mechanical Engineering degree from Uni-

versité Laval (Québec, Qc, Canada) and has worked as R&D
Engineer in the medical device industry (AltertekBio, Sainte-
Foy, Qc, Canada) prior to attending graduate school. Following
his Ph.D. studies in Biomedical Engineering at Université Laval
and at the Georgia Institute of Technology (Atlanta, GA, USA),
he was awarded Postdoctoral Research Fellowships at the Mas-
sachusetts Institute of Technology (Harvard-MIT Division of
Health Sciences and Technology, Cambridge, MA, USA) and at the Wyss Institute for
Biologically Inspired Engineering (Harvard Medical School, Boston, MA, USA). After
completion of his postdoctoral research, he joined the Personnel Protection Section
as Staff Scientist for Defense Research and Development Canada (DRDC-Valcartier,
Qc, Canada), before accepting the position of Scientific Liaison Officer at the Quebec
Center for Functional Materials (CQMF, Qc, Canada). He also served as the elected
Chair of Scientific and Professional Development of TERMIS-North America from 2010
to 2013 and holds a Research Professor position at the Department of Surgery (Fac-
ulty of Medicine) of Université Laval. Dr. Gauvin is co-author of over 35 peer-reviewed
journal articles, 4 review articles and 3 book chapters. He presented his work at over
50 national and international conferences and is a reviewer for numerous scientific
journals in the fields of biomedical engineering and material sciences.
R. Gauvin
1 Introduction
Since its beginning, materials science has evolved from the use of inert structural ma-
terials to materials with tailored properties which allow reactive capacities, such as
the intrinsic ability to respond to stimuli and environmental changes, and activation
of specific functions according to these changes. Improvements and innovations in
the fields of chemistry, physics, and engineering have allowed better understanding
of the structure-property-performance relationships and precise control of the com-
position of materials. As a result, considerable developments have been made in the
engineering of innovative functional materials, and research is currently addressing
numerous fields such as photovoltaics, batteries, electrolytes, supercapacitors, en-
ergy conversion, biomaterials, tissue engineering, medical imaging, nanotechnology,
microfluidics, and computer modeling, to name just a few. Functional materials are
therefore expected to have a considerable impact on many aspects of our lives such as
energy, transportation, life sciences, and environment. As applied research tends to
be a very dynamic and innovation-driven field, writing a book on functional materi-
als is therefore a never-ending enterprise, as a new technology appears on the market
every month. This book is designed in such a way that the reader will learn about
functional materials and the types of applications they are designed for. Thus, it will
be possible to correlate emerging functional materials with the underlying concepts
that were involved in their design, although these might still be developed in the fu-
ture. The book is divided into three sections and covers the synthesis and applications
of functional materials (Part I), novel materials for energy applications (Part II), and
new trends in functional materials for sustainable development and biomedical ap-
plications (Part III).
Part I comprises six chapters focusing on the various approaches to preparation
and synthesis of organic and inorganic functional materials. Since the most promising
technologies for increased efficiency and improved reliability involve controlling the
composition and the microstructure of novel materials, it is essential for the reader
to understand the underlying concepts directing the synthesis and the generation of
their assembly. In Chapter 2, the fundamental concepts of colloids and polymeriza-
tion in dispersed medium are introduced. A colloid is a dispersion of very fine objects
in a fluid, these objects being solids, liquids or gas, and the corresponding colloidal
dispersion being referred to as a suspension, an emulsion or foam. Polymer colloids
are used for a large number of applications, ranging from coatings, adhesives, inks,
impact modifiers, drug-delivery vehicles, etc. The domain of colloidal stabilization
is therefore a fascinating and vibrant area of science and this chapter presents im-
portant technological advances combining physical chemistry and fluid dynamics re-
lated to colloids and polymer chemistry. Chapter 3 illustrates the tremendous influ-
ence fullerene and graphene have had on various technologies in recent years. These
2 | Introduction

carbon-based materials are undoubtedly amongst the most-studied materials in both

academic and industrial laboratories since the discovery of fullerenes in 1985, carbon
nanotubes in 1991, and graphene in 2004. Due to their great structural diversity, nu-
merous applications can be envisioned for these materials, and recent synthetic meth-
ods for the design of functional mesostructured materials are presented. Some per-
spectives of these applications in catalysis, selective sorption, and biomedical devices
are reviewed, as well as numerous methods of modification available for modulation
of the surface properties, introduction of functionalities, and control of size and shape
of the particles of these mesoporous solids. Chapter 4 presents an overview of materi-
als of a porous structure with features in the nanometer range which have emerged
as key elements in the development of future technologies including miniaturized
electronics, magnetic and optical devices, environmentally-friendly catalysts, mate-
rials for pollutant removal, biocompatible implants, and drug delivery systems. The
nanopore size range offers vast potential for elaborate functional systems with tailored
properties, as the size and volume of the pores in a given material have a profound
influence on its final properties, such as adsorption/desorption capability, diffusion
mechanisms, storage capacity, density, and mechanical stability. A few examples are
presented, such as nanoporous materials used as highly selective sorbents, selective
membranes, systems for energy storage or energy conversion, recyclable solid cata-
lysts, low k-dielectrics, sensors, biomaterials for drug delivery, and contrast agents for
medical imaging. Chapter 5 focuses on the properties of nanomaterials for the synthe-
sis and design of functional materials based on the fundamental structures of these
particles. The fabrication of nanostructures using “top-down” and “bottom-up” ap-
proaches is described. As their name implies, top-down methods involve the creation
of nanosized entities from larger blocks of matter, including mechanical size reduc-
tion by crushing and grinding, as well as more sophisticated lithographic techniques.
Bottom-up approaches, on the other hand, seek to build nanostructures from smaller
components, typically atoms or molecules, and therefore frequently involve elements
of self-assembly or supramolecular chemistry. Chapter 6 covers the synthesis of conju-
gated polymers for organic electronics applications. These polymers have recently re-
ceived a great deal of attention since they combine the features of metals or inorganic
semiconducting materials (excellent electrical and optical properties), with those of
synthetic polymers (flexibility, ease of processing and low cost). This synergy makes
these functional materials useful in existing optoelectronic devices and creates com-
pletely new technological opportunities. For instance, polymeric semiconductors are
considered to be one of the most promising materials for lowering the cost of solar
energy. The electrical conductivity of these polymer materials has therefore enabled
a wide range of technologies such as photovoltaics, field-effect transistors, and light
emitting diodes, which are expected to foster the upcoming era of plastic and print-
able electronics. Chapter 7 covers the computational approaches and theoretical tools
which are used to investigate the composition and performance of various functional
materials in silica. This chapter demonstrates that a complex structure can be stud-
1 Introduction | 3

ied as an assembly of a multitude of building blocks, and that optimization of these

functional units can lead to improved performance of the macroscopic properties of a
material. Moreover, the versatility of the available chemical processes combined with
theoretical modeling enables the design of compounds with well-defined and tunable
Part II addresses common challenges regarding the development of functional
materials for energy applications including improved performance, conversion effi-
ciency, energy density, and sustainability. Since the efficient use of energy is directly
linked to its production and conservation, energy must ideally be stored for further
use or made available for portable applications once it has been generated from a
readily available source. Thus, there is a tremendous need for electrochemical storage
devices such as high energy density batteries and supercapacitors. Similarly, the stor-
age of hydrogen is also of great interest as it represents the principal fuel source for
electric vehicles. As fuel cells allow electrochemical oxidation of hydrogen-rich fuel
into electrical current, they produce clean energy directly in the form of current and
do not require recharging. Therefore, the drive for innovation in sustainable, clean,
efficient, and portable energy is now stronger than ever. Chapter 8 discusses electro-
chemical energy storage systems such as batteries and supercapacitors. The ability to
efficiently store and retrieve electrical energy is at the heart of the mobile revolution
and is currently of great concern globally. Efficient and low-cost energy storage sys-
tems will be required to meet this challenge. Two of the candidates most likely to meet
these high power, high energy density requirements are optimized lithium-ion batter-
ies and advanced electrochemical capacitors. This chapter presents the fundamental
concepts associated with electrochemical energy storage systems using the thermo-
dynamics, kinetics, structure, and mass transport properties of battery materials. In
Chapter 9 the design of ionic liquid electrolytes for lithium-ion batteries is explained
to highlight the fact that although the redox behaviour and the kinetics of dissolved
species are important, the electrolyte plays a crucial role in the electrochemistry of ad-
vanced systems such as energy storage devices. Indeed, the design of the electrolyte
will affect the transport of lithium ions and the range of operational voltage, thereby
affecting the power and energy density values of the system. Ionic liquids have the
potential to play an important role in the development of electrolytes, which will in-
crease the overall performance and improve the efficiency of energy storage devices.
Continuing with this topic, Chapter 10 focuses on solid electrolytes, which are materi-
als capable of conducting ions used in many electrochemical devices including batter-
ies, fuel cells, sensors, electrolyzers, and electrodialysis systems for water purification
and saltwater desalination. These electrochemical cells are devices capable of pro-
ducing electrical energy from spontaneous chemical reactions. For example, proton
exchange membrane fuel cells are considered attractive power sources for portable
applications and for the automotive industry. Nevertheless, these systems still suffer
from limitations, discussed in this chapter, hindering the competition of this technol-
ogy with lead-acid batteries, fossil fuels, and internal combustion engines. Chapter 11
4 | Introduction

introduces the basic concepts related to supercritical adsorption of hydrogen on mi-

croporous adsorbents. Physical adsorption on microporous adsorbents is widely used
in a number of industrial physicochemical processes such as gas separation and pu-
rification or catalytic support. Materials-based storage of gases such as hydrogen can
be achieved through chemical binding or through absorption inside a solid metal hy-
dride matrix. Activated carbons have been proposed as a means of storing gases such
as natural gas at room temperature, and have been widely studied for hydrogen stor-
age applications, due to their availability and the possibility of tailoring their porous
structures through various chemical and physical treatments. For hydrogen, storage
applications require low temperature operation and highly microporous activated car-
bon matrices with high specific surface areas, due to the small value of the binding
energies between carbon structures and molecular hydrogen. This chapter will cover
the thermodynamic description of these systems, which regulates the energy scale re-
quired for charging and discharging gases through these devices.
Part III groups together the design, synthesis and fabrication of novel functional
materials according to new trends in sustainable development and biomedical appli-
cations. Chapter 12 gives a detailed overview of advanced materials used in biomed-
ical applications. More specifically, it aims to summarize the fields of implants and
prostheses from a materials science and engineering point of view. Since biomaterials
are at the interface between medicine, life sciences, and materials science, the num-
ber of strategies for their design and engineering is almost infinite, spanning from
the control of surface roughness, the conjugation of signal peptides for cell adhesion
and drug encapsulation, to biodegradable materials which will be replaced during
the healing process. The synthetic and biological materials and devices available to
repair or replace damaged tissues and organs are reviewed herein. Following on from
the previous chapter, an introduction to the synthesis of functionalized nanoparticles
for bioimaging and diagnostic applications is provided in Chapter 13. The develop-
ment of magnetic nanoparticles as contrast agents for vascular, molecular, and cel-
lular magnetic resonance imaging (MRI) applications has been exponential during
the last decade. Nanoparticles generating contrast in MRI is one of the most direct
applications of nanotechnology. This chapter is an introduction to the principles un-
derlying the performance of nanoparticulate-based MRI contrast agents. It addresses
the main considerations guiding the design and physicochemical characterization of
magnetic nanoparticles. The fundamental aspects of nanoparticle magnetism and re-
laxometric characterization are discussed, as well as examples of applications in bi-
ological models. While introducing the fundamentals of microfluidics, Chapter 14 re-
views important considerations regarding the synthesis of synthetic and biological
functional materials on the microscale. The topics covered range from the introduction
of key concepts in microfluidics and reactor fabrication to their utilisation in the study
and synthesis of new micromaterials for various applications. Chapter 15 provides a
concise description of natural and synthetic polypeptides, from peptides to globular
and fibrous proteins, for the development of functional materials. Proteins and pep-
1 Introduction | 5

tides are promising building blocks for the conception of materials, as they can be
tailored for a broad variety of functions, structures and properties, possess the ability
to self-assemble into complex architectures, respond to specific environmental stim-
uli, bind to specific receptors and ligands, and resist mechanical stresses and strains.
Through several examples, general principles and important areas of research regard-
ing polypeptide-based materials are described, as well as their considerable potential
in various, and possibly unanticipated, application fields. Chapter 16 introduces as-
pects of chemistry and physics of coatings reinforced with nanoparticles, especially
for wood substrates intended for common indoor and outdoor use. Nanoparticle rein-
forcement can improve resistance to UV degradation, water absorption, and wear and
tear without affecting the most important characteristics of paints and coatings: bril-
liance, color and transparency. This applied research example provides a good illus-
tration of how even the simplest everyday paint or varnish requires extensive research
and design in order to be competitive and display optimal properties.
Functional materials are at the center of most technology advancements. This
book covers three important fields in the development of functional materials: energy,
environment, and biomedical applications. These topics are explained and discussed
from an experimental and theoretical perspective, as an understanding of material
properties is fundamental to the development of novel functionalities and technolo-
gies. It is therefore believed that this book will be of great interest to any scientist keen
to broaden their knowledge in both the fundamentals and applications of functional
Part I: Functional materials: Synthesis
and applications
A. Al Shboul, F. Pierre, and J. P. Claverie
2 A primer on polymer colloids: structure, synthesis
and colloidal stability
2.1 Introduction

A colloid is a dispersion of very fine objects in a fluid [1]. These objects can be solids,
liquids or gas, and the corresponding colloidal dispersion is then referred to as sus-
pension, emulsion or foam. Colloids possess unique characteristics. For example, as
their size is smaller than the wavelength of light, they scatter light. They also offer a
large interfacial surface area, meaning that interfacial phenomena are of paramount
importance in these dispersions. The weight of each dispersed particle being small,
gravity and buoyancy forces are not sufficient to counteract the thermal random mo-
tion of the particle, named Brownian motion (in tribute to the 19th century botanist
Robert Brown who first characterized it). The particles do not remain in a dispersed
state indefinitely: they will sooner or later aggregate (phase separation). Thus, the col-
loidal state is in general metastable and colloidal stability is one of the key features to
take into account when working with colloids.
Among all colloids, the polymer colloid family is one of the most widely inves-
tigated [2]. Polymer colloids are used for a large number of applications, ranging
from coatings, adhesives, inks, impact modifiers, drug-delivery vehicles, etc. The
particles range in size from about 10 nm to 1 000 nm (1 μm) in diameter. They are
usually spherical, but numerous other shapes have been observed. Polymer colloids
are not uncommon in nature. For example, natural rubber latex, the secretion of the
Hevea brasiliensis tree, is in fact a dispersion of polyisoprene nanoparticles in wa-
ter. Synthetic polymer colloids, also called synthetic latexes, play a prominent role
in industrial chemistry. Interest in synthetic latexes developed during the Second
World War, when the Japanese Navy threatened access to natural Hevea, an impor-
tant raw material for tire manufacturing at that time. It appeared judicious to produce
synthetic polymers under the same aspect, so that downstream operations on the
elastomer processing units could remain unchanged. This led to the development of
the emulsion polymerization process, one of the most versatile polymerization pro-
cesses [3, 4]. The words latex, polymer colloids, and dispersed polymer nanoparticles
are used interchangeably for any kind of stable colloidal submicronic polymer dis-
persions in a solvent, which in the majority of cases is water. In this chapter we will
present a few salient features of polymer colloid structure, followed by data on the
synthesis of these colloids, and finally we will give several key points on colloidal
10 | Part I Functional materials: Synthesis and applications

2.2 Polymer colloids inside out

2.2.1 How many polymer chains per particle?

Let us consider a polymeric sphere of diameter (dp ) = 20 nm, constituted of poly-

styrene (molecular weight 105 g/mol). Thus, the weight of a single polymer chain is
m = 105 /Na = 1.7 ⋅ 10−19 g, where (Na ) is the Avogadro number. The volume (V) of one
spherical nanoparticle is V = π/6 d3p = 4.2 ⋅ 10−21 l. Knowing that polystyrene den-
sity (ρ) is ρ = 1 040 g l−1 , the average number of polymer chains per polymer particle is
Vρ/m = 26. The average number of chains per particle for representative particle sizes
and polymer molecular weights is listed in Table 2.1. Core-shell particles, whereby a
core particle of a given material is engulfed in a shell of another component, form an
important class of polymer colloids. For example, hybrid nanoparticles made up of an
inorganic core and a polymer shell find wide application in various fields of materials
science such as optics, catalysis, microelectronics, biology, and medicine [5]. As an-
other example, let us consider a 20 nm diameter silica particle surrounded by a 5 nm
polymeric layer (molecular weight 105 g/mol), which contains 61 polymer chains. At
first glance, this number may seem large in comparison to the 26 chains constituting
a 20 nm particle. However, it should be remembered that most of the weight of a par-
ticle resides in its outer layer. For core-shell particles, the volume of the shell becomes
larger than the volume of the core when the shell thickness is 12.5 % of the core dia-

Table 2.1. Number of polymer chains per particle for various polymer nanoparticles.

Particle Outer Polymer Number of

diameter molecular chains per
(nm) weight particle

Spherical, polystyrene 20 105 26

Spherical, polystyrene 100 10 3 200
Spherical, polystyrene 20 10 260
Core-shell 30 105 61
Core = silica (20 nm)
Shell = polystyrene

2.2.2 How many particles?

Three main parameters characterize a polymer dispersion, namely the particle diam-
eter in nanometers (dp ), the solid content (SC), i.e., the weight percentage of solid in
the dispersion, and the particle number (Np ), which is the number of polymer parti-
cles per unit volume of dispersion. These three parameters are interrelated and they
2 A primer on polymer colloids: structure, synthesis and colloidal stability | 11

cannot be varied independently of one another. Simple geometrical arguments can be

used to demonstrate that:
Np = 6 ⋅ 1025 3 , (2.1)
dp ρ
where ρ is the density in g/l of the polymer particle in water. Usually, Np will consist
of between 1013 and 1019 particles/l, which correspond respectively to dilute and con-
centrated dispersions. The final use of the polymer dispersion dictates what the val-
ues of Np , SC, and dp should be. For example, when the polymer dispersion is dried
to form a polymer film, the solid content should be as high as possible in order for the
film to form rapidly. Dispersions with an SC of 60 % and higher are used in the man-
ufacture of aqueous paints (such as acrylic paints). For polymers with a density close
to one, volume fraction and SC are similar. For a monodisperse particle distribution
(i.e., particles all have the same diameter), the theoretical maximum volume fraction is
74 %, which corresponds to a cubic close-packed arrangement of spheres (also called
face-centered cubic, or FCC) [6]. As the nanoparticle arrangement is similar to that
of molecules in a crystal, this state is referred to as a colloidal crystal (Fig. 2.1). The
periodic arrangement of particles and the voids in between them acts as a diffraction
grating for light, with the resulting effect that colloidal crystals are brightly colored.
This explains why they are also called photonic crystals [7]. The majority of colloidal
crystals are prepared with inorganic materials such as silica, although polymer col-
loidal crystals have also been developed [8, 9].

Fig. 2.1. Scanning electron microscopy images of

crystalline assemblies of 100 nm polystyrene nano-
500 nm particles (reprinted with permission from [10]).

When latex is dried, water first evaporates until the particles are in close contact
(stage I of film formation, Fig. 2.2). If the particles are monodisperse and hard (high
glass transition temperature, Tg ), and if the arrangement is devoid of defects (a con-
dition which is practically difficult to achieve), then a colloidal crystal is formed (see
Fig. 2.1). If the particles are deformable (low Tg ), the particles are compressed into
a rhombic dodecahedral arrangement (stage II). Since Tg is low, the polymer from
one particle will diffuse in adjacent particles, resulting in the formation of a continu-
ous film (stage III). The three stages of film formation represent a simplified view of
12 | Part I Functional materials: Synthesis and applications

Stage I

Stage II
particle deforms

Stage III

Fig. 2.2. Schematic representation of the film formation process from a latex and freeze-fracture
transmission electron micrograph image of a latex film at stage II (reprinted with permission
from [12]). The marker bar represents 370 nm.

the formation of a film from latex. The film formation process is in fact much more
complex [11], a fact with important repercussions in the domain of coating technology.
Returning to a dispersion of polymer nanoparticles, we have seen that for a vol-
ume fraction of 74 %, monodisperse latex takes the form of a crystal-like solid. Prac-
tically, the latex does not flow anymore and has an infinitely high viscosity as soon
as the volume fraction reaches 64 % [13, 14]. At this point, the spheres are randomly
packed, and long-range motions are impossible. However, free flowing latexes with
volume fractions higher than 64 % can nonetheless be obtained. In this case, the la-
tex must have a polydisperse size distribution, whereby small particles can be lodged
in the voids created by large particles [15].

2.2.3 Are the chains immobile within the nanoparticle?

When Tg of the polymer is above room temperature, the polymer chains are frozen and
the chains can be considered immobile within the time frame of the experiment. When
Tg is below room temperature, various models such as the Rouse model or the repta-
tion model, can be used to estimate the self-diffusion coefficient (D) of the chains. Let
us consider a case for which the value of D = 10−20 m2 /s [16], which corresponds to a
low value of D that would either be encountered for a high molecular weight chain, or
at temperatures just above Tg . The diffusion length scale (L) can be approximated as
L = (DT)0.5 , where (T) is the diffusion time. In a time interval T = 1 month, the diffu-
sion length scale L will be 161 nm, which is larger than the size of the particles that we
will consider in this chapter. Thus, above Tg the chains diffuse freely within the time
scale of the experiment, whereas below Tg the chains are immobile. So the remain-
ing question is whether the Tg of the nanoparticle corresponds to the Tg of the bulk
polymer. Before answering this question, one should consider the environment of the
nanoparticle. The solvent in which the nanoparticle is suspended (such as water, for
2 A primer on polymer colloids: structure, synthesis and colloidal stability | 13

example) can act as a plasticizer, resulting in a lower Tg . Hydroplasticization is gener-

ally thought to be responsible for a decrease in Tg of approximately 12°C for styrenic
and acrylic polymer nanoparticles suspended in water [17]. As a result of the polymer
chains being confined in a small nanoparticle, Tg deviates from the bulk value. This
so-called confinement effect is due to the combination of interfacial and size effects.
Interfacial effects are influenced by the specific interaction between the particle sur-
face and its environment, whereas size effects occur when the particle size is of the
same order of magnitude as the characteristic size of the region in which cooperative
motion occurs. Although no exact number can be given, one would expect that size
effects are important for nanoparticles smaller than 10 nm. Numerous observations on
polymeric thin films also confirm that size effects are only significant for films thinner
than approximately 20 nm [18]. However, data on the Tg of polymeric nanoparticles are
still very much controversial. For example, recent data indicate that in water, the Tg
of polystyrene nanoparticles with a diameter of 90 nm is 58°C lower than bulk Tg [19].
The relatively simple picture presented in the above paragraph only applies to
simple nanoparticles composed of a single homopolymer. Usually one would prepare,
characterize, and utilize nanoparticles made up of a large number of components, ei-
ther under the form of blends of homopolymers or of copolymers with various archi-
tectures. This point is introduced in the following paragraph.

2.2.4 Morphology of polymeric nanoparticles

As polymers are rarely miscible, phase separation often occurs within nanoparticles,
leading to the formation of a nanostructured object with a specific morphology – that
is to say, with a specific arrangement of each of the polymeric phases. Unraveling the
morphology of a polymer nanoparticle is not a simple task. It is often achieved by
transmission electron microscopy (TEM) with selective staining of one of the phases,
although other techniques such as nuclear magnetic resonance (NMR, [20]) or differ-
ential scanning calorimetry (DSC) have also been used [21].
Starting with a simple blend of two non-compatible polymers, numerous mor-
phologies can be obtained such as core-shell particles, hemispherical particles like
half-moons, particles with various inclusions, etc. (Fig. 2.3). Among all of these, kinet-
ically trapped morphologies should be distinguished from thermodynamic morpholo-
gies. The latter is usually attained when the particle has been exposed to temperatures
above Tg , as in this case the chain’s mobility is high enough to reach the lowest en-
ergy conformation. Thermodynamically speaking, the system is constituted of three
phases: two polymers and the continuous phase. In a seminal study on the coales-
cence of immiscible droplets [22], Torza and Mason defined spreading coefficients Sij
Si = 𝛾jk − (𝛾ik + 𝛾ij ), (2.2)
14 | Part I Functional materials: Synthesis and applications

whereby 𝛾ij defines the interfacial tension between the i and j phases. If the continu-
ous phase is numbered 2 and the polymer phase 1 is the one for which 𝛾12 > 𝛾23 , then
S1 < 0. There are only three possible sets of values for Si , corresponding to three dif-
ferent equilibrium morphological configurations (Fig. 2.3).

1 3

S1 < 0 S1 < 0 S1 < 0

S2 < 0 S2 < 0 S2 > 0
S3 > 0 S3 < 0 S3 < 0
Complete engulfing Partial engulfing No engulfing

Fig. 2.3. Possible equilibrium configurations corresponding to the three sets of Si . The continuous
medium is phase 2.

In fact, the analysis by Torza and Mason relies on simplified scenarios. A more rigorous
approach, developed by Sundberg et al. [23], indicates that equilibrium morphologies
depend not only on the exact values of the respective surface tensions 𝛾ij , but also on
the amount of each polymer and the surface coverage of the particle by a surfactant
(Fig. 2.4). We will see below that surfactants are indispensable to impart colloidal sta-
bility to the polymer nanoparticles.

0% 10% 15% 25% 50% 100%

Fig. 2.4. Predicted morphologies of a polymer particle consisting of two polymers (polystyrene in
black and poly-n-butylmethacrylate in grey) suspended in water with various surface coverage by
sodium dodecyl sulfate (% surface saturation). Reprinted with permission from [23].

As mentioned above, the main technique for unraveling the morphological features
of polymer nanoparticles is TEM. A selective positive stain is used to discriminate one
of the polymer phases. For example, osmium tetroxide can be used to stain polymers
containing benzene rings or isolated double bonds. For simple cases (such as core
shell particles), it is possible to assess the morphology without sectioning the particles
in thin layers with a microtome. However, for more complex cases the particles must
be microtomed after embedding in a hard resin or the morphology may be wrongly
2 A primer on polymer colloids: structure, synthesis and colloidal stability | 15

Fig. 2.5. Simulated TEM pictures of a single

composite particle consisting of polystyrene
in black and poly-n-butylmethacrylate in grey,
microtomed at a thickness of 90 nm, as pre-
dicted by the software UNHLATEXTM_EQMORPH
(Advanced Polymer Laboratory, University of
New Hampshire). Reprinted with permission
250 nm from [23].

attributed. TEM pictures must also be analyzed with great care. For example, Fig. 2.5
illustrates all possible TEM pictures of a single type of nanoparticle, depending on its
preparation and its orientation under the TEM beam.
The morphological features of particles consisting of two homopolymers are very
different from those obtained from polymers with more complex architectures. Among
these, amphiphilic diblock copolymers, which consist of a hydrophobic and a hy-
drophilic block, have been the most studied [24, 25]. Such copolymers, comprised of
a hydrophilic and a hydrophobic block self-assemble in water in order to shield the
hydrophobic units from water. Water is then qualified as a selective solvent, as it is a
good solvent for one part of the block and a poor solvent for the other one. Organic
solvents can also be selective, resulting in reverse structures with hydrophilic units
on the inside and hydrophobic on the outside. Several equilibrium morphologies are
possible, the most classical ones being polymeric spherical and cylindrical micelles
and vesicles (Fig. 2.6). The self-assembly of amphiphilic polymers is dictated by the
value of the critical packing parameter (P), where P = V/(a0 lc ), where V is the volume
of the hydrophobic block, lc is the end-to-end distance between both extremities of
the hydrophobic block, and a0 is the cross-section between both blocks. This param-
eter, first introduced by Israelachvili to predict the assembly of non-polymeric sur-
factants [26], is a nondimensional number which measures the curvature of the cone
generated by the surface enveloping both blocks. Low values of P are obtained for
asymmetric block copolymers formed by a short hydrophobic and a long hydrophilic
block. They self-assemble into spherical polymeric micelles, whereby the hydrophilic
block is solvated, whereas the hydrophobic one forms the core of the micelle. When the
hydrophobic block becomes longer, packing into a sphere is impossible and the mi-
celles become elongated (so-called cylindrical micelles). Eventually, for a high pack-
ing number, lamellar phases are formed, whereby two diblock copolymers in a head-
to-tail conformation generate a locally planar arrangement. These lamellar structures
can fold in on themselves to form a vesicle. Polymer vesicles are often called polymer-
somes [27]. Importantly, at the length-scale of the polymer, the vesicle interface is flat,
and the curvature of the interface is nearly null, thus the diameter of the vesicle is in-
16 | Part I Functional materials: Synthesis and applications

Spherical Cylindrical ‘Polymersomes’

micelles micelles



High Medium Low

curvature curvature curvature
P≤⅓ ⅓≤P≤½ ½≤P≤1

Fig. 2.6. Various self-assembled structures formed by amphiphilic block copolymers in a block-
selective solvent. Reprinted with permission from [28].

dependent of the structure of the block copolymer and mainly depends on preparation
conditions (stirring, temperature, ionic strength, etc.).
The value of the packing parameter can only be used to predict equilibrium mor-
phology. Practically, this condition is achieved when the amphiphilic chains are able
to diffuse through the solvent and probe several conformations until they form the
structure of lowest energy. This is the case when the amphiphilic polymer has nonzero
water solubility (when the self-assembly is performed in water), thus for polymers
with a rather long hydrophilic block. In this case, self-assembly (Fig. 2.6) occurs when
the concentration of polymers in water is above a critical concentration, called criti-
cal assembly concentration (CAC), in reference to the critical micellar concentration
(CMC) observed for small molecule surfactants. It should however be mentioned that
CAC are usually very low (frequently in the μmol/l or below), whereas CMC are usu-
ally higher. Various techniques have been devised to precisely measure CAC, the most
classical being monitoring the change of the fine structure fluorescent emission band
of an organic fluorophore (pyrene), for various concentrations of polymer [29]. Below
the CAC, the polymer chains are dissolved as individual entities, whereas above CAC,
self-assembled have a structured form, but a fraction of the polymer chains (at con-
centration CAC) is solubilized in water. In the latter case, the soluble polymer chains
and self-assembled chains exchange during the experiment. Does this mean that self-
assembled structures are always dynamic and exchange with dissolved chains? Not at
all, it is in fact possible to prepare self-assembled polymer colloids for polymers with
virtually zero solubility in the solvent. Various techniques exist and will be described
below, but it is important to remember that the resulting self-assembled object may be
in a kinetically trapped conformation.
2 A primer on polymer colloids: structure, synthesis and colloidal stability | 17

Fig. 2.7. Self-assembly of block-copolymers

into star polymeric micelles and crew-cut
Star micelles Crew-cut micelles polymeric micelles.

Polymeric spherical micelles, the most commonly encountered self-assembled poly-

meric object, can thus be classified into two distinct categories using the terminology
first introduced by Eisenberg (Fig. 2.7, [30, 31]). Star micelles in water are composed
of a long hydrophilic and a short hydrophobic block, are in dynamic exchange with
freely dissolved polymer chains, and adopt a thermodynamically favored conforma-
tion. They have a low but nonzero CAC, and they form spontaneously when the poly-
mer is added to water. By contrast, crew-cut micelles, consisting of a long hydrophobic
and a short hydrophilic block, are kinetically frozen (Fig. 2.7). Their CAC is virtually
zero and no dynamic exchange process occurs within a reasonable time period. Lastly,
they can only be formed by an indirect method (see below).

2.3 Preparation of polymer nanoparticles

There are so many methods of preparing polymer nanoparticles that an entire book
would not be sufficient to cover them all. This section will present two main classes
of preparation techniques, the first belonging to the class of heterophase polymer-
izations, and the second to the class of self-assembly. Within heterophase polymer-
izations [32], three techniques will be introduced: emulsion, miniemulsion and mi-
croemulsion polymerizations. Other techniques, such as suspension and dispersion
polymerizations will not be covered as they usually lead to micron-size particles. Het-
erophase polymerization (Table 2.2) is the class of polymerization reactions in which a
polymer is formed in a nonsolvent (typically water). Most of these polymerizations are
radical polymerizations because free radicals, unlike carbanions and carbocations, do
not react with water or other solvents currently employed for polymerization. Emul-
sion polymerization is a process whereby a monomer which is mostly water insoluble
is polymerized in water using a water-soluble radical initiator. The resulting product
is a latex, that is to say a dispersion of polymer nanoparticles in water. This is the main
18 | Part I Functional materials: Synthesis and applications

polymerization technique for obtaining polymer nanoparticles. Besides conventional

emulsion polymerization, miniemulsion polymerization and microemulsion polymer-
ization also lead to the formation of latexes. If the monomer is soluble in the solvent
but not the polymer, the process is referred to as dispersion polymerization. If both
monomer and radical initiators are insoluble in water, then large particles are ob-
tained – this process is referred to as suspension polymerization.

Table 2.2. Main features of heterophase polymerization.

Polymer- Emulsion Dispersion Suspension

ization Conventional Microemulsion Miniemulsion
Continuous water water water organic water
Phase solvents or
Monomer – sparingly – in micelles – in pre- – soluble – low sol-
soluble in formed in the ubility in
water stable continuous water
– in droplets droplets phase – droplets
– in particles
Initiator water-soluble water-soluble water-soluble soluble in the soluble in the
or insoluble continuous monomer
Dispersant ionic/nonionic ionic/nonionic ionic/nonionic soluble hydrosoluble
surfactant surfactant surfactant polymer polymer
Product stable latex stable stable latex stable latex suspension
10 to 500 nm transparent 50 to 500 nm 0.1 to 0.5 μm; (> 5 μm)
latex dispersions
10 to 20 nm < 20 μm

2.3.1 Emulsion polymerization

Emulsion polymerization is a remarkably simple experiment to conduct which usu-

ally leads to a high yield of monodisperse nanoparticles [2]. Numerous variants ex-
ist, such as, for example, the seeded emulsion polymerization, whereby small nano-
particle seeds are used as nuclei for polymerization, thus superseding the need for
nucleation. By using several monomers, a very large number of particle morphologies
can be attained. Emulsion polymerization is one of the major processes used in indus-
try for the production of polymers. Among the monomers polymerized in emulsion, let
us cite the acrylic monomers, fluorinated and chlorinated monomers such as vinyli-
dene chloride, vinyl chloride or tetrafluoroethylene, vinyl acetate and its copolymers,
styrene, and butadiene [4].
2 A primer on polymer colloids: structure, synthesis and colloidal stability | 19

Emulsion polymerization presents numerous advantages:

– As water has a very high heat capacity, the dispersed system allows good control
of the exothermicity of the polymerization reaction.
– The viscosity in the latex remains low, even when SC reaches 50 % or above, in
contrast with solution polymerization, where viscosity skyrockets as SC increases.
– Emulsion polymerization yields high molecular weight polymers, usually higher
than solution or bulk polymerization under similar conditions. This is due to the
fact that the radicals are compartmentalized in separate particles. Their life span
is usually longer than in bulk.
– The rates of emulsion polymerizations are usually higher than solution polymer-
– Emulsion polymerization is a very handy way of manufacturing sticky rubbers
or other viscoelastic compounds, since the particles are surrounded by a layer of
emulsifier that prevents them from coagulating.
– Last but not least, emulsion polymerization does not require any organic solvent.

There is of course a price to pay for all these advantages:

– First, latexes are very fragile and are prone to flocculation (coagulation). Latexes
are always stabilized by convenient surfactants, necessary to retain the colloidal
state (see below). However, the surfactant can be undesirable in the final product.
– One key step during the polymerization process is nucleation, which is often
found to be highly variable and difficult to reproduce. This often results in latexes
presenting variable particle sizes.

Let us consider a typical emulsion polymerization recipe: 400 g of an organic mono-

mer, 1 000 g of water, 4 g of a surfactant, for example sodium dodecyl sulfate (SDS),
and 1 g of a water soluble radical initiator such as potassium persulfate. As the cmc
of SDS is 2.4 g/l at room temperature, then approximately 4 − 2.4 = 1.6 g of SDS are
engaged in forming micelles. Since on average each SDS micelle is constituted of
50 molecules, there are 6.7 ⋅ 1019 micelles in the system. These micelles are swollen
by the monomer: each gram of micelle can accommodate around 0.5 g of styrene
within its core. Furthermore, a small amount of styrene (0.3 g) is simply dissolved
in water (styrene has a low but nonzero aqueous solubility). Thus, from the initial
400 g of styrene monomer, approximately 399 g are neither in the aqueous phase
nor in the micelles. If the system is not stirred, a separate organic phase will ap-
pear on top of the aqueous phase, but under mechanical stirring large droplets (1–10
microns in diameter) stabilized by a small amount of surfactant will form. If the
water-soluble radical polymerization initiator is added at this point, emulsion poly-
merization is triggered. A series of colloidal events ensue which were first described by
Harkins [33], and which is referred to as the mechanism of emulsion polymerization
(Fig. 2.8).
20 | Part I Functional materials: Synthesis and applications

Stage 1 I : Initiator
I M : Monomer
: Surfactant

Monomer swollen
micelle 2-10 nm

Monomer swollen
polymer particle
Stage 2



Monomer M M
droplet MM
1-10 microns I
Stage 3



Fig. 2.8. The three stages in Harkins’ model for emulsion polymerization.

Stage I: Nucleation
The radical initiator, dissolved in water, decomposes to form free radicals which sub-
sequently form free radical oligomers by reacting with the low amount of monomers
dissolved in the aqueous phase. Each time a monomer is inserted, the oligomer be-
2 A primer on polymer colloids: structure, synthesis and colloidal stability | 21

comes less and less water-soluble and eventually diffuses to the organic phase. As-
suming that the monomer droplets have an average diameter of 10 microns, and using
equation (2.1), one finds that there are 2 ⋅ 1012 droplets in the system. The total sur-
face area provided by all these droplets is 650 m2 . By contrast, the 6.7 ⋅ 1019 micelles,
with an average diameter of 4 nm, present a surface area of 3 400 m2 . Statistically, the
radicals are captured by the micelles which are swollen by monomers, and on poly-
merization, the micelle is transformed into a growing polymer particle. The formation
of polymer particles is referred to as nucleation which, according to this model, occurs
in the micelles and is a case of micellar nucleation. Generally, only a small fraction of
the micelles are used for particle nucleation, the rest serving as a surfactant reservoir
to stabilize growing particles. When no more micelles are present, the nucleation is
terminated. At this point there are typically 1014 –1018 particles per liter of emulsion
and the overall conversion is of the order of 1–10 %. Nucleation can also take place
when the surfactant concentration is below CMC. This is referred to as homogeneous
nucleation. In this case, the radical oligomer grows into the aqueous phase; it even-
tually precipitates, forming a very small particle (so-called primary particle), which
serves as a nucleus for a polymer particle. Homogeneous nucleation leads to latexes
with a lower Np , usually consisting of between 1013 –1015 particles per liter.

Stage II: Steady state

Once nucleation is terminated, Np remains constant. Polymerization occurs within the
particle where monomer concentration is high (the polymer is swollen by the mon-
omer). Each time a monomer is polymerized in a polymer particle, another one dif-
fuses from the aqueous phase to the particle and yet another diffuses from a mono-
mer droplet to the aqueous phase. Thus, the monomer droplets act as monomer reser-
voirs and the monomer concentration in the particle (Cp ) remains constant during this
stage. Two antagonistic factors control Cp : the swelling of the polymer particle by the
monomer, which is entropically favorable for the polymer, and the increase of particle
surface upon swelling by the monomer, which is associated with an increase in sur-
face energy. These two contributions are balanced in the Morton–Kaizermann–Altier
2V0 𝛾
0 = ln(1 − φp ) + φp + χφ2p + , (2.3)
where χ is the Flory—Huggins interaction parameter between the polymer and the
monomer, r the radius of the particle, φp the volume fraction of the polymer in the
particle, and V° the molar volume of the monomer. From this equation, φp can be cal-
culated, followed by Cp (Table 2.3, [34]).
During stage II, Cp and Np remain constant. Therefore, the rate of polymerization
is constant; this stage is the steady state period of the polymerization. Similar to the
nucleation state, radicals formed in the aqueous phase enter the particles at regular
intervals. If the particle already contains a radical, then chances are that the two radi-
22 | Part I Functional materials: Synthesis and applications

Table 2.3. Monomer solubility in polymer particles (Cp ) and in water (Cw ) for the emulsion polymer-
ization of common monomers.

Monomer Cp (mol/l) Cw (mol/l)

Styrene 5.5 0.0043
Methyl methacrylate (MMA) 6.6 0.15
Butyl methacrylate (BMA) 3.8 0.0025
Butyl acrylate (BUA) 5.0 0.0064
Vinyl acetate (VAC) 7.5 0.5

cals, being in close proximity, will terminate. Thus, except for very large polymer par-
ticles, the particle can only contain 0 or 1 radical, and the average number of radicals
per particle is n = 12 . In certain cases (for example after a transfer to monomer), radi-
cals can exit the particle before another radical entry occurs, and in this case, n ≺ 12 .
Large particles can accommodate two or more radicals without termination, and in
this case, n ≻ 12 .

Stage III
At this point the monomer droplets have disappeared and the monomer concentration
in the particle diminishes. Often, a gel effect (Trommsdorff effect) occurs: in short, the
polymerization rate increases significantly due to a drastic increase of n (radical ter-
mination is slowed by the high internal viscosity of the particle). Furthermore, as the
monomer is consumed in the polymer particles, the particles shrink during this stage.
Although emulsion polymerization is by far the most widely practised hetero-
phase emulsion polymerization technique, it also suffers from intrinsic limitations.
For example, its mechanism implies that the monomer diffuses through the aqueous
phase. Therefore, monomers with extremely low aqueous solubility (Table 2.3) are of-
ten difficult to polymerize. In this case miniemulsion polymerization should be used.

2.3.2 Miniemulsion polymerization

The use of miniemulsions to form polymer nanoparticles was pioneered by Ugelstad

and El-Aasser [35]. Miniemulsions are dispersions of critically stabilized oil in wa-
ter droplets prepared by shearing a system containing oil, water, a surfactant, and
a hydrophobe [36–38]. In miniemulsion polymerization, the monomer is emulsified to
form a stable miniemulsion ranging in size from 100 to 500 nm. Each of these droplets
becomes an independent nanoreactor for the polymerization, and the resulting poly-
mer particles form an exact replica of the initial nanodroplets.
One of the key features of miniemulsion polymerization is the colloidal stability of
the initial droplets. In conventional oil-in-water emulsions, droplets of various sizes
2 A primer on polymer colloids: structure, synthesis and colloidal stability | 23

are formed and rapidly phase separate into one separate oil-phase when stirring is
stopped. Ostwald ripening, the diffusion of the oil phase through the aqueous phase
(Fig. 2.9) is responsible for this gradual coarsening of the emulsion. This process is
driven by the difference in Laplace pressure between droplets having different radii:
the oil diffuses from the smallest to the largest droplets. Besides Ostwald ripening,
mass transfer between droplets can occur when they collide. This collision mechanism
is most conspicuous in highly sheared systems and leads to either fusion or fission of
the droplets.

Ostwald ripening

Diffusion through
continuous phase



Fig. 2.9. Schematic representations of Ostwald ripening and the fission and fusion mechanisms of

Ostwald ripening can be effectively suppressed by the addition of a small amount of

a hydrophobic compound (called hydrophobe) to the oil (Fig. 2.10). The hydrophobe
cannot diffuse from one droplet to another and is trapped in each droplet. If the emul-
sified oil was to diffuse from a small to a larger droplet, the hydrophobe concentration
would increase in the smaller droplet (Fig. 2.10), which is not favorable thermodynam-
ically due to osmotic pressure increase. Therefore, the presence of the hydrophobe in
each droplet efficiently prevents Ostwald ripening. Hexadecane is the most commonly
used hydropobe, but numerous other hydrophobes have been shown to successfully
prevent Ostwald ripening [36–37]. Similar to direct miniemulsion, osmotic pressure
in reverse miniemulsion (water-in-oil emulsion) can effectively be suppressed by an
agent soluble only in the aqueous phase, a so-called “lipophobe”, such as an inor-
ganic salt [39].
24 | Part I Functional materials: Synthesis and applications

The formulation of a miniemulsion is relatively simple: nearly any type of oil and
surfactant can be miniemulsified to lead to metastable droplets with sizes usually
ranging from 50–500 nm. The dispersion requires a high-energy mechanical device,
usually a probe sonicator is used, but other devices such as high pressure homogeniz-
ers (for example a commercial microfluidizer), can also be employed.


Fig. 2.10. Suppression of Ostwald ripening by addition of a hydrophobe to the emulsified oil.

Miniemulsion polymerization has become increasingly popular in recent years be-

cause it is an extremely versatile tool for preparing polymer nanoparticles and also
for encapsulating hydrophobic compounds within polymer particles. Encapsulation
of lipophilic drugs [40], bactericides [41], functional oligomers [38], and even inor-
ganic particles [42] has been achieved by this method. However, miniemulsion is not
an ideal technique for small nanoparticles (d < 50 nm). Recently, much effort has
been made to prepare miniemulsions which don’t require sonication of high-shear
homogenization [43].

2.3.3 Microemulsion polymerization

In microemulsion, the monomer is initially confined to micelles and no droplets are

present. Polymerization occurs within the micelle. It has been estimated that approxi-
mately one out of ten micelles is nucleated and becomes a polymer particle, the other
micelles serving as monomer reservoir and surfactant reservoir. The resulting disper-
sion is thus formed of small polymer particles which are nonetheless larger than the
original micelles. Their size ranges from 10–50 nm [44–45]. Microemulsion polymer-
ization prevents several specific characteristics which are worth mentioning. Firstly,
due to their very small size (Fig. 2.11), the nanoparticle dispersion is translucent and
does not significantly scatter light. Polymer molecular weights are also usually very
high; for a system composed of small particles, the number of particles is very high
(see equation (2.1)), and once a radical enters a micelle, another radical entry into the
same micelle leading to termination is unlikely. As a result, most if not all particles
contain one radical during polymerization, and as Np is very high, polymerization is
extremely rapid.
2 A primer on polymer colloids: structure, synthesis and colloidal stability | 25


Fig. 2.11. Nanolatexes of polyethylmethacrylate (EMA), polystyrene (S), butyl methacrylate (BMA),
and 2-ethylhexylmethacrylate (EHMA) obtained by microemulsion polymerization. For these nanola-
texes, a cobalt-based catalyst was used to limit polymer molecular weight. Reprinted with permis-
sion from [46].

In contrast to metastable miniemulsions, which are only obtained after a high-energy

mixing step, microemulsions are thermodynamically stable and form spontaneously.
However, their formulation is particularly difficult. They require large amounts of sur-
factant, often larger amounts than monomer. In the ternary phase diagram (water,
monomer, surfactant), the stability region for the microemulsion is extremely narrow.
Thus, a slight change in monomer concentration, temperature or other operating pa-
rameters can result in loss of the microemulsion [47, 48].

2.3.4 Self-assembly in selective solvents

As mentioned above, block copolymers with long hydrophobic sequences form crew-
cut micelles upon self-assembly in water. However, due to the lack of exchange
mechanism, the self-assembly process is not spontaneous and it requires an indi-
rect preparation route. The method generally employed consists of transferring the
block copolymer from a good solvent, in which both blocks are soluble, to a selec-
tive solvent, in which only one of the blocks is soluble. The main requirement is that
the good solvent and the selective solvent be miscible. Among the usual combina-
tions of good and selective solvents, one finds THF/water or DMF/water. Once the
self-assembly step is performed, the good solvent is removed, either by dialysis or
by repeated centrifugations, although care should be taken to avoid flocculation of
the nanoparticles during centrifugation. During the addition of the selective solvent,
the solvation environment changes: it is first a good solvent, then an intermediate
quality solvent, whereby micellization and rapid exchange between micellar and
free polymer occurs, and finally it becomes a highly selective solvent, whereby the
micelle configuration is frozen [30, 31]. Ternary phase diagrams for the polymer and
26 | Part I Functional materials: Synthesis and applications

both solvents are often extremely complex. Depending on the experimental design
(polymer concentration and rate of selective solvent addition), a given trajectory will
be followed on the phase diagram when the selective solvent is added [49], and var-
ious zones of predominance will be visited. The result of the self-assembly process
corresponds to the first structure for which exchange between micellar and dissolved
chains is slow relative to the duration of the experiment.

2.4 Colloidal stabilization

As mentioned above, polymer nanoparticles are not colloidally stable unless they are
stabilized, for example by a surfactant. Several mechanisms can be used to provide
colloidal stability; they are based on (1) electrostatic stabilization, (2) steric stabi-
lization, and (3) depletion mechanism. When the system is not colloidally stable it
will coagulate or flocculate, resulting in the formation of coagulum, or floc. Both
terms refer to the formation of particular aggregates, although for polymer colloids,
the word coagulum is usually employed for compact aggregates which cannot be re-
dispersed, whereas flocs are loosely bound aggregates, which in some cases can be
redispersed into free flowing colloids. Although the colloidal stabilization of polymer
nanoparticles at low solid content does not present any major difficulty, at high solid
content (above 50 %), systems are prone to flocculate and one cannot rely on a sin-
gle mechanism (electrostatic, steric, depletion) to impart colloidal stabilization. As
a last note, several manufacturers commercially produce surfactant-free polystyrene
latexes, for example as a tool for calibrating light-scattering instruments. Surfactant-
free does not mean charge-free (residual charges from the emulsion polymerization
initiator are present on the particle surface), and this low number of charges is suffi-
cient to impart colloidal stability.

2.4.1 Electrostatic stabilization

Electrostatic stabilization is the predominant strategy used to stabilize polymer col-

loids in aqueous media. It is effective as soon as charges are immobilized at the sur-
face of the nanoparticle. This can be achieved in a variety of ways: most commonly
a charged surfactant will be added to the colloidal dispersion. In the case of the an-
ionic surfactant sodium dodecyl sulfate (C12 H25 –O–SO−3 Na+ ), the hydrophobic tail
(C12 H25 ) adsorbs on the surface of the particle, leaving the polar sulfate head (OSO−3 )
at the interface between the particle and the aqueous phase. The sodium counter ion
is ‘free’ in water. Thus, the polymer nanoparticle is decorated by negative charges via
the adsorption of the surfactant (sulfate heads, Fig. 2.12). Note that in this case, the
adsorbed surfactant is in dynamic equilibrium with the dissolved surfactant, there-
fore they are not static punctual charges per se. However, the mathematical descrip-
2 A primer on polymer colloids: structure, synthesis and colloidal stability | 27

tion of electrostatic stabilization is not affected by the presence of this dynamic equi-
The simplest (and inaccurate) explanation for the electrostatic stabilization mech-
anism is the presence of an electrostatic repulsion between the two negatively charged
nanoparticles (assuming that an anionic surfactant is used). This is far from satisfying
and more information can be gleaned by considering the electrical potential of a sin-
gle charged spherical nanoparticle immersed in water (Fig. 2.12). Due to the presence
of surface charges, ions of opposite charge are condensed (adsorbed) at the proximity
of the particle surface. The liquid layer where these ions are located is called the Stern
layer. Further away from the Stern layer, the ions are attracted by their close neighbors,
but they can also exchange with freely diffusing ions in the solution. Far away from
the particle surface, the spatial distribution of cations and anions is homogeneous
(concentration c∞ ), and the attraction (or repulsion) by the charged particle surface
is entirely screened. Thus, an intermediate layer exists between the Stern layer and
the free solution containing ions of low (but non-zero) mobility, which feel the elec-
trostatic potential generated by the surface charge. It is referred to as the diffuse layer
(Fig. 2.12), and the Stern and diffuse layers combined form the double layer. Let us now
consider two charged nanoparticles approaching one another. When the diffuse lay-
ers interpenetrate, the concentration of ions in the overlap region becomes larger than
the concentration of ions which are free in solution, resulting in an increase of osmotic
pressure. The system will thus tend to separate the two charged nanoparticles in order
to recover osmotic equilibrium. Therefore, the ‘force’ which repels two charged nano-
particles is principally osmotic in nature and is often called an electro-osmotic force.
By solving the Poisson–Boltzmann equation using a number of assumptions [50],
one finds that the electrical potential (V(r)) generated from the surface is maximal at
the surface and decreases exponentially with r, where r is the distance from the sur-
face of the nanoparticle (Fig. 2.12). The characteristic decay length of this exponential
is 1/κ, which is referred to as the Debye length of the diffuse layer. The quantity 1/κ
is often used as a measurement of the thickness of the double layer. The electrostatic
potential has nearly entirely vanished at a distance 2/κ from the surface and thus, the
distance at which the surfaces of two approaching nanoparticles start to electrostat-
ically interact is 4/κ. The mathematical expression of the Debye length (Fig. 2.12) is
relatively simple and only takes temperature and ionic strength of the medium into
account, with a characteristic exponent of −1/2. Thus, the higher the ionic strength,
the faster the surface charges will be screened by other charges and the steeper the de-
cay of the electrostatic potential. In Table 2.4, values of Debye lengths are consigned
for various electrolytes and concentrations. For divalent and trivalent cations, the De-
bye length becomes very small, even at moderate ionic strength; thus nanoparticles
can approach each other at very close proximity without any electrostatic repulsion.
In these conditions of ionic strength, the electrostatic stabilization is ineffective. It
is thus not surprising that Al3+ -containing salts such as aluminum sulfate are used
industrially as flocculants. Inversely, at very low ionic strength, the diffuse layer ex-
28 | Part I Functional materials: Synthesis and applications

tends very far into the solution, and even at low solid content the diffuse layers of
each polymer nanoparticle interpenetrate. As the ions and the surrounding hydrat-
ing water in the diffuse layer have low mobility, the viscosity of the latex increases
significantly. Such an effect is referred to as an electroviscous effect. A spectacular ex-
periment for demonstrating this effect consists of deionizing the continuous phase of
a monodisperse latex with an equimassic mixture of acidic and basic ion exchange
resins. When the ionic strength is low enough, the latex looks like an opalescent vis-
cous gel which will immediately return to liquid on addition of a pinch of salt.

V(r) 2 
1 –kr Particle
V(r) = 64 RT c∞ ϒ 0 e
k Stern layer
1 RTεε0 0.301 Diffuse layer
= ≈
K F 2 c∞ zi
–ζ i

Buffer concentration c∞


Fig. 2.12. Representation of the double layer around a charged nanoparticle (charge = z) and
electrostatic potential V(r) (in absolute value) versus distance from the nanoparticle surface (r).
The potential depends on the temperature (T), the concentration of ions in the solution (c∞ ), and
the term 𝛾o = tan h (zFΨ/4RT) where (Ψ) is the potential at the surface and (F) the Faraday constant.
The Debye length (1/κ) depends on the ionic strength, defined as I = Σ(c∞ z2i ), where zi is the re-
spective charge of each ion in solution. The potential (ζ) corresponds to the value of the electro-
static potential at the interface between the Stern and the diffuse layer.

Table 2.4. Debye length in nm versus electrolyte concentration.

Electrolyte Debye length (1/κ) in nm

(mol/l) NaCl MgCl2 Al(NO3 )3
10 96.2 55.5 39.2
10−4 30.4 17.5 12.4
10−3 9.6 5.6 3.9
10−2 3.1 1.8 1.2
10−1 1 0.5 0.3
2 A primer on polymer colloids: structure, synthesis and colloidal stability | 29

So far, only the repulsive electro-osmotic force was considered to describe the in-
teraction of charged nanoparticles. However, other interactions are at stake and a sim-
plified description of all the interactions between two charged nanoparticles was pro-
posed by Derjaguin and Landau [51], Verwey and Overbeek [52], in what constitutes
the DLVO theory. The interaction potential (Fig. 2.13) between two polymer particles
is given by the summation of an attractive van der Waals potential, Va , the repulsive
electro-osmotic potential, Vr described above, and a repulsive very short action Born
potential, Vb , which prevents the two nanoparticles from interpenetrating. The over-
all potential presents two minima (Fig. 2.13). A deep primary minimum is located at a
short distance which corresponds to the distance at which the two nanoparticles are in
close contact. When the particles fall into this minimum, they are (irreversibly) coag-
ulated. A maximum height (Vm ) is present at intermediate distance. Lastly, a shallow
secondary minimum is present at large distance The energy difference ΔVf represents
the energy barrier that a particle must cross in order to reach the primary minimum.
The higher this barrier, the more stable the colloid. It is generally accepted that under
normal conditions a barrier of ΔVf = 15 RT is sufficiently high for the colloid to be
stable. As the Debye length is shorter with higher ionic strength, the increase in ionic
strength is one of the key factors responsible for a decrease of ΔVf and therefore for
the loss of colloidal stabilization.

V maximum

V = Vr + Va + Vb

1 –kr
Vr = 64RT c∞ ϒ 0 e
Vm Va =
Vb ΔVf

Va mininum


Fig. 2.13. Interaction energy versus H, where H is the distance separating two spheres. Vb is the
Born repulsive potential, Va is the van der Waals attractive potential (lower dashed line), and Vr is
the electrostatic potential (upper dashed line). The radius of each sphere is R and the composite
Hamaker constant (which takes both the sphere and the continuous medium into account) is A.
30 | Part I Functional materials: Synthesis and applications

2.4.2 Steric stabilization

Another method of stabilizing latex particles is to physically prevent their approach

by introducing a hairy layer of solvated polymer at their surface. If the hairy layers
are sufficiently thick, the two spheres will only be able to approach each other up to a
distance at which the van der Waals attraction is small relative to thermal energy (kT).
For a particle of 20 nm size (Fig. 2.14), a layer a few nanometers thick will be sufficient.
However, for larger particles, very thick layers need to be employed as the reach of the
van der Waals attraction increases with particle size. Thus, it is easier to stabilize small
particles by a steric mechanism. From Fig. 2.14, one may infer the size of the hairy
layer which is necessary to impart steric stabilization. It is possible to convert layer
thickness into polymer molecular weight, provided specific information on the poly-
mer conformation in the solvent is available. Such information can be gathered ex-
perimentally (using viscosimetric or light scattering experiments) or theoretically (via
atomistic calculations). The rule of thumb is that the radius of gyration of the chain is
Rg ≈ 0.05 MW0.5 . For a particle of diameter 200 nm, a hairy layer of 6.6 nm provides a
stabilization of 2 kT, which corresponds to a molecular weight of 17 000 g/mol.
When sterically stabilized particles approach each other, interpenetration of the
hairy layers results in a restriction of their accessible configurations, which is unfavor-
able in terms of entropy. Furthermore, the concentration of hairs in the overlap region
increases, which results in a local increase of osmotic pressure. To maintain osmotic
equilibrium, the solvent will tend to decrease the local concentration of hairs, which
results in separation of the particles. Thus, steric stabilization is osmotic in nature [53].
Steric stabilization can be achieved by several means. For example, non-ionic sur-
factants, consisting of a hydrophobic unit and a pegylated water-soluble polymer can
impart steric stabilization. The reader should be aware that these surfactants exhibit
a cloud point, that is to say a temperature above which they are not soluble [54]. In-
deed, at low temperature, the pegylated chain is soluble in water due to the presence of
hydrogen bonds between water and the oxygen atom of the pegylated chain. This rep-
resents an enthalpic contribution to the heat of dissolution of the polymer. At higher
temperature, the entropic cost associated with the liberation of the hydrogen bonded
water molecules offset the enthalpic gain and the pegylated chain becomes insoluble
in water. Cloud points are dependent on the nature and concentration of the surfac-
tant and the nature of the ions present in the continuous phase.
Another method of achieving steric stabilization relies on the use of water-soluble
polymers which are introduced during polymerization and act as chain transfer
agents. For example, polyvinyl pyrollidone (PVP) is susceptible to hydrogen abstrac-
tion (chain transfer) in α to the carbonyl group. During an emulsion polymerization,
the hydrogen abstraction step is followed by the insertion of hydrophobic monomers,
resulting in the formation of a graft amphiphilic copolymer on a PVP backbone. This
amphiphilic graft polymer, formed in situ during polymerization, serves as steric
2 A primer on polymer colloids: structure, synthesis and colloidal stability | 31





–10 dp = 20 nm
–12 dp = 200 nm
dp = 500 nm


0 5 10 15 20
r (nm)

Fig. 2.14. Effect of the particle diameter on the van der Waals attraction between two particles. The
potential of interaction was calculated using the analytical expression for van der Waals potential
presented in [55]. The horizontal line corresponds to an attraction potential of 2 kT. A Hamaker con-
stant of 0.95 ⋅ 10−20 J, corresponding to the Hamaker constant of polystyrene in water, was used to
create this graph.

Steric stabilization is usually considered to be insensitive to ionic strength but sensi-

tive to temperature. Unlike electrostatic stabilization, steric stabilization is also effec-
tive in organic mediums, provided the hairy layer is soluble in the solvent. Finally, in
aqueous mediums, a very efficient method of stabilizing a colloid consists of the elec-
trosteric mechanism, whereby the hairy layer is constituted of a polyelectrolyte [56],
as it combines the features of electrostatic and steric stabilization.

2.4.3 Depletion stabilization

In the depletion stabilization mechanism, the stabilizing polymer does not inter-
act directly with the nanoparticle surface, but is simply dissolved in the continuous
medium. Thus, the choice of stabilizer is greatly simplified, as its only structural
requirement is that it be soluble in the solvent. However, this mechanism is to be
handled with great care, as the same molecule can act as either stabilizer or floccu-
lant [57]. The stabilization mechanism occurs when the medium is concentrated in
stabilizer (or dilute in nanoparticles). Inversely, depletion occurs when the medium
is dilute in polymer (or concentrated in nanoparticles). Practically, stabilization by
depletion occurs at stabilizer concentrations which are so high that it is difficult to use
under most experimental conditions. The reverse is unfortunately true: it is a com-
mon observation that flocculation of a stable dispersion is triggered upon addition of
a soluble polymer.
32 | Part I Functional materials: Synthesis and applications

When a polymer is dissolved in the solvent, its center of mass cannot approach
the surface of the nanoparticle (Fig. 2.15). Therefore, each nanoparticle is surrounded
by an imaginary layer which cannot be visited by the center of mass of the dissolved
polymer. The thickness of this depletion layer is equal to Rg [58]. When two nano-
particles approach one another, their depletion layers overlap. In the overlap volume,
the depletion layer is shared by two particles, with the consequence that solvent has
been released outside the depletion layer. Thus, the volume of solvent accessible to
the polymer increases. Lowering the polymer concentration results in a decrease of its
chemical potential, which is favorable thermodynamically. This is the basis for deple-
tion flocculation.


Particle Particle Particle Particle

Depletion zone

d : Inter-particle distance 2 1 Depletion

Fig. 2.15. Sketch of the interaction potential between two spheres immersed in a solvent containing
a polymer (radius of gyration Rg ) dissolved in the solvent.

Let us now consider the case where the nanoparticle suspension is very dilute. When
two nanoparticles approach each other, they will have to first repulse (or exclude) dis-
solved polymer chains before the depletion layers start to overlap. Compression of
the polymer chains starts to occur when the inter-particle distance is equal to 2 Rg .
When the inter-particle spacing is further reduced, the conformation of the polymer
chain is changed from a random coil to a conformation which is more geometrically
constrained and less entropically favorable. Thus, an energy barrier must be crossed
in order for the depletion layers to overlap and the particles to flocculate. According
to Napper, if this barrier is higher than 20 kT [53], then the colloid is protected from
flocculation. However, this is only the case if the concentration of polymer dissolved in
2 A primer on polymer colloids: structure, synthesis and colloidal stability | 33

the solvent is high and if the polymer molecular weight is high. Indicative calculations
are presented in [53] for a polystyrene latex stabilized by a non-ionic surfactant with
17 ethylene oxide units. Depletion stabilization occurs when a 1 000 g/mol polyethy-
lene glycol polymer is added at a concentration of 380 g/l or above. Any lower concen-
tration results in destabilization by depletion flocculation. If a 10 000 g/mol polyethy-
lene glycol is used instead, stabilization occurs for concentration greater than 55 g/l.
Obviously such high concentrations are impractical, as they will lead to very high vis-

2.4.4 Future directions

The field of colloidal stabilization is a fascinating and vibrant area of science which
combines physical chemistry and fluid dynamics. For example, recently a new form of
particle stabilization, named haloing, was discovered. It occurs when an uncharged
nanoparticle is immersed in a medium containing very small, highly charged nano-
particles [59]. The domain also has exceedingly important technological repercus-
sions, ranging from advanced materials to drug delivery. The small chapter presented
here only scratches the surface of the field, and the interested reader should consult
the references listed below.


[1] Morrison ID, Ross S. Colloidal dispersions: suspensions, emulsions, and foams. John Wiley &
Sons, New York, USA, 2002.
[2] Fitch RM. Polymer Colloids: A comprehensive introduction. Academic Press, San Diego, USA,
[3] Lovell PA, El-Aasser M. Wiley: Emulsion polymerization and emulsion polymers. John Wiley &
Sons, Chichester, UK, 1997.
[4] van Herk, AM. Chemistry and technology of emulsion polymerization. Blackwell, Oxford, UK,
[5] Bourgeat-Lami E. Organic–inorganic nanostructured colloids. J. Nanosci. Nanotechnol.
2 (2002), 1–24.
[6] Manoharan V, Pine D. Building materials by packing spheres. MRS Bull. 2004, 91–5.
[7] Joannopoulos J, Johnson S, Winn J, Meade R. Photonic crystals: molding the flow of light.
Princeton University Press, Princeton, USA, 2011.
[8] Edrington AC, Urbas AM, DeRege P, et al. Polymer-based photonic crystals. Adv. Mat.
13(6) (2001), 421–5.
[9] Li Z, Wang J, Song Y. Self-assembly of latex particles for colloidal crystals. Particuology.
9 (2011), 559–65.
[10] Gates B, Qin D, Xia Y. Assembly of nanoparticles into opaline structures over large areas. Adv.
Mat. 11 (1999), 466–9.
[11] Keddie JL, Routh AF. Fundamentals of latex film formation, processes and properties. Springer,
Dordrecht, the Netherlands, 2010.
34 | Part I Functional materials: Synthesis and applications

[12] Wang Y, Kats A, Juhue D, Winnik MA, Shivers RR, Dinsdale CJ. Freeze-fracture studies of latex
films formed in the absence and presence of surfactant. Langmuir. 8 (1992), 1435–42.
[13] Greenwood R, Luckham PF, Gregory T. The effect of diameter ratio and volume ratio on the
viscosity of bimodal suspensions of polymer lattices. J. Colloid Interface Sci. 191 (1997), 11–21.
[14] Torquato S, Truskett T, Debenedetti P. Is random close packing of spheres well defined? Phys.
Rev. Lett. 84 (2000), 2064–7.
[15] Schneider M, Claverie J, Graillat C, McKenna TF. High solids content emulsions. I. A study of
the influence of the particle size distribution and polymer concentration on viscosity. J. Appl.
Polym. Sci. 84 (2002), 1878–96.
[16] Bartels CR, Crist B, Graessley WW. Self-diffusion coefficient in melts of linear polymers: chain
length and temperature dependence for hydrogenated polybutadiene. Macromolecules.
17 (1984), 2702–8.
[17] Jiang B, Tsavalas J, Sundberg D. Measuring the glass transition of latex-based polymers in the
hydroplasticized state via differential scanning calorimetry. Langmuir. 26 (2010), 9408–15.
[18] Forrest J, Dalnoki-Veress K, Stevens J, Dutcher J. Effect of free surfaces on the glass transition
temperature of thin polymer films. Phys. Rev. Lett. 77 (1996), 2002–5.
[19] Zhang C, Guo Y, Priestley RD. Glass Transition temperature of polymer nanoparticles under soft
and hard confinement. Macromolecules. 44 (2011), 4001–6.
[20] Landfester K, Dimonie VL, El-Aasser MS. The evaluation of the size and the structure of the
interphase in composite particles containing a macromonomer studied by solid-state NMR.
Macromol. Chem. Phys. 203 (2002), 1772–80.
[21] Stubbs JM, Sundberg DC. Measuring the extent of phase separation during polymerization
of composite latex particles using modulated temperature DSC. J. Polym. Sci. Pol. Phys.
43 (2005), 2790–806.
[22] Torza S, Mason SG. Coalescence of two immiscible liquid drops. Science. 163 (1969), 813–4.
[23] Sundberg DC, Durant YG. Latex Particle morphology, fundamental aspects: A review. Polym.
Reac. Eng. 11 (2003), 379–432.
[24] Förster S, Antonietti M. Amphiphilic block copolymers in structure-controlled nanomaterial
hybrids. Adv. Mat. 10 (1998), 195–217.
[25] Riess G. Micellization of block copolymers. Prog. Polym. Sci. 28 (2003), 1107–70.
[26] Israelachvili JN. Intermolecular and surface forces, revised 3rd edition. Academic Press,
Waltham, USA, 2011.
[27] Discher BM. Polymersomes: tough vesicles made from diblock copolymers. Science.
284 (1999), 1143–6.
[28] Blanazs A, Armes SP, Ryan AJ. Self-assembled block copolymer aggregates: From micelles to
vesicles and their biological applications. Macromol. Rapid Comm. 30 (2009), 267–77.
[29] Tsitsilianis C, Iliopoulos I, Ducouret G. An associative polyelectrolyte end-capped with
short polystyrene chains. Synthesis and rheological behavior. Macromolecules. 38 (2000),
[30] Zhang L, Eisenberg A. Multiple morphologies and characteristics of “Crew-Cut ” micelle-like
aggregates of polystyrene-b-poly ( acrylic acid ) diblock copolymers in aqueous solutions.
Macromolecules. 24 (1996), 3168–81.
[31] Zhang L, Eisenberg A. Multiple morphologies of “crew-cut” aggregates of polystyrene-b-
poly(acrylic acid) block copolymers. Science. 268 (1995), 1728–31.
[32] Antonietti M, Tauer K. 90 Years of polymer latexes and heterophase polymerization: more vital
than ever. Macromol. Chem. Phys. 204 (2003), 207–19.
[33] Harkins WD. A general theory of the mechanism of emulsion polymerization. J. Am. Chem. Soc.
69 (1947), 1428–44.
2 A primer on polymer colloids: structure, synthesis and colloidal stability | 35

[34] Gilbert RG. Emulsion polymerization: a mechanistic approach. Academic Press, Oxford, UK,
[35] Ugelstad J, El-Aasser MS, Vanderhoff JW. Emulsion polymerization: initiation of polymerization
in monomer droplets. J. Polym. Sci.: Polym. Lett. 11 (1973), 503–13.
[36] Asua JM. Miniemulsion polymerization. Prog. Polym. Sci. 27 (2002), 1283–346
[37] Schork FJ, Luo Y, Smulders W, Russum JP, Butté A, Fontenot K. Miniemulsion polymerization.
Adv. Polym. Sci. 175 (2005), 129–255.
[38] Crespy D, Landfester K. Miniemulsion polymerization as a versatile tool for the synthesis of
functionalized polymers. Beilstein. J. Org. Chem. 6 (2010), 1132–48.
[39] Landfester K, Willert M, Antonietti M. Preparation of polymer particles in nonaqueous direct
and inverse miniemulsions. Macromolecules. 33 (2000), 2370–6.
[40] Huang C-Y, Chen C-M, Lee Y-D. Synthesis of high loading and encapsulation efficient paclitaxel-
loaded poly(n-butyl cyanoacrylate) nanoparticles via miniemulsion. Int. J. Pharm. 338 (2007),
[41] Zhang M, Cabane E, Claverie J. Transparent antifouling coatings via nanoencapsulation of a
biocide. J. Appl. Polym. Sci. 105 (2007), 3824–33.
[42] Tiarks F, Landfester K, Antonietti M. Preparation of polymeric nanocapsules by miniemulsion
polymerization. Langmuir. 17 (2001), 908–18.
[43] Cheng S, Ting SRS, Lucien FP, Zetterlund PB. Size-tunable nanoparticle synthesis by RAFT poly-
merization in CO2 -induced miniemulsions. Macromolecules. 45 (2012), 1803–10.
[44] Capek I. Microemulsion polymerization of styrene in the presence of anionic emulsifier. Adv.
Colloid Interface Sci. 82 (1999), 253–73.
[45] Chow PY. Microemulsion polymerizations and reactions. Adv. Polym. Sci. 175 (2005), 257–98.
[46] Smeets NMB, McKenna TFL. The synthesis of translucent polymer nanolatexes via microemul-
sion polymerization. J. Colloid Interface Sci. 383 (2012), 28–35.
[47] Holtzscherer C, Candau F. Application of the cohesive energy ratio concept (CER) to the forma-
tion of polymerizable microemulsions. Colloids Surf. 29 (1988), 411-23.
[48] Baglioni M, Rengstl D, Berti D, Bonini M, Giorgi R, Baglioni P. Removal of acrylic coatings
from works of art by means of nanofluids: understanding the mechanism at the nanoscale.
Nanoscale. 2 (2010), 1723–32.
[49] Hanley KJ, Lodge TP, Huang C-I. Phase behavior of a block copolymer in solvents of varying
selectivity. Macromolecules. 33 (2000), 5918–31.
[50] Lyklema J. Fundamentals of interface and colloid science / Volume IV, Particulate colloids.
Academic Press, San Diego, USA, 2005.
[51] Derjaguin BV, Landau LD. No Title. Acta Physicochim USSR. 14 (1941), 633.
[52] Verwey E, Overbeek JT. Theory of the stability of lyophobic colloids. Dover Publications, Dover
USA, 1999.
[53] Napper DH. Polymeric stabilization of colloidal dispersions. Academic Press, London, UK,
[54] Nozary S, Modarress H, Eliassi A. Cloud-point measurements for salt + poly(ethylene gly-
col) + water systems by viscometry and laser beam scattering methods. J. Appl. Polym. Sci.
89 (2003), 1983–90.
[55] Hiemenz PC, Raj Rajagopalan. Principles of colloid and surface chemistry, Third edition. Ta-
ble 10.4, p 486. Marcel Dekker, New York, USA, 1997.
[56] Fritz G, Schädler V, Willenbacher N, Wagner NJ. Electrosteric stabilization of colloidal disper-
sions. Langmuir. 18 (2002), 6381–90.
[57] Feigin RI, Napper DH. Depletion stabilization and depletion flocculation. J. Colloid Interface
Sci. 75 (1980), 525–41.
36 | Part I Functional materials: Synthesis and applications

[58] Lekkerkerker H, Tuinier R. Colloids and the depletion interaction. Springer, Dorchester, UK,
[59] Tohver V, Smay JE, Braem a, Braun PV, Lewis J a. Nanoparticle halos: a new colloid stabilization
mechanism. Proc. Natl. Acad. Sci. 98, 8950–4, USA, 2001.
S. Rondeau-Gagné and J.-F. Morin
3 Synthesis, functionalization and properties of
fullerenes and graphene materials
3.1 Introduction

Carbon-based materials are undoubtedly amongst the most studied materials in both
academic and industrial laboratories since the discovery of fullerenes in 1985 [1]. Prior
to this discovery, three carbon allotropes were known, namely amorphous carbon,
graphite, and diamond. These rather abundant naturally occurring materials have
been used by human kind for centuries for different applications, ranging from energy
production to fabrication of luxury goods. Although very useful, these carbon mate-
rials were perceived as low-tech and not much research regarding their physical or
chemical properties was undertaken. However, everything changed very rapidly with
the discovery of fullerene in 1985 and carbon nanotubes in 1991 [2], and a renaissance
of carbon materials took place in many scientific areas. These new carbon allotropes
showed very interesting properties, especially electronic ones, due to their delocalized
π-electron system. While the research and development related to fullerenes and car-
bon nanotubes were at their prime, graphene, a layer of graphite one atom thick, was
isolated for the first time in 2004 [3], giving new reasons for scientists to be excited
about new forms of carbon materials. With such a diversity of structures, a plethora
of applications can be envisioned for these materials, including electronic, biomed-
ical, energy, aerospace, water treatment, and so on. In this chapter, we will discuss
the synthesis and chemical modifications of two of these materials, more particularly
fullerene-C60 and graphene. A brief discussion of their applications in materials sci-
ence is also presented.

3.2 Fullerenes

Discovered in 1985 by Robert Curl, Harold Kroto, and Richard Smalley, fullerenes have
brought about a revolution in materials science, chemistry, and physics [4, 5]. With
their amazing properties, these carbon clusters have stimulated research and incorpo-
rated into different types of devices, often leading to increased efficiency and unique
properties [6, 7]. Among this family, C60 is particularly interesting, since it is symmet-
rical, reactive towards different kinds of reagents, and can be prepared in relatively
large quantities (the most abundant of all fullerenes). Therefore, it is not surprising
that an increasing number of scientific papers on C60 synthesis and characterization
have been published over the past two decades. The chemistry of this new carbon
allotrope is wide-ranging and extends from chemical functionalization to encapsula-
38 | Part I Functional materials: Synthesis and applications

tion of atoms and molecules. In this section, the chemistry of fullerene C60 and the
properties that make it a major building block in modern chemistry will be looked at
in detail.

3.2.1 General considerations

Fullerenes are a new family of compounds consisting entirely of sp2 -hybridized car-
bon atoms. From a structural point of view, fullerene C60 is perfectly spherical and
consists of twelve pentagons and twenty hexagons. However, depending on their
size, fullerenes can adopt different shapes. For example, the shape of C70 is ovoid
rather than spherical. Also, fullerenes can have from twenty to a hundred carbon
atoms [8–10]. The first mention of fullerenes in the literature was by Osawa in 1970, as
he predicted that the corannulene was, in fact, a sub-structure of C60 [11]. However,
it was only fifteen years later that his hypothesis was confirmed by Curl, Kroto, and
Smalley in an experiment aiming at “[. . . ] understanding the mechanism by which long-
chain carbon molecules are formed in interstellar space and circumstellar shells” [1].
By vaporizing a rotating graphite target with a laser beam, fullerenes were produced
and detected by mass spectrometry (peaks at m/z = 720.7 and 841.8 for C60 and C70 ,
respectively). The importance of this discovery for materials science and nanotech-
nology was rewarded by the Nobel committee in 1996 [12].

Fig. 3.1. C60 (a) and its Schlegel re-

(a) (b) presentation (b).

Although its single and double bonds alternate, C60 is not aromatic. Despite the pres-
ence of 32 aromatic rings, C60 has 60 π-electrons, and therefore does not meet the
criteria of the Hückel rule: 2(N + 1)2 with N = 0, 1, 2, 3, . . . , for three-dimensional
molecules [13]. As an indication of this anti-aromatic character, C–C bond junctions
[6 : 6] are slightly smaller (1.40 Å) than the C–C bonds of the [5 : 6] junctions (1.45 Å),
meaning that the [6 : 6] junctions and [5 : 6] junctions have double and single bond
character, respectively. The van der Waals diameter of C60 is 1.1 nm. Table 3.1 summa-
rizes some interesting properties of this carbon allotrope [14].
3 Synthesis, functionalization and properties of fullerenes and graphene materials | 39

Table 3.1. Some interesting properties of C60 .

Inner diameter 3.48 Å

Crystalline network FCC
Density 1.72 g/cm
Thermal conductivity 0.4 W/mK
Electronic affinity 2.65 eV
1st ionization potential 7.58 eV
Band gap 1.7 eV
Solubility o-DCB (25°C) 27 mg/mL
Solubility CS2 (25°C) 12 mg/mL
Solubility toluene (25°C) 3.2 mg/mL

3.2.2 Synthesis and purification of fullerenes

Synthesis of fullerenes can be achieved using various techniques and processes. Gen-
erally, these techniques require sublimation of simple carbon feedstock such as gas
(methane, ethylene, and acetylene), solvents (methanol, ethanol, and hydrocarbons)
or solid (graphite) at high temperature [15]. The formation mechanism of fullerenes is
still a source of debate among the scientific community, but the most plausible hypoth-
esis seems to be nucleation of carbon particles [16]. Production of fullerenes always
results in the formation of several types of fullerenes, which differ in size. However,
C60 and C70 are the two major isomers formed after pyrolysis. Interestingly, it is possi-
ble to control the C60 /C70 ratio by changing various synthetic parameters. Nowadays,
the most commonly used production route for the large-scale synthesis of fullerene is
the flame method. This method requires installations allowing control of multiple pa-
rameters, such as pressure, flame temperature, concentration of carrier, gas and oxy-
gen/carbon ratio. The flame production method consists of the sublimation of carbon
rods with a flame fed with benzene or acetylene. On sublimation, carbon grows around
a center of nucleation and forms different types of fullerenes within the residual soot
from which they can then be isolated. For best results, it is important to introduce a
certain amount of carrier gas such as H2 or O2 . In the best production conditions, the
mass percentage of fullerene collected is about 20%. It is worth mentioning that some
companies have been using this method to produce a total of 40 metric tons of C60
per year [17].
Following the production of fullerenes, a problem arises: purification. Indeed,
given that the fullerenes obtained have very similar polarity, size, and low solubility
in most common solvents, purification so as to obtain pure fullerene clusters is very
difficult. Several purification methods, such as high-performance liquid chromatog-
raphy (HPLC), have been developed in order to isolate the fullerene C60 [18, 19]. This
technique is the most frequently used for the purification of different-sized fullerenes.
It is also the only known method for separation of larger fullerenes (> C76 ). However,
this method is very expensive when hundreds of milligrams or even grams of mate-
40 | Part I Functional materials: Synthesis and applications

C60, C70, C>76 C70 DBU + C60 (99,5%)
C>76 DBU

Fig. 3.2. Complexation of C≥70 with DBU.

rial are needed. In addition, modern chromatography columns cannot separate more
than one gram per hour, which makes it a time-consuming method. Thus, purification
of large quantities of fullerenes is very difficult. To overcome this problem, a tech-
nique using a chemical agent, 1,8-Diazabicyclo[5.4.0]undec-7-ene (DBU), has been de-
veloped, enabling the isolation of pure C60 with excellent yields [20]. This technique
relies on complexation between large fullerenes (≥C70 ) with low reduction potential
and DBU, leading to a precipitation of the C≥70 ⋅ DBU complex and leaving the C60 in
solution. Figure 3.2 presents the general complexation of C≥70 with DBU.

3.2.3 Chemical and physical properties of C60

Because of its structure, its incredible electronic and electrochemical properties, and
its specific reactivity, C60 is very attractive to the scientific community. Its most remark-
able property is undoubtedly its triply degenerated LUMO orbital, which can accom-
modate up to six electrons upon electrochemical reduction [21]. This property makes
C60 a unique electron acceptor for different applications, especially in electronics.
Thus, as shown in Fig. 3.3, six reversible redox processes can be observed in controlled
electrochemical experiments (cyclic voltammetry) [22]. It is important to mention that
C70 has a doubly degenerated LUMO orbital which can accept up to four electrons.
However, the level of the nondegenerated LUMO + 1 is sufficiently close in energy to
accept two additional electrons, which gives C70 similar electronic properties to those
of C60 .
This remarkable electron affinity is still not fully understood and continues to be
investigated by the scientific community. Studies on bowl-shaped subunits of C60 are
particularly useful and may help in gaining insights into the electronic properties of
fullerenes [23]. The study of fullerene subunits also allows more understanding of
their reactivity. As a simple model, the fullerene C60 can be considered a collection
of pyracyclene units, as shown in Fig. 3.4. The pyracyclene unit is anti-aromatic (4n π-
electrons). By accepting two electrons, anionic units are formed, each having a 4n +2
electrons structure which leads to aromaticity and thereby enhances stability. The for-
mation of cyclopentadienide units is thought to give to C60 its “n-type” character. The
3 Synthesis, functionalization and properties of fullerenes and graphene materials | 41

c60 at –10°c

10 µA

5 µA

–1.0 –2.0 –3.0

Potential (Volts vs Fc/Fc+)

Fig. 3.3. Electrochemical reduction of C60 as observed by cyclic voltammetry (CV) [22].

formation mechanism of cyclopentadienide via direct electron transfer or nucleophilic

reaction is shown in Fig. 3.4 [24].
Since the electronic properties of fullerenes are those most exploited, several stud-
ies have been conducted to control and modify them, especially the LUMO energy level
of C60 , in order to improve its performance in various types of electronic devices. This
is particularly important in solar cells, in which C60 derivatives are used as electron-
accepting units [24]. Modulation of the LUMO energy level of C60 can be accomplished

+2 e–



Fig. 3.4. Formation of cyclopentadienide unit via direct electron transfer or nucleophilic addition.
42 | Part I Functional materials: Synthesis and applications

through various covalent and supramolecular functionalizations [25–30]. Generally,

the electronic level of the orbitals in a fullerene can be modified by incorporating dif-
ferent electron-accepting or electron-withdrawing groups. A strong electron-donating
group increases the electron density of C60 , which thereby increases the LUMO energy
level. An electron-accepting substituent has exactly the opposite effect. In addition,
a through-space electronic phenomenon called periconjugation also allows modula-
tion of the electronic properties of the fullerene [31]. A through-space electron transfer
is possible by an orbital overlapping between a substituent and C60 , resulting in an in-
crease or decrease of the electron density of C60 . Thus, the direct consequence of this
change in electron density is the modulation of electronic levels.

3.2.4 Chemical functionalization of C60

Since the discovery of fullerenes several studies have been conducted to improve their
solubility in common solvents. As mentioned previously, a second objective of the
functionalization of fullerenes is modulation of their properties, mainly the electronic
properties. In order to make these changes, chemists have attempted to use conven-
tional tools of organic chemistry. However, functionalization does not proceed with
the same ease on every fullerene. In fact, fullerenes of different sizes have different
reactivity toward organic reactions, and C60 is one of the most reactive fullerenes. Re-
action on its highly curved structure promotes the release of cyclic strain. Moreover,
as a result of being perfectly spherical and symmetric, it is possible to perform regios-
elective reactions on the fullerene cage, which greatly facilitate product purification.
The main functionalization methods are [3 + 2] and [4 + 2] cycloaddition reactions,
as well as nucleophilic reactions.
One of the most useful functionalization reactions on C60 is the [3 + 2] cycloaddi-
tion. The most famous reaction is the Prato reaction [32, 33]. Using C60 in the presence
of an amino acid, usually a sarcosine derivative and an aldehyde, cycloaddition takes
place at a [6,6] junction of the fullerene core, resulting in the formation of a fulleropyr-
rolidine. The general Prato reaction and its mechanism are shown in Fig. 3.5.
The Prato reaction is a very useful reaction for the formation of several deriva-
tives because it has important advantages over other methods. Essentially, this reac-
tion does not require the use of toxic reagents, catalysts or unstable organometallic
reagents. Some side products may be formed during the reaction, which usually leads
to a 60–80% conversion of C60 into fulleropyrrolidine (30 to 40% yield overall). Prop-
erties of the parent fullerenes can be modulated through two side groups, which can be
of different chemical nature. This reaction is mainly used in the synthesis of function-
alized polymer [34], donor-acceptor complexes [35], functionalization of peptides [36],
and synthesis of water-soluble derivatives of C60 [37]. It is noteworthy that fulleropy-
rrolidines can also be obtained by thermal ring opening of aziridine [38].
3 Synthesis, functionalization and properties of fullerenes and graphene materials | 43

N R1

H R1 2) C60, toluene, reflux

N R1

H R1 H R1
N + Cycloaddition [3+2]
H R2 N H N
O H R1 + O R2 +

Fig. 3.5. The Prato reaction.

Another widely used type of fullerene functionalization is the [4 + 2] cycloaddition.

Such functionalization can be carried out via three main methods: by generating a
nucleophilic carbon from a α-halo ester, by the thermal addition of diazo compounds
followed by thermolysis (or photolysis), or by a carbene addition. One of these three
methods is particularly interesting, and is called the Bingel reaction [39]. This reaction
requires the use of a halogenated diester which, in the presence of a strong base and
C60 , forms a fullerene derivative bearing a cyclopropane moiety. This type of derivative
is called methanofullerene. The general reaction is shown in Fig. 3.6.

O O NaH, C60

RO OR Toluene, reflux

Fig. 3.6. The Bingel reaction.

Functionalization of C60 by this method has several advantages. First, given that the
synthesis of diester compounds is relatively easy, the reaction is very versatile and
can be applied to a variety of compounds. Another advantage is that the ester groups
can easily be post-functionalized to form more complex architectures. In addition, the
synthetic yields are generally very good (50% or more), and this reaction occurs regios-
44 | Part I Functional materials: Synthesis and applications

electively at the [6,6] junctions of the fullerene core. It is noteworthy that the α-halo
ester may be formed in situ from a diester and a dihalogen (Br2 or I2 ) in the presence
of DBU.
Nucleophilic addition is also another very interesting method of C60 functional-
ization [40–43]. This reaction is much less common in the literature and is often used
strictly for synthetic ease. Nonetheless, this method allows the preparation of several
types of useful derivatives whose properties and functions depend on the atom di-
rectly attached to C60 . In most cases, the nucleophilic species is an acetylenic or ben-
zylic carbon, but it can also be a heteroatom such as oxygen or nitrogen. An example
of a nucleophilic substitution reaction on C60 with an acetylide derivative is shown in
Fig. 3.7.

R Li E
R R2

Fig. 3.7. Nucleophilic substitution of an acetylenic carbon on C60 .

Usually, the nucleophilic moiety is formed on a terminal alkyne in the presence of a

strong base, such as the non-nucleophilic lithium hexamethyldisilylamide (LHMDS),
or n-butyllithium. The resulting carbanion generated can be added to C60 , thus trans-
ferring the negative charge onto the cage, more specifically onto the adjacent carbon
of the newly formed bond. An electrophilic species must then be added to quench
this anion. The electrophilic species can be of various natures such as a proton, an
alkyl chain, a benzoyl group, etc. Similar to functionalization methods previously pre-
sented, the nucleophilic addition has several advantages. First, this method allows a
wide range of derivatives from readily available terminal alkyne derivatives to be ob-
tained. Given that two different side groups (nucleophiles and electrophiles) can be
added to C60 in a single reaction, the number of possible combinations is astonishing.
Nucleophilic addition also allows the synthesis of derivatives with very good yields
(40–60%). Purification of the reaction media is also simplified, as few side products
are formed during the reaction. There is, however, the possibility that polyaddition
products will be formed. This side reaction can be minimized by controlling the re-
action temperature and the number of equivalents of acetylide in the reaction media.
The reaction is also very difficult to control, and a difference of one minute between
various additions of reagents may significantly affect the reaction yield.
Although nucleophilic addition is a promising class of functionalization reac-
tions for fullerenes, another type of reaction leading to water-soluble polysubstituted
derivatives has proven to be very useful for biological applications [44, 45]. Experi-
3 Synthesis, functionalization and properties of fullerenes and graphene materials | 45

mental conditions are found to be relatively harsher than in other types of functional-
ization, but the products obtained have interesting properties and specific reactivity.
In a first step, fullerene is halogenated using high-pressure F2 gas for a long period
of time. Thereafter, by adding a nucleophile and an organic catalyst, it is possible
to carry out nucleophilic substitution on the fullerene cage. However, the rules of
nucleophilic substitution are somewhat changed with a fullerene. In fact, given the
shape of the C60 cage, the normal SN 2 reaction is geometrically impossible. Thus, the
reaction mechanism is still controversial and many studies have been conducted to
determine it precisely. Since the reaction requires long exposure to a gaseous halo-
gen, few manipulations are necessary. It is also possible to introduce a wide variety
of nucleophilic species, which allows several types of derivatives with completely
different properties, depending on the nucleophile chemical nature, to be obtained.
However, the exact number of substituents inserted onto the fullerene cage is very
difficult to control and may vary significantly, which is a major drawback when pure
materials are needed, as is the case for biological applications. Currently, the best
estimate of substitution ratio comes from the study by nuclear magnetic resonance
(NMR) spectroscopy.
Encapsulation of various metals and molecules in the C60 core is also an in-
creasingly popular functionalization method. This technique leads to endohedral
species, called endofullerenes, with interesting properties and rather unusual struc-
tures [46, 47]. Generally used with higher fullerenes (C80 and above), encapsulation
of small molecules and metals is also feasible with C60 [48]. By opening the C60 cage
using laser or chemicals, it is possible to introduce small molecules such as H2 or
He. It is also possible to introduce lanthanides and metals. These new endohedric
fullerenes have a very interesting advantage. Since the introduction of molecules into
the core does not alter the electronic structure, there is no loss of π electrons and,
thereby, the electronic properties of pristine C60 are conserved [49]. This last point is
the main reason why formation of endofullerenes is an increasingly popular avenue
for fullerene functionalization, especially in organic electronics.

3.2.5 Applications

As shown previously, fullerenes have a number of interesting properties. It is there-

fore not surprising to see fullerenes in several complex architectures and in a variety
of applications ranging from organic electronics to biotechnology. We present below
some of the most interesting applications for which fullerenes are becoming important
building blocks and components.
Given its strong n-type character, C60 is the building block of choice in organic
electronics, particularly in organic photovoltaics [50, 51]. In a bulk heterojunction so-
lar cell (Fig. 3.8), a blend made of a p-type, light-harvesting polymer (electron-donor)
and a methanofullerene, the C61 -phenylbutyricmethyl ester (PCBM), which serves as
46 | Part I Functional materials: Synthesis and applications

electron-acceptor [52–54]. When the polymer absorbs a photon from the solar spec-
trum, an exciton is formed. This exciton can migrate into the polymer-PCBM blend
by means of various photophysical processes and thereby generate an electric cur-
rent. This type of solar cell is increasingly used, and their conversion efficiency is con-
stantly increasing. In the field of organic photovoltaics, C60 is also increasingly used
as an electron acceptor in artificial photosynthesis systems [55, 56]. These types of
system are studied to gain insights into the photophysical processes of charge trans-
fer/separation and energy generation. Understanding these phenomena makes the
creation of more efficient and green devices to collect solar energy for generating elec-
tricity possible.

– +


Plastic foil


(a) PCBM (b) PC61BM

Fig. 3.8. (a) Architecture of a bulk heterojunction solar cell and (b) structure of the PCBM
methanofullerene commonly used as electron acceptor.

The fullerene C60 is also widely used in molecular electronics. Given its ability to ac-
cept multiple electrons reversibly, and its high stability versus redox processes, C60 is
a promising building block for the fabrication of monomolecular transistors and the
construction of flash memory [57]. Thus, it has been demonstrated that self-assembled
monolayers containing ethynyl-bridged C60 can be formed on gold (Fig. 3.9). These
monolayers can store two charges in a stable way, making them excellent surface ca-
pacitors, which paved the way to surface molecular electronics.
Fullerenes are also found in rather unusual applications in nanoscience. These
new applications exploit several structural properties of fullerenes, such as their
spherical structure. Due to this spherical structure, C60 is often used as a “wheel” in
nanocars and nanovehicles (Fig. 3.10) [58, 59]. By creating different models of these
nanocars and studying their behavior on a gold surface with a scanning tunneling mi-
croscope (STM), it was recently demonstrated that fullerenes do not slide but literally
roll over the surface. This result helps gain understanding of molecular movement
and how to control it in order to further develop useful nanomachines.
3 Synthesis, functionalization and properties of fullerenes and graphene materials | 47




Orbital energy (eV)

H3C H3C H3C H3C H3C H 3C –3.2
SH –5.6

2 3 4 5 6 C60 3 5 1 2 4 6

Fig. 3.9. Fullerene/thiol-terminated molecules used in flash memory devices [57].

OC10H21 OC10H21

H21C10O H21C10O H
H21C10O OC10H21 OC10H21

H21C10O OC10H21 2

H OC10H21 OC10H21 H21C10O

H 2
H H21C10O
H21C10O H21C10O 2
OC10H21 H

Fig. 3.10. Some examples of C60 -containing nanocars.

Finally, applications have even been found in medicine and pharmacology for full-
erene C60 [60, 61]. Despite significant solubility problems often resulting in accumula-
tion in the kidneys, C60 and its water-soluble derivatives can be used in medicine for
the treatment of particular diseases such as AIDS [62], in medical imaging [63], and as
an antioxidant [64]. In the case of diseases like AIDS, it has been shown that C60 can
cause inhibition of the protease enzyme, which significantly reduces the replication
rate of these viruses. In medical imagery, endohedral derivatives represent an inter-
esting avenue, due to the possibility of including heavy ions such as Gd, currently
used in magnetic resonance imaging (MRI), in their core. C60 has also demonstrated
interesting antioxidant properties and cancer cytotoxicity, mainly due to its ability to
generate singlet oxygen.

3.3 Graphene

Unlike carbon nanotubes and fullerenes, graphene, an all-carbon 2D polymer, is a nat-

urally abundant material, which makes it very attractive and has prompted a rapid
48 | Part I Functional materials: Synthesis and applications

technology market penetration [65–67]. However, graphene does not exist as such,
but rather as stacks of sheets commonly known as graphite. The molecular structure
of graphite is represented in Fig. 3.11. Graphene exists in different forms, including
one-atom thick layer, double layers, and several-layers, and each of them has par-
ticular optical and electronic properties which will not be discussed here. In some
ways, a one-atom thick graphene sheet can be seen as an unzipped carbon nanotube,
although the “graphene nanoribbon (GNR)” would be more appropriate to describe
this material in this case, since the unzipping of a carbon nanotube would lead to a
graphene sheet with a finite width.


0.34 nm

Fig. 3.11. Molecular structure of graphite.

Defect-free graphene is a semimetal, zero gap semiconducting material owing to the

alternation of single and double carbon bonds which allow electron delocalization
all along the structure of the graphene sheet [68]. The high value of electron mobil-
ity measured at room temperature (∼ 15 000 cm2 V−1 s−1 ) is a few orders of magnitude
higher than silicon, making graphene highly valuable for electronic applications such
as transparent conducting electrodes, energy conversion, circuitry, transistors, capac-
itors, and so on [69].
In order to be isolated, graphene sheets must be separated from each other
through exfoliation. This process is not simple, since rather strong van der Waals
interactions between the graphene sheets have to be broken, which requires quite a
large amount of energy. Graphite can be exfoliated using mechanical methods, such
as the famous “scotch-tape” method [3], or by using solvents whose surface energies
match that of graphene [70]. However, none of the methods of producing graphene
from graphite yield large amount of materials, thus limiting the scope of this strat-
egy for many applications. Another way of producing graphene is to prepare it from
the bottom-up using reactive carbon species as feedstock. The so-called “physical
3 Synthesis, functionalization and properties of fullerenes and graphene materials | 49

method” can yield very high quality graphene with tunable properties such as size,
shape, and physical properties, although some issues need to be addressed regarding
purity and batch-to-batch variability. In order to make highly pure nanographenes
(NG) and nanoribbons (GNR), an “all-organic” approach involving synthetic organic
chemistry can be used.
In the following section, we will briefly describe the more common methods, both
“physical” and “all-organic” approaches, for the production of graphene and GNR.
We will also present some of the most useful strategies to functionalize them cova-
lently before we end this chapter by briefly presenting different areas of application
for graphene.

3.3.1 Production of graphene

Physical methods
One of the most popular methods of producing graphene from carbonaceous feed-
stock is chemical vapour deposition (CVD) [71]. In this technique, graphene is pro-
duced when a transition metal surface is heated to 750°C and higher, depending on
the gas, and exposed to hot hydrocarbon gases such as methane or acetylene. Using
the CVD method, large sheets of highly pure graphene, mostly single-layer, can be
produced. When proper metallic substrate is used, the graphene sheets can be trans-
ferred to other substrates such as silicon for electronic applications. When liquid pre-
cursors such as hexane are used as the carbon feedstock on a polycrystalline copper
foil, graphene sheets of a few centimeters can be prepared and transferred easily to
other substrates [72].
Inspired by the production of carbon nanotubes, researchers used the arc dis-
charge method to produce graphene. This method of production has the advantage
of repairing graphene while it is formed, in addition to decreasing significantly
the amount of amorphous carbon produced during the process. In this method, a
graphite-based anode and cathode are exposed to a 100-ampere current, leading
to a discharge which produces highly reactive carbon species. This process can be
achieved under hydrogen (H2 ) [73] or air [74] to produce graphene in optimized con-
Another way of producing graphene is to use epitaxial growth of metal or silicon
carbide (SiC) substrate. In the latter technique, SiC is heated to an elevated tempera-
ture (> 1 200°C) under vacuum in order to displace the Si atoms, whose sublimation
rate is higher than carbon, thus concentrating carbon atoms at the surface. The car-
bon atoms rearrange at high temperature to form uniform sheets of graphene [75].
Similarly, metallic surfaces can be use as a seed for the growth of graphene of quality,
allowing the study of its intrinsic physical properties [76].
Important methods for the production of graphene do not involve the transfor-
mation of a carbon feedstock, but rather the transformation of naturally occurring
50 | Part I Functional materials: Synthesis and applications

graphite. When graphite is used as a starting material, the challenge is to break the van
der Waals interactions between the graphene sheets in order to individualize them.
One of the most efficient techniques is the oxidation of graphite to form graphene ox-
ide, followed by a reduction process to produce individualized graphene sheets [77].
This method allows the low-cost production of a very large quantity of graphene at
once, although the quality of the final material is far from that obtained by other phys-
ical methods. Basically, natural graphite is heated in a strong oxidizing acidic mixture
to produce exfoliated sheets of graphene oxide (GO) which can be solubilized or sus-
pended in different solvents, including water (Fig. 3.12; [78]). Then, GO can be reduced
using thermal annealing or a chemical reducing agent such as hydrazine, NaBH4, vita-
min C or pyrogallol to produce graphene. The quality of the final material depends on
the conditions used, but usually a significant amount of oxygen is still present after the
reduction process. Thus, this production method can be used for a limited number of
potential applications and is unlikely for the production of electronic-grade graphene.
For this reason, discussion of this technique will not be extended further here. How-
ever, GO is a very useful intermediate in the covalent functionalization of graphene




Graphene oxide Fig. 3.12. Preparation of GO from graphite [78].

3 Synthesis, functionalization and properties of fullerenes and graphene materials | 51

Another method which uses graphite as the starting material is exfoliation. The
greatest advantage of this technique over the oxidation/reduction process described
above is that graphene sheets are not chemically transformed, so the sp2 carbon atom
network is not disrupted, thus keeping the band structure and the electronic prop-
erties intact. In this regard, exfoliated graphite opens the way to large-scale produc-
tion of electronic-grade graphene without a complex production setup. The strategy
behind this technique is to use a stabilizer which will force the graphene sheets to
separate by forming van der Waals interactions that are at least as strong as the inter-
sheet interactions. Hence, stabilizers which show a high affinity for graphene have to
be used. Among the most efficient are stabilizers with π-conjugated moieties such as
perylene [79] and tetracyanoquinodimethane [80], and a polar group which enables
dispersion in a polar solvent. More classical surfactants such as sodium dodecylben-
zene sulfonate (SDBS) also give good results [81].




Stabilized Graphene

Fig. 3.13. Dispersion of graphene sheets through exfoliation.

Solvent molecules can also be used to exfoliate graphite efficiently. Solvents such
as ortho-dichlorobenzene (ODCB) [82], N-methylpyrrolidone (NMP) [70], and perflu-
orinated solvents [83] are examples of solvents which have been used to disperse
graphene sheets. An external source of energy such as heat or ultrasound is often
required to make this process work.

All-organic methods
Although graphene with a high level of purity can be obtained via the above-men-
tioned physical methods, none of them allow perfect control over the size and shape
of the resulting sheets [84–86]. For this reason, chemists and materials scientists have
been attempting for decades to produce well-defined graphene nanosheets with cus-
tomized properties using of organic chemistry, although research still needs to be done
52 | Part I Functional materials: Synthesis and applications

in this area to make the whole process more efficient in terms of time and quantity of
materials produced.
One way of producing graphene nanosheets or nanoribbons is by preparing a
polyphenylene-type dendrimer or polymer which is rigidified by intermolecular cross-
linking reactions known as Scholes reactions, as shown in Fig. 3.14 [87–90]. In this
case, owing to its finite width, the resulting GNR is a semiconductor. Besides, because
alkyl chains can be installed into the periphery of the nanoribbons, the resulting ma-
terials are soluble in common organic solvent, thus enabling their characterization.
This good solubility also allows the semiconducting GNR to be processed from solu-
tion to form thin films on various kinds of substrates, opening the way for the use
of GNRs in organic electronics. Many other types of nanoribbons with various band
gap and carbon atom configuration have been synthesized using this intramolecular
cross-linking approach.
In addition to nanoribbons, finite graphene nanosheets can be constructed using
the intramolecular cross-linking approach (Fig. 3.15; [91]). Again, the band gap of these
semiconducting materials can be directly modulated by the size of the nanosheets.
The addition of functional groups such as ester, alkyne, ether, triarylamine, aryl, and
transition metals to the periphery of graphene nanosheets is another efficient strategy
for modulating the electronic properties [88].

3.3.2 Graphene in energy conversion devices

One of the most promising applications of graphene in electronic devices is its use
as transparent electrode to eventually replace commonly used semiconductors such
as ITO, which suffers from significant disadvantages: rarity of indium, chemical sen-
sitivity, and poor mechanical properties [92, 93]. The high transparency (97.7% for a
single-layer sheet), high electron mobility, excellent chemical stability, and mechani-
cal properties of graphene make it a “near-perfect” material for electrode fabrication.
The applications that would benefit the most from all the qualities of graphene are
undoubtedly organic solar cells (OSCs) and light-emitting diodes (LEDs), in which
transparent electrodes are necessary to let the light in to and out of the devices. For
example, Peumans and coworkers replaced ITO by solution-processed graphene in
P3HT-based bulk heterojunction (BHJ) solar cells and observed that the devices suf-
fered only a slight decrease in conversion efficiency compared to ITO-based devices
(η = 0.4% and 0.8% for graphene and ITO, respectively; [94]). The efficiency can be
increased (η = 1.2%) when CVD graphene is used, since the sheet resistance is lower
than for solution-processed graphene [95]. As for BHJ solar cells, graphene can also be
used as different components of DSSCs. More specifically, graphene can be used as a
transparent counter electrode to replace the expensive Pt/ITO often used to inject elec-
trons into the electrolytes. When redox couples other than I−3 /I2 are used, graphene
counter electrodes show higher catalytic activity than Pt/ITO [96].
3 Synthesis, functionalization and properties of fullerenes and graphene materials | 53

a b
Br Br + (HO)2B Br Br x x
2 3 4 x= TMS
5 x= H


x y

x y

Fig. 3.14. Preparation of a graphene nanoribbon through an all-organic process [90].

Besides electrode fabrication, graphene can be used as electron acceptor (or n-

type materials) in BHJ solar cells, although its transparency might be an important
drawback to consider. By replacing PCBM with graphene in a blend with poly(3-
octylthiophene) (P3OT), efficiency of η = 1.4% was obtained [97]. Dai and coworkers
showed that covalent functionalization of graphene with C60 led to an electron accep-
tor with improved properties compared to C60 or graphene alone [98].
Graphene oxide can also be used as hole-transporting material (HTL) on top of
the ITO electrode in BHJ solar cells. In addition, to change the work function of ITO
to better match that of active materials, GO protects the surface of ITO from chemi-
cal deterioration due to ambient conditions [93]. In a P3HT-based device, efficiency of
up to η = 3.5% was measured, which is almost the same as that measured for a similar
device using PEDOT:PSS as the HTL material [99].
54 | Part I Functional materials: Synthesis and applications

1 2 3

–12 H –56 H –106 H

4 5 6

Fig. 3.15. Preparation of a graphene nanosheet through an all-organic process [91].


Although carbon nanomaterials have their place in a huge number of different tech-
nological applications, one can expect to witness even more research activities in
this area in the years to come. Some challenges still need to be overcome in order
for these nanomaterials to be widely used in commodity devices. Purity, uniformity,
batch-to-batch consistency, processability, and price are important issues which must
be addressed in the near future to go beyond the simple academic object of study.
Moreover, real efforts have to be made to better control their properties by develop-
ing synthetic methods which will allow controlled introduction of chemical functions.
This could give rise to new electronic and optical properties from which unique appli-
cations could be developed.


[1] Kroto, H. W.; Heath, J. R.; O’brien, S. C.; Curl, R. F.; Smalley, R. E. C60 : Buckminsterfullerene.
Nature 318, 1985, 162–163.
[2] Ijima, S. Helical microtubules of graphitic carbon. Nature 354, 1991, 56–58.
[3] Novoselov, K. S.; Geim, A. K.; Morozov, S. V.; Jiang, D.; Zhang, Y.; Dubonos, S. V.; Grigorieva,
I. V.; Firsov, A. A. Electric field effect in atomically thin carbon films. Science 306, 2004,
3 Synthesis, functionalization and properties of fullerenes and graphene materials | 55

[4] Walton, H. W. K. D. R. M. The fullerenes, New Horizons for the Chemistry, Physics and Astro-
physics of Carbon; Cambridge University Press, 1993.
[5] Curl, R. F. Dawn of the fullerenes: experiment and conjecture. Mod. Phys. 69, 1997, 691–702.
[6] Wudl, F. Fullerene materials. J. Mater. Chem. 12, 2002, 1959–1963.
[7] Thompson, B. C.; Frechet, J. M. J. Polymer–fullerene composite solar cells. Angew. Chem. Int.
Ed. 47, 2008, 58–77.
[8] Kroto, H.W. Smaller carbon species in the laboratory and space. Int. J. Mass Spectrom. Ions
Process. 138, 1994, 1–15.
[9] van Orden, A.; Saykally, R. J. Small carbon clusters: spectroscopy, structure, and energetics.
Chem. Rev. 98, 1998, 2313–2357.
[10] Prinzbach, H.; Weiler, A.; Landenberger, P.; Wahl, F.; Wörth, J.; Scott, L. T.; Gelmont, M.; Ole-
vano, D.; Issendorff, B.V. Gas-phase production and photoelectron spectroscopy of the small-
est fullerene, C20. Nature, 407, 2000, 60–63.
[11] Osawa, E. Superaromaticity. Chem. Abstr. 74, 1971, 75698.
[12] Kroto, H. W.; Heath, J. R.; Obrien, S. C.; Curl, R. F.; Smalley, R. E. Symmetry, space, stars, and
C60 (Nobel lecture). Nature, 318, 1985, 162–163.
[13] Hirsch, A.; Chen, Z.; Jiao, H. Spherical aromaticity in Ih symmetrical fullerenes: the 2(N + 1)2
rule. Angew. Chem. Int. Ed. 39, 2000, 3915–3917.
[14] Taylor, R. The Chemistry of Fullerenes; World Scientific Publishing: Vol. 4, 1995.
[15] Kratschmer, W.; Lamb, L. D.; Fostiropoulos, K.; Huffman, D. R. C60 : a new form of carbon.
Nature, 347, 1990, 354–358
[16] Howard, J. B.; Lafleur, A. L.; Makarovsky, Y.; Mitra, S.; Pope, C. J.; Yadav, T. K. Fullerenes syn-
thesis in combustion. Carbon, 30, 1992, 1183–1201.
[17] Tremblay, J.-F. Fullerenes by the ton. Chem. Eng. News, 81, 2003, 13–14.
[18] Klute, R. C.; Dorn, H. C.; Mcnair, H. M. HPLC separation of higher (C84+ ) fullerenes. J. Chro-
matogr. Sci. 30, 1992, 438–442.
[19] Kimata, K.; Hirose, T.; Moriuchi, K.; Hosoya, K.; Araki, T.; Tanaka, N. High-capacity station-
ary phases containing heavy atoms for HPLC separation of fullerenes. Anal. Chem. 67, 1995,
[20] Nagata, K. D., E.; Kikuchi, Y. and Hashiguchi, M. Efficient and scalable method for [60]
fullerene separation from a fullerene mixture: selective complexation of fullerenes with DBU
in the presence of water. Org. Process Res. Dev. 9, 2005, 660–662.
[21] Guldi, D. M.; Prato, M. Excited-state properties of C60 fullerene derivatives. Acc. Chem. Res. 33,
2000, 695–703.
[22] Echegoyen, L.; Echegoyen, L. E. Electrochemistry of fullerenes and their derivatives. Acc.
Chem. Res. 31, 1998, 593–601.
[23] Wu, T.-C.; Chen, M.-K.; Lee, Y.-W.; Kuo, M.-Y.; Wu, Y.-T. Fragments of C70 or higher fullerenes:
synthesis, structural analysis, and inversion dynamics. Angew. Chem. Int. Ed. 52, 2013,
[24] Wudl, F. The chemical properties of buckminsterfullerene (C60 ) and the birth and infancy of
fulleroids. Acc. Chem. Res. 25, 1992, 157–161.
[25] Kooistra, F. B.; Knol, J.; Kastenberg, F.; Popescu, L. M.; Verhees, W. J. H.; Kroon, J. M.; Humme-
len, J. C. Increasing the open circuit voltage of bulk-heterojunction solar cells by raising the
LUMO level of the acceptor. Org. Let. 9, 2007, 551–554
[26] Ross, R. B.; Cardona, C. M.; Guldi, D. M.; Sankaranarayanan, S. G.; Reese, M. O.; Kopidakis,
N.; Peet, J.; Walker, B.; Bazan, G. C.; Van Keuren, E.; Holloway, B. C.; Drees, M. Endohedral
fullerenes for organic photovoltaic devices. Nat. Mater. 8, 2009, 208–212.
56 | Part I Functional materials: Synthesis and applications

[27] Rondeau-Gagné, S.; Curutchet, C.; Grenier, F.; Scholes, G.; Morin, J.-F. Synthesis, characteri-
zation and DFT calculations of new ethynyl-bridged C60 derivatives. Tetrahedron, 66, 2010,
[28] Rondeau-Gagné, S.; Lafleur-Lambert, A.; Soldera, A.; Morin, J.-F. Ethynyl-bridged fullerene
derivatives: effect of the secondary group on electronic properties. New J. Chem. 35, 2011,
[29] Lafleur-Lambert, A.; Rondeau-Gagné, S.; Morin, J.-F. Synthesis and characterization of
a new ethynyl-bridged C60 derivative bearing a diketopyrrolopyrrole moiety. Tetrahedron
Let. 52, 2011, 5008–5011.
[30] Backer, S. A.; Sivula, K.; Kavulak, D. F.; Frechet, J. M. J. High efficiency organic photovoltaics in-
corporating a new family of soluble fullerene derivatives. Chem. Mater. 19, 2007, 2927–2929.
[31] Hamasaki, R.; Ito, M.; Lamrani, M.; Mitsuishi, M.; Miyashita, T.; Yamamoto, Y. Nonlinear optical
studies of fullerene–arylethyne hybrids. J. Mater. Chem. 13, 2003, 21–26.
[32] Wudl, F.; Sukuki, T.; Prato, M. Probing the properties of C60 through fulleroids ABC61 . Synth.
Met. 59, 1993, 297–305.
[33] Maggini, M.; Scorrano, G.; Prato, M. Addition of azomethine ylides to C60 : synthesis, char-
acterization, and functionalization of fullerene pyrrolidines. J. Am. Chem. Soc. 115, 1993,
[34] Cravino, A.; Zerza, G.; Neugebauer, H.; Maggini, M.; Bucella, S.; Menna, E.; Svensson, M.;
Andersson, M. R.; Brabec, C. J.; Sariciftci, N. S. Electrochemical and photophysical properties
of a novel polythiophene with pendant fulleropyrrolidine moieties: toward “double cable”
polymers for optoelectronic devices. J. Phys. Chem. B. 106, 2002, 70–76.
[35] Guldi, D. M.; Maggini, M.; Scorrano, G.; Prato, M. Intramolecular electron transfer in
fullerene/ferrocene based donor-bridge-acceptor dyads. J. Am. Chem. Soc. 119, 1997,
[36] Pantarotto, D.; Bianco, A.; Pellarini, F.; Tossi, A.; Giangaspero, A.; Zelezetsky, I.; Briand,
J. P.; Prato, M. Solid-phase synthesis of fullerene-peptides. J. Am. Chem. Soc. 124, 2002,
[37] Da Ros, T.; Prato, M.; Carano, M.; Ceroni, P.; Paolucci, F.; Roffia, S. Enhanced acceptor charac-
ter in fullerene derivatives. Synthesis and electrochemical properties of fulleropyrrolidinium
salts. J. Am. Chem. Soc. 120, 1998, 11645–11648.
[38] Kuwashima, S. Y.; Kubota, M.; Kushida, K.; Ishida, T.; Ohashi, M.; Nogami, T. Synthesis and
structure of nitrene-C60 adduct C60 NPhth (Phth = Phthalimido). Tet. Let. 35, 1994, 4371–4374.
[39] Bingel, K. Cyclopropanierung von Fullerenen. Chem. Ber. 126, 1993, 1957–1959.
[40] Shirai, Y.; Zhao, Y. M.; Cheng, L.; Tour, J. M. Facile synthesis of multifullerene-OPE hybrids via
in situ ethynylation. Org. Let. 6, 2004, 2129–2132.
[41] Anderson, H. L.; Faust, R.; Rubin, Y.; Diederich, F. Fullerene–acetylene hybrids: on the way to
synthetic molecular carbon allotropes. Angew. Chem. Int. Ed. 33, 1994, 1366–1368.
[42] Komatsu, K.; Murata, Y.; Takimoto, N.; Mori, S.; Sugita, N.; Wan, T. S. M. Synthesis and proper-
ties of the first acetylene derivatives of C60 . J. Org. Chem. 59, 1994, 6101–6102.
[43] Murata, Y.; Motoyama, K.; Komatsu, K. Synthesis, properties, and reactions of a stable car-
banion derived from alkynyldihydrofullerene: 1-octynyl-C60 carbanion. Tetrahedron, 52, 1996,
[44] Taylor, R.; Holloway, J. H.; Hope, E. G.; Avent, A. G.; Langley, G. J.; Dennis, T. J.; Hare, J. P.;
Kroto, H. W.; Walton, D. R. M. Nucleophilic substitution of fluorinated C60 . J. Chem. Soc. Chem.
Comm. 1992, 665–667.
[45] Chiang, L. Y.; Bhonsle, J. B.; Wang, L. Efficient one-flask synthesis of water-soluble [60]
fullerenols. Tetrahedron, 52, 1996, 4963–4972.
3 Synthesis, functionalization and properties of fullerenes and graphene materials | 57

[46] Shinohara, H.; Yamaguchi, H.; Hayashi, N.; Sato, H.; Ohkohchi, M.; Ando, Y.; Saito, Y. Isola-
tion and spectroscopic properties of scandium fullerenes (Sc2@ C74, Sc2@ C82, and Sc2@
C84). J. Phys. Chem. 97, 1993, 4259–4261.
[47] Bethune, S.; Johnson, J. S.; Salem, J. R.; De Vries, M. S.; Yannoni, C. S. Atoms in carbon cages:
the structure and properties of endohedral fullerenes. Nature, 366, 1993, 123–128.
[48] Murata, M.; Murata, Y.; Komatsu, K. Surgery of fullerenes. Chem. Comm. 2008, 6083–6094.
[49] Ross, R. B.; Cardona, C. M.; Guldi, D. M.; Sankaranarayanan, S. G.; Reese, M. O.; Kopidakis,
N.; Peet, J.; Walker, B.; Bazan, G. C.; Van Keuren, E.; Holloway, B. C.; Drees, M. Electronic struc-
tures and spectral properties of endohedral fullerenes. Nat. Mater. 2009, 8, 2009, 208–212.
[50] S.; Neugebauer, H.; Sariciftci, N. S. Conjugated polymer-based organic solar cells. Chem. Rev.
107, 2007, 1324–1338.
[51] Brabec, C. J.; Gowrisanker, S.; Halls, J. J.; Laird, D.; Jia, S.; Williams, S. P. Polymer–fullerene
bulk-heterojunction solar cells. Adv. Mater. 22, 2010, 3839–3856.
[52] Hummelen, J. C.; Knight, B. W.; Lepeq, F.; Wudl, F.; Yao, J.; Wilkins, C. L. Preparation and char-
acterization of fulleroid and methanofullerene derivatives. J. Org. Chem. 60, 1995, 532–538.
[53] Chirvase, D.; Parisi, J.; Hummelen, J. C.; Dyakonov, V. Influence of nanomorphology on the
photovoltaic action of polymer–fullerene composites. Nanotech, 15, 2004, 1317–1323.
[54] Janssen, R. A. J.; Hummelen, J. C.; Lee, K.; Pakbaz, K.; Sariciftci, N. S.; Heeger, A. J.; Wudl,
F. Photoinduced electron transfer from π-conjugated polymers onto Buckminsterfullerene,
fulleroids, and methanofullerenes. J. Chem. Phys. 103, 1995, 788–793.
[55] Imahori, H.; Sakata, Y. Fullerenes as novel acceptors in photosynthetic electron transfer. Eur. J.
Org. Chem. 1999, 2445–2457.
[56] Kuciauskas, D.; Lin, S.;Seely, G. R.; Moore, S. L.; Moore, T. A.; Gust, D. Photoinduced electron
transfer in carotenoporphyrin-fullerene triads: Temperature and solvent effects. J. Phys. Chem.
100, 1996, 15926–15932.
[57] Shirai, Y.; Guerrero, J. M.; Sasaki, T.; He, T.; Ding, H.; Vives, G.; Yu, B. C.; Cheng, L.; Flatt, A. K.;
Taylor, P. G.; Gao, Y.; Tour, J. M. J. Org. Chem. 74, 2009, 7885–7897.
[58] Shirai, Y.; Morin, J.-F.; Sasaki, T.; Guerrero, J. M.; Tour, J. M. Recent progress on nanovehicles.
Chem. Soc. Rev. 35, 2009, 1043–1055.
[59] Vive, G.; Tour, J. M. Synthesis of single-molecule nanocars. Acc. Chem. Res. 42, 2009,
[60] Pantarotto, D.; Tagmatarchis, N.; Bianco, A.; Prato, M. Synthesis and biological properties of
fullerene-containing amino acids and peptides. Rev. Med. Chem. 4, 2004, 805–814.
[61] Jensen, A. B.; Wilson, S. R.; Schuster, D. I. Biological applications of fullerenes. Bioorg. Med.
Chem. 4, 1996, 767–779.
[62] Bosi, S.; Da Ros, T.; Spalluto, G.; Prato, M. Fullerene derivatives: an attractive tool for biologi-
cal applications. Eur. J. Med. Chem. 38, 2003, 913–923.
[63] Wharton, T.; Wilson, L. J. Highly-iodinated fullerene as a contrast agent for X-ray imaging.
Bioorg. Med. Chem. 10, 2002, 3545–3554.
[64] Gharbi, N.; Pressac, M.; Hadchouel, M.; Szwarc, H.; Wilson, S. R.; Moussa, F. [60] Fullerene is a
powerful antioxidant in vivo with no acute or subacute toxicity. Nano Lett. 5, 2005, 2578–2585.
[65] Guo, S.; Dong, S. Graphene nanosheet: synthesis, molecular engineering, thin film, hybrids,
and energy and analytical applications. Chem. Soc. Rev. 40, 2011, 2644–2672.
[66] Novoselov, K. S.; Fal’ko, V. I.; Gelbert, P. R.; Schwab, M. G.; Kim, K. A roadmap for graphene.
Nature 490, 2012, 192–200.
[67] James, D. K.; Tour, J. M. Graphene: Powder, flakes, ribbons, and sheets. Acc. Chem. Res. ASAP
Article DOI : 10.1021/ar300127r.
[68] Mas-Ballesté, R.; Gomez-Navarro, C.; Gomez-Herrero, J.; Zamora, F. 2D materials: to graphene
and beyond. Nanoscale 3, 2011, 20–30.
58 | Part I Functional materials: Synthesis and applications

[69] Geim, A. K.; Novoselov, K. S. The rise of graphene. Nature Mat. 6, 2007, 183–191.
[70] Hernandez, Y.; Nicolosi, V.; Lotya, M.; Blighe, F. M.; Sun, Z.; De, S.; McGovern, I. T.; Holland, B.;
Byrne, M.; Gun’ko, Y. K.; Bolanf, J. J.; Niraj, P.; Duesberg, G.; Krishnamurthy, S.; Goodhue, R.;
Hutchinson, J.; Scardaci, V.; Ferrari, A. C.; Coleman, J. N. High-yield production of graphene by
liquid-phase exfoliation of graphite. Nature Nanotech. 3, 2008, 563–568.
[71] For representative example, see: Li, X.; Cai, W.; An, J.; Kim, S.; Nah, J.; Yang, D.; Piner, R.; Ve-
lamakanni, A.; Jung, I.; Tutuc, E.; Banerjee, S. K.; Colombo, L.; Ruoff, R. F. Science, 324, 2009,
[72] Srivastava, A.; Galande, C.; Ci, L.; Song, L.; Rai, C.; Jariwala, D.; Kelly, K. F.; Ajayan, P. M.
Novel liquid precursor-based facile synthesis of large-area continuous, single, and few-layer
graphene films. Chem. Mater. 22, 2010, 3457–3461.
[73] Subrahmanyam, K. S.; Panchakarla, L. S.; Govindaraj, A.; Rao, C. N. R. Simple
method of preparing graphene flakes by an arc-discharge method. J. Phys. Chem. C, 113, 2009,
[74] Wang, Z.; Li, N.; Shi, Z.; Gu, Z. Low-cost and large-scale synthesis of graphene nanosheets by
arc discharge in air. Nanotechnology, 21, 2010, 175602 (4pp).
[75] Shivaraman, S.; Barton, R. A.; Yu, X.; Alden, J.; Herman, L.; Chandrashekhar, M. V. S.; Park, J.;
McEuen, P. L.; Parpia, J. M.; Craighead, H. G.; Spencer, M. G. Free-standing epitaxial grapheme.
Nano Lett., 9, 2009, 3100–3105.
[76] Pletikosic, I.; Kralj, M.; Pervan, P.; Brako, R.; Coraux, J.; N’Diaye, A. T.; Busse, C.; Michely, T.
Dirac cones and minigaps for graphene on Ir(111). Phys. Rev. Lett. 102, 2009, 056808 (4pp).
[77] Chen, D.; Feng, H.; Li, J. Graphene oxide: preparation, functionalization, and electrochemical
applications. Chem. Rev. 112, 2012, 6027–6053.
[78] Su, C.; Loh, K. P. Carbocatalysts: graphene oxide and its derivatives. Acc. Chem. Res. ASAP
Articles DOI : 10.1021/ar300118v
[79] Englert, J. M.; Röhrl, J.; Schmidt, C. D.; Graupner, R.; Hundhausen, M.; Hauke, F.; Hirsch, A.
Soluble Graphene: generation of aqueous graphene solutions aided by a perylenebisimide-
based bolaamphiphile. Adv. Mater. 21, 2009, 4265–4269.
[80] Hao, R.; Qian, W.; Zhang, L.; Hou, Y. Chem. Commun. 2008, 6576–6578.
[81] Lotya, M.; Hernandez, Y. Aqueous dispersions of TCNQ-anion-stabilized graphene sheets.
King, P. J.; Smith, R. J.; Nicolosi, V.; Karlsson, L. S.; Blighe, F. M.; De, S.; Wang, Z.; McGovern, I.
T.; Duesberg, G. S.; Coleman, J. N. J. Am. Chem. Soc. 131, 2009, 3611–3620.
[82] Hamilton, C. E.; Lomeda, J. R.; Sun, Z.; Tour, J. M.; Barron, A. R. High-yield organic dispersions
of unfunctionalized graphene. Nano Lett., 9, 2009, 3460–3462.
[83] Bourlinos, A. B.; Georgakilas, V.; Zboril, R.; Steriotis, T. A.; Stubos, A. K. Liquid-Phase exfolia-
tion of graphite towards solubilized graphenes. Small, 5, 2009, 1841–1845.
[84] E. T. Chernick and R. R. Tykwinski. Carbon-rich nanostructures: the conversion of acetylenes
into materials. J. Phys. Org. Chem., 26, 2013, 742–749.
[85] Rondeau-Gagné, S.; Morin, J.-F. Preparation of carbon nanomaterials from molecular precur-
sors. Chem. Soc. Rev. Advance Article DOI: 10.1039/c3cs60210a.
[86] Morin J.-F, Oligoyne derivatives as reactive precursors for the preparation of carbon nanomate-
rials. Synlett, 2013 DOI: 10.1055/s-0033-1339680.
[87] Wu, J.; Pisula, W.; Müllen, K. Graphenes as potential material for electronics. Chem. Rev. 107,
2007, 718–747.
[88] Dössel, L.; Gherghel, L.; Feng, X.; Müllen, K. The graphene nanoribbons by chemists– nanome-
ter sized, soluble and defect-free. Angew. Chem. Int. Ed. 50, 2011, 1–5.
[89] Schwab, M. G.; Narita, A.; Hernandez, Y.; Balandina, T.; Mali, K. S.; De Feyter, S.; Feng, X.;
Müllen, K. Structurally defined graphene nanoribbons with high lateral extension. J. Am. Chem.
Soc. 134, 2012, 18169–18172.
3 Synthesis, functionalization and properties of fullerenes and graphene materials | 59

[90] Wu, J.; Gherghel, L.; Watson, M. D.; Li, J.; Wang, Z.; Simpson, C.; Kolb, U.; Müllen, K. From
branched polyphenylenes to graphite ribbons. Macromolecules 36, 2003, 7082–7089.
[91] Simpson, C. D.; Mattersteig, G.; Martin, K.; Gherghel, L.; Bauer, R. E.; Räder, H. J.; Müllen, K.
Nanosized molecular propellers by cyclodehydrogenation of polyphenylene dendrimers. J. Am.
Chem. Soc. 126, 2004, 3139–3147.
[92] Wan, X.; Huang, Y.; Chen, Y. Focusing on energy and optoelectronic applications: A journey for
graphene and graphene oxide at large scale. Acc. Chem. Res. 45, 2012, 598–607.
[93] Chen, D.; Zhang, H.; Liu, Y.; Li, J. Graphene and its derivatives for the development of solar
cells, photoelectrochemical, and photocatalytic applications. Energy Environ. Sci. 6, 2013,
[94] Wu, J.; Becerril, H. A.; Bao, Z. N.; Liu, Z. F.; Chen, Y.; Peumans, P. Organic solar cells with
solution-processed graphene transparent electrodes. Appl. Phys. Lett. 92, 2008, 263302.
[95] De Arco, L. G.; Zhang, Y.; Schlenker, C.; Ryu, K.; Thompson, M. E.; Zhou, C. Continuous, highly
flexible, and transparent graphene films by chemical vapor deposition for organic photo-
voltaics. ACS Nano 4, 2010, 2865–2873.
[96] Kavan, L.; Yum, J. H.; Grätzel, M. Graphene nanoplatelets outperforming platinum as the
electrocatalyst in co-bipyridine-mediated dye-sensitized solar cells. Nano Lett., 11, 2011,
[97] Liu, Z.; Liu, Q.; Huang, Y.; Ma, Y.; Yin, S.; Zhang, X.; Sun, W.; Chen, Y. Organic photovoltaic
devices based on a novel acceptor material: grapheme. Adv. Mater. 20, 2008, 3924–3930.
[98] Yu, D.; Park, K.; Durstock, M.; Dai, L. Fullerene-grafted graphene for efficient bulk heterojunc-
tion polymer photovoltaic devices. J. Phys. Chem. Lett. 2, 2011, 1113–1118.
[99] Li, S. S.; Tu, K. H.; Lin, C. C.; Chen, C. W.; Chhowalla, M. Solution-processable graphene oxide
as an efficient hole transport layer in polymer solar cells. ACS Nano, 4, 2010, 3169–3174.
J. Florek, R. Guillet-Nicolas, and F. Kleitz
4 Ordered mesoporous silica: synthesis
and applications
4.1 Introduction

Porous materials were initially defined in terms of their adsorption properties, and
thus distinguished by their pore size range. Pore size usually refers to pore width, i.e.,
the diameter or distance between opposite walls in a solid. According to the IUPAC
definition [1], porous solids are then divided into 3 classes: microporous (< 2 nm),
mesoporous (2–50 nm) and macroporous (> 50 nm) materials. Additionally, the term
“nanoporous”, which refers to pores in the nanometer size range (< 100 nm), is in-
creasingly being used.
Materials with pores in the nanometer range have emerged as key elements for
the development of future technologies including miniaturized electronics, magnetic
and optical devices, environmentally friendly catalysts, materials for pollutant re-
moval, biocompatible implants, and drug delivery systems [2–4]. The nanopore size
range offers vast potential for the construction of elaborate functional systems with
tailor-made properties. As a few examples, nanoporous materials may be used as
highly selective sorbents, permselective membranes, systems for energy storage or
energy conversion, recyclable solid catalysts, low k-dielectrics, sensors, biomateri-
als for drug/gene delivery or medical imaging, etc. Furthermore, nanoscale pores
enable confinement effects which can restrict the growth of crystals and quantum
objects, shift the phase behavior of fluids, or create compatible hybrid interfaces.
Nano-reactors able to perform size- and shape-selective chemical conversions are
also being developed on the basis of nanoporous solids. In such systems, coopera-
tive/complementary chemical processes may be realized in the confined space of the
nanopores, acting as spatially-functionalized cavities in analogy to enzymatic active
sites. However, it is important to keep in mind that the size and volume of the pores
in a given material have a profound influence on the final properties of the solid,
such as adsorption-desorption processes, diffusion mechanisms, storage capacity,
size exclusion, confinement, density, mechanical stability, etc.
Many types of synthetic nanoporous materials, such as ordered microporous and
mesoporous materials, controlled pore glasses, gels, pillared clays, anodic alumina,
porous polymers, carbon nanotubes, and so forth, have been intensively studied [5–7].
In general, the main attribute is that nanoporous materials exhibit high surface area,
and the only way to generate such materials with the desired surface area is through
a structuring of the solid at the nanometer level. In many cases, this is achieved with
procedures which rely on structuring through the addition of porogen species or tem-
plates, which can be single organic molecules or supramolecular aggregates, most
62 | Part I Functional materials: Synthesis and applications

frequently. In addition, some other methods are based on the use of solid templates,
namely the nanocasting techniques [8–10]. All of these templating pathways are used
to synthesize nanoporous materials with a high level of control over their structural
and textural properties. These categories of solids are generally viewed as high per-
formance materials for applications in catalysis, separation or storage. As discussed
by Schüth and Schmidt [11], their regular pore system with its exceedingly high sur-
face area can be used to introduce guests (e.g., molecules, clusters, macromolecules,
or particles) that are stabilized by the solid framework and spatially organized, which
enables the development of functional materials.
In particular, the use of supramolecular assemblies (i.e., micellar aggregates) as
structure-directing agents (SDAs) allowed the synthesis of a new family of mesoporous
silica and aluminosilicate compounds, which was designated M41S [12, 13]. These
solids were the first examples of solids exhibiting ordered arrangements of mesopores
with a narrow pore size distribution. This discovery was a major breakthrough, which
then opened up a whole field of research, and great new possibilities in many areas
of chemistry and materials science. In this chapter, the authors focus first on general
aspects related to the synthesis and functionalization of ordered mesoporous silica
(OMS) materials. Then, a few recent developments concerning the use of function-
alized mesoporous silica will be presented, with special emphasis on modern drug
delivery and separation applications.

4.2 Ordered mesoporous silica (OMS)

In the area of ordered mesoporous materials, silica-based systems are still the most
widely studied. There are several reasons for this: the great variety of possible struc-
tures, the precise control of the hydrolysis-condensation reactions possible due to the
lower reactivity (compared to transition metals for example), enhanced thermal sta-
bility, and a great variety of functionalization methods available. In addition, meso-
porous silica is fairly biocompatible, which is of great importance for biomedical ap-
plications (see below, Section 4.5).
The characteristic approach for the synthesis of ordered mesoporous materials
is the use of liquid crystal-forming templates which enable the specific formation of
pores with a predetermined size. Among the different materials reported in 1992 by
the scientists at Mobil Corporation [12], the one named MCM-41 (Mobil Composition of
Matter №41) exhibits a highly ordered hexagonal array of cylindrical mesopores with
a relatively narrow pore size distribution. This new type of solid is thus characterized
by periodic arrangements of pores, but the framework pore walls are built of amor-
phous silica. In general, the combination of powder X-ray diffraction (XRD), transmis-
sion electron microscopy (TEM), and gas physisorption analysis enables reliable char-
acterization of ordered mesoporous materials (Fig. 4.1). In particular, the hexagonal
arrangement of uniform pores of MCM-41 can be clearly visualized by TEM (Fig. 4.1(a)).
4 Ordered mesoporous silica: synthesis and applications | 63

The Mobil synthesis performed in alkaline medium led to three well-defined struc-
tures: MCM-41, MCM-48, and MCM-50. Vartuli et al. [14] showed that the surfactant-
to-silica mole ratio is a critical variable in the formation of M41S materials. Us-
ing tetraethoxysilane (TEOS) with cetyltrimethylammonium chloride (CTAC, C16 H33
(CH3 )3 NCl), they found that progressively increasing the surfactant-to-silica molar
ratio from 0.5 to 2.0 resulted in hexagonal (< 1), cubic Ia3̄d (1–1.5), lamellar (1.2–2),
and uncondensed cubic octamer (> 2) mesoscale structures. The structure of MCM-48
belongs to the Ia3̄d space group (Fig. 4.1c), which has also been found in the binary
water/cetyltrimethylammonium bromide (CTAB) system [15]. This three-dimensional
(3D) porous structure is considered to be bicontinuous with a simplified representa-
tion of two 3D mutually intertwined networks of rods [16, 17]. The unit cell parameter
measured for cubic MCM-48 usually ranges in between 8 nm and 10 nm. MCM-50 is a
lamellar mesostructure in the as-synthesized form. However, removal of the template
results in the collapse of this layered structure unless the material has previously been
Since the discovery of the MCM family in the early 90s, considerable progress has
been achieved regarding the synthesis, characterization, and porosity control of or-
dered mesoporous silica. An impressive diversity of synthesis approaches have been
developed, which now enables the formation of various OMS (e.g., MSU [19], SBA-15
[20], SBA-16 [20], FDU [21], KIT-6 [22], etc.) with a great variety of morphological, struc-
tural, and textural properties [23, 24]. However, for all the materials, the concepts in-
volved in their synthesis usually remain quite similar and can be seen mainly as a com-
bination of three key features: (1) surfactant, co-surfactant, solvent, and co-solvent
types, (2) controlled polymerization of suitable inorganic species, and (3) interactions
between the inorganic (silica) precursor and the organic templating agents.

4.2.1 Principle of synthesis

For the synthesis of OMS, two main routes are possible. These are based on the in-
teractions between organic moieties, known as structure-directing-agents (SDAs) or
templates, and the silica precursors, which ultimately lead to the formation of an or-
dered hybrid mesophase (as illustrated in Fig. 4.2). After extended polymerization and
condensation of the silicates (or poly-silicic acid species for low pH synthesis condi-
tions), mesoporosity is then created through the removal of the template.
While the general concept is similar, there are differences between these two path-
– In pathway A, i.e., the cooperative self-assembly route, the mesophase is created
through the addition of the silica precursor to the micellar system. Prior to mix-
ing, no liquid crystalline phase exists; only isotropic micelles are present in the
solution. The self-assembly reorganization of the micelles occurs in situ, cooper-
64 | Part I Functional materials: Synthesis and applications



Volume adsorbed (cm3/g)

400 0.20

Dv(d) (cm3/A/g)

100 Ads. 0.00

0 1 2 34 5 678
Des. Pore width (nm)
20 nm 0.0 0.2 0.4 0.6 0.8 1.0
(a) (b)

2 3 4 5 6 2 3 4 5 6
2Θ 2Θ

Fig. 4.1. (a) Transmission electron microscopy image of MCM-41. Reprinted with permission
from [24]. (b) Nitrogen adsorption-desorption isotherms (−196°C) and corresponding NLDFT pore
size distribution for calcined MCM-41 silica (BET surface area: 1 070 m2 g−1 . Total pore volume:
0.92 cm3 g−1 ; NLDFT pore size: 4.1 nm). Reprinted with permission from [18]. (c) Examples of powder
X-ray patterns obtained for ordered mesoporous silica mesophases shown with their pore topology.
On the left: MCM-41 with p6mm symmetry, on the right: MCM-48 with Ia3̄d symmetry. Reprinted
with permission from [24].

atively with the formation and polymerization of the inorganic network around
them, leading ultimately to a highly organized hybrid mesophase.
– In pathway B, i.e., the true liquid-crystal templating (TLCT), a pre-formed liquid
crystalline phase is used as a template for the infiltration of the silica precursors.
The polymerization of the inorganic network is then triggered around this micellar
organization. This synthesis pathway was successfully used by Attard et al. [25] to
synthesize OMS with various mesostructures, and especially mesoporous silica
4 Ordered mesoporous silica: synthesis and applications | 65

Most often, however, the so-called cooperative self-assembly takes place between
the templating species and the mineral network precursors with synchronized self-
assembly and inorganic network formation, yielding highly organized mesoscopic
architectures. This cooperative formation mechanism, which was first proposed by
Stucky, Schüth and co-workers, is now largely accepted by most researchers as an
explanation for the mesophase formation [24, 26–28].

Liquid solution Mixture of solution and precipitation


Inorganic Species
Cooperative Cooperative aggregation Liquid crystal Further polymerization
A nucleation and phase separation formation with and condensation of
molecular inorganics inorganics
Template Elimination
Mesoporous framework
of final product

Liquid crystal Incorporation Transformation Template

formation of inorganics’ of precursors to elimination
precursor aimed materials

Fig. 4.2. The two main synthesis pathways for the formation of ordered mesoporous materials:
A, cooperative self-assembly and B, true liquid-crystal templating. Reprinted with permission
from [23].

Cooperative self-assembly and hybrid interfaces

In practice, synthesis starts with the dissolution of an amphiphilic molecule, i.e., sur-
factant that can be ionic or non-ionic, in water. Here, the surfactant-inorganic hy-
brid mesophase forms cooperatively from the species present in solution which are
not in a liquid crystalline state prior to mixing of the precursors. Quaternary cationic
surfactants, Cn H2n+1 N(CH3 )3 Br (n = 8–22), are the most common for the synthesis
of ordered mesoporous silica materials with pores ≤ 5 nm. The commercially avail-
able C16 TAB is widely used for the synthesis of M41S silicas. In order to synthesize
materials with larger pores, swelling agents (SAs) can be used, but they generally
lead to poorly reproducible syntheses and low-quality materials due to the hetero-
geneous dispersion of the SAs in the system. Alternatively, amphiphilic non-ionic tri-
block copolymers were also proposed as self-assembling SDAs, and these have be-
come more and more popular as they allow the design of OMS with larger pores and
thicker walls in acidic or neutral media [19, 29]. Nowadays, the commercially available
66 | Part I Functional materials: Synthesis and applications

Pluronics™ ((EO)x -(PO)y -(EO)x ) are widely used for the syntheses of large pore ordered
mesoporous silicas, e.g., SBA-15, SBA-16 or KIT-6 (vide infra).
When dissolved in aqueous solutions, if proper conditions are met, ionic surfac-
tants or block copolymers assemble into isotropic micelles due to their amphiphilic
behavior. Alkoxysilanes, e.g., TEOS, are then added to this stock solution. Inorganic
silica precursors are then catalytically hydrolyzed producing silanol groups due to the
non-neutral nature of the media. Silanols further condense, forming polymeric silica-
based species which aggregate around the micellar structures. The cooperative self-
assembly process eventually leads to a supramolecular templating, which results in
the hybrid mesophase formation (Fig. 4.2A).
Achieving a well-defined segregation of the organic and inorganic domains at the
nanometric scale plays an essential role in the synthesis of OMS. Indeed, it is the na-
ture of this hybrid interface that governs the assembly process, and thus the overall
quality of the resulting material. The key thermodynamic factors affecting the forma-
tion of a hybrid interface were identified by Huo et al. [27, 28]. In their model, described
as charge density matching, the free energy of the mesostructure formation (ΔGms ) is
considered the sum of the free energy contributions of the inorganic-organic interface
(ΔGinter ), the inorganic framework (ΔGinorg ), the self-assembly of the organic molecules
(ΔGorg ), and a contribution of the solution (ΔGsol ). The final hybrid mesophase consists
of the ordered arrangement with the lowest interface energy.
As explained previously, in the cooperative assembly route the template concen-
tration may be well below that necessary for obtaining liquid crystalline assemblies
or even micelles. Therefore, the creation of a compatible hybrid interface between the
inorganic walls and the organic templates (ΔGinter ) is essential for the generation of
a well-ordered mesoscopic hybrid structure with appropriate curvature. From the ki-
netic point of view, the formation of an organized hybrid mesostructure results thus
from a well-balanced combination of inorganic polymerization, organization of the
SDA, and organic-inorganic phase separation. Hence, two aspects are essential to fine-
tune the mesophase formation: the reactivity of the inorganic precursors (polymer-
ization rate, isoelectric point, etc.) and the interactions involved in generating the
hybrid interface. Here, a generalized cooperative mechanism of formation can be de-
scribed based on the specific (electrostatic) interactions between the inorganic pre-
cursor (I) and the surfactant head group (S). Soler-Illia et al. [30] summarized six
possible cooperative interaction pathways for the assembly of the hybrid mesophase,
which were originally proposed by Huo et al. [27, 28] and later extended by Pinnavaia
[19, 31]. As presented in Fig. 4.3, in the simplest case, the ionic surfactant and the in-
organic species are oppositely charged, leading to electrostatic interactions (S+ I− or
S− I+ ). Since the synthesis of MCM-41 and MCM-48 silicas is carried out in strongly al-
kaline media, silica oligomers have a negative charge that interacts with the positive
charge of the cationic CTAB molecules (S+ I− mechanism). These two direct routes can
be completed by two indirect ones, i.e., interactions between surfactants and inor-
ganic moieties with the same charges through the use of a counter-ion [28]. For ex-
4 Ordered mesoporous silica: synthesis and applications | 67

S–I+ S+I–

M O H+ O M O
M +
O H O–
M O H M O–

S+X–I+ S–M+I–

O M O H+ M O
M +
O H O–
M O M O–
Fig. 4.3. Schematic representation of the dif-
S (IX) O ferent types of organic-inorganic hybrid inter-
O M O M faces for mesostructure formation. S corre-
M O M O H+ sponds to the surfactant species, and I to the
inorganic framework. M+ and X− are correspond-
M M +
OH O H ing counter-ions. Solvent molecules are repre-
sented as triangles. Reprinted with permission
H from [30].

ample, the S+ X− I+ pathway can explain the mesophase formation under acidic con-
ditions and in the presence of halide anions (X− = Cl− , Br− ), whereas the S− M+ I−
route is characteristic of base-catalyzed synthesis in the presence of alkaline metal
ions (M+ = Na+ , K+ ).
Moreover, following the early development of the so-called MSU synthesis which
was performed in neutral conditions using non-ionic block copolymers, a novel as-
sembly approach was proposed and denoted No Io [19]. In this pathway, mesophase
formation is driven by hydrogen-bonding forces. Later on, Zhao et al. proposed the
use of non-ionic triblock copolymers (i.e., the Pluronics) and acidic conditions to syn-
thesize SBA-15 silica and other related large pore materials [20]. However, in this case,
as the silicic acid species are positively charged and the block copolymer could also
be positively charged, the No Io route was thus derived to a (No H+ )(X− I+ ) pathway. Fi-
nally, a last concept, the S0 I0 interaction is associated with materials called HMS [31],
where silicate species interact with a neutral SDA through hydrogen-bonding between
the hydroxyl groups of hydrolyzed silicate species and polar amine head groups.

Hydrothermal treatment and template removal

Most common syntheses are carried out in aqueous media at temperatures close to
room temperature (< 60°C), depending on the type of material. After this step, a hy-
drothermal treatment (HT) or aging can be performed for over a few hours or up to
several days in order to achieve a more complete condensation of the silica frame-
68 | Part I Functional materials: Synthesis and applications

work. Even if this aging step is not always mandatory, it usually greatly improves the
quality and organization of the resulting OMS [32, 33]. Moreover, mesophase tailor-
ing can be performed during this treatment [22, 34, 35]. Even mesophase transitions
are possible in some cases, because variations in the charge density of the silicate-
based framework may occur upon condensation [26]. Furthermore, various organics
(i.e., co-surfactants) can also be added to the mother liquid in order to modulate the
“soft” mesostructure of the material during this stage [24].
After aging, the resultant product is cooled, recovered by filtration or centrifuga-
tion if the particles are nanosized, and properly dried in air. The mesoporous mate-
rial is finally obtained after the removal of the organic template. The most common
method for the elimination of the organic moieties is calcination (at temperatures
usually above 500°C)[36]. Calcination allows the complete removal of organic species
from surfactant-templated silicas. However, when as-made materials containing large
amounts of organics are calcined, some carbon deposits or coke formation may be ob-
served. To avoid this contamination, calcination should be performed under sufficient
air flow with slow heating rates (1°C min−1 ) and an extended period of heating once
the 500–550°C plateau is reached (4 to 8 h). Calcination is the most efficient and conve-
nient method for the removal of organic templates, but the silanol condensation which
occurs during this process has two major impacts on the resulting mesostructure: (1) a
pronounced decrease of the unit cell of the material (shrinking), and (2) an increase
in surface hydrophobicity. Framework condensation can be limited by adjusting the
temperature and the duration of the hydrothermal treatment [36]. Furthermore, for
copolymer-templated materials especially (e.g., SBA-type and KIT-6 silicas), a brief ex-
traction step using an ethanol/HCl mixture can be performed prior to calcination, in
order to improve efficiency and limit detrimental exothermic effects [37, 38]. In addi-
tion, the surface silanol density can be easily restored after calcination by performing
controlled acidic (aqueous) treatments [39, 40], in order to carry out more efficient
post-grafting procedures [41]. Note that calcination is often recommended when OMS
are designed for biomedical applications because it removes all trapped SDA organic
species. This aspect is of tremendous importance because residual amounts of free
surfactants (e.g., CTAB), which typically remain after liquid extraction, were found
to be toxic [42, 43]. On the other hand, calcined OMS were found to be “reasonably”
biocompatible [44, 45].
Other methods for template removal include liquid extraction [46], acid treat-
ments [47], H2 O2 oxidation under microwave irradiation [48], supercritical CO2 extrac-
tion [49] and ozone treatment [50]. Each method causes noticeable variations in the
final properties of the porogen-free materials. For example, prolonged (12–24 hours)
or multiple liquid extraction steps (e.g., Soxhlet extraction) can be used in the case
of organic-inorganic hybrid materials in which some organic functionalities must be
preserved after template removal. In contrast, H2 O2 oxidation under microwave irra-
diation cannot be used with functionalized organic–inorganic hybrids, but it enables
complete removal of the template while generating porous materials with higher pore
4 Ordered mesoporous silica: synthesis and applications | 69

volume and a more hydrophilic surface than calcined counterparts, and this in a very
short time (< 15 min) [51].

4.2.2 Mesostructure diversity and tailoring

Tailoring of the textural and structural properties of the material, i.e., pore size, pore
shape, and connectivity, is essential, especially regarding the potential application
of OMS in catalysis, selective sorption, sensing technologies, and so on. Various pa-
rameters may be tuned, but the most important are the choice and ratios of reactants,
synthesis time and temperature, and the use of additives. By carefully adjusting these
parameters, a large diversity of materials can be synthesized. The most frequent or-
dered mesoporous silica materials are compiled in Table 4.1, along with their struc-
tural characteristics.

Table 4.1. Structural parameters of the most common ordered mesoporous silicas (OMS).

OMS SDA system Type of Pore Typical References

interaction mesostructure pore size (nm)
MCM-41 Cn (CH3 )3 N+ Br− or Cl− S+ I− 2D hexagonal 1.5–10.0 [12–14]
8 ≤ n ≤ 18 p6mm
MCM-48 Cn (CH3 )3 N+ Br− or Cl− S+ I− 3D cubic Ia3̄d 1.5–4.6 [14, 26, 37]
12 ≤ n ≤ 18
SBA-1 Cn (C2 H5 )3 N+ Br− or Cl S+ X− I+ 3D cubic Pm3̄n 1.5–3.0 [27, 28]
12 ≤ n ≤ 18
SBA-3 Cn (CH3 )3 N+ Br− or Cl S+ X− I+ 2D hexagonal 1.5–3.5 [33, 52]
12 ≤ n ≤ 18 p6mm
SBA-12 Brij 76 (C18 EO10 ) (So H+ )(X− I+ ) 3D hexagonal 3.0–5.0 [53, 61]
SBA-15 P123 (EO20 PO70 EO20 ) (No H+ )(X− I+ ) 2D hexagonal 4.0–15.0 [20, 37, 61]
SBA-16 F127 (EO106 PO70 EO106 ) (No H+ )(X− I+ ) 3D cage-like cubic 4.7–12.0 [35, 54, 61]
F127 + P123 Im3̄m
F127 + BuOH
KIT-6 P123 + BuOH (No H+ )(X− I+ ) 3D cubic Ia3̄d 4.0–12.0 [22, 55]
FDU-1 B50-6600 (No H+ )(X− I+ ) 3D cage-like cubic 8.0–14.0 [21, 56]
(EO39 BO47 EO39 ) Fm3̄m
FDU-12 F127 (No H+ )(X− I+ ) 3D cage-like cubic 6.0–12.5 [57, 58]
or KIT-5 F127 + TMB Fm3̄m
MSU-H P123 No Io 2D hexagonal 7.5–12.0 [59, 60]

TMB = trimethylbenzene
70 | Part I Functional materials: Synthesis and applications

Triblock copolymer-templated large pore silica

SBA-15 (Santa Barbara Acid №15). A breakthrough in the preparation of ordered meso-
porous silica was made by Zhao and Stucky in 1998 [20, 61], who used poly(ethylene
oxide)-poly(propylene oxide)-poly(ethylene oxide) triblock copolymers for the synthe-
sis of the large pore SBA-15 material. The synthesis is simple and based on the use of
organic silica sources, such as TEOS or tetramethoxysilane (TMOS), in combination
with diluted acidic aqueous solution of a Pluronic-type triblock copolymer (2–7 wt%
in water), such as P123 (EO20 –PO70 –EO20 ). As discussed above, the hybrid interface
formation is here suggested to follow a (N0 H+ ) (X− I+ ) model, since the block copolymer
could be positively charged under the reaction conditions.
Evidently, the use of triblock copolymers expands the accessible range of meso-
pore sizes. Mesoporous silicas obtained with such copolymers usually exhibit uniform
large pores with diameters well above 5 nm and quite thick walls, the latter providing
high thermal stability and improved hydrothermal stability compared to OMS synthe-
sized with ionic surfactants, e.g., M41S materials [62]. To compare, MCM-41 materials
usually showed a wall thickness of about 1 nm. SBA-15 silica can be synthesized with
pore sizes ranging between 5 nm and 12 nm and thick walls (3.0–6.5 nm in width), de-
pending on the reagent ratios, pH, and aging temperature (Figs. 4.4(a), (c); [38]). This
material exhibits a large surface area of around 800-1 000 m2 g−1 and pore volume
up to 1.5 cm3 g−1 . A TEM image showing the hexagonal structure of SBA-15 is presented
in Fig. 4.4(a).
SBA-15 silica is of growing interest for a wide range of applications (e.g., sorbent,
support for catalysts and biomolecules, nanoreactor, solid template, etc.). At first,
SBA-15 was thought to be a large pore equivalent of MCM-41, which has unconnected
mesoporous cylindrical channels. However, studies showed that the pore size distri-
bution of SBA-15 is rather bimodal, whereby the larger, hexagonally ordered structural
mesopores are connected by smaller pores (micropores or small mesopores) located
inside the silica walls [37, 63–66]. These pores are not ordered and most probably orig-
inate from the penetration of the PEO blocks of the copolymer inside the silica frame-
work. Owing to the interaction of the hydrophilic chains of the P123 copolymer with
the polymerizing silica species during the mesophase formation of SBA-15, some EO
groups are occluded in the silica walls. After removal of P123, SBA-15 exhibits therefore
a secondary pore system in its framework wall (i.e., intra-wall pores). These intra-wall
pores are usually in the micropore-small mesopore (∼ 2–3 nm) range, but the actual
size and associated volume are highly dependent on the details of the synthesis (see

SBA-16. In the family of copolymer-templated materials, ordered mesoporous silicas

consisting of interconnected large cage-like pores are also of significant interest. A sil-
ica mesophase related to SBA-15 is the material designated SBA-16 (cubic Im3̄m sym-
metry), which is synthesized in a similar way, but using a different nonionic triblock
copolymer, e.g., Pluronic F127 (EO106 –PO70 –EO106 ). This large pore silica consists of
4 Ordered mesoporous silica: synthesis and applications | 71

50 nm 100 nm

(a) (b)

900 0.12
Dv(d) (cm3/A/g)

0.10 130°C
Volume adsorbed (cm3/g)

Volume adsorbed (cm3/g)

700 0.06
600 0.02
500 0.00 300
0 2 4 6 8 10 12 14
400 Pore width (nm)
0 0
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
P/P0 P/P0
(c) (d)

Fig. 4.4. (a) TEM image showing the 2D hexagonal arrangement of the mesopores in SBA-15.
Reprinted with permission from [2]. (b) TEM image of cage-like SBA-16 silica aged at 100°C, viewed
along the [111] direction. Reprinted with permission from [2]. (c) Nitrogen adsorption-desorption
isotherm (−196°C) and corresponding NLDFT pore size distribution for SBA-15 (BET surface area:
875 m2 g−1 ; Total pore volume: 1.26 cm3 g−1 ; NLDFT pore size: 8.5 nm). Reprinted with permission
from [18]. (d) Nitrogen adsorption-desorption isotherms (−196°C) of SBA-16 samples aged at 100°C
or 130°C, as indicated. Insert shows a scheme of the pore structure. Reprinted with permission
from [54].

spherical cavities of 6–11 nm in diameter organized in a body-centered cubic (bcc) ar-

ray, and the cavities are 3D interconnected through mesoporous openings of 2–4 nm
(Figs. 4.4(b), (d)) [54, 61, 67]. Pluronic F127 presents a high hydrophilic to hydrophobic
volume ratio (high EO/PO ratio), which is favorable for the formation of highly curved
globular micelles under aqueous conditions.
72 | Part I Functional materials: Synthesis and applications

KIT-6 (Korea Institute of Technology №6). Another member of the family of triblock
copolymer-based silica mesophases is the large pore equivalent of MCM-48 known as
KIT-6. One of the easiest methods of generating this interesting mesophase was intro-
duced in 2003 by Kleitz et al. [22, 55], who used a blend of Pluronic P123 and n-butanol
for the structure-direction, along with a fine tuning of the acid concentration. This
KIT-6 silica material exhibits a structure with cubic Ia3̄d symmetry, and the pore net-
work topology can be described as an interpenetrating bicontinuous network of highly
interconnected channels (shown schematically in Fig. 4.5; [68]). The mesopore struc-
ture of KIT-6 is thus 3D interconnected and built of two continuous ordered channel
systems separated by a silica wall which follows the infinite periodic minimal surface
(IPMS) called the Gyroid surface (G) [22, 55, 69]. The porosity of KIT-6 is quite similar
in nature to that of SBA-15, although subtle differences have been observed [70]. KIT-6
silica has high pore volume and large accessible pores tailored between 5 and 12 nm,
with additional intra-wall pores as well [22, 51, 55, 69, 71]. Several other methods of
producing a large pore cubic Ia3̄d silica have been reported [72–74].


(a) 50 nm (b)

Fig. 4.5. (a) Representative TEM of mesoporous KIT-6 silica. Shown is a view along the [111] direc-
tion. (b) Representation of the Gyroid G infinite periodic minimal surface, which is followed by the
silica walls in KIT-6. Also shown is an alternative representation of the Ia3̄d structure, as two inter-
woven networks of branched cylindrical channels. The G surface separates the two sub-frameworks
of rod-like mesopores. Reprinted with permission from [70].

Some tools for tailoring structure and porosity

A method of tuning the pore size of surfactant-directed inorganic materials is to sim-
ply change the length of the surfactant carbon chain. Usually, a linear relationship is
observed between pore size and length of the carbon chain of a molecular template.
With Cn TAB (n = 8–18), the pore size of the as-synthesized MCM-41 material increases
by about 0.45 nm when increasing n by two carbon atoms. Kruk et al. [75, 76] also
confirmed that the pore size of calcined MCM-41 and MCM-48 materials increases al-
most linearly using Cn TAB surfactants with chain lengths of 8 to 16 carbons. How-
ever, this simple strategy is applicable only as long as the surfactant is soluble and
4 Ordered mesoporous silica: synthesis and applications | 73

leads to the formation of a mesophase. It seems that the shortest chain surfactant from
which a mesophase could be created is with n = 8. On the other hand, long-chain sur-
factants (n > 20) are not easily available and are practically insoluble in water, and
the mesophases obtained are sometimes rather poorly ordered [12, 13, 77]. Similarly,
changes in molecular geometry and chain length of nonionic block copolymers per-
mit fine tuning of the pore size of large pore mesoporous silica. There, the adjustment
of pore size can be continuously performed by varying the concentration of SDA and
changing the composition of the copolymer or the block size [61, 78, 79]. Indeed, the
ratio of hydrophilic to hydrophobic blocks (EO/PO) in the block-copolymer can be de-
cisive for the nature of the mesostructure. In general, lowering this ratio results in the
formation of lamellar mesostructures, while higher ratios favor the packing of spher-
ical micelles into cubic mesophases [79]. For instance, triblock copolymers possess-
ing long hydrophilic chains (i.e., high EO/PO ratio), such as F127, lead to materials
with highly curved cage-like pores (e.g., SBA-16, KIT-5; [58]). In these cases, simulta-
neous tailoring of the cage dimensions and pore openings of these cage-like silicas
is also feasible by using copolymer blends (P123 mixed with F127), both under con-
trol of synthesis temperature and time [35]. The EO/PO ratio of the copolymer has a
marked influence on pore size and wall thickness of the resulting materials. Alfreds-
son et al. investigated the influence of the variation in block copolymer composition
in the synthesis of SBA-15 [79, 80]. Their results established that, for synthesis condi-
tions where a hexagonal mesostructure is obtained, an increase in PO chain length
resulted in larger pores. On the contrary, an increase in EO chain length led to thicker
Another convenient way to tailor the pore size of an OMS is to vary the temper-
ature and duration of the HT. Applying aging treatments at different temperatures
and for prolonged periods (from 24 hours up to several days) can efficiently modu-
late the nature of the mesophase. This type of treatment can either be performed di-
rectly in the mother liquid or at a different pH in fresh solutions (typically water or
alcohol). For instance, MCM-41 silica could be restructured at elevated temperatures
in its mother liquid, resulting in pore size expansion from 3.5 to 6 nm [34, 81]. More-
over, this approach usually results in a material with enhanced stability and higher
structural quality owing to denser walls [82]. This improved stability could arise from
increased condensation of the silanols within the silicate framework, i.e., better silica
polymerization, leading to less silanol groups, thus less shrinkage occurring during
calcination, and thicker walls. In the case of large pore OMS prepared with triblock
copolymers, such as SBA-15 or KIT-6, both synthesis and aging temperature strongly
influence the mesostructure formation of these OMS, and this by altering micelle hy-
drophobicity and silica condensation. The solution reactions leading to mesophase
formation are usually performed within a range of 35°C to 45°C depending on the block
copolymer. Increasing the synthesis temperature within this range and beyond ren-
ders the EO groups more hydrophobic, leading to micelles with larger hydrophobic
core volume and smaller hydrophilic regions [78]. The temperature and duration of
74 | Part I Functional materials: Synthesis and applications

the hydrothermal aging, which is then applied after this first synthesis step, are also
critical. The aging temperature is usually between 40°C to 150°C (often 90–100°C).
The mesopore size of SBA-15 is easily tailored from about 5 nm up to 12 nm simply by
increasing the aging temperature from 60°C to 140°C. Also, substantially larger pore
volumes are obtained and the nature of the intra-wall porosity is drastically modi-
fied, depending on the HT temperature applied (Fig. 4.6, [22, 55, 63, 66, 83–89]). It
is proposed that SBA-15 prepared with aging between 35°C and 60°C exhibits micro-
pores with no apparent connection between mesopores. In contrast, at 100°C, SBA-15
shows both the presence of micropores and larger connections between the ordered
mesopores. At 130°C, the material shows no more micropores, but much larger pore
interconnections are present [66]. Note that an aging of 12 hours to several days is nor-
mally required to produce silica materials with satisfying quality. Similar to SBA-15,
the mesopore size of the cubic Ia3̄d KIT-6 silica can be varied within a comparable
range of diameters [22]. Also, the size of the spherical mesopores of cage-like materi-
als (e.g., SBA-16, FDU-1, KIT-5) can be tailored by applying different aging temperatures
and times. However, in this case, not only the main mesocage is enlarged, but the pore
openings of the cages also become wider upon prolonged hydrothermal treatment.
Fine tuning of pore size and mesostructure can also be performed by adjusting
the solution pH upon addition of given amounts of acid or base during synthesis.
This is, for example, well-established for MCM-41 and MCM-48 syntheses [90, 91]. It is
usually explained by the strong influence that the solution pH has on the degree of
condensation and polymerization of the inorganic oligomeric species, on the charge
density of the polyelectrolyte inorganic species involved, and on the micellar organi-
zation. Indeed, as reported by Ryoo, adjusting the pH of the reaction mixture in situ
during the synthesis of MCM-48 favors the formation of cubic mesophase [92]. More-
over, the materials (MCM-type) prepared with careful monitoring and adjustment of
the pH often demonstrate a high degree of long-range order [91]. For block copolymer-
templated syntheses, usually performed in acidic media, pH also has great impor-
tance. For SBA-15, it is the key parameter to control the overall kinetics of the syn-
thesis. SBA-15 can be synthesized in various acid concentrations [38, 93, 94], however,
some noticeable differences in the structural order, particle morphology, and poros-
ity features of the resulting solids have been observed. Syntheses performed with acid
concentration ≥ 1.5 mol l−1 led to very rapid precipitation [61]. Furthermore, high acid
content may somewhat influence the micellar organization [95], although little effect
on the micelle shape has been observed, before addition of TEOS, with [HCl] up to
2 mol l−1 [93].
The addition of electrolyte salts, such as NaCl or KCl, obviously affects surfactant
packing and the interactions between surfactant molecules and silica. The surface
charge density of the surfactant micelles can be modified by adsorbed counter-ions.
The surfactant molecules may then self-assemble into a mesophase with, for exam-
ple, a lower surface charge density, and mesophase transitions could take place upon
modification of the surface curvature. Usually, inorganic salts have a strong influence
4 Ordered mesoporous silica: synthesis and applications | 75

Front view Side view

(a) No connections

(b) Connections 2–3 nm diameter

(c) Large mesoporous connections

Fig. 4.6. The effects of hydrothermal aging temperature on the pore structure of SBA-15-type ma-
terials: (a) low temperature aging (35–60°C); main mesopores: 5–6 nm - wall thickness: 4 nm –
micropore volume: ∼ 0.3 cm3 g−1 ; (b) aging at 80–100°C; main mesopores: 7–9 nm – wall thickness:
3.2 nm – micropore volume ∼ 0.1 cm3 g−1 ; (c) high temperature aging (> 120°C); main mesopores:
> 9 nm – wall thickness: 2 nm – no micropores. Reprinted from [70].

on the values of CMC (critical micellar concentration) and CMT (critical micellar tem-
perature) of the triblock copolymer micelles, which can both be decreased or increased
upon salt addition. Salting-out electrolytes (lyotropic ions) such as KCl, NaCl or K2 SO4 ,
are not adsorbed in the copolymer micelles. These salts dehydrate the hydrophilic por-
tion of the block copolymer, inducing a pronounced reduction in the preferential inter-
facial curvature of the micelles. On the other hand, salting-in electrolytes (hydrotropic
ions) are adsorbed in the micelles and tend to inhibit their growth, which could thus
increase the preferential interfacial curvature [73, 96]. The Zhao group was the first
to use the salting-out effects caused by the addition of electrolytes to triblock copoly-
mer solutions to produce various well-defined mesostructures [97]. It should be kept
in mind, however, that not only the aggregation behavior of the copolymer micelles
is affected by salt additions, but the presence of electrolytes will influence hydrolysis
and condensation, and the kinetics of aggregation of the inorganic species.
76 | Part I Functional materials: Synthesis and applications

Dissolving hydrophobic additives inside the core of the micelles is largely ex-
ploited to increase the pore size of mesoporous silicas. They can alter the interface
energy of the system, ultimately leading to changes in numerous features, e.g., micelle
shape and size, mesophase transition, enlargement of mesopores and/or variation in
the morphology of the final products [23, 24]. Trimethylbenzene (TMB) has been one
of the most widely used additives [33, 98], although aliphatic hydrocarbons such as
hexane have been used as well [99]. For example, it was shown that the pore size of
MCM-41 can be altered in a controlled manner between 2 nm and 10 nm by addition
of TMB [13]. An almost linear relationship was found between TMB concentration
and final pore size. Hydrocarbons or hydrophobic aromatics are regarded as swelling
agents that are preferentially solubilized in the core of the micelles. In contrast, co-
surfactant molecules, such as short-chain n-alcohols or n-amines, are accumulated
in the palisade layer of the micellar aggregates and therefore induce more intricate
effects, whereby both the mesophase behavior and the d-spacing of the mesoscop-
ically ordered material can be affected [100]. The MCM-48 syntheses performed by
Ryoo et al. [101] (addition of EtOH) and by Schumacher et al. [102] (addition of tri-
ethylamine) are good illustrations of the complex roles of these additives.
The use of TMB to swell the pores of the triblock copolymer-based OMS is also
widespread. SBA-15 materials with pore sizes of 15–16 nm can be obtained with this
additive. However, the quantity introduced has to be cautiously controlled for reten-
tion of the ordered mesoporous mesostructure [103–105]. The silica materials exhibit-
ing pores reaching 30 nm which were obtained by addition of TMB to the synthe-
sis mixture were in fact disordered foam-like structures [105]. Also, additives and HT
modulation (time and temperature) can be synergistically combined to tailor poros-
ity of SBA-15 [85]. TMB can also be used to control the phase transition in the tri-
block copolymer F127 system, leading to large pore OMS [106]. Similarly, for block
copolymer-templated OMS synthesized in acidic media, the use of co-surfactant addi-
tives can lead to complex systems [107, 108]. The synthesis of KIT-6 silica is a perfect
example. This material is uniquely obtained in very high phase purity by carefully con-
trolling both the addition of a co-surfactant (n-butanol) and the acidity of the mother
liquid ([HCl] < 0.8 M) [22]. In fact, in these syntheses using n-butanol, the adjustment
of the HCl concentration to 0.3–0.7 M is a prerequisite for directing the formation of a
given silica-based mesophase. In this way, it actually became possible to synthesize
several large pore cubic silica mesophases (Ia3̄d, Im3̄m, and Fm3̄m) in a wide range of
reagent compositions (see Fig. 4.7 for the diagram of product phases, as a function of
reagent ratios; [109]).
4 Ordered mesoporous silica: synthesis and applications | 77

H + + +
H + +
2.4 H + +
+ + +
TEOS (molar ratio)

H + +

1.6 H H +

H +
1.2 H + + + +
H + + +
+ + Ia-3d
+ H 2-D hex
0.8 + Disordered
Mixed with 2-D hex to Ia-3d
Mixed Ia-3d to disordered
0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8
(a) BuOH (molar ratio)

F Fm-3m
D Disordered
2.5 D D D D
TEOS (molar ratio)

2.0 D D D D D


1.5 F
– D D D
Im3 m


Fm3 m
1.0 F F
F 2-D hex
0.0 0.5 1.0 1.5 2.0 2.5 3.0
(b) BuOH (molar ratio)

Fig. 4.7. Diagrams of mesophase structures synthesized using blends of triblock copolymer and
n-butanol. The diagrams are established according to XRD measurements. (a) Each sample is
prepared with a molar ratio of 0.017 P123/x TEOS/y BuOH/1.83 HCl/195 H2 O. Reprinted with per-
mission from [22]. (b) Each sample is prepared with a molar ratio of 0.0035 F127/x TEOS/y BuOH/
0.91 HCl/117 H2 O. Reprinted with permission from [109].
78 | Part I Functional materials: Synthesis and applications

4.3 Functionalization of ordered mesoporous silica

The available methods for the functionalization of ordered mesoporous materials are
vast, and a considerable number of studies have been dedicated to potential appli-
cations of functionalized mesoporous materials, especially in heterogeneous cataly-
sis. This section is not intended to be a comprehensive survey of the topic; instead it
should only provide a brief overview of the various strategies which have been devel-
oped to modify mesoporous materials.
OMS are promising in many applications because of their unique porous proper-
ties. However, in order to be useful, they usually need to be functionalized. Depending
on the requirements of the synthesis and/or the targeted application, various strate-
gies to introduce useful functions to OMS can be considered. It is important to note that
only the most common methods will be presented in this brief section. As illustrated
in Fig. 4.8, there are two main strategies available to integrate (organic or inorganic)
functionalities to OMS [41, 110].
– Post-synthetic modification, which is usually performed by grafting or impregna-
tion/adsorption methods on porogen-free OMS. This approach allows functional-
ization of the pores and the external surface.
– Direct addition of the functionalities during the synthesis of OMS. This “one-pot”
process is also described as co-condensation. In this method, both the silica walls
and pores can be modified. A variation of this method consists of the sequen-
tial addition of organosilanes or metal-precursors at different synthetic stages.
This approach may allow control of the spatial localization of the functionalities
[111, 112].

Both methods have their own advantages and drawbacks [110, 111, 113]. Post-grafting
functionalization relies on the reaction of organosilanes (e.g., amino-, thio- , phospho-
silanes, etc.) with the free silanol groups of the pore surface. Indeed, even after calci-
nation at 550°C, the pristine OMS is not fully condensed and some silanols are still
available for grafting even if their density is quite low (1–2 SiOH nm−2 ; [24]). Post-
grafting is usually performed by treating the OMS powder in organic solutions of the
organosilane under reflux for a prolonged period. This method has several advan-
tages: ease of implementation, it does not alter the mesostructure, and it also offers
a great versatility in the choice of introduced groups. However, in some cases, a pref-
erential reaction of organosilanes at the pore entrance can be observed, leading to
an inhomogeneous distribution of functionalities. If very large molecules are grafted,
some pore blocking may occur [110].
In contrast, ordered mesoporous organosilicas can be obtained by co-condensa-
tion of tetraalkoxysilanes (TEOS or TMOS) and terminal trialkoxyorganosilanes of the
type (R󸀠 O)3 Si-R (where R󸀠 is either methyl or ethyl, and R is a non-hydrolyzable organic
group). Here, the functional groups are most often placed dangling on the surface, but
also partly inside the framework walls [114–117]. In fact, the organoalkoxysilane plays
4 Ordered mesoporous silica: synthesis and applications | 79

Direct synthesis or post synthesis

M+ exchange
grafting L Silyalation Enzyme
L encapsulation
Immobilization particle
Non-silica material
Surface coating hybrid framework

Mesoporous materials High surface area (1000 m2/g)

(MCM–41, –48, SBA–15, etc) Narrow pore size distribution
Thermal stability

Fig. 4.8. Schematic representation of the different methods available for the functionalization
of OMS. Reprinted with permission from [110].

two roles, since it acts as a building block in the inorganic structure, co-condensing
with the tetraalkoxysilane precursor, and it supplies organic functionality. A wide va-
riety of functional groups can be incorporated using this method (e.g., vinyl, phenyl,
aminopropyl, imidazole, cyanopropyl, mercaptopropyl, etc.). The mercaptopropyl
groups are especially interesting since these groups can subsequently be oxidized
with nitric acid and/or H2 O2 to yield sulfonic acid groups [118]. However, the choice
of a suitable organosilane precursor is usually limited by the conditions of synthesis,
and the template removal must be performed by solvent extraction or careful acid
The impregnation/adsorption pathway is also an important post-synthesis meth-
od for modifying OMS with organics or inorganic species. This technique is based on
the capillary introduction of a volatile solvent containing the precursors or molecules
of interest inside the mesopores of the solid. After evaporation of the solvent, the func-
tional group is then chemically linked (grafted) to the OMS by a subsequent thermal
treatment. Different techniques can be used [111, 119], but the incipient wetness, which
employs a minimal amount of solvent at the limit of the powder wetness, has been
shown to be very effective [120]. As a nice example, Choi et al. have successfully ap-
plied this technique to confine polymerization of vinyl monomers (e.g., styrene, acry-
lates) selectively on the mesopore surface of SBA-15 silica [121]. The incipient wetness
80 | Part I Functional materials: Synthesis and applications

is a very versatile tool which usually leads to a rather homogeneous distribution of

functional species inside the pore network of the OMS [120, 122].

4.4 Morphology control

Morphology control is indispensable in many of the advanced applications envi-

sioned for functional mesoporous materials [123]. Perm-selective membranes, micro-
spheres or monoliths are important for sorption, separation and chromatography
purposes. Porous thin films or fibrous structures are relevant for electronics, optics,
low k-dielectrics, and sensing applications. Colloidal particles or nano-spheres are
preferred for biomedical systems to be used in drug delivery or magnetic resonance
imaging (MRI) with contrast agents. The first ordered mesoporous materials which
were synthesized were typically finely divided powders consisting of small particles
(< 10 μm) with no well-defined morphology. Since then, a wide variety of shapes,
including thin films, (nano)spheres, fibers, tubes, macroporous-mesoporous mono-
liths, and many other complex morphologies have been described for ordered meso-
porous materials (Fig. 4.9; [124–135]). Mesoporous solids with controlled macroscale
morphology can either be designed by processing conditions such as dip-coating,
spin-coating or emulsion templating, or alternatively, formed spontaneously through
self-organization processes which are mostly based on kinetic regimes. Due to the
amorphous nature of the silica walls, simultaneous modulation of both the mesoscale
(hybrid mesophase) and macroscale (particle size and shape) is possible during syn-
thesis. However, it has to be kept in mind that formation of the mesophase and growth
of the morphology influence one another and cannot be seen as separate aspects [23].
With the ongoing emergence of complex nanostructures, controlling both mesopore
structures and morphologies of MSNs at the nanoscale is not straightforward and
requires a thorough understanding of the chemistry involved [136].
Mesoporous particles with spherical morphology are easily synthesized under al-
kaline aqueous conditions. For instance, Huo and Schüth [137] reported in 1997 the
preparation of hard transparent spheres from an emulsion at room temperature. Inge-
niously, Grün and Unger modified the Stöber synthesis of monodisperse spheres per-
formed in the presence of ethanol and ammonium hydroxide [138] and could success-
fully prepare almost monodispersed mesoporous MCM-41 and MCM-48 spheres [102].
Also, pseudomorphic transformation of commercially available pre-shaped spherical
silica particles (5 to 800 μm) can also be used to produce ordered mesoporous MCM-41
and MCM-48-like particles with a spherical morphology. In this latter method, amor-
phous silica is progressively and locally dissolved under mild alkaline conditions and
re-precipitated at the same rate in the presence of the surfactant, without modifying
the global spherical morphology [139, 140]. Finally, the synthesis of hollow particles
is achievable, for example, by spray-drying techniques, based on very rapid solvent
evaporation and retention of the pre-formed shapes [141].
4 Ordered mesoporous silica: synthesis and applications | 81

(a) (b)

50 μm 1 μm
(c) (d)

50 nm 50 nm

Fig. 4.9. Illustration of possible OMS morphologies: (a) SEM images of mesoporous silica fibers
(image: R. Guillet-Nicolas, F. Kleitz, U Laval), (b) ordered mesoporous silica colloidal spheres (im-
age: R. Guillet-Nicolas, F. Kleitz, U Laval), (c) TEM images of ordered mesoporous MCM-48 silica
nanospheres (image: R. Guillet-Nicolas, F. Kleitz, U Laval), and (d) representative TEM image of an
ordered mesoporous silica thin film. Reprinted with permission from [126].

Among these, morphologies, spheres, and especially nanospheres, are most interest-
ing for the biomedical world because these objects may interact well with cells and
do not exhibit sharp edges or preferential faces [142]. Most of the current mesoporous
spherical particle syntheses are actually derived from the seminal work of Grün, which
described the synthesis of colloidal spheres of MCM-41 and MCM-48 [102, 143]. In 2001,
Cai et al. [144] and Mann et al. [145] both reported the first successful syntheses of indi-
vidualized MCM-41 nanoparticles. However, the term Mesoporous Silica Nanoparticles
(MSNs) was popularized in 2003 by Victor Lin, who published the first facile prepa-
ration of functionalized MCM-41 nanoparticles for drug delivery applications [146].
This synthesis provided homogeneous spherical particles with a diameter ≤ 200 nm
and good porosity features, making them excellent candidates for cellular applica-
tions. Following this breakthrough, many efforts were devoted to developing MSNs
with controllable particle and pore sizes [136]. In particular, in 2008, Kim et al. re-
ported the first synthesis of MCM-48 nanospheres with high pore ordering and highly
uniform particle size (Fig. 4.10(c); [147]). By using a modified version of the Stöber syn-
thesis, and Pluronic F127 as a “particle-designer” agent in alkaline media again, they
successfully formed monodisperse MSNs with controllable sizes within the range of
70–500 nm [148].
However, the main drawback of the most common MSN protocols remains the
relatively small pore sizes of the synthesized particles. Indeed, because almost all
82 | Part I Functional materials: Synthesis and applications

syntheses are performed with CTAB-like molecules as SDAs, a maximum pore size of
4–4.5 nm is achieved. Swelling agents can be used to reach larger pores, but as for the
classical OMS, they lead to a decrease or a loss in mesostructure ordering [149]. How-
ever, this latter aspect is not necessarily of critical importance for most biomedical
applications. Triblock copolymers may also be suggested as SDAs to produce particles
with larger pore size, while keeping mesoscopic ordering. Unfortunately, obtaining
MSNs in acidic medium is more difficult than in alkaline medium (more complex in-
teractions occurring during synthesis, and difficulties associated with kinetic control).
In this area, He et al. [150] and Kim et al. [151] recently proposed two different pathways
leading to pellet-shaped SBA-15 silica (large pores) of 300–600 nm (see Fig. 4.10). Both
methods are based on restricting the growth of the SBA-15 particles. The first group
used ZrIV multivalent metal ions to “cut” the micellar aggregates in situ, whereas the
latter employed Pluronic P104 as SDA, and specific synthesis conditions to favor initial
nucleation and growth of primary particles while limiting further aggregation [152]. Al-
ternatively, by mixing fluorocarbon-surfactant (e.g., FC-4) with the classical Pluronics,
Han et al. successfully synthesized a new family of large pore MSNs which were called
IBN [153]. These new materials could have great potential for biomedical applications
as they exhibit spheroidal shapes with required particle dimensions (Fig. 4.10(c)) and
fairly large pores (> 8 nm; [154, 155]).

(a) (b) (c)

200 nm 200 nm 100 nm

Fig. 4.10. TEM images of large pore ordered MSNs: (a) SBA-15 pellets obtained using the proto-
col described by He et al. [150], (b) Kim et al. [151] (images: R. Guillet-Nicolas, F. Kleitz, U Laval),
and (c) IBN-like material obtained by Hartono et al. Reprinted with permission from [155].

4.5 Selected applications of functionalized ordered

mesoporous silica

Applications of mesoporous materials have been considered in many areas, includ-

ing catalysis, optoelectronics, sensors, sorption, biomedical materials, environmen-
tal remediation, green chemistry, and most recently, energy storage and conversion
[5, 156–159]. In this chapter, emphasis is placed on recent developments in the area
of sorbents and materials for chromatographic extraction, as well as new innovative
drug delivery systems.
4 Ordered mesoporous silica: synthesis and applications | 83

4.5.1 Functionalized MSNs as controlled drug delivery platforms

Targeted drug delivery is one of the greatest challenges in modern medicine. To ad-
dress the limitations of conventional drug delivery systems, mesoporous silica nano-
particles (MSNs) are considered robust inorganic alternatives to polymeric nano-
particles, owing to their high porosity, biocompatibility, and ease of modification.
MSNs have thus captivated a lot of interest worldwide and have emerged as promis-
ing carrier materials for controlled and vectorized delivery of drugs [160–175]. Their
stable mesoporous structure and well-defined surface properties make mesoporous
silicas good matrices to host a wide variety of drugs and biologically active species
for local and controlled drug delivery applications [149, 176–179]. Another attractive
advantage is that amorphous silica is fairly degradable in aqueous solution, and thus
problems related to the removal of the material after use can be avoided. Moreover,
there are large numbers of silanol groups covering the mesoporous silica walls which
are susceptible of undergoing chemical or biochemical functionalization, which is
one of the key aspects for the prospect of biological applications of these materials.
An ideal drug delivery system should enable efficient healing at the lowest drug
concentration and dosage frequency, while being both patient-friendly and safe. Ow-
ing to their outstanding features, which allow both the loading of various drugs or
bioactive species and the adequate functionalization needed for in vitro and in vivo
purposes, stimuli-responsive MSNs can be seen as truly promising carriers for deliv-
ering precise doses of drugs to targeted sites. Moreover, by combining diagnosis and
therapeutic tools, theranostic MSN-based platforms may be designed [164, 166].
The concept of controlled drug delivery comprises 4 major steps:
(1) The fabrication of a biocompatible device that will efficiently encapsulate high
loading of the desired drug(s).
(2) No premature release prior to reaching the target location is needed in order to
protect the healthy organs and/or cells, i.e., avoidance of side effects due to non-
specific interactions, and prevent the decomposition/denaturing of the drugs.
(3) Efficient and sustained release of the drug at the targeted location.
(4) Easy clearance of the biocompatible and/or biodegradable carrier by natural

Such a strategy is expected to enhance the drug efficiency while minimizing the re-
quired quantities owing to enhanced bioavailability at the key location [180–182].
Nevertheless, the key challenge still remains to simultaneously achieve precise tar-
geting and proper colloidal stability, especially in real physiological media [122, 183].
To achieve these goals, several controlled drug delivery systems (CDDS) have been pro-
posed based on various external or internal stimuli, e.g., temperature, time, chemical
reactions, enzymes or pH, to name a few [184–189].
With MSNs, the sequestration of the drug inside the porous network is usually
realized through the functionalization of the inner and/or outer pore surface with var-
84 | Part I Functional materials: Synthesis and applications

ious barriers acting as “gates”, such as proteins, polymers, macrocycles, or even nano-
particles. These “gatekeepers” respond to a specific chemical or environmental stim-
ulus which will induce their (reversible) removal upon exposure, hence triggering the
drug release. Barriers are of prime importance as immediate drug release is commonly
observed after administration in sink conditions, when drugs are simply adsorbed into
non-modified MSNs [177, 190]. However, the release behavior of pristine MSNs should
be considered with great care as it is also dependent on the conditions chosen to sim-
ulate the different body fluids [191]. The main MSN-CDDS are summarized in Table 4.2.
Among these, the systems based on drug release triggered by pH variations have been
widely investigated. Indeed, since the pH variations within the body and/or cells are
well-known, they can be advantageously used for target drug delivery. For example,
the pH difference between normal and cancer cells may be used for the specific target-
ing of tumor cells using MSNs [192].
Because of the extreme importance of cancer therapy, most of the research involv-
ing MSNs has so far been dedicated to the release of compounds in an acidic environ-
ment, i.e., cancer cells where pH is mildly acidic [166, 182, 189]. However, MSN-CDDS
could also be synthesized for drug delivery applications where release is triggered
at neutral or physiological pH. This feature is, for instance, of high interest for oral
delivery applications because the pH in the human gastrointestinal tract naturally
varies, i.e., the stomach is highly acidic (pH = 1.2) compared to the small intestine
(pH = 6.5–7.0) and colon (pH = 7.0–8.0), making a neutral pH-triggered approach a
smart strategy for oral delivery of drugs into the intestine. This is highly desirable as
it improves patient compliance and convenience [193, 194].
In 2007, Kawi reported a simple and fast method for encapsulating protein-loaded
NH2 –SBA-15 with polyacetic acid, creating a smart pH-responsive protein delivery
system for the first time [195]. This material exhibited almost no premature release
in pH = 1.2 (less than 2%) during the first five hours, being compatible with the
United States Pharmacopeia and the National Formulary (USP–NF) guidelines for
gastro-resistant compounds, i.e., less than 10% drug release after 2 hours in gastric
conditions [196]. However, in this example, protein release at pH = 7.4 was only 40%
after 35 hours, limiting somewhat the efficiency of this system. This low release in
physiological conditions was linked to the poor colloidal and chemical stability of the
materials. Another method of generating pH-responsive MSN-CDDS is to incorporate
positive charges into the mesopore channels of MSNs by means of trimethylammo-
nium (TA) groups [197]. These groups allow efficient adsorption of anionic molecules
and minimize their release under acidic pH owing to unfavorable electrostatic interac-
tions. At neutral pH the strong electrostatic repulsions then trigger a sustained release
of the loaded drug. This original system showed excellent drug sequestration ability
in an acidic environment and appreciable drug release in physiological media. How-
ever, if such materials are not coated or properly encapsulated, the drug loaded inside
the pores might be denatured by the acidity of the stomach environment, ultimately
altering the bioactivity of the molecule once released in the intestine.
4 Ordered mesoporous silica: synthesis and applications | 85

Table 4.2. Some controlled drug delivery systems (CDDS) using MSNs, reported in the literature.
Adapted with permission from [172].

Class Examples Schematic structure

Structure I Au Nps
Nanoparticles CdS Nps
Functiionalized NPs External
Fe3 O4 Nps
Drug molecule stimuli
ZnO Nps
Modification group Mesoporous silica
Structure II Cyclodextrin (CD) α-CD CB[6]
Macrocyclic Cucurbit[6]uril O N N
organic (CB[6]) O 6
O O stimuli
molecules Dibenzo-24-crown-8 BD24C8 O
Structure III Linear polyamine Linear molecules
Linear Saccharide
molecules derivative
Linear polymer External
stimuli or
Peptide sequence
Structure IV Polymer layers
Multilayer Biomolecules
shell coating Polyelectrolyte stimuli
Mesoporous silica
Structure V Functional molecule
Pore Polymer
modification Azobenzene External External
derivatives impeller stimuli stimuli

(a) Functional molecule (b) Nanoimpeller

or polymer

Furthermore, this approach can only be used with anionic compounds. In 2011, the
Cheng group [198] improved this system by introducing a hydrazone bond to synthe-
size MSN–hydrazone–TA materials. The TA groups can further be eliminated through
the progressive hydrolysis of hydrazone bonds in gastric pH conditions, leading to
a rapid and complete release of the drug only once the intestine has been reached.
This time-dependent and pH-sensitive MSN-CDDS was shown to enable accurate de-
livery of therapeutic drugs to the targeted tissue (colon), while limiting premature
release during gastric emptying. Here, high biocompatibility was found and no cy-
totoxicity, even at a high nanoparticle concentration, i.e., 500 μg ml−1 . In another
example, Sun et al. [199] also developed a pH-responsive oral delivery system, this
time by exploiting coordination between an anti-tumor-active poly-oxo-metalate and
amino-functionalized MCM-41 mesoporous silica nanosphere. Their results indicate
zero premature release in stomach and small intestine conditions, whereas release
was observed in colon conditions due to the pH increase. In this system, the gating
effect is believed to be based on the rupture of the coordinate bond between the metal
86 | Part I Functional materials: Synthesis and applications

centers and the amino groups in mildly basic conditions. Importantly, no cytotoxicity
was observed with healthy human amniotic cells. Unfortunately, the polyoxometalate
clusters only showed modest inhibition when tested on human malignant tumor cell
lines derived from different tissues.
From the perspective of delivery of active compounds through the gastrointestinal
tract, the use of nanocapsules or coatings would be highly beneficial to completely
isolate and protect the active compound from the acidity of the stomach. As such,
one could argue that focus should be placed on the development of simple, efficient,
biocompatible and/or biodegradable materials. Furthermore, in vivo colloidal and
chemical stability of functionalized MSNs is a critical aspect of this technology. In this
case, addressing the issue of premature release of bioactive compounds loaded in the
nanocarriers is extremely critical, especially with costly drugs. In most reported cases,
the drug is just physically adsorbed into the pores, which may cause undesired leach-
ing of the cargo before it reaches the target site making these systems less effective
in therapeutic targeting. In this context, Kleitz and Qiao described a new system in
which an azo prodrug is covalently bound to MSNs and the drug is released via reduc-
tion of the azo bond by the enzyme azoreductase, which is present in the microflora
of the colon. Hence, enzyme-responsive MSNs have been designed for a site-specific
delivery to the colon, which is achieved by a combination of passive targeting via
the nanoparticles and selectivity of the cargo itself towards the colonic azoreductase.
In this system, sulfasalazine (SZ, a prodrug firstline therapy for inflammatory bowel
disease (IBD)) was covalently attached to the mesopore surface of MCM-48 silica
nanospheres acting as the enzyme-responsive carrier, and the molecule could then
be reduced to 5-ASA and sulfapyridine by the azo-reductase bacteria inside a simu-
lated colon medium (see Fig. 4.11). The chemical binding of the prodrug on the pore
surface prevents premature release of the drug. Monodisperse MCM-48 nanospheres
(∼ 150 nm) were chosen as an inorganic scaffold due to their proven superiority in
adsorption and release, owing to their 3D continuous pore network and their very
high surface area. This system responded very well to the enzyme and no release was
observed before the particles were exposed to the simulated colon medium. The cova-
lent attachment of SZ onto nanoparticles thus clearly ensures zero release in stomach
(pH 1.2) or intestine (pH 7.4), and at the same time this system also has great potential
for delivering therapeutics to the right locations i.e., inflamed tissues or cancer cells,
therefore reducing side effects and improving bioavailability of the drugs. With this
type of carrier, one could hope that high concentrations of drug-loaded nanoparticles
will reach the inflamed tissues and cells, with high in vivo efficacy ensured by zero
premature release.
Leading candidates for the next generation therapeutic carriers may also be pH-
responsive systems. As discussed above, this approach has received a lot of attention
for cancer treatment application, owing to the noticeable pH difference between nor-
mal and cancer cells [197, 200, 201]. Particularly interesting are strategies based on the
use of proteins as stimuli-responsive gating systems [202, 203]. Among their advan-
4 Ordered mesoporous silica: synthesis and applications | 87

Toluene N N
Me N
Me O Si I N N
0 N
̶OH = O Si I = N S NH =
50nm O HO N 0

Enzyme NH2 NH2

Azo mediated NH2 NH2 NH2
release NH2
N N reductase N

NH2 = Sulfapyridine NH2 = 5-ASA


Fig. 4.11. (a) Synthesis of the enzyme-responsive drug delivery system based on mesoporous silica
nanoparticles with 3D pore structure. (b) Representation of the process of enzymatic release of
5-ASA and sulfapyridine from the nanoreactors in the presence of azoreductase bacteria. Reprinted
with permission from [188].

tages, these materials are characterized by high biocompatibility and biodegradabil-

ity, abundance of reactive groups which can be used for chemical modifications, i.e.,
grafting [204, 205]. Similarly, being able to enhance colloidal stability in the desired
release media constitutes a major improvement, as it will lead to a more homogeneous
and efficient delivery of the active compound to the target location.
In this area, our laboratory evaluated the potential of the nutraceutical MSN
conjugates as a new pH-responsive oral drug delivery system showing low toxic-
ity and sufficient colloidal stability (Fig. 4.12; [206]). In this work, nutraceutical-
functionalized mesoporous silica particles have been developed for the protection and
site-specific release of gastro-sensitive compounds via the oral route. This innovative
approach makes use of the fully biocompatible and biodegradable nutraceutical β-
lactoglobulin, which serves as a pH-responsive gating device. This nano-conjugation
led to significantly enhanced colloidal stability (a prerequisite for most drug delivery
or cell tracking applications), while the functionality of β-lactoglobulin was fully pre-
served, i.e., the gelation and gating effect was observed, and thus it could be used for
protecting and/or delivering bioactive components (i.e., via pH-controlled delivery).
Nutraceutical proteins, such as β-lactoglobulin (a member of the albumin family,
present in cow’s milk) or soy protein isolates, clearly offer the advantages of low cost
88 | Part I Functional materials: Synthesis and applications

and excellent biocompatibility. By using bio-conjugation chemistry smartly, it was

possible to synthesize a pH-responsive system showing limited premature release in
acidic media (mimicking stomach acidity). In contrast, when the particles where sus-
pended in physiological intestine conditions (pH 7.4), sustained release was observed
for 24 hours, together with enhanced colloidal stability over 48 hours. Moreover, the
differences in drug release which were observed at pH = 5 could also make this
system very attractive from the perspective of cancer therapeutics. In addition, since
β-lactoglobulin is a major by-product of the dairy industry, its production is quite cost-
effective compared to other complex and expensive systems. As such, this approach
is hence also inscribed in a strategy of smart recycling of industrial waste.

APTES grafting loading

Toluene Hexane/DMSO
110°C, overnight RT, 24 hours

pH responsive Succinylated MES/
release β-lactoglobulin DMSO-MES
grafting RT, 2 hours

pH > 5 pH < 5

: –NH2 : Succynilated β-Lactoglobulin (pH > 5)

: Ibuprofen (IBU)/Acridine (ACR) : Succynilated β-Lactoglobulin (pH < 5)

Fig. 4.12. Schematic representation of the post grafting, bio-functionalization, and pH-responsive
release of a drug/dye from β-lactoglobulin-modified-MSNs. Reprinted with permission from [206].

4.5.2 Functionalized mesoporous materials for extraction chromatography (EXC)


Historically, the first system for adsorption of heavy metal ions based on mesoporous
silica supports was used for the trapping of mercury [207]. It was proven that silica ma-
terials containing thiols (or thiol derivatives) were particularly effective as sorbents for
the capture of environmentally harmful Hg2+ . Soon after, several other silica materi-
4 Ordered mesoporous silica: synthesis and applications | 89

als, modified mainly with amino- or phosphonate-groups, were applied for adsorb-
ing metals, such as Cu2+ , Zn2+ , Ni2+ , or Fe3+ [208, 209]. Later it was proposed that
silica-based inorganic-organic hybrid sorbents could also be versatile for sequestrat-
ing/extracting radionuclides [210, 211]. For instance, functionalized MCM-41 materi-
als grafted with acetamide-phosphonate ligands have shown promising properties for
Am3+ and Pu3+ decontamination from acidic aqueous solutions [212].
Evidently, the development of new sequestration materials for the nuclear in-
dustry is essential from a health and environmental protection point of view, as they
not only provide an appropriate solution to nuclear wastes, but also enable their
detection if coupled with the appropriate analytical methodologies. The incident in
Fukushima Dai-Ichi, Japan, and the long-term consequences of this event are con-
vincing reminders that sequestration/detection of long-lived radionuclides is critical.
The expected consequences of radiological/nuclear events and their aftermath on
agriculture, population, and the environment have led the scientific community to
rethink its approach to monitoring methods based on radiochemical separation [213].
With this objective in mind, our laboratory described a simple and effective function-
alization of large pore 3D cubic mesoporous KIT-6 silica to produce a new family of
selective sorbents for radionuclides [214, 215]. Using a simple one-step post-grafting
method, phosphonate groups were chemically anchored to the silica surface, provid-
ing a highly selective and durable functionalization for the extraction of actinides
(see Fig. 4.13).
Extraction experiments performed with this material demonstrated extremely
rapid kinetics (< 1 min) and high selectivity towards actinide extraction, especially
uranium (VI). The unique spatial configuration of the 3D cubic mesoporous hybrid
induced much higher selectivity and adsorption capacity than other sorbents, and the
material is applicable over a range of conditions (pH = 4, room temperature) which
are relevant for real environmental analysis. In addition to these unique extraction
performances, synthesis of the mesoporous hybrid EXC chromatography materials is
very simple and easily scalable, which represents a true alternative for the replace-
ment of currently non-recyclable commercial EXC resins at an acceptable cost. The
high extraction efficiency of the functionalized KIT-6 sorbent for U (VI) at relatively
neutral pH could also lead to alternative analytical strategies for uranium wastewa-
ter management, and environmental and biological monitoring from the perspective
of anthropogenic contamination. In addition, such materials are mandatory for all
aspects of the nuclear fuel cycles, from mining (treatment of wastewater) to fuel
reprocessing (separation of impurities).
Well-defined mesoporous materials can also provide opportunities in the highly
valuable area of rare-earth elements [216–218]. Nowadays, the importance of rare-
earth elements (REEs) in the global economy is booming as they are used in numerous
advanced technologies more and more. High purity REEs are needed for the produc-
tion of magnets, chemical sensors or lasers, computers, plasma screens, cell phones,
cameras, and so forth; however, natural REE resources are either very limited or their
90 | Part I Functional materials: Synthesis and applications

–OH CH3 –O CH3
H3C O Toluene,
O Si O reflux, 24h O
–OH + –O Si
N2 atm
–OH O –O O

SPE cartridge
mg U g–1 sample


(b) U extraction capacity

Fig. 4.13. (a) Schematic representation of the synthesis of the extracting agent-functionalized
SBA-15-P/KIT-6-P materials. (b) Comparison of the extraction capacities of different phosphonate-
functionalized materials and photograph of an SPE cartridge assembled with the KIT-6-based sor-
bent. Reprinted with permission from [215].

extraction/separation processes are not acceptable in terms of sustainable develop-

ment [219, 220]. Therefore, it is certainly of value to develop efficient sorbents for the
extraction and valorization of REEs, also from alternative sources, for instance, indus-
trial and mining wastes. Commercially, extraction and purification of REEs is based on
multiple liquid-liquid extraction (LLE) or chromatographic-based resin separations.
However, these approaches are hampered by several issues. Because of the subtle dif-
4 Ordered mesoporous silica: synthesis and applications | 91

ferences between the various REEs, extraction and purification of REEs is very time-
consuming and involves multiple extraction/purification steps. In order to substan-
tially improve this process and provide a greener alternative to LLE, novel functional
nanoporous hybrid materials were proposed demonstrating enhanced selectivity to-
wards heavier REEs in comparison to commercially available products. In this case,
KIT-6 silica was modified with the diglycolylamide (DGA) ligand to generate an effi-
cient sorbent for REEs extraction applications (Fig. 4.14; [221]). For this, the DGA lig-
and was chemically grafted to the silica surface, which enabled the resulting hybrid
materials to be cycled and regenerated, a key feature in the development of sustain-
able and cost-effective resin-type materials minimizing waste production. The choice
of KIT-6 silica was motivated here by the highly interconnected nature of the porous
network of this material, which is expected to reduce the risk of pore blocking and to
be beneficial for diffusion of liquids through the system, all being obvious advantages
in chromatographic processes. Extraction of REEs was tested in a solid-liquid system,
and distribution coefficients (named Kd ) from batch extraction tests were obtained.
Clearly, these new sorbents showed excellent stability upon recycling and demon-
strated greater selectivity than commercially available DGA resins under the extrac-
tion conditions tested. Most importantly, some of the new sorbents exhibited a much
higher affinity for the separation of heavier lanthanides (i.e., yttric earths), being most
relevant for the electronics industry. These hybrid sorbents also showed specificity to-
wards Eu and Gd and low competitive behavior with other non-lanthanide trivalent
ions and actinides, which are problematic in the commercial extraction of REEs. The
perspective for this new system will be to minimize the number of extraction steps
used for the purification of REEs as far as possible. Functionalized nanoporous mate-
rials with a high surface area and high pore volume can offer high contact efficiency
with solutions and high adsorption capacities while preserving adequate flow and
transport properties if properly structured. Using such materials may indeed enable
the substantial reduction of the number of steps needed for separation of these critical
elements, and thus decrease both the required time and the waste production.

4.5.3 Mesoporous organic-inorganic hybrid membranes for water desalination

Nowadays, water scarcity, brought about by population growth and industrializa-

tion, is one of the major challenges facing society. Desalination of brackish or sea-
water is one of the most effectively implemented solutions. However, an alternative
process to the conventional desalination technologies, e.g., distillation and reverse
osmosis (RO), is the thermally-driven membrane distillation (MD; [222]). Typically,
such a membrane maintains water at the pore entrance and allows water evapo-
ration, leaving the non-volatile salts behind [205]. Membrane distillation has some
benefits, although its industrial application remains low. This lack of commercial
success is mostly the result of the dominance of reverse osmosis processes, mem-
92 | Part I Functional materials: Synthesis and applications

DGA-functionalized KIT-6

50 nm

Fig. 4.14. On the left: high-resolution SEM image of the functional organic-inorganic hybrid KIT-6.
On the right: diglycolylamide(DGA)-modification of the surface of KIT-6 silica to generate the meso-
porous rare-earth element (REE) sorbents. Reprinted with permission from [221].

brane flux decay, and the use of macroporous (0.2–0.7 mm), hydrophobic polymeric
membranes [224, 225]. These hydrophobic polymeric membranes are often plagued
by fouling and pore wetting issues, and thus the development of new membrane ma-
terials which will overcome these limitations is necessary. In this area, membranes
based on nanoporous, inorganic-organic hybrid materials represent a potential al-
ternative, with adequate chemical and thermal stability and appropriate pore struc-
ture [226, 227]. Within this context, the use of nanoporous organosilica thin-film
membranes (∼ 20 cm2 in size) with highly ordered pores (∼ 2 nm) has been suggested
(Fig. 4.15; [228]). The mesoporous hybrid membranes were prepared by the dip-coating
technique using 1,2-bis(triethoxysilyl)ethane, as a single bridged silicon source, in the
presence of Pluronic F68 (E080 P030 EO80 ). After adequate drying, the films were then
calcined in air at 300°C to preserve the organic groups in the framework walls.
These mesoporous films exhibit excellent desalination performance in the case of
synthetic salt solutions with feed temperatures between 20 and 60°C (using a vac-
uum MD process). These membranes produced pure water across a large range of
salt concentrations (10–150 g l−1 NaCl) at average temperatures ≤ 60°C, without ex-
hibiting the usual degradation. Furthermore, the results revealed excellent salt rejec-
tion (> 99.9%) and good water fluxes (up to 13 kg m−2 h−1 at 60°C). Here, it was hy-
pothesized that the organic moieties placed within the siloxane framework conferred
enough hydrophobicity to the pore walls to form a liquid/vapor interface at the pore
entrance, whilst the small mesopore size was crucial in preventing pore wetting. This
most recent development could indeed open up a potentially scalable process for fab-
ricating high-performance membranes for efficient water desalination.
4 Ordered mesoporous silica: synthesis and applications | 93

Salt water Membrane Fresh water

H2O Na+ Cl–


20 100

Water flux, kg m–2 hr–1


Salt Rejection, %

Flux(OS) 94
Salt Rejection(OS) 92
Salt Rejection(PS)
0 90
0 20 40 60 80 100 120 140
Time, min

Fig. 4.15. (a) Schematic representation of the water desalination process by membrane distilla-
tion using a periodic mesoporous organosilica (the mesoporous structure of the hybrid mem-
brane is shown by TEM). (b) Comparison of organosilica (OS) membrane (square symbols) and
pure silica (PS) membrane (cross symbols) in 50 g l−1 feed concentration run at 60°C. Filled sym-
bols represent water fluxes and open symbols represent salt rejection. Reprinted with permission
from [228].


The authors would like to thank NSERC (Canada) and FQRNT (Province of Quebec) for
their financial support.
94 | Part I Functional materials: Synthesis and applications


[1] K.S.W. Sing, D.H. Everett, R.H.W. Haul, L. Moscou, R.A. Pierotti, J. Rouquerol, T. Siemie-
niewska, Pure Appl. Chem. 1985, 57, 603.
[2] K. Klabunde, R. Richards, (Eds.) Nanoscale Materials in Chemistry, 2nd edition, Wiley-VCH
Publishers, Weinheim, Germany, 2009.
[3] G. Cao, Nanostructures and Nanomaterials, Synthesis, Properties and Applications, Imperial
College Press, London, 2004.
[4] C.P. Poole Jr, F.J. Owens, Introduction to Nanotechnology, John Wiley and Sons, New Jersey,
[5] F. Schüth, K.S.W. Sing, J. Weitkamp (Eds.), Handbook of Porous Solids, Wiley-VCH Publishers,
Weinheim, Germany, 2002.
[6] G.Q. Lu, X.S. Zhao (Eds.), Nanoporous Materials: Science and Engineering (Series on
Chemical Engineering), Imperial College Press, London, 2005.
[7] G. Ertl, H. Knözinger, F. Schüth, J. Weitkamp (Eds.), Handbook of Heterogeneous Catalysis,
2nd edition, Wiley-VCH Publishers, Weinheim, Germany, 2008.
[8] R. Ryoo, S.H. Joo, S. Jun, J. Phys. Chem. B 1999, 103.
[9] A.-H. Lu and F. Schüth, Adv. Mater. 2006, 18, 1793.
[10] M. Tiemann, Chem. Mater. 2008, 20, 961.
[11] F. Schüth, W. Schmidt, Adv. Mater. 2002, 14, 629.
[12] C.T. Kresge, M.E. Leonowicz, W.J. Roth, J.C. Vartuli, J. S. Beck, Nature 1992, 359, 710.
[13] J.S. Beck, J.C. Vartuli, W.J. Roth, M.E. Leonowicz, C.T. Kresge, K.D. Schmitt, C.T.-W. Chu, D.H.
Olson, E.W. Sheppard, S.B. McCullen, J.B. Higgins, J.L. Schlenker, J. Am. Chem. Soc. 1992,
114, 10834.
[14] J.C. Vartuli, K.D. Schmitt, C.T. Kresge, W.J. Roth, M.E. Leonowicz, S.B. McCullen, S.D. Hellring,
J.S. Beck, J.L. Schlenker, D.H. Olson, E.W. Sheppard, Chem. Mater. 1994, 6, 2317.
[15] X. Auvray, C. Petipas, R. Anthore, I. Rico, A.J. Lattes, J. Phys. Chem. 1989, 93, 7458.
[16] R. Schmidt, M. Stöcker, D. Akporiaye, E.H. Tørstad, A. Olsen, Microporous Mater. 1995, 5, 1.
[17] V. Alfredsson, M.W. Anderson, Chem. Mater. 1996, 8, 1141.
[18] H. Staub, I. Del Rosal, L. Maron, F. Kleitz, F.-G. Fontaine, J. Phys. Chem. C. 2012, 116, 25919.
[19] S.A. Bagshaw, E. Prouzet, T.J. Pinnavaia, Science 1995, 269, 1242.
[20] D. Zhao, J. Feng, Q. Huo, N. Melosh, G.H. Frederickson, B.F. Chmelka, G.D. Stucky, Science
1998, 279, 548.
[21] C.Z. Yu, Y. Yu, D.Y. Zhao, Chem. Commun. 2000, 575.
[22] T.W. Kim, F. Kleitz, B. Paul, R. Ryoo, J. Am. Chem. Soc. 2005, 127, 7601.
[23] Y. Wan, D.Y. Zhao, Chem. Rev. 2007, 107, 2821.
[24] F. Kleitz, Ordered Mesoporous Materials in Handbook of Heterogeneous Catalysis, 2nd edi-
tion, Chapter 8, 168. G. Ertl, H. Knötzinger, F. Schüth, J. Weitkamp, (Eds.), VCH-Wiley), Wein-
heim, 2008.
[25] G.S. Attard, J.C. Glyde, C.G. Göltner, Nature 1995, 378, 366.
[26] A. Monnier, F. Schüth, Q.S. Huo, D. Kumar, D. Margolese, R.S. Maxwell, G.D. Stucky,
M. Krishnamurty, P. Petroff, A. Firouzi, M. Janicke, B.F. Chmelka, Science 1993, 261, 1299.
[27] Q.S. Huo, D.I. Margolese, U. Ciesla, P. Feng, T.E. Gier, P. Sieger, R. Leon, P.M. Petroff, F.
Schüth, G.D. Stucky, Nature 1994, 368, 317.
[28] Q.S. Huo, D.I. Margolese, U. Ciesla, D.G. Demuth, P. Feng, T.E. Gier, P. Sieger, A. Firouzi,
B.F. Chmelka, F. Schüth, G.D. Stucky, Chem. Mater. 1994, 6, 1176.
[29] G.J. A.A. Soler-Illia, E.L. Crepaldi, D. Grosso, C. Sanchez, Curr. Opin. Colloid In. 2003, 8, 109.
[30] G.J. A.A. Soler-Illia, J. Patarin, B. Lebeau, C. Sanchez, Chem. Rev. 2002, 102, 4093.
[31] P.T. Tanev, T.J. Pinnavaia, Science 1995, 267, 865.
4 Ordered mesoporous silica: synthesis and applications | 95

[32] J. Patarin, B. Lebeau, R. Zana, Curr. Opin. Colloid In. 2002, 7, 107.
[33] Q.S. Huo, D.I. Margolese, G.D. Stucky, Chem. Mater. 1996, 8, 1147.
[34] D. Khushalani, A. Kuperman, G.A. Ozin, K. Tanaka, N. Coombs, M.M. Olken J. Garcés, Adv.
Mater. 1995, 7, 842.
[35] T.-W. Kim, R. Ryoo, M. Kruk, K.P. Gierszal, M. Jaroniec, S. Kamiya, O. Terasaki, J. Phys. Chem. B
2004, 108, 11480.
[36] F. Kleitz, W. Schmidt, F. Schüth, Microporous Mesoporous Mater. 2003, 65, 1.
[37] M. Kruk, M. Jaroniec, C.H. Ko, R. Ryoo, Chem. Mater. 2000, 12, 1961.
[38] M. Choi, W. Heo, F. Kleitz, R. Ryoo, Chem. Commun. 2003, 1340.
[39] E.G. Langeroudi, F. Kleitz, M.C. Iliuta, F. Larachi, J. Phys. Chem. C 2009, 113, 21866.
[40] V. Reichhardt, R. Guillet-Nicolas, M. Thommes, B. Klösgen, T. Nylander, F. Kleitz, V. Alfreds-
son, Phys. Chem. Chem. Phys., 2012, 14, 5651.
[41] F. Hoffmann, M. Cornelius, J. Morell, M. Fröba, Angew. Chem. Int. Ed. 2006, 45, 3216.
[42] Q. He, J. Shi, F. Chen, M. Zhu, L. Zhang, Biomaterials 2010, 31, 3335.
[43] Y.-S. Lin, C.L. Haynes, J. Am. Chem. Soc. 2010, 132, 4834.
[44] M. Al Shamsi, M.T. Al Samri, S. Al-Salam, W. Conca, S. Shaban, S. Benedict, S. Tariq, A.V.
Biradar, H.S. Penefsky, T. Asefa, A.-K. Souid, Chem. Res. Toxicol. 2010, 23, 1796.
[45] S. Al-Salam, G. Balhaj, S. Al-Hammadi, M. Sudhadevi, S. Tariq, A.V. Biradar, T. Asefa, A.-K.
Souid, Toxicol. Sci. 2011, 122, 86.
[46] S. Hitz, R. Prins, J. Catal. 1997, 168, 194.
[47] C.M. Yang, B. Zibrowius, W. Schmidt, F. Schüth, Chem. Mater. 2003, 15, 3739.
[48] B. Tian, X. Liu, C.Z. Yu, F. Gao, Q. Luo, S. Xie, B. Tu, D.Y. Zhao, Chem. Commun. 2002, 1186.
[49] S. Kawi, M.W. Lai, Chem. Commun. 1998, 1407.
[50] M.T.J. Keene, R. Denoyel, P.L. Llewellyn, Chem. Commun. 1998, 2203.
[51] A. Rumplecker, F. Kleitz, E.L. Salabas, F. Schüth, Chem. Mater. 2007, 19, 485.
[52] Q.S. Huo, R. Leon, P.M. Petroff, G.D. Stucky, Science 1995, 268, 1324.
[53] Y. Sakamoto, I. Diaz, O. Terasaki, D.Y. Zhao, J. Perez-Pariente, J.M. Kim, G.D. Stucky, J. Phys.
Chem. B 2002, 106, 3118.
[54] F. Kleitz, T. Czuryszkiewicz, L.A. Solovyov, M. Lindén, Chem. Mater. 2006, 18, 5070.
[55] F. Kleitz, S.H. Choi, R. Ryoo, Chem. Commun. 2003, 2136.
[56] J.R. Matos, M. Kruk, L.P. Mercuri, M. Jaroniec, L. Zhao, T. Kamiyama, O. Terasaki, T.J. Pinna-
vaia, Y. Liu, J. Am. Chem. Soc. 2003, 125, 821.
[57] J. Fan, C.Z. Yu, F. Gao, J. Lei, B. Tian, L. Wang, Q. Luo, B. Tu, W. Zhou, D.Y. Zhao, Angew. Chem.
Int. Ed. 2003, 42, 3146.
[58] F. Kleitz, D. Liu, G.M. Anilkumar, I.S. Park, L.A. Solovyov, A.N. Shmakov, R. Ryoo, J. Phys.
Chem. B 2003, 107, 14296.
[59] S.S. Kim, T.R. Pauly, T.J. Pinnavaia, Chem. Commun. 2000, 1661.
[60] S.S. Kim, A. Karkambar, T.J. Pinnavaia, M. Kruk, M. Jaroniec, J. Phys. Chem. B 2001, 105, 7663.
[61] D. Zhao, Q. Huo, J. Feng, B. F. Chmelka, G. D. Stucky, J. Am. Chem. Soc. 1998, 120, 6024.
[62] K. Cassiers, T. Linssen, M. Mathieu, M. Benjelloun, K. Schrijnemakers, P. van der Voort,
P. Cool, E.F. Vansant, Chem. Mater. 2002, 14, 2317.
[63] R. Ryoo, C.H. Ko, M. Kruk, V. Antochsuk, M. Jaroniec, J. Phys. Chem. B 2000, 104, 11465.
[64] M. Impéror-Clerc, P. Davidson, A. Davidson, J. Am. Chem. Soc. 2000, 122, 11925.
[65] P.I. Ravikovitch, A.V. Neimark, J. Phys. Chem. B 2001, 105, 6817.
[66] A. Galarneau, H. Cambon, F. DiRenzo, R. Ryoo, M. Choi, F. Fajula, New J. Chem. 2003, 27, 73.
[67] Y. Sakamoto, M. Kaneda, O. Terasaki, D.Y. Zhao, J.M. Kim, G.D. Stucky, H.J. Shin, R. Ryoo,
Nature 2000, 408, 449.
[68] Y. Sakamoto, T.W. Kim, R. Ryoo, O. Terasaki, Angew. Chem. Int. Ed. 2004, 42, 5231.
96 | Part I Functional materials: Synthesis and applications

[69] Y. Doi, A. Takai, Y. Sakamoto, O. Terasaki, Y. Yamauchi, K. Kuroda, Chem. Commun. 2010, 46,
[70] F. Kleitz, F. Bérubé, R. Guillet-Nicolas, C.M. Yang, M. Thommes, J. Phys. Chem. C 2010, 114,
[71] T.-W. Kim, L.A. Solovyov, J. Mater. Chem. 2006, 16, 1445.
[72] X. Liu, B. Tian, C. Yu, F. Gao, S. Xie, B. Tu, R. Che, L .-M. Peng, D. Zhao, Angew. Chem. Int. Ed.
2002, 41, 3876.
[73] Y.Q. Wang, C.M. Yang, B. Zibrowius, B. Spliethoff, M. Lindén, F. Schüth, Chem. Mater. 2003,
15, 5029.
[74] K. Flodström, V. Alfredsson, N. Källrot, J. Am. Chem. Soc. 2003, 125, 4402.
[75] M. Kruk, M. Jaroniec, A. Sayari, J. Phys. Chem. B 1997, 101, 583.
[76] M. Kruk, M. Jaroniec, R. Ryoo, S.H. Joo, Chem. Mater. 2000, 12, 1414.
[77] U. Ciesla, F. Schüth, Microporous Mesoporous Mater. 1999, 27, 131.
[78] J.M. Kim, Y. Sakamoto, Y.K. Hwang, Y.U. Kwon, O. Terasaki, S.-E. Park, G.D. Stucky J. Phys.
Chem. B 2002, 106, 2552.
[79] P. Kipkemboi, A. Fogden, V. Alfredsson, K. Flodström, Langmuir 2001, 17, 5398.
[80] K. Flodström, V. Alfredsson, Microporous Mesoporous Mater. 2003, 59, 167.
[81] K. Kushalani, A. Kuperman, G. A. Ozin, K. Tanaka, J. Garces, M. M Olken, N. Coombs, Adv.
Mater. 1995, 7, 842.
[82] M. Grün, K.K. Unger, A. Matsumoto, K. Tsutsumi, Microporous Mesoporous Mater. 1999, 27,
[83] A. Galarneau, M. Nader, F. Guenneau, F. DiRenzo, A. Gedeon, J. Phys. Chem. C 2007, 111,
[84] P.F. Fulvio, S. Pikus, M. Jaroniec, J. Mater. Chem. 2005, 15, 5049.
[85] J. Fan, C.Z. Yu, L. Wang, B. Tu, D.Y. Zhao, Y. Sakamoto, O. Terasaki, J. Am. Chem. Soc. 2001,
123, 12113.
[86] B. Coasne, A. Galarneau, R.J.M. Pellenq, F. Di Renzo, Chem. Soc. Rev. 2013, 42, 4141.
[87] C.J. Gommes, H. Friedrich, M. Wolters, P.E. de Jongh, K.P. de Jong, Chem. Mater. 2009, 21,
[88] H. Tüysüz, C.W. Lehmann, H. Bongard, B. Tesche, R. Schmidt, F. Schüth, J. Am. Chem. Soc.
2008, 130, 11510.
[89] M. Thommes, Chem. Ing. Tech. 2010, 82, 1059.
[90] R. Ryoo, J.M. Kim, Chem. Commun. 1995, 711.
[91] K.J. Edler, J.W. White, Chem. Mater. 1997, 9, 1226.
[92] R. Ryoo, S.H. Joo, J.M. Kim, J. Phys. Chem. B 1999, 103, 7435.
[93] S. Manet, J. Schmitt, M. Impéror-Clerc, V. Zholobenko, D. Durand, C.L.P. Oliviera, J.S. Peder-
sen, C. Gervais, N. Baccile, F. Babonneau, I. Grillo, F. Meneau, C.J. Rochas, J. Phys. Chem. B
2011, 115, 11330.
[94] R. Guillet-Nicolas, F. Bérubé, T.W. Kim, M. Thommes, F. Kleitz, Stud. Surf. Sci. Catal. 2008,
174, 141.
[95] B. Yang, C. Guo, S. Chen, J. Ma, J. Wang, X. Liang, L. Zheng, H. Liu, J. Phys. Chem. B 2006, 110,
[96] A. Kabalnov, U. Olsson, H. Wennerström, J. Phys. Chem. 1995, 98, 6220.
[97] C.Z. Yu, B. Tian, J. Fan, G.D. Stucky, D.Y. Zhao, J. Am. Chem. Soc. 2002, 124, 4556.
[98] C. Boissière, M.A.U. Martines, M. Tokumoto, A. Larbot, E. Prouzet, Chem. Mater. 2003, 15,
[99] M. Lindén, P. Ågren, S. Karlsson, P. Bussian, H. Amenitsch, Langmuir 2000, 16, 5831.
[100] F. Kleitz, J. Blanchard, B. Zibrowius, F. Schüth, P. Ågren, M. Lindén, Langmuir 2002, 18, 4863.
[101] J.M. Kim, S.K. Kim, R. Ryoo, Chem. Commun. 1998, 259.
4 Ordered mesoporous silica: synthesis and applications | 97

[102] K. Schumacher, M. Grün, K.K. Unger, Microporous Mesoporous Mater. 1999, 27, 201.
[103] J.S. Lettow, Y.J. Han, P. Schmidt-Winkel, P. Yang, D.Y. Zhao, Langmuir 2002, 16, 8291.
[104] J. Sun, H. Zhang, D. Ma, Y. Chen, X. Bao, A. Klein-Hoffmann, N. Pfander, D.S. Su, Chem.
Commun. 2005, 5343.
[105] P. Schmidt-Winkel, W.W. Lukens, D.Y. Zhao, P. Yang, B.F. Chmelka, G.D. Stucky, J. Am. Chem.
Soc. 1999, 121, 254.
[106] D.H. Chen, Z. Li, Y. Wan, X.J. Tu, Y.F. Shi, Z.X. Chen, W. Shen, C.Z. Yu, B. Tu, D.Y. Zhao, J. Mater.
Chem. 2006, 16, 1511.
[107] P. Holmqvist, P. Alexandridis, B. Lindman, Macromolecules 1997, 30, 6788.
[108] P. Holmqvist, P. Alexandridis, B. Lindman, J. Phys. Chem. B 1998, 102, 1149.
[109] F. Kleitz, T.W. Kim, R. Ryoo, Langmuir 2006, 22, 440.
[110] A. Taguchi, F. Schüth, Microporous Mesoporous Mater. 2005, 77, 1.
[111] F. Bérubé, B. Nohair, F. Kleitz, S. Kaliaguine, Chem. Mater. 2010, 22, 1988.
[112] V. Clauda, A. Schlossbauer, J. Kecht, A. Zurner, T. Bein, J. Am. Chem. Soc. 2009, 131, 11361.
[113] A. Stein, B.J. Melde, R.C. Schroden, Adv. Mater. 2000, 12, 1403.
[114] M.H. Lim, A. Stein, Chem. Mater. 1999, 11, 3285.
[115] J.A. Melero, R. van Grieken, G. Morales, Chem. Rev. 2006, 106, 3790.
[116] E.L. Burkett, S.D. Sims, S. Mann, Chem. Commun. 1996, 1367.
[117] D.J. Macquarrie, Chem. Commun. 1996, 1961.
[118] D. Margolese, J.A. Melero, S.C. Christiansen, B.F. Chmelka, G.D. Stucky, Chem. Mater. 2000,
12, 2448.
[119] H. Yen, Y. Seo, R. Guillet-Nicolas, S. Kaliaguine, F. Kleitz, Chem. Commun. 2011, 47, 10473.
[120] Z.H. Luan, E.M. Maes, P.A.W. van der Heide, D.Y. Zhao, R.S. Czernuszewicz, L. Kevan, Chem.
Mater. 1999, 11, 3680.
[121] M. Choi, F. Kleitz, D.N. Liu, H.Y. Lee, W.S. Ahn, R. Ryoo, J. Am. Chem. Soc. 2005, 127, 1924.
[122] R. Guillet-Nicolas, J.-L. Bridot, Y. Seo, M.-A. Fortin, F. Kleitz, Adv. Funct. Mater. 2011, 21, 4653.
[123] R.C. Hayward, P. Alberius-Henning, B.F. Chmelka, G.D. Stucky, Microporous Mesoporous
Mater. 2001, 44-45, 619.
[124] S.H. Tolbert, A. Firouzi, G.D. Stucky, B.F. Chmelka, Science 1997, 278, 264.
[125] P. Yang, T. Deng, D. Zhao, P. Feng, D. Pine, B.F. Chmelka, G.M. Whitesides, G.D. Stucky,
Science 1998, 282, 2244.
[126] H. Miyata, K. Kuroda, Chem. Mater. 2000, 12, 49.
[127] H. Fan, S. Reed, T. Bear, R. Schunk, G. P. López, C.J. Brinker, Microporous Mesoporous Mater.
2001, 44-45, 625.
[128] J.-H. Smått, S. Schunk, M. Lindén Chem. Mater. 2003, 15, 2354–2361.
[129] F. Maddox Sayler, M. G. Bakker, J.-H.Smått, M.Lindén, J. Phys. Chem. C 2010, 114, 8710.
[130] P. Yang, D.Y. Zhao, B.F. Chmelka, G.D. Stucky, Chem. Mater. 1998, 10, 2033.
[131] N.A. Melosh, P. Lipic, F.S. Bates, F. Wudl, G.D. Stucky, G.H. Fredrickson, B.F. Chmelka, Macro-
molecules 1999, 32, 4332.
[132] K. Yano, Y. Fukushima, J. Mater. Chem. 2004, 14, 1579.
[133] C. Boissière, D. Grosso, A. Chaumonnot, L. Nicole, C. Sanchez, Adv. Mater. 2011, 23, 599.
[134] C. Sanchez, C. Boissière, D. Grosso, C. Laberty, L. Nicole, Chem. Mater. 2008, 20, 682.
[135] F. Kleitz, F. Marlow, G.D. Stucky, F. Schüth, Chem. Mater. 2001, 13, 3587.
[136] S.-H. Wu, C.-H. Mou, H.-P. Lin, Chem. Soc. Rev. 2013, 42, 3862.
[137] Q. Huo, J. Feng, F. Schüth, G. D. Stucky, Chem. Mater. 1997, 9, 14.
[138] W. Stöber, A. Funk, E. Bohn, J. Colloid Interfaces Sci. 1968, 26, 62.
[139] T. Martin, A. Galarneau, F. DiRenzo, F. Fajula, D. Plee, Angew. Chem. Int. Ed. 2002, 41, 2590.
[140] A. Galarneau, J. Iapichella, K. Bonhomme, F. DiRenzo, P. Kooyman, O. Terasaki, F. Fajula, Adv.
Funct. Mater. 2006, 16, 1657.
98 | Part I Functional materials: Synthesis and applications

[141] P.J. Bruinsma, A.Y. Kim., J. Liu, S. Baskaran, Chem. Mater. 1997, 9, 2507.
[142] I.I. Slowing, B.G. Trewyn, S. Giri, V.S.-Y. Lin, Adv. Funct. Mater. 2007, 17, 1225.
[143] M. Grün, I. Lauer, K.K. Unger, Adv. Mater. 1997, 9, 254.
[144] Q. Cai, Z.-S. Luo, W.-Q. Pang, Y.-W. Fan, X.-H. Chen, F.-Z. Cui, Chem. Mater. 2001, 13, 258.
[145] C.E. Fowler, D. Khushalani, B. Lebeau, S. Mann, Adv. Funct. Mater. 2001, 13, 649.
[146] B.G. Trewyn, I. Slowing, S. Giri, H.-T. Chen, V.S.-Y. Lin, Acc. Chem. Res. 2007, 40, 846.
[147] T.-W. Kim, P.-W. Chung, I.I. Slowing, M. Tsunoda, E.S. Yeung, V.S.-Y. Lin, Nano Lett. 2008, 8,
[148] T.-W. Kim, P.-W. Chung, V.S.-Y. Lin, Chem. Mater. 2010, 22, 5093.
[149] M.-H. Kim, H.-K. Na, Y.-K. Kim, S.-R. Ryoo, H.S. Cho, K.E. Lee, H. Jeon, R. Ryoo, D.-H. Min,
ACS Nano 2011, 5, 3568.
[150] Q. He, J. Shi, J. Zhao, Y. Chen, F. Chen, J. Mater. Chem. 2009, 19, 6498.
[151] T.W. Kim, I.I. Slowing, P.-W. Chung, V.S.Y. Lin, ACS Nano 2011, 5, 360.
[152] P. Linton, V. Alfredsson, Chem. Mater. 2008, 20, 2878.
[153] Y. Han, J.Y. Ying, Angew. Chem. Int. Ed. 2005, 44, 288.
[154] F. Gao, P. Botella, A. Corma, J. Blesa, L. Dong, J. Phys. Chem. B 2009, 113, 1796.
[155] S.B. Hartono, W. Gu, F. Kleitz, J. Liu, L. He, A.P.J. Middelberg, C.Z. Yu, G.Q. Lu, S.Z. Qiao,
ACS Nano 2012, 6, 2104.
[156] M.E. Davis, Nature 2002 , 417, 813.
[157] A. Corma, Chem. Rev. 1997, 97, 2373.
[158] C. Perego, R. Millini, Chem. Soc. Rev. 2013, 42, 3956.
[159] A. Walcarius, Chem. Soc. Rev. 2013, 42, 4098.
[160] A. Bernardos, E. Aznar, M.D. Marcos, R. Martínez-Máñez, F. Sancenón, J. Soto, J.M. Barat,
P. Amorós, Angew. Chem. Int. Ed. 2009, 48, 5884.
[161] F. Muhammad, M. Guo, W. Qi, F. Sun, A. Wang, Y. Guo, G. Zhu, J. Am. Chem. Soc. 2011, 133,
[162] E. Aznar, M.D. Marcos, R. Martínez-Máñez, F. Sancenón, J. Soto, P. Amorós, C. Guillem,
J. Am. Chem. Soc. 2009, 131, 6833.
[163] C. Coll, L. Mondragón, R. Martínez-Máñez, F. Sancenón, M. D. Marcos, J. Soto, P. Amorós,
E. Pérez-Payá, Angew. Chem. Int. Ed. 2011, 50, 2138.
[164] J.M. Rosenholm, C. Sahlgren, M. Lindén, Curr. Drug Targets 2011, 12, 1166.
[165] A.E. Garcia-Bennett, Nanomedecine 2011, 6, 867.
[166] V. Mamaeva, C. Sahlgren, M. Lindén, Adv. Drug Deliver. Rev. 2013, 65, 689.
[167] D. Tarn, C.E. Ashley, M. Xue, E.C. Carnes, J.I. Zink, C.J. Brinker, Accounts Chem. Res. 2013, 46,
[168] W.X. Mai, H. Meng, Integr. Biol. 2013, 5, 19.
[169] N.-T. Chen, S.-H. Cheng, J.S. Souris, C.-T. Chen, C.-Y. Mou, L.-W. Lo, J. Mater. Chem. B 2013, 1,
[170] F. Tang, L. Li, S. Chen, Adv. Mater. 2012, 24, 1504.
[171] Y.-W. Yang, Med. Chem. Commun. 2011, 2, 1033.
[172] P. Yang, S. Gai, J. Lin, Chem. Soc. Rev. 2012, 41, 3679.
[173] C. Park, K. Oh, S.C. Lee, C. Kim, Angew. Chem. Int. Ed. 2007, 46, 1455.
[174] F. Muhammad, M. Guo, W. Qi, F. Sun, A. Wang, Y. Guo, G. Zhu, J. Am. Chem. Soc. 2011, 133,
[175] J.M. Rosenholm, V. Mamaeva, C. Sahlgren, M. Lindén, Nanomedecine 2012, 7, 111.
[176] M. Hartmann, Chem. Mater. 2005, 17, 4577.
[177] J. Andersson, J. Rosenholm, S. Areva, M. Lindén, Chem. Mater. 2004, 16, 4160.
[178] A. Datt, I. El-Maazawi, S. C. Larsen, J. Phys. Chem. C 2012, 116, 18358.
[179] J.M. Rosenholm, C. Sahlgren, M. Lindén, Nanoscale 2010, 2, 1870.
4 Ordered mesoporous silica: synthesis and applications | 99

[180] R. Langer, Nature 1998, 392, S5.

[181] F. Tang, L. Li, D. Chen, Adv. Mater. 2012, 24, 1504.
[182] Q. He, J. Shi, Adv. Mater. 2014, 26, 391.
[183] L. Pan, Q. He, J. Liu, Y. Chen, M. Ma, L. Zhang, J. Shi, J. Am. Chem. Soc. 2012, 134, 5722.
[184] N.K. Mal, M. Fujiwara, Y. Tanaka, Nature 2003, 421, 350.
[185] S. Angelos, N.M. Khashab, Y.-W. Yang, A. Trabolsi, H.A. Khatib, J.F. Stoddart, J.I. Zink,
J. Am. Chem. Soc. 2009, 131, 12912.
[186] J.-H. Park, Y.-H. Lee, S.-G. Oh, Macromol. Chem. Phys. 2007, 208, 2419.
[187] A. Popat, J. Liu, G.Q. Lu, S.Z. Qiao, J. Mater. Chem. 2012, 22, 11173.
[188] A. Popat, B.P. Ross, J. Liu, S. Jambhrunkar, F. Kleitz, S.Z. Qiao Angew. Chem. Int. Ed. 2012, 51,
[189] A. Popat, S.B. Hartono, F. Stahr, J. Liu, S.Z. Qiao, C.Q. Lu, Nanoscale 2011, 3, 2801.
[190] M. Manzano, V. Aina, C.O. Aréan, F. Balas, V. Cauda, M. Colilla, M.R. Delgado, M. Vallet-Regì,
Chem. Eng. J. 2008, 137, 30.
[191] M. Van Speybroeck, R. Mellaerts, T.D. Thi, J.A. Martens, J. Van Humbeeck, P. Annaert, G. Van
Den Mooter, P. Augustijns, J. Pharm. Sci. 2011, 100, 4864.
[192] I.-J. Fang, B.G. Trewyn, Nanomedecine 2012, 508, 41.
[193] H. Chen, R. Langer, Adv. Drug Deliver. Rev. 1998, 34, 339
[194] L.S. Liu, M.L. Fishman, J. Kost, K.B. Hicks, Biomaterials 2003, 24, 3333.
[195] S.W. Song, K. Hidajat, S. Kawi, Chem. Commun. 2007, 4396.
[196] US Pharmacopoeia and national formulary, Drug release, delayed-release (enteric-coated)
articles and general drug release standard, 2017, Rockville, MD: USP Convention Inc., 2002.
[197] C.-H. Lee, L.-W. Lo, C.-Y. Mou, C.-S. Yang, Adv. Funct. Mater. 2008, 18, 3283.
[198] S.-H. Cheng, W.-N. Liao, L.-M. Chen, C.-H. Lee, J. Mater. Chem. 2011, 21, 7130.
[199] G. Sun, Y. Chang, S. Li, Q. Li, R. Xu, J. Gu, E. Wang, Dalton Trans. 2009, 38, 4481.
[200] S. Angelos, E. Choi, F. Vögtle, L. De Cola, J. I. Zink, J. Phys. Chem. C 2007, 111, 6589.
[201] Y. Ma, L. Zhou, H. Zheng, L. Xing, C. Li, J. Cui, S. Che, J. Mater. Chem. 2011, 21, 9483.
[202] D.M.R. Georget, S.A. Barker, P.S. Belton, Eur. J. Pharm. Biopharm. 2008, 69, 718.
[203] X. Liu, Q. Sun, H. Wang, L. Zhang, J.-Y. Wang, Biomaterials 2005, 26, 109.
[204] X. Zhang, M. Oulad-Abdelghani, A.N. Zelkin, Y. Wang, Y. Haîkel, D. Mainard, J.-C. Voegel,
F. Caruso, N. Benkirane-Jessel, Biomaterials 2010, 31, 1699.
[205] M. Subirade, J. Gueguen, K.D. Schwenke, J. Colloid Interface Sci. 1992, 152, 442.
[206] R. Guillet-Nicolas, A. Popat, J.-L. Bridot, G. Monteith, S.Z. Qiao, F. Kleitz, Angew. Chem. Int.
Ed. 2013, 52, 2318.
[207] X. Feng, G.E. Fryxell, L.Q. Wang, A.Y. Kim, J. Liu, K.M. Kemner, Science 1997, 276, 923.
[208] L. Bois, A. Bonhommé, A. Ribes, B. Pais, G. Raffin, F. Tessier, Colloids Surf. A 2003, 221, 221.
[209] K.F. Lam, K. L. Yeung, G. McKay, Environ. Sci. Technol. 2007, 41, 3329.
[210] P. Makowski, X. Deschanels, A. Grandjean, D. Mayer, G. Toquer, F. Goettmann, New. J. Chem.
2012, 36, 531.
[211] P. Makowski, A. Thomas, P, Kuhn, F. Goettmann, Energy Environ. Sci. 2009, 2, 480.
[212] G.E. Fryxell, Y. Lin, S. Fiskum, J.C. Birnbaum, H. Wu, K. Kemner, S. Kelly, Environ. Sci. Technol.
2005, 39, 1324.
[213] D. Butler, L. Stricker, Nature 2011, 472, 274.
[214] P.J. Lebed, K. de Souza, F. Bilodeau, D. Larivière, F. Kleitz, Chem. Commun. 2011, 47, 11525.
[215] P.J. Lebed, J.-D. Savoie, J. Florek, F. Bilodeau, D. Larivière, F. Kleitz, Chem. Mater. 2012, 24,
[216] X. Kang, C. Li, Z. Cheng, P. Ma, Z. Hou, J. Lin, WIREs Nanomed Nanobiotechnol. 2014, 6, 80.
[217] J. Feng, H. Zhang, Chem. Soc. Rev. 2013, 42, 387.
100 | Part I Functional materials: Synthesis and applications

[218] M.S. Moorthy, P.K. Tapaswi, S.S. Park, A. Mathew, H.-J. Cho, C.-S. Ha, Microporous Meso-
porous Mater. 2013, 180, 162.
[219] S. Massari, M. Ruberti, Resources Policy 2013, 38, 36.
[220] C.K. Gupta, N. Krishnamurthy, Int. Mater. Rev. 1992, 37, 197.
[221] J. Florek, F. Chalifour, F. Bilodeau, D. Larivière, F. Kleitz, Adv. Funct. Mater. 2014, 24, 2668.
[222] M. Khayet, Adv. Colloid Interface Sci. 2011, 164, 56.
[223] M.S. El-Bourawi, Z. Ding, R. Ma, M. Khayet, J. Membr. Sci. 2006, 285, 4.
[224] Z. Steiner, J. Miao and R. Kasher, Chem. Commun. 2011, 47, 2384.
[225] J.-P. Mericq, S. Laborie, C. Cabassud, Water Res. 2010, 44, 5260.
[226] H.L. Castricum, A. Sah, R. Kreiter, D.H.A. Blank, J.F. Vente, J.E. Ten Elshof, Chem. Commun.
2008, 1103.
[227] N. Liu, R. A. Assink, C.J. Brinker, Chem. Commun. 2003, 370.
[228] Y.T. Chua, C.X.C. Lin, F. Kleitz, X.S. Zhao, S. Smart, Chem. Commun. 2013, 49, 4534.
A. Ritcey
5 Nanoparticles: Properties and applications
5.1 Introduction

The ever-increasing importance of nanoparticles to the development of functional ma-

terials is incontestable. A simple keyword search of the scientific literature [1] reveals
the accelerating growth of nanoparticle research over the past decade and, as illus-
trated in Fig. 5.1, nearly 60 000 publications treating this subject appeared in 2012
alone. New scientific journals dedicated to the field of nanoscience have also been
created [2] and are rapidly establishing their place among high impact periodicals.


Number of publications





2000 2002 2004 2006 2008 2010 2012
Publication year

Fig. 5.1. Annual number of publications retrieved by a keyword search for “nanoparticles” [1].

Within this context, the following pages can only scratch the surface of the title subject
and cannot in any way be considered a complete review of the literature. It is hoped,
however, that the selected examples will provide the reader with a valid apprecia-
tion of the enormous potential of functional nanoparticles in materials science. In ad-
dition, this chapter treats some fundamental issues of nanoparticle preparation and
handling through a general approach that is relevant to many diverse specific systems.

5.2 Synthetic methods

Most introductions to the fabrication of nanostructures begin with a reference to the

contrasting “top-down” and “bottom-up” approaches, terms first applied to the field
102 | Part I Functional materials: Synthesis and applications

of nanoscience by the Foresight Institute in 1989 [3]. As their name implies, top-down
methods involve the creation of nanosized entities from larger blocks of matter. Exam-
ples include mechanical size reduction by crushing and grinding as well as more so-
phisticated lithographic techniques. Bottom-up approaches, on the other hand, seek
to build nanostructures from smaller components, typically atoms or molecules, and
therefore frequently involve elements of self-assembly or supramolecular chemistry.
In the case of nanoparticles, the large majority of preparation methods are based on
controlled precipitation, that is, the nucleation and growth of particles from a super-
saturated solution. Supersaturation can be achieved by a chemical reaction, such as
reduction, hydrolysis, condensation or decomposition, which converts a soluble pre-
cursor to the desired, sparingly soluble, material.

5.2.1 Particle nucleation and growth

Any solution process that leads to the generation of a sparingly soluble species will
result in the eventual nucleation of a second, typically solid, phase. The thermody-
namics of nucleation includes two considerations: the decrease in free energy associ-
ated with the phase transition (usually crystallization), and the increase in free energy
associated with the necessary creation of the solid-liquid interface. The free energy of
nucleation can therefore be written as

ΔGnucleation = n ⋅ ΔḠ crystallization + ΔGsurface , (5.1)

where ΔḠ crystallization is the molar free energy of crystallization and n is the number of
moles of the substance i in the nucleus. ΔḠ crystallization depends on the concentration
of i, [i], according to
ΔḠ crystal = RT ln , (5.2)
and thus becomes increasingly negative as the concentration exceeds saturation at
[i]sat .
The second contribution to ΔGnucleation can be expressed as

ΔGsurface = σ ⋅ A, (5.3)

where σ is the interfacial energy and A is the surface area of the nucleus. This compo-
nent is always positive. Nucleation can only occur if the crystallization process liber-
ates sufficient energy to cover the cost of creating the new interface between the par-
ticle seed and the solution. Importantly, although the absolute value of both ΔGsurface
and ΔGcrystallization increase with nucleus size, the former is proportional to the surface
area, whereas the latter is proportional to the volume. In the simple case of a spherical
nucleus of radius r, ΔGsurface varies with r2 and ΔGcrystallization with r3 . These dependen-
cies (dashed lines) are illustrated schematically in Fig. 5.2 and lead to the presence
5 Nanoparticles: Properties and applications | 103

of a maximum, labeled rcr , in the sum of the two contributions. Nuclei smaller than
the critical size rcr are thermodynamically unstable and will spontaneously disappear
as indicated by the left-pointing arrow. Nuclei larger than the critical size, however,
will spontaneously grow, since in this size regime increasing size corresponds to a
decrease in free energy, as indicated by the arrow to the right.


Fig. 5.2. Free energy of nucleation as a function

rcr Nucleus
size of nucleus size. Contributions from surface and
crystallization energies are shown as dashed
lines. rcr denotes the critical nucleus size and the
ΔGcrystallization arrows indicate the direction of spontaneous size

How can particle nucleation occur if small nuclei are unstable and disappear rather
than grow? The answer lies in random concentration fluctuations which allow for the
spontaneous formation of aggregates that, upon formation, already exceed the crit-
ical size. Clearly, the probability of the formation of stable nuclei will therefore de-
crease with increasing critical size. As the concentration is increased beyond satura-
tion, ΔGcrystallization becomes increasingly negative, permitting the formation of stable
nuclei of smaller size. Nucleation, in turn, becomes more probable. These thermo-
dynamic considerations explain the general behavior illustrated by the well-known
LaMer diagram provided in Fig. 5.3 [4, 5].
The process of nanoparticle formation can be divided into three phases. In the
early stages, immediately following the initiation of the reaction responsible the gen-
eration of the particle forming species, the concentration is lower than the solubility
and there is thus no thermodynamic driving force for the formation of a second phase.
As the reaction proceeds, the concentration increases, eventually reaching saturation.
Particle nucleation, however, does not occur immediately because of the free energy
cost of creating a new interface. As outlined above, nucleation requires that the so-
lution become super-saturated to the point where ΔḠ crystallization , as given by equa-
tion (5.2), is sufficient to cover the cost of the new surface. This point is labeled the
critical nucleation concentration in Fig. 5.3. Once the critical concentration is reached,
nucleation begins, removing material from solution. As nucleation proceeds, the so-
lution concentration decreases and eventually falls below the critical value for nu-
clei formation. After this point, no new nuclei are formed and any material remaining
in solution or generated by further reaction can only contribute to particle growth.
The concentration profile of the LaMer diagram has important implications for the
104 | Part I Functional materials: Synthesis and applications


Critical nucleation


Pre- Nucleation Time

Fig. 5.3. LaMer diagram illustrating the evolution of solution concentration during the nucleation
and growth process.

size distribution of the resulting particle population. Conditions which restrict the du-
ration of the nucleation stage favor populations of low polydispersity. Simultaneous
nucleation, followed by growth in a common medium, will lead to particles of quasi-
identical size.

5.2.2 Synthesis in inverse micelles

In the presence of an appropriate surfactant, small amounts of water can be dispersed

in an organic medium to form inverse micelles. As illustrated in Fig. 5.4, an inverse
micelle can be viewed as a nanometer-sized water droplet surrounded by a surfactant
layer. The capacity to dissolve reactants in the water core, combined with the constant
exchange of the aqueous phase between micelles during collisions, allows for chemi-
cal reactions to be carried out in the confined volume of the enclosed water reservoir.
In general, if the reaction leads to a solid product, nanoparticles are obtained. This
method has been employed for the preparation of nanoparticles from a diverse variety
of materials, including metals, metal oxides, and even organic compounds [6]. The in-
verse micelle approach typically offers exceptional control of particle size, which can
be conveniently modulated through the variation of simple system parameters such
as the relative quantities of solvent, surfactant, and reagents. Furthermore, in certain
cases, essentially monodisperse size-distributions are obtained. For example, we have
employed the inverse micelle approach to prepare single crystal yttrium fluoride nano-
particles with the exquisite size control illustrated in Fig. 5.5 [7, 8].
5 Nanoparticles: Properties and applications | 105

H2 O
Fig. 5.4. Schematic representa-
tions of (a) an inverse micelle,
and (b) content exchange during
(a) (b) micellar collisions.

Fig. 5.5. (H3 O)Y3 F10 nanocrystals prepared by the in-

50 nm X300000 1300x1030 pixels 4/29/2004
4.16852 s
verse micelle method [7, 8].

5.3 Particle aggregation and stabilization of

colloidal suspensions

The free energy of the interface between two immiscible phases is always positive. If
this were not the case, the free energy of the system would decrease with increasing
interfacial area and the phases would spontaneously divide into progressively smaller
and smaller domains, leading to miscibility. Since interfacial energy is positive, sus-
pensions of nanoparticles are necessarily thermodynamically unstable. Kinetically
stable colloidal suspensions can, however, be obtained by ensuring sufficient inter-
particle repulsion. Common stabilization strategies fall into two main categories; elec-
trostatic and steric.
Electrostatic stabilization involves the introduction of a surface charge to the
nanoparticles and is primarily restricted to aqueous systems, since solvents other
than water are unable to support the necessary charge separation. Many inorganic
nanoparticles, such as those composed of sparingly soluble salts or metal oxides, are
naturally charged when dispersed in water. Surface charge is frequently determined
by equilibria controlling ion adsorption and dissolution or the protonation and depro-
tonation of surface moieties. In other cases, surface charge can be introduced by the
adsorption of charged species, such as the citrate ion. Electrostatic repulsion between
particles of like charge serves as a barrier to aggregation and, if sufficiently strong
with respect to the van der Waals attraction, can result in a stable suspension.
The interplay of van der Waals attractions and electrostatic repulsion was quan-
titatively captured by Derjaguin, Verwey, Landau, and Overbeek in what is commonly
106 | Part I Functional materials: Synthesis and applications

known as DVLO theory [9]. The basic outcome of DVLO calculations is illustrated in
Fig. 5.6. Both electrostatic repulsion, shown as the dashed line, and van der Waals
attractions, represented by the dotted line, become stronger as particles are brought
closer together. Importantly, the exact way in which each of the two forces varies with
separation distance is different: interparticle van der Waals attractions vary inversely
to distance squared, whereas electrostatic repulsion follows an exponential depen-
dency. The net force is weakly attractive at large distances. However, as the particles
approach each other, electrostatic repulsion increases faster than the attractive force.
This leads to the appearance of a repulsive barrier in the net interaction curve. Parti-
cles able to cross this barrier will reach very small separation distances where attrac-
tive forces dominate and aggregation will occur.

Electrostatic repulsion

Repulsive barrier
Interaction energy

Van der Waals attraction

Interparticle distance

Fig. 5.6. Interaction energy between two particles as a function of separation. The net interaction,
shown as the solid line, is the sum of electrostatic repulsion and attractive van der Waals forces.

The stability of a colloidal suspension therefore depends on the height of the repul-
sive barrier with respect to the energy of particle collisions. If particles collide with
sufficient energy to pass the repulsive barrier, they will fall into the attractive well and
aggregate. Once in contact, there is no driving force for spontaneous re-dispersion,
and in this respect, aggregation is irreversible. Collision energy is determined by the
kinetic energy of the colliding particles, which depends on temperature and follows
a Boltzman distribution centered at 3/2 kT. Importantly, at a given temperature, colli-
sions are not all of the same energy and there is a nonzero probability that some will
have sufficient energy to pass the barrier. In fact, given enough time, all of the particles
in a suspension will eventually reach the aggregated state. The stability of a colloidal
suspension therefore depends on the time scale of observation, and for most applica-
tions colloids are typically considered to be kinetically stable if the repulsive barrier
5 Nanoparticles: Properties and applications | 107

exceeds 10 kT. The thermodynamically stable state remains the aggregated one; the
repulsive barrier simply serves to delay the inevitable.
The second common strategy for stabilization employs steric repulsion. In this
approach, polymer chains are attached to the particle surface, either through chem-
ical grafting or simple adsorption. As two particles come together, a repulsive force
will be felt at the point of chain overlap. Although conceptually simple, as illustrated
by Fig. 5.7, repulsive contributions from both osmotic pressure and chain deforma-
tion must be taken into account, and the quantitative treatment of steric stabilization
rapidly becomes complex [10]. Steric stabilization can be employed in both aqueous
and organic media, and with the appropriate choice of surface ligand nanoparticles
can be dispersed in any carrier liquid.

Fig. 5.7. Schematic illustration of steric stabilization

through surface grafting of polymer chains.

5.4 Colloidal quantum dots

Among the panoply of functional nanoparticles, it is probably the family known as

quantum dots which exhibits the most striking example of size-dependent proper-
ties. Quantum dots are nanocrystals of inorganic semi-conductor materials which pos-
sess unique optical properties arising from the electronic confinement imposed by the
small size of the particles [11, 12].
The electronic structure of bulk semiconductors is characterized by two distinct
bands, the valence band which contains the electrons that bond the material together,
and, at higher energy, the conduction band, which is essentially devoid of electrons at
low temperature. Energies between the valence band and the conduction band, that
is, within the so-called band gap, are forbidden. Electrons can, however, be promoted
from the valence band across the band gap into the conduction band, if supplied with
sufficient energy. For a bulk semiconductor, optical absorption will therefore begin at
the band gap energy and, because of the band nature of the excited state, the optical
absorption spectrum will appear as a continuum above the band gap. In contrast to the
band structure of bulk semiconductors, quantum dots are characterized by a series of
discrete electronic quantized levels, and their optical spectra correspondingly exhibit
discrete electronic transitions.
The promotion of an electron from the valence band to the conduction band re-
sults in the creation of an electron-hole pair known as an exciton. The electron-hole
pair is bound by Coulombic interactions which are relatively small when compared
108 | Part I Functional materials: Synthesis and applications

to kT and, in a bulk semiconductor, considerable charge separation is possible. In

the case of a quantum dot, particle size is reduced below the characteristic distance
of charge separation for the electron-hole pair and confinement effects become pro-
nounced [13]. The energies of the electron and hole levels are sensitive to the degree of
confinement and, as a result, the optical absorption spectra of quantum dots depend
strongly on their size. As the exciton becomes increasingly confined, the energy of the
primary transition increases and the corresponding absorption band moves to shorter
wavelengths, as illustrated for PbSe in Fig. 5.8.

8.1 nm
Absorbance (arbitrary units)

6.5 nm

4.8 nm

4.6 nm

3.6 nm

3.3 nm
800 1200 1600 2000 2400 Fig. 5.8. Absorption spectra of PbSe quantum dots of
Wavelength (nm) various sizes. Reprinted with permission from [13].

Quantum dots have received considerable attention as functional materials because

of their luminescence properties. Here, exciton confinement also plays an important
role by greatly enhancing the probability of radiative recombination. Radiative recom-
bination is the return of an electron from the conduction band to the valence band ac-
companied by emission of a photon. The wavelength of emission is thus determined
by the band gap and is highly size-dependent. This is illustrated in Fig. 5.9 for a series
of CdSe quantum dots, removed from the reaction medium at different times during
particle synthesis [14]. As the particles grow larger, the emission spectra systemati-
cally shift to longer wavelengths.
Early research on quantum dots primarily involved materials exhibiting absorp-
tion and emission within the visible region of the spectrum. As recently discussed
by Shirasaki et al. [15], quantum dots show great potential as emitters in thin film
light-emitting devices. For this application, quantum dots offer several important ad-
vantages with respect to organic fluorophores. For example, the characteristic narrow
emission band of quantum dots assures high color purity, and the tunability of the
5 Nanoparticles: Properties and applications | 109

0.5 min
10 min
15 min

Normalized PL Intensity
60 min
120 min

400 450 500 550 600 650 700
Wavelength (nm)

Fig. 5.9. Temporal evolution of the photoluminescence spectra of crude solutions of colloidal
CdSe nanocrystals during their growth at 280°C. Adapted with permission from [14].

emission color through particle size allows for the fabrication of multicolor devices
from a single material. More recently, research activity has been increasingly focused
on the development of narrow band gap materials which emit in infrared [16–18]. Cur-
rent efforts aim at infrared emission for applications in bioimaging, telecommunica-
tions, and solar cells.
A solar cell is a device which converts the sun’s energy to electricity. The avail-
ability of efficient solar cells, amenable to large scale production at reasonable cost,
will clearly have major positive repercussions for the environmental impact of future
human activity and economic development. Numerous designs for solar cells have
been proposed and all involve the following three basic processes: the absorption of
light to generate electron-hole pairs, the separation of charge carriers, and the cap-
ture of the charges by external electrodes. One of the primary limitations of solar cell
performance is the low efficiency of the primary photoconversion step [13]. The solar
spectrum is relatively broad, and only photons absorbed by the solar cell can lead to
charge separation. The incorporation of narrow gap quantum dots in hybrid solar cells
can thus lead to improvements in light harvesting in the infrared region [19]. Quantum
dots are also being developed for novel solar energy conversion strategies. For exam-
ple, cell performance can be enhanced through a process known as multiple exciton
generation [20]. In this process, single photons with energy surpassing twice that of
the band gap can produce two or more electron-hole pairs. This leads to improved cell
efficiency, since in the absence of multiple exciton generation, the excess energy of
high energy photons is lost as heat. Significantly, the confinement of charge carriers
110 | Part I Functional materials: Synthesis and applications

greatly enhances the probability of multiple exciton generation, leading to exciting

prospects for the application of quantum dots in next generation solar cells [20].
Clearly the role of quantum dots as functional materials extends beyond the lim-
ited examples cited here. Furthermore, current research involves particles of increas-
ingly complex composition and structure [18], as well as new paradigms, such as the
control of properties through ligand-mediated processes [21]. The interested reader is
encouraged to consult more complete treatises on the subject.

5.5 Metal nanoparticles

Metal nanoparticles are ubiquitous and have been known since antiquity. The unique
optical properties of colloidal noble metals, exploited in ancient times to stain glass,
are today at the basis of exciting new developments in both fundamental science and
the design of novel optical materials.
The interaction of light with metallic nanostructures is dominated by the collec-
tive excitation of free electrons, known as the surface plasmon; illustrated in Fig. 5.10.
The optical excitation of plasmon is the most efficient process by which light interacts
with matter [22] and confers the ability to concentrate and manipulate light at dimen-
sions below the diffraction limit onto metal nanostructures. Plasmonics is a rapidly
growing field of research and has recently been identified [23] as a potential pervasive
technology, offering opportunities to achieve unprecedented optical functionalities.
Electric field

Fig. 5.10. Schematic representation of the polarization of metal nanoparticles

by excitation of the surface plasmon.

Although light of all frequencies can induce electron oscillations in metals, the inter-
action is particularly efficient at the characteristic plasmon resonance frequency. Plas-
mon frequency depends on a variety of factors, including the identity of the metal, the
dielectric constant of the surrounding medium, and particle size [24–26]. The size de-
pendence of plasmon frequency is illustrated in Fig. 5.11 for a series of gold particles.
The surface plasmons of neighboring nanoparticles in close proximity can couple, and
for this reason assemblies of metal nanoparticles have plasmon frequencies which dif-
fer from those of the isolated constituent particles. The coupled plasmon resonance
5 Nanoparticles: Properties and applications | 111

60 nm

8 nm

4.6 nm
120 nm
Absorbance (a.u.)

2.5 nm

Fig. 5.11. Plasmon extinction spectra

of aqueous suspensions of gold nano-
300 400 500 600 700 particles of varying size. Adapted with
Wavelength (nm) permission from [26].

depends on the distance of separation in a known way, and can thus be exploited to
probe nanosystems through a novel application known as the plasmon ruler [27–29].
Such a ruler allows determination of separation distances in media (solution) which
are not accessible with electron or scanning probe microscopy.
The surface plasmon is probably best known for its role in surface-enhanced spec-
troscopies. The surface oscillation of electric charges generated by plasmon excita-
tion leads to significant enhancement in the local electromagnetic field [30], which in
turn modifies the probability of optical transitions for molecules located within the
affected region. For example, in the case of surface-enhanced Raman scattering, nor-
mally weak Raman signals can be enhanced by many orders of magnitude [31], leading
to such exceptional sensitivities that single molecule detection becomes possible [32].
The intense local fields generated by excitation of nanoparticle plasmon can
also modify the fluorescence properties of nearby molecules through a phenomenon
known as metal-enhanced fluorescence [33]. The exact mechanism of fluorescence
enhancement is not fully understood and probably involves more than one photo-
physical process. Firstly, the intense local field associated with the surface plasmon
can increase the excitation of flourophores in proximity. Additional enhancement
effects are also possible through interactions between the excited state fluorophore
and the metal nanoparticle. In particular, if the resonance frequencies are matched,
the excited state fluorophore can couple with a neighboring metal nanoparticle to
112 | Part I Functional materials: Synthesis and applications

excite the plasmon. The plasmon can then relax radiatively to emit light. With this
mechanism, the emission spectrum will be identical to that of the fluorophore, even
though the nanoparticle is the emitting species. Typically, metal-enhanced fluores-
cence has a much shorter lifespan than that of the fluorophore alone, supporting
the conclusion that the nanoparticle is the emitting species. It has been suggested
that it is best to think of the fluorophore-metal complex as the emitting species [34].
Plasmon-enhanced fluorescence offers exiting possibilities in the development of
ultra-sensitive methods, probes, and devices for biomedical detection.
The majority of scientific publications in the field of nanoparticle plasmonics in-
volve either silver or gold. This is, in part, because of the ease of particle synthesis,
with gold and silver nanoparticles being readily accessible by the reduction of water-
soluble salts. Furthermore, nanoparticles of these two metals exhibit resonance plas-
mon frequencies located within the visible region of the spectrum. Silver has the more
intense plasmon of the two [35], but gold is biocompatible and more stable in bio-
logical media. Despite the prevalence of silver and gold, current research is looking
beyond these two metals for next generation technologies [36, 37].
Metal nanoparticles are currently being employed in applications other than
those based on their plasmonic properties. For example, silver nanoparticles have
important antibacterial and anti-fungal properties [38] and are being incorporated
into a growing number of consumer products such as antiseptic sprays, antimicro-
bial bandages, and sports clothing. Metals are also known to catalyze a number of
important organic reactions. Since surface area plays a predominant role in catalytic
activity, the increased specific surface area associated with decreasing particle size
can improve efficiency, and supported metal nanoparticles are widely used as solid
catalysts [39, 40].

5.6 Metal oxide nanoparticles

The wide range of properties and applications related to this diverse class of nanoma-
terials will be illustrated by consideration of three specific examples; titanium dioxide
(TiO2 ), iron oxide (Fe3 O4 and Fe2 O3 ), and silicon dioxide, or silica (SiO2 ).

5.6.1 Titanium dioxide

Titanium dioxide is highly reflective in the visible region of the spectrum, and for this
reason has been long utilized as a pigment material [41]. Small particles of titanium
dioxide are the brightest of all commercial white pigments, and the 2010 global pro-
duction of nanoscale TiO2 is estimated at 5 000 metric tons [42]. The primary applica-
tion is in paints and other opaque coatings, but TiO2 is also a frequent component of
personal care products such as toothpaste and sunscreen.
5 Nanoparticles: Properties and applications | 113

Transition metal oxides are semi-conductors, and thus share some of the elec-
tronic properties and applications described above for quantum dots. TiO2 is a wide
band gap material and absorbs UV light in the 280–400 nm wavelength range [41].
Upon absorption, an electron is elevated from the valence band to the conduction
band to generate an electron-hole pair, and TiO2 nanoparticles have been employed
to enhance photoconversion in solar cells [43]. The electron-hole pair can also partic-
ipate in charge transfer to chemical species adsorbed at the semiconductor surface,
leading to chemical reactions through a process known as photocatalysis [44]. The
excited state conduction band electron can serve to reduce an accepter molecule,
whereas the valence level hole can accept an electron from a neighboring donor
species. These photoinduced electron transfer processes are governed by the posi-
tion of the semiconductor bands with respect to the redox potentials of the surface
adsorbed molecules. In the case of TiO2 , the position of the valence band leads to
a particularly high oxidation potential for the photoinduced holes. The holes can
thus readily react with adsorbed water to produce hydroxyl radicals, which in turn
are strongly oxidizing and able to convert organic compounds to CO2 and water. For
this reason, TiO2 nanoparticles can be used as antibacterial [45] and water treatment
agents [46], and have even been demonstrated to split water into hydrogen and oxy-
gen [47]. Since the photocatalytic process requires migration of electrons and holes to
the TiO2 surface, nanoparticles, with their characteristically large surface to volume
ratio, show greater catalytic activity than the corresponding bulk material.

5.6.2 Iron oxide

The second important category of metal oxide nanoparticles is that composed of the
iron oxides. These functional materials are primarily of interest for their magnetic
properties. A variety of oxides are known, differing in the oxidation state of iron and
the crystal structure. The two main forms are magnetite (Fe3 O4 ), which contains both
Fe (II) and Fe (III) in a 1 : 2 ratio, and the gamma phase of iron (III) oxide (𝛾-Fe2 O3 ),
known as maghemite.
As is the case for other transition metals, the magnetic moments of iron atoms
and ions arise from the presence of unpaired electrons in the d orbitals. Within small
regions of a bulk material, the individual neighboring atomic magnetic moments are
generally aligned with one another to form a magnetic domain. The magnetization
within each domain points in a uniform direction, but the magnetization of differ-
ent domains may point in different directions, as illustrated in Fig. 5.12. The domain
structure can be modified by the application of an external magnetic field to increase
magnetization in a given direction. If the domains do not return to a random state after
removal of the external field, a permanent magnet will result.
114 | Part I Functional materials: Synthesis and applications

Fig. 5.12. Schematic representation of a multidomain magnetic


The novel properties of magnetic nanoparticles appear when particle size is reduced to
below domain size, which is of the order of 100 nm [48]. Small iron oxide nanoparticles
are monodomain and exhibit what is known as superparamagnetic behavior.
The direction of magnetization within a domain can randomly flip at a character-
istic frequency determined by what is known as the Néel relaxation time. At a given
temperature, the Néel relaxation time decreases with decreasing particle size and, in
the case of small nanoparticles, is generally much shorter than typical observation
times. Under these conditions, known as the superparamagnetic state, the magnetiza-
tion of the nanoparticle averages to zero. However, when an external magnetic field is
applied to an assembly of superparamagnetic nanoparticles, their magnetic moments
tend to align along the applied field, leading to a net magnetization. Superparamag-
netic nanoparticles can be dispersed in a carrier liquid to form what is known as a fer-
rofluid [49]. When placed in a magnetic field gradient, a ferrofluid moves to the region
of highest flux. This means that ferrofluids can be precisely positioned and shaped
by an external magnetic field. Ferrofluids are found in a large number of commercial
devices, including audio speakers, vacuum seals, switches, and sensors [50]. We have
recently demonstrated the fabrication of a ferrofluidic deformable liquid mirror, based
on the system shown in Fig. 5.13 [51, 52].
Clearly, the role of magnetic nanoparticles as functional materials is not restricted
to their use in ferrofluids. In particular, iron oxide nanoparticles are being increasingly
employed in medical applications, including, for example, their use as contrast agents
in magnetic resonance imaging, [53] as discussed elsewhere in this volume.

Fig. 5.13. Magnetically deformable mirror prepared

by coating an appropriate ferrofluid with a reflective
monolayer of silver nanoparticles. The central defor-
mation is generated by the magnetic field from an
underlying permanent magnet.
5 Nanoparticles: Properties and applications | 115

5.6.3 Silica

Nanoparticles composed of silica have a large number of applications in materials

science. Silica is a very versatile substance offering a number of attractive proper-
ties, including, for example, chemical and thermal stability and biocompatibility
[54]. Furthermore, the surface of silica can be readily functionalized with organic or
biologically active molecules, a key requirement for biological imaging or sensing
One important example of the use of silica is in the preparation of lumines-
cent nanoparticles by the encapsulation of fluorophores [55, 56]. Luminescent nano-
particles are used in many fields of application, the most prevalent being bioimaging.
Fluorescence organic molecules typically exhibit enhanced properties when dis-
persed in the silica matrix, which protects them against both collisional quenching
and photodegradation [57]. In addition, for applications in living systems, nano-
particles have been reported to offer significant advantages over organic molecules
with respect to resistance to metabolic disintegration, low toxicity, and adequate
bioavailability [58]. Fluorescent doped silica nanoparticles have been demonstrated
as promising probes for the detection of specific ionic species [57], intercellular sens-
ing [55], cellular imaging [56], and even the diagnosis of cancer [59].
One very important class of silica nanoparticles are the mesoporous materi-
als [60, 61]. The high surface area of these materials makes them inherently inter-
esting for catalysis. They are also receiving a great deal of attention as drug delivery
platforms [62], and can be designed to offer sophisticated functionalities such as
stimulus-responsive gated release [63]. Since the preceding chapter of this volume is
dedicated to mesoporous materials, no further details will be provided here.

5.7 Polymeric nanoparticles

Polymer microbeads and nanoparticles are used on a very large scale as latex paints.
As aqueous suspensions, latexes are environmentally friendly, practically odorless
and faster drying than their oil-based counterparts. Latexes are typically obtained
by the well-known process of emulsion polymerization [64]. More recently, polymer
nanoparticles have been attracting additional interest as host matrices for the prepa-
ration of novel hybrid materials. Polymers offer great versatility as host matrices. The
large inventory of existing polymers means that a number of properties, such as polar-
ity, elasticity and refractive index, can all be adjusted to meet specific requirements.
The formation of polymer nanoparticles is achieved by the dispersion of the mon-
omer into small droplets prior to the initiation of the polymerization reaction. Many
considerations influence the exact choice of the method of monomer dispersion, but
for the preparation of nanosized particles, miniemulsion polymerization is often the
method of choice [65]. The process of miniemulsion polymerization is illustrated in
116 | Part I Functional materials: Synthesis and applications

Fig. 5.14. In contrast to emulsion polymerization, in which the monomer must diffuse
through the continuous phase to feed particle growth, in the miniemulsion process,
all of the monomer is pre-dispersed as nanodroplets prior to the initiation step. Thus,
in miniemulsion polymerizations, each droplet acts as an independent preformed
nanoreactor. Miniemulsion polymerization is therefore particularly well-suited to the
preparation of doped nanoparticles, since dopants can be dissolved in the monomer
at the desired concentration before dispersion [65]. For example, we have employed
this approach for the preparation of polystyrene nanoparticles doped with a lumi-
nescent lanthanide complex, as illustrated in Fig. 5.15 [66]. As reviewed elsewhere
[65], this approach can be extended beyond molecular dopants, and miniemulsion
polymerization is finding increasing interest for the encapsulation of solid nano-


Emulsion Miniemulsion


Polymer nanoparticles

Fig. 5.14. Schematic depiction of miniemulsion polymerization.

Fig. 5.15. Transmission electron microscope image of

x60 000 polystyrene nanoparticles prepared by miniemulsion
0.2 μm
polymerization. Reprinted with permission from [66].
5 Nanoparticles: Properties and applications | 117

5.8 Advanced architectures and hybrid systems

The examples provided above illustrate the immense diversity of functional nano-
particles and their role in developing technologies. For the most part, the present
discussion has, however, been limited to monolithic particles. In reality, nanoparticle
architecture is becoming increasingly sophisticated, and a cursory review of cur-
rent scientific literature is sufficient to ascertain the growing trend toward multi-
component, multi-function materials, such as core-shell and doped nanoparticles.
For example, metal cores have been coated with fluorescent dye-functionalized silica
shells to combine plasmonic and luminescence properties in nanoparticles designed
for biomedical detection [67]. Core-shell architectures have also been employed to
facilitate energy transfer and upconversion luminescence in lathanide-doped ma-
terials [68]. These are only two examples of the exciting future of nanoparticles as
functional materials. The possibilities are essentially endless. Promising new ma-
terials are being reported daily and any attempt to adequately review the scientific
literature in this field becomes outdated before even going to press.


[1] SciFinder, American Chemical Society, Chemical Abstracts Service, Columbus, Ohio 1999. consulted June 26, 2013.
[2] See for example titles such as ACS Nano, Nano Letters, Small, Nature Nanotechnology and the
Journal of Nanoparticle Research.
[3] Nanotechnology and Enabling Technologies, Foresight Briefing #2, The Foresight Institute,
1989., consulted January 7, 2014.
[4] Lamer V, Dinegar R. Theory, Production and Mechanism of Formation of Monodispersed Hy-
drosols. J Amer Chem Soc 1950; 72: 4847–4854.
[5] Sugimoto T. Nucleation and Growth of Monodispersed Particles: Mechanisms, Encyclopedia of
Surface and Colloid Science, Taylor & Francis, 2006: 1, 4257–4270.
[6] Destrée C, Nagy JB. Mechanism of formation of inorganic and organic nanoparticles from
microemulsions. Adv Colloid Interface Sci 2006; 123–126: 353–367.
[7] Lemyre J-L, Ritcey AM. Synthesis of Lanthanide Fluoride Nanoparticles of Varying Shape and
Size. Chem Mater 2005; 17: 3040–3043.
[8] Lucier BEG, Johnston KE, Arnold DC, Lemyre J-L, Beaupreì A, Blanchette M, Ritcey AM, Schurko
RW. Comprehensive Solid-State Characterization of Rare Earth Fluoride Nanoparticles. J Phys
Chem C 2014; 118: 1213–1228.
[9] Ohshima H. The DLVO Theory of Colloidal Stability. In Electrical Phenomena at Interfaces and
Biointerfaces: Fundamentals and Applications in Nano-Bio- and Environmental Sciences,
Chapter 3. First Edition. Ohshima. H, Ed., John Wiley & Sons, Inc. 2012.
[10] Israelachvili J. Intermolecular & Surface Forces, Second Edition, pp 293-298, Academic Press,
San Diego, 1992.
[11] Norris, DJ. Electronic Structure in Semiconductor Nanocrystals: Optical Experiment. In Semi-
conductor and Metal Nanocrystals, Synthesis and Electronic and Optical Properties; Klimov VI,
Ed.; Chapter 2. Marcel Dekker, Inc., New York, 2004.
118 | Part I Functional materials: Synthesis and applications

[12] Brus LE. A simple model for the ionization potential, electron affinity, and aqueous redox po-
tentials of small semiconductor crystallites. J Chem Phys 1983;79: 5566–5571.
[13] Semonin OE, Luther JM and Beard MC. Quantum dots for next generation Photovoltaics. Mater
Today 2012; 15(11): 508–515.
[14] de Mello Donega C, Hickey SG, Wuister SF, Vanmaekelbergh D, Meijerink A. J Phys Chem B
2003; 107: 489–496.
[15] Shirasaki Y, Supran GJ, Bawendi MG, Bulović V. Emergence of colloidal quantum-dot light-
emitting technologies. Nat Photonics 2013; 7: 13–23.
[16] Kershaw SV, Susha AS, Rogach AL. Narrow bandgap colloidal metal chalcogenide quantum
dots: synthetic methods, heterostructures, assemblies, electronic and infrared optical proper-
ties. Chem Soc Rev 2013; 42: 3033–3087.
[17] Gaponik N, Hickey SG, Dorfs D, Rogach AL, Eychmüller A. Progress in the Light Emission of
Colloidal Semiconductor Nanocrystals. Small 2010; 6(13): 1364–1378.
[18] Samokhvalov P, Artemyev M, Nabiev I. Basic Principles and Current Trends in Colloidal Synthe-
sis of Highly Luminescent Semiconductor Nanocrystals. Chem Eur J 2013; 19: 1534–1546.
[19] Talapin DV, Lee J-S, Kovalenko MV, Shevchenko EV. Prospects of Colloidal Nanocrystals for
Electronic and Optoelectronic Applications. Chem Rev 2010; 110: 389–458.
[20] Nozik AJ, Beard MC, Luther JM, Law M, Ellingson RJ, Johnson JC. Semiconductor Quantum Dots
and Quantum Dot Arrays and Applications of Multiple Exciton Generation to Third-Generation
Photovoltaic Solar Cells. Chem Rev 2010: 110: 6873–6890.
[21] Knowles KE, Frederick MT, Tice DB, Morris-Cohen AJ, Weiss EA. Colloidal Quantum Dots:
Think Outside the (Particle-in-a-) Box. J Phys Chem Lett 2012; 3: 18–26.
[22] Malynych S, Chumanov G. Light-Induced Coherent Interactions between Silver Nanoparticles
in Two-Dimensional Arrays. J Amer Chem Soc 2003; 125; 2896–2898.
[23] Schuller JA, Barnard ES, Cai W, Jun YC, White JS, Brongersma ML. Plasmonics for extreme light
concentration and manipulation. Nat Mater 2010; 9: 193–204.
[24] Henry A-I, Bingham JM, Ringe E, Marks LD, Schatz GC, Van Duyne RP. Correlated Structure and
Optical Property Studies of Plasmonic Nanoparticles. J Phys Chem C 2011; 115: 9291–9305
[25] Schatz GC, Kelly KL, Coronado E, Zhao LL. The Optical Properties of Metal Nanoparticles: The
Influence of Size, Shape, and Dielectric Environment. J Phys Chem B 2003; 107: 668–677.
[26] Hodak JH, Henglein A, Hartland GV. Photophysics of Nanometer Sized Metal Particles: Electron-
Phonon Coupling and Coherent Excitation of Breathing Vibrational Modes. J Phys Chem B
2000; 104: 9954–9965.
[27] Reinhard BM, Siu M, Agarwal H, Alivisatos P, Liphardt J. Calibration of Dynamic Molecu-
lar Rulers Based on Plasmon Coupling between Gold Nanoparticles. Nano Lett 2005; 5:
[28] Jain PK, Huang W, El-Sayed MA. On the universal scaling behavior of the distance decay of
plasmon coupling in metal nanoparticles pairs: A plasmon ruler equation. Nano Lett 2007; 7:
[29] Yang L, Wang H, Yan B, Reinhard BM. Calibration of Silver Plasmon Rulers in the 1–25 nm Sep-
aration Range: Experimental Indications of Distinct Plasmon Coupling Regimes. J Phys Chem C
2010; 114: 4901–4908.
[30] Stockman, MI. Nanoplasmonics: The physics behind the applications. Phys Today 2011; 64 (2):
[31] Rycenga M, Camargo PHC, Li W, Moran CH, XiaY. Understanding the SERS Effects of Single
Silver Nanoparticles and Their Dimers, One at a Time. J Phys Chem Lett 2010; 1: 696–703.
[32] Nie S, Emory SR. Probing Single Molecules and Single Nanoparticles by Surface-Enhanced
Raman Scattering. Science 1997; 275: 1102–1106.
5 Nanoparticles: Properties and applications | 119

[33] Lakowicz JR, Ray K, Chowdhury M, Szmacinski H, FuY, Zhang J, Nowaczyk K. Plasmon-
controlled fluorescence: a new paradigm in fluorescence spectroscopy. Analyst, 2008; 133:
[34] Lakowicz JR. Principles of Radiative Decay Engineering: Metal-Enhanced Fluorescence. In Fluo-
rescence Spectroscopy, Third Edition, Chapter 25. Springer Science 2006.
[35] Mahmoud M, El-Sayed MA. Different Plasmon Sensing Behavior of Silver and Gold Nanorods.
J Phys Chem Lett 2013; 4: 1541–1545.
[36] DeSantis CJ, Weiner RG, Radmilovic A, Bower MM, Skrabalak SE. Seeding Bimetallic Nano-
structures as a New Class of Plasmonic Colloids. J Phys Chem Lett 2013; 4: 3072–3082.
[37] Naik GV, Shalaev VM, Boltasseva A. Alternative Plasmonic Materials: Beyond Gold and Silver.
Adv Mater 2013; 25(24): 3264–3294.
[38] Liu L, Yang J, Xie J, Luo Z, Jiang J, Yang YY, Liu S. The potent antimicrobial properties of cell pen-
etrating peptide-conjugated silver nanoparticles with excellent selectivity for Gram-positive
bacteria over erythrocytes. Nanoscale 2013; 5: 3834–3840.
[39] Jin R. The impacts of nanotechnology on catalysis by precious metal nanoparticles.
Nanotechnol Rev 2012; 1: 31–56.
[40] Reetz MT. Size-selective Synthesis of Nanostructured Metal and Metal Oxide Colloids and
Their Use as Catalysts. In Nanoparticles and Catalysis, Astruc D, Ed., Chapter 8. Wiley-VCH
Weinheim, 2008.
[41] Lan Y, Lu Y, Ren Z. Mini review on photocatalysis of titanium dioxide nanoparticles and their
solar applications. Nano Energy 2013; 2:1031–1045.
[42] Weir A, Westerhoff P, Fabricius L, von Goetz N. Titanium Dioxide Nanoparticles in Food and
Personal Care Products. Environ Sci Technol 2012; 46(4): 2242–2250.
[43] Genovese MP, Lightcap IV, Kamat PV. Sun-Believable Solar Paint. A Transformative One-Step
Approach for Designing Nanocrystalline Solar Cells. ACS Nano 2012; 6(1): 865–872.
[44] Linsebigler AL, Lu G, Yates JT Jr. Photocatalysis on TiOn Surfaces: Principles, Mechanisms, and
Selected Results. Chem Rev 1995; 95: 735–758.
[45] Wolfrum EJ, Huang J, Blake DM, Maness P-C, Huang Z, Nefiest J, Jacoby WA. Photocatalytic
Oxidation of Bacteria, Bacterial and Fungal Spores, and Model Biofilm Components to Carbon
Dioxide on Titanium Dioxide-Coated Surfaces. Environ Sci Technol 2002; 36: 3412–3419.
[46] Reisch MS. Workhorse pigment titanium dioxide is also the stuff that dreams are made of.
Chem Eng News 2003; 81(13): 13.
[47] Fujishima A, Honda K. Electrochemical Photolysis of Water at a Semiconductor Electrode.
Nature 1972; 238: 37–38.
[48] Sarkar D, Mandal M, Mandal K. Domain controlled magnetic and electric properties of variable
sized magnetite nanohollow spheres. J Appl Phys 2012; 112: 064318.
[49] Rosensweig, RE. Ferrohydrodynamics, Cambridge University Press: London, 1985.
[50] Raj K, Moskowitz R. Commercial Applications of Ferrofluids. J Magn Magn Mater 1990; 85:
[51] Déry J-P, Borra EF, Ritcey AM. An ethylene glycol based ferrofluid for the fabrication of magneti-
cally deformable liquid mirrors. Chem Mater 2008; 20: 6420–6426.
[52] Ritcey AM, Borra E. Magnetically Deformable Liquid Mirrors from Surface Films of Silver Nano-
particles. Chem Phys Chem 2010; 11: 981–986.
[53] Na HB, Song IC, Hyeon T. Inorganic Nanoparticles for MRI Contrast Agents. Adv Mater 2009; 21:
[54] Asefa T, Tao Z. Biocompatibility of Mesoporous Silica Nanoparticles. Chem Res Toxicol 2012;
25: 2265–2284.
[55] Ruedas-Rama MJ, Walters JD, Orte A, Hall EAH. Fluorescent nanoparticles for intracellular sens-
ing: A review. Anal Chim Acta 2012; 751: 1–23.
120 | Part I Functional materials: Synthesis and applications

[56] Wang K, He X, Yang X, Shi H. Functionalized Silica Nanoparticles: A Platform for Fluorescence
Imaging at the Cell and Small Animal Levels. Acc Chem Res 2013; 46(7): 1367–1376.
[57] Montalti M, Rampazzo E, Zaccheroni N, Prodi L. Luminescent chemosensors based on silica
nanoparticles for the detection of ionic species. New J Chem 2013; 37: 28–34.
[58] Sharma P, Brown S, Walter G, Santra S, Moudgil B. Nanoparticles for bioimaging. Adv. Colloid
Interface Sci. 2006; 123-126: 471–485.
[59] Arap W, Pasqualini R, Montalti M, Petrizza L, Prodi L, Rampazzo E, Zaccheroni N, Marchiò S.
Luminescent silica nanoparticles for cancer diagnosis. Curr Med Chem 2013; 20(17): 2195–211.
[60] Chen Y, Chen H-R, Shi J-L. Construction of Homogenous/Heterogeneous Hollow Mesoporous
Silica Nanostructures by Silica-Etching Chemistry: Principles, Synthesis, and Applications.
Acc Chem Res 2014; 47(1): 125–137.
[61] Wu S-H, Mou C-Y, Lin H-P. Synthesis of mesoporous silica nanoparticles. Chem Soc Rev 2013;
42(9): 3862–3875.
[62] Argyo C, Weiss V, Braeuchle C, Bein T. Multifunctional Mesoporous Silica Nanoparticles as a
Universal Platform for Drug Delivery. Chem Mater 2014; 26(1): 435–451.
[63] Nadrah P, Planinsĕk O, Gabersčĕk M. Stimulus-responsive mesoporous silica particles.
J Mater Sci 2014: 49: 481–495.
[64] Landfester K. Synthesis of colloidal particles in miniemulsions. Annu Rev Mater Res 2006; 36:
[65] Landfester K, Weiss CK. Encapsulation by Miniemulsion Polymerization. Adv Polym Sci 2010;
229: 1–49.
[66] Desbiens J, Bergeron B, Patry M, Ritcey AM. Polystyrene nanoparticles doped with a lumines-
cent europium complex. J Colloid Interface Sci 2012; 376 (1): 12–19.
[67] Lessard-Viger M, Rioux M, Rainville L, Boudreau D. Nano Lett 2009; 9(8): 3066–3071.
[68] Wang F, Deng R, Wang J, Wang Q, Han Y, Zhu H, Chen X, Liu X. Tuning upconversion through
energy migration in core-shell nanoparticles. Nat Mater 2011; 10: 968–973.
N. Allard and M. Leclerc
6 Conjugated polymers for organic electronics
Conjugated polymers have received a great deal of attention from both academic and
industrial laboratories, since they combine the best features of metals or inorganic
semiconducting materials (excellent electrical and optical properties), with those of
synthetic polymers (mechanical flexibility, simple processing, low cost). This syn-
ergy makes these functional materials useful in existing optoelectronic devices, and
creates completely new technological opportunities. For instance, polymeric semicon-
ductors are considered to be one of the most promising materials to lower the cost of
solar energy. Since they can be dissolved in common solvents, processing techniques
such as inkjet printing, spin coating and large scale roll-to-roll coating techniques,
widely used for thermoplastics, are now available for the low-cost production of
printed photovoltaic devices. All of these properties will help to establish the so-
called plastic electronics era, which includes light-emitting diodes and field-effect

6.1 Introduction

There is hardly an aspect of our lives which is not touched by polymeric materials. In
the fields of optics and electronics, synthetic polymers are traditionally used in ap-
plications such as packaging, electrical insulators, and photoresists. However, new
opportunities have emerged with the development of electroactive and photoactive
conjugated polymers. This chapter will describe several examples which demonstrate
the tremendous innovation potential of these novel functional materials.
As with many other important discoveries, the development of the field of con-
jugated polymers was somehow accidental. A student of Professor Hideki Shirakawa
at the Tokyo Institute of Technology was working on the preparation of polyacety-
lene (Fig. 6.1) by the Ziegler–Natta polymerization method. By mistake, this student
prepared a multimolar concentration of the catalyst, instead of the usual millimolar
concentration, and obtained a thin polyacetylene film which looked like a metallic
foil instead of the usual dark, powdery material. During a visit to Japan, Professor
A.G. MacDiarmid (Department of Chemistry, University of Pennsylvania) met Profes-
sor Shirakawa and was fascinated by this shiny polymeric film. He invited him to come
to Philadelphia to investigate this new form of polyacetylene in more detail. In 1977,
in collaboration with Professor A.J. Heeger, a colleague from the Physics Department,
this team reported that upon partial oxidation with iodine or bromine (so-called
doping reaction), the conductivity of the intrinsically semiconducting polyacetylene
increased more than a millionfold (up to 100 S/cm) [1]. These days, electrical con-
ductivities of up to 105 S/cm are now obtained with some highly crystalline forms of
122 | Part I Functional materials: Synthesis and applications

polyacetylene. The delocalized electronic structure (alternation of single and double

bonds) of polyacetylene is responsible for the good intrachain and interchain mobility
of the charge carriers (radical cations or their negative analogs) created upon doping.
This delocalized (conjugated) structure is also responsible for a strong absorption
in the UV-visible range. This group (Shirakawa, MacDiarmid, Heeger) eventually re-
ceived the 2000 Nobel Prize in Chemistry for this discovery [2–4]. Unfortunately, this
first conducting polymer is still very difficult to process and unstable in the presence
of oxygen.
Rapidly, many scientists investigated more stable conjugated polymers. For in-
stance, in the early 1980s, a lot of studies were devoted to electropolymerized poly-
thiophene [5], polypyrrole [5, 6], and polyaniline [7] (Fig. 6.1). This method has the
advantage of being capable of preparing thin films of these infusible and insoluble
rigid-rod conjugated polymers in one step. Depending upon the electrochemical po-
tential, it is possible to switch between the undoped semiconducting state and the
oxidized conducting state. Interestingly, these redox processes are also accompanied
by color changes (electrochromism). However, the ultimate goal remains the devel-
opment of polymeric materials which combine the electrical and optical features of
metals or semiconductors with the processing advantages and mechanical properties
of traditional polymers.

n n H n n
Polyacetylene Polythiophene Polypyrrole Polyaniline

Fig. 6.1. First generation of conjugated polymers.

6.2 Processable conjugated polymers

The attempts to fulfill these requirements led to the development of a second genera-
tion of processable conjugated polymers. A first example of processable polyacetylene
was demonstrated by Edwards and Feast in 1980 [8]. Using ring-opening metathesis
polymerization (ROMP), a soluble polymeric non-conjugated precursor could be con-
verted into a polyacetylene thin film and a volatile by-product upon heating (Fig. 6.2).
This two-step procedure was not perfect, but it was the first example of a processable
and potentially conjugated polymer. A similar approach was developed for the prepa-
ration of poly(para-phenylene) (PPP) [9].
A breakthrough occurred in 1985–86 with the syntheses of highly conjugated and
processable poly(3-alkylthiophene)s [10]. The fact that these five-membered rings can
exhibit an anti co-planar conformation (Fig. 6.3) reduces the steric hindrance devel-
oped by the presence of the solubilizing side-chains. Thus, the coexistence of both pla-
narity (important to keep the delocalized and conjugated structure) and processability
6 Conjugated polymers for organic electronics | 123



Fig. 6.2. Synthesis of polyacetylene from a processable precursor.

was demonstrated for the first time. For instance, it is interesting to note that process-
able alkyl-substituted polyacetylenes, polyanilines, and poly(para-phenylene)s are
non-planar and have poor electrical properties. Some researchers describe these
types of substituted conjugated macromolecules as hairy-rod polymers, this configu-
ration facilitating the dispersion and interactions with the solvent. Following these
first studies on poly(3-alkylthiophene)s, it became quite clear that the synthesis of
well-defined head-to-tail coupled poly(3-alkylthiophene)s, which should yield to the
lowest steric hindrance from the side chains and possibly a more efficient three-
dimensional packing, would lead to a significant improvement in the performance of
these polymeric materials [11]. Therefore, in attempts to bring more reliable synthetic
procedures to the field of electronic materials, a variety of synthetic tools have been
implemented, allowing significant advances in this research field. Among other ad-
vantages, these investigations led to the first preparations of well-defined regioregular
poly(3-alkylthiophene)s by McCullough and Rieke in 1992 (Fig. 6.3, [12, 13]). Further,
these relatively complicated polymerization procedures have been optimized and sim-
plified, leading to the Grignard metathesis method (GRIM, [14]). For instance, this new
method eliminates the need for highly reactive metals and can be performed at low
temperatures. Among all poly(3-alkylthiophene)s investigated, regioregular poly(3-
hexylthiophene), or (P3HT), has become the polymer of choice, exhibiting the best
electrical and optical properties in addition to adequate processability. The synthesis
of air-stable, semi-transparent, highly conducting poly(3,4-ethylenedioxythiophene),
or (PEDOT), is another example of rational design of a conjugated polymer with op-
timized structural and physical properties [15]. Both P3HT and PEDOT are currently
the most utilized conjugated polymers and are commercially available from various
In the meantime, the focus shifted from the synthesis of highly conductive poly-
mers to the design of stable semiconducting polymers through collaborations between
physicists and engineers. This new driving force was based on the aim to initiate the
so-called plastic electronics era, where micro-electronic devices could be printed
on different substrates using practical processing methods. For this purpose, the
electronic properties of the existing polymers have to be chemically tuned to obtain
low bandgap semiconducting materials. This bandgap tuning can be achieved in
various ways. For instance, the rigidification of the conjugated backbone, the intro-
duction of electron-withdrawing or electron-donating side groups, and the increase
of the quinoid (versus aromatic) character of the main chain have been widely used
124 | Part I Functional materials: Synthesis and applications

Oxidation R
R or
electropolymerization S
S n
R Kumada
(Mc Cullough) S
Br [Ni]cat. S n
R Negishi
(Rieke) S
Br S Br S n
R Kumada
Br Br S n Fig. 6.3. Synthesis of processable
S [Ni]cat.
R poly(3-alkylthiophene)s.

in the past few years to modulate the bandgap of such conjugated polymers [16].
However, one of the most efficient approaches involves the utilization of electron-
rich or electron-poor units leading to alternating push-pull architecture. As shown
in Fig. 6.4, when these two different moieties are combined in an alternating copoly-
mer, hybridization occurs between the molecular orbitals of the electron-rich unit
and of the electron-poor unit, leading to a significant modulation of the copolymer
bandgap. Usually, the energy level of the highest occupied molecular orbital (HOMO)
of the resulting copolymer is determined by the electron-rich unit, while the energy
level of the lowest unoccupied molecular orbital (LUMO) is mainly influenced by the
electron-poor unit.


Eg Electron-poor

Fig. 6.4. Hybridization of the HOMO-LUMO energy
D D-A A levels in donor-acceptor (push-pull) copolymers.
6 Conjugated polymers for organic electronics | 125

N N Suzuki S S
(RO)2B B(OR)2 + S S
N Br Br [Pd]cat. n
R3Sn S S SnR3 Stille S S
Si S
Br S Br [Pd]cat. n

R R Si
+ S S n
R S Br S [Pd]cat.

Fig. 6.5. Typical polymerization reactions for conjugated alternating copolymers.

Over the years, a third generation of semiconducting copolymers have therefore been
synthesized to satisfy the needs of various electronic applications [17]. Unfortunately,
there are relatively few synthetic methods which allow efficient preparation of such
alternating copolymers; most of these copolymers being obtained via Suzuki [18] or
Stille [19] cross-coupling polymerization (Fig. 6.5). These state-of-the-art methods gen-
erally involve numerous synthetic steps and organometallic reagents which give rise
to metal waste and various by-products. Therefore cheaper, environmentally friendly
and more efficient synthetic procedures would clearly be a great asset for the sustain-
able preparation of affordable conjugated polymers. In order to solve these problems,
the utilization of the latest synthetic developments in organic chemistry, termed di-
rect (hetero)arylation, is very appealing. These new reactions allow for the forma-
tion of carbon-carbon bonds between (hetero)arenes and aryl halides. Hence, they
do not require organometallic intermediates, thereby significantly reducing both syn-
thetic steps and cost (see Fig. 6.5). Moreover, this new direct (hetero)arylation poly-
merization (DHAP) method, which only produces acid as a by-product, has recently
attracted a lot of interest [20]. Although these reactions are not easily controlled for
compounds having more than one type of C-H bond, they can sometimes lead to the
synthesis of branched or crosslinked polymers. A fine-tuning of the reaction condi-
tions and careful choice of the monomers may suppress these unwelcome side re-
actions. All of these step-growth polymerization methods (Suzuki, Stille, DHAP) can
lead to number-average molecular weights of up to 100 kDa/mol in the best case, and
must exhibit polydispersity indexes of about 2 according to the Carother equation. Fi-
nally, it is important to note that most recently-developed monomers (e.g. carbazoles,
dithienosiloles, thienopyrrolodiones, etc.) bear flexible side-chains which are far from
the conjugated backbone, reducing the steric hindrance created within these hairy-rod
chemical structures.
126 | Part I Functional materials: Synthesis and applications

These chemical tools allowed chemists and engineers to synthesize a large num-
ber of semiconducting polymers for the manufacture of diverse electronic devices
such as light-emitting diodes, field-effect transistors, solar cells, and a continuously
growing list of other applications. For instance, water-soluble luminescent conjugated
polymers have allowed the specific, rapid and ultra-sensitive detection of unlabeled
DNA, RNA, or proteins. This relatively young field of chemical and biological sensors
based on semiconducting conjugated polymers is expanding rapidly but will not be
further discussed in this chapter. For those interested in these biomedical applica-
tions, the authors recommend several reviews and the references therein [21, 22].

6.3 Applications in renewable energy

6.3.1 Organic solar cells

Solar energy represents an attractive solution to fulfill our needs regarding green and
renewable energy, while protecting the environment at the same time. For example,
one should be aware that the sun provides in one hour what humanity consumes dur-
ing one year. However, the large scale development of this technology is somewhat
limited by the relatively high cost of producing and installing a solar panel versus its
moderate efficiency. Many studies suggest that the cost of a solar panel must be de-
creased by a factor of 5 to 10 in order to compete with other already known energy
sources such as hydro, nuclear, wind or fossil. Electroactive and photoactive conju-
gated polymers are considered to be one of the most promising ways to lower the
price of solar energy. They can be solubilized in common solvents and processed us-
ing techniques such as inkjet printing, spin coating and large scale roll-to-roll coating
techniques, which are widely used for thermoplastics and are available for the man-
ufacture of various devices. In particular, bulk heterojunction (usually a blend of a
polymeric p-type semiconductor with an n-type fullerene derivative) solar cells offer
great opportunities, with power conversion efficiency (PCE) now approaching thresh-
old values (a PCE of 10 % with a lifespan of 5–10 years) for large scale commercial-
ization. Indeed, the most efficient polymeric materials can now reach power conver-
sion efficiencies of over 9 % in a single-layer configuration [23]. These new functional
materials should help to initiate new and competitive manufacturing of photovoltaic
Different types of device architecture have been utilized for the manufacture of
organic photovoltaic devices. As mentioned above, the most common configuration is
called the bulk heterojunction (BHJ) approach (Fig. 6.6). Glass is usually employed as
common substrate, although plastics can be used to generate flexible devices. A layer
of semi-transparent indium-tin oxide (ITO), which acts as the anode, is deposited on
the substrate. After deposition of the anode, PEDOT, a semi-transparent and highly
conductive polymer (described in Section 6.1), is spin-coated from a water-based so-
6 Conjugated polymers for organic electronics | 127

lution on top of ITO for the purpose of enhancing the hole transport and improving
contacts between the ITO electrode and the organic BHJ layer. Then the active layer,
which is a bulk heterojunction of a donor (p-type semiconductor) and an acceptor (n-
type semiconductor), usually from an organic solution containing both components,
is deposited over the PEDOT layer. Such a structure provides large surface area in-
terfaces which allow the generation of a charge separated state with efficient sub-
sequent charge migration through 3-D percolation networks. The donor is usually a
p-type semiconducting conjugated polymer, while the acceptor is generally a process-
able fullerene derivative such as [6,6]-phenyl-C61 -butyric acid methyl ester (PCBM) or
its C71 analog; fullerenes being known to support stable and mobile negative charge
carriers (see Chapter 3). Finally, a reflective metallic layer acting as the cathode is gen-
erally evaporated or printed on top of the assembly.

Polymer: PCBM

Fig. 6.6. Standard architecture for an organic
Glass or plastic solar cell.

As shown in Fig. 6.7, upon illumination, the semiconducting BHJ active layer absorbs
some of the photons leading to the formation of excitons (Step 1). For instance, these
excitons may come from the excitation of an electron from the HOMO of the polymer to
its LUMO, creating bound electron-hole pairs (Step 2). These excitons can then migrate
about 10–20 nm within the polymer phase (related to their lifespan and mobility) to
reach the acceptor phase. The nanoscale BHJ morphology is particularly helpful for
this purpose: at this boundary, the electron in the LUMO of the polymer is transferred
to the LUMO of the acceptor (Step 3). After this photo-induced electron transfer, the
positive charge carriers and the electrons, which are no longer held by Coulombic in-
teractions, travel through their respective percolated phases to reach their respective
electrodes (Step 4). It is important to mention that a similar mechanism is observed
when light is absorbed by the fullerene derivative or any other n-type semiconductor;
the only major difference being the fact that the electron transfer involves the HOMO
of the semiconductors. Interestingly, despite the relative complexity of all these in-
terfaces and physical mechanisms, internal quantum efficiency of close to 100 % has
been reported, implying that for some optimized nanoscale morphology, essentially
every absorbed photon results in a separated pair of charge carriers which are col-
lected at the electrodes [24].
128 | Part I Functional materials: Synthesis and applications

1. LUMO 2. LUMO –

hv AI AI

Polymer PCBM Polymer PCBM

4. LUMO 3. LUMO –


+ +
Polymer PCBM Polymer PCBM

Fig. 6.7. General working mechanism of an organic solar cell.

6.3.2 Conjugated polymers for organic solar cells

While the manufacture and working mechanism of an organic solar cell may seem rel-
atively simple, finding the optimal conjugated polymer and obtaining the right mor-
phology (which depends upon the blend ratio, solvent, deposition technique, temper-
ature, etc.) is still a big challenge. For instance, as a first requirement, the semicon-
ducting polymer must be stable in air. Consequently, its HOMO energy level should
be under −5.2 eV, which is about the air-oxidation threshold. Also, as mentioned pre-
viously, there must be an electron transfer between the LUMO of the donor and the
LUMO of the acceptor. To make this electron transfer possible, there must be an ener-
getic difference of about 0.3 eV between these LUMO energy levels (the same is true
when the electron transfer takes place through the HOMO energy levels). In addition,
it is preferable for the HOMO energy level of the polymer to be as low as possible, since
the open circuit voltage (Voc ) is mostly defined by the difference between the LUMO of
the acceptor and the HOMO of the polymer. Moreover, since the polymer is the main
light absorber when blended with fullerene derivatives, it is necessary that the poly-
mer absorbs as much sunlight as possible. Thus, bandgaps between 1.2 and 1.9 eV are
desired, since they correspond to the maximum intensity of the solar emission spectra
and allow for the best compromise between optimized photocurrent and photovoltage
(Fig. 6.8).
6 Conjugated polymers for organic electronics | 129

OCH3 Energy (eV)

4.00 3.00 2.50 2.00 1.75 1.50 1.25 1.00

Photon flux AM 1.5 (m–2 .s–1 .nm–1)

Polymer Ideal Zone
Donor 18
Lumo Acceptor
> 0.3 eV –4.3 eV
Band gap 2x1018
1.2–1.9 eV

–6.1 eV










Wavelength (nm)

Fig. 6.8. Properties needed for conjugated polymers and photon flux spectra of the sun.

Once a good polymer/fullerene composite with the right morphology is found, a device
is created and tested to determine the PCE. For this purpose, the electrodes are con-
nected and light is beamed through the ITO electrode using a solar simulator (AM1.5G)
imitating the sun irradiation at a tilted angle of 37 degrees (Pin is of 100 mW/cm2 ).
At the same time, a potential sweeping is applied. After data processing, a typical
J-V curve from which the fill factor (FF), the short-circuit current density (Jsc ), and the
open-circuit voltage (Voc ) can be obtained, is used to calculate the PCE of the device
(Fig. 6.9). Pmax represents the maximum power which can be produced by such a solar
cell. This last parameter is important for the characterization of the fill factor, which
describes the overall quality of the device.
Current density (mA/cm2)

VPmax VOC Pout FF x JSC x VOC

PCE = =
Pin Pin

JPmax x VPmax
JPmax FF =
Pmax Jsc x VOC
Potential (V)

Fig. 6.9. Typical J-V curve obtained for a photovoltaic device.

130 | Part I Functional materials: Synthesis and applications

Hundreds of polymeric structures have been synthesized and investigated in the

last few years, but only three main polymers (P3HT [25, 26], PCDTBT [25, 27], and
PTB7 [23]) have demonstrated good and reproducible PCEs, with a lifespan of up to
7 years for PCDTBT (Fig. 6.10). However, large, stable, and low-cost modules will have
to be developed and tested under “real” conditions in order to reach commercial ap-


H9C4 C2H5 O C2H5

C6H13 N N S
S n n S S n
PCE = 6.5% PCE = 7.9% H9C4 C2H5 PCE = 9.2%

Fig. 6.10. Most widely studied conjugated polymers for organic solar cells.

6.4 Applications in micro-electronics

6.4.1 Field-effect transistors

Field-effect transistors (FETs) are electronic switches widely utilized in electronic de-
vices such as displays, computer logics, radio frequency identification tags (RFID),
etc. The vast majority of these devices are currently made from monocrystalline, poly-
crystalline or amorphous silicon. In FETs, one of the main performance criteria is the
charge mobility which can be simply related to the on/off switching speed of the de-
vice. Higher switching speeds permit applications in high performance electronics.
These widely used silicon compounds have shown charge carrier mobility of up to
900 cm2 /V s for monocrystalline silicon [28] used in high performance electronic de-
vices, and up to 1 cm2 /V s for amorphous silicon [28] utilized in standard electronic
Although silicon-based devices perform very well, they are relatively difficult to
produce and cannot be used in flexible devices, whereas conjugated polymer-based
devices can be solution-processed at low temperatures and easily coated onto flexible
devices. Despite these great possibilities, polymer-based FETs will probably only com-
pete with amorphous silicon. Indeed, amorphous or semi-crystalline semiconducting
polymers do not exhibit the necessary long-range organization to lead to very high
charge carrier mobility.
In most cases, an FET is constituted of three electrodes: the gate, the source,
and the drain. It is also composed of a dielectric layer and of a semiconducting layer.
Figure 6.11 shows the different configurations for organic FETs (OFETs). The two most
6 Conjugated polymers for organic electronics | 131

D S Semiconductor
Semiconductor D S
Dielectric Dielectric
Gate Electrode Gate Electrode
(a) (b)

Gate Gate
Gate dielectric Dielectric
Semiconductor D S
D S Semiconductor
Substrate Substrate
(c) (d)

Fig. 6.11. Schematic description of OFETs with (a) bottom-gate top-contact (BG/TC), (b) bottom-gate
bottom-contact (BG/BC), (c) top-gate bottom-contact (TG/BC), and (d) top-gate top-contact (TG/TC)
structures [29].

frequent configurations are bottom-gate top-contact (BG/TC; (a)) and bottom-gate

bottom-contact (BG-BC; (b)) due to their relatively simple production.
The general working mechanism of an OFET starts by applying a potential to the
gate while a potential is constantly applied between the source and the drain (see
Fig. 6.12). If a negative potential is applied to the gate (VGS ), positive charge carriers
are formed at the interface between the semiconducting polymer and the dielectric
layer. Then, due to the potential between the source electrode and the drain electrode,
these positive charge carriers travel through the semiconducting layer, forming a p-
type OFET. The drain-source current (IDS ) first follows ohmic behavior as a function
of the applied drain-source potential (VDS ), then saturates at high voltages. If a pos-
itive potential is applied at the gate, negative charge carriers are formed at the poly-
mer/dielectric interface creating n-type OFET.
These OFETs are mainly characterized by three values: the charge carrier mobil-
ity (μ), the Ion /Ioff ratio, and the threshold voltage (VT ). The charge mobility refers to
the speed at which the charges formed in the semiconductor layer travel between the
source and the drain electrodes. These values can be determined from I-V curves, sim-
ilar to those reported in Fig. 6.12. The higher the charge mobility, the faster switching
between the on and off states of the OFET can occur. This charge mobility is numbered
either in hole mobility or electron mobility depending on the potential applied at the
gate. The Ion /Ioff ratio refers to the ratio between the current when the transistor is at
the on-state and when it is at the off-state. While multiple switching happens in the
device even if there is no potential applied at the gate, residual charges are still in
the semiconducting layer, creating an electric current at the off state. Therefore, it is
132 | Part I Functional materials: Synthesis and applications

important to have Ion /Ioff ratios as high as possible so that clear switching is always
possible. The threshold voltage corresponds to the minimum potential which needs
to be applied at the gate to obtain switching. It is necessary from a commercial point
of view to have a low and stable threshold voltage. The electronic properties of the
polymeric semiconductor and the nature of the dielectric layer play a major role in the
values of the threshold voltage.

–60 –20V
–40 –100V


0 –20 –40 –60 –80 –100

VDS/V Fig. 6.12. Typical I-V curve of a p-type OFET.

6.4.2 Conjugated polymers for field-effect transistors

Conjugated polymers need to possess some critical properties in order to be good can-
didates for OFET applications. First, the semiconducting polymers must be stable over
time and, as mentioned earlier, must possess a HOMO energy level lower than −5.2 eV
to be air-stable. Stability is important, but the main factor which allows conjugated
polymers to attain high mobility is their ability to form highly organized films. It has
been shown that the closer the π-stacking is, the better the charge carrier mobility.
The strategy for obtaining a polymer with good packing properties is to use planar
conjugated units which will drive efficient π-stacking. To obtain processable materi-
als, alkyl chains are added along the conjugated backbone. Obviously, adding alkyl
chains is not necessarily good for the π-stacking efficiency, but the goal is to find a
compromise between the alkyl chains needed for solubility and the π-stacking in or-
der to obtain good mobility when solution-processed. In addition to polymer chain
stacking, there is also the point of how these stacks of polymer chains are oriented
6 Conjugated polymers for organic electronics | 133

relative to the substrate. It has also been shown that the packing of these polymers
should be oriented perpendicular to the surface in order to permit good charge circu-
lation through the source and drain electrodes [30]. See Fig. 6.13 for examples of the
best polymers reported so far for OFETs. Obviously they exhibit the right combination
of processability, π-stacking, and polymer chain orientations onto the substrates.

C10H21 C8H17
S H13C6 F F
S S S S n
O S n n S O
H17C8 H33C16 C16H33 C6H13
C10H21 C8H17


μh= 10.5 cm2V–1s–1 μh= 5.5 cm2V–1s–1 μe= 2.36 cm2V–1s–1

Fig. 6.13. Efficient conjugated polymers for p-type and n-type OFETs [31–33].

6.5 Applications in lighting

6.5.1 Light-emitting diodes

Light-emitting diodes (LEDs) are mostly used as lighting sources. Today, they can be
found in a large variety of electronic devices acting as electronic displays or light
indicators. The first LED was demonstrated in 1936 by George Destriau in the labo-
ratory of Marie Curie. When a potential is applied to ZnS:P powder placed between
two electrodes, the emission of a low intensity red light can be observed. That dis-
covery led to the development of a large variety of red, orange, yellow, and green
inorganic light-emitting diodes. In 1962, General Electric commercialized the first in-
organic LED which was based on phosphorous doped gallium arsenide (GaAs:P) [34].
Light-emission from an organic material was only reported for the first time in 1963
by Martin Pope and his collaborators at New York University. They demonstrated
that an anthracene crystal placed between two silver electrodes could emit blue
light [35]. Unfortunately, the lifespan of this type of device was far shorter than those
obtained with inorganic-based devices. The real interest in organic material for LED
came from the research of Ching W. Tang and Steven Van Slyke at Eastman Kodak
Research Laboratories in 1987. They invented an organic LED (OLED) based on a thin
evaporated layer of tris(8-hydroxyquinoline)aluminum (AlQ3 ), which led to high elec-
troluminescence yield and interesting lifespans [36]. It was only in 1990 that the first
polymeric LED (PLED) based on a π-conjugated polymer was demonstrated, when
Sir Richard Friend and his collaborators at the University of Cambridge observed a
134 | Part I Functional materials: Synthesis and applications

green light emission from a thin film of poly(para-phenylenevinylene), or PPV, placed

between two electrodes [37]. Since that discovery, a large number of small molecules,
oligomers, and polymers have been synthesized and used in light-emitting devices.
It is important to recall here that OLEDS are generally prepared using small vacuum-
processed organic molecules, whereas polymers have the advantage of being solution
processable, providing better opportunities for making large-area devices. Today,
many electronic devices made of OLEDs such as flat screen televisions, mp3 players,
digital cameras, and smartphones are commercially available and a number of com-
panies, such as Universal Display, Cambridge Display Technology (CDT), and Philips
produce lighting sources from conjugated small molecules, oligomers or polymers.
A large number of different configurations exist for OLEDs and PLEDs resulting
in different lifespans and efficiency. The simplest device configuration, and the first
which was demonstrated, consists of a light-emitting material sandwiched between
two electrodes. Electroluminescence is defined by the emission of light induced by
charge recombination. Therefore, in order for LEDs to function, charges must be in-
jected into the material. An electron is removed from the HOMO of the light-emitting
material at the cathode, creating a positive charge carrier. Inversely, at the cathode
an electron is inserted in the LUMO of the light-emitting material creating a negative
charge carrier. The hole and the electron travel to their opposite electrodes under the
influence of the applied electric field. They can also recombine to form an excited state
commonly called exciton. That exciton can be either in the singlet state, where the
spins of the electron/hole pairs are in the opposite direction, or in the triplet state
where the spins of the electron/hole pairs are in the same direction. The singlet state
is the only one which can lead to light emission in fluorescent conjugated polymers
(Fig. 6.14). However, phosphorescent materials can help to increase the efficiency of
such devices. It is interesting to note that light-emitting diodes behave essentially in
the opposite manner to photovoltaic devices; in other words, they create light from
electricity whereas solar cells produce electricity from the light source.






Fig. 6.14. Schematic working mechanism for LEDs.

6 Conjugated polymers for organic electronics | 135

In the case of the simple device made of only two electrodes and the light-emitting
material, the efficiency of the charge recombination is very low since the charge can
directly travel to the respective opposite electrode, leading to poor electrolumines-
cence. A large number of device architectures have been developed in order to contain
the electrons and holes in the light-emitting material. First, the light-emitting layer is
“sandwiched” between a hole-transporting layer and an electron-transporting layer to
maximize the containment of the charge carriers in the emitting layer (Fig. 6.15). Fur-
thermore, in order to contain even more charge carriers within the emitting layer, the
previous structure was optimized by adding a hole and electron blocking layer around
the emitting layer (Fig. 6.15). Containing the charges in the light-emitting material al-
lows for better efficiency and stability of the OLEDs and PLEDs. Such a multi-layered
structure is more adequate for vacuum-processed molecules than solution-processed
polymers and can explain the superior properties of OLEDs versus PLEDs. However,
the development of so-called orthogonal solvents (for instance, aqueous versus non-
aqueous polymeric solutions) could lead to multi-step depositions and lead to more
efficient PLEDs.


Electron transporting layer

Cathode Hole blocking layer

Electron transporting layer Light-emitting material

Cathode Light-emitting material Electron blocking layer
Light-emitting material Hole transporting layer Hole transporting layer
Anode Anode Anode
Substrate Substrate Substrate

Efficiency and stability

Fig. 6.15. Examples of architectures for OLEDs and PLEDs.

6.5.2 Conjugated polymers for light-emitting diodes

In order to commercially produce viable PLEDs made of conjugated polymers, they

must be soluble enough to be solution-processed and chemically stable over time.
A high glass transition temperature is also needed, since heat can be produced in the
device during its use and the morphology of the polymer film has to remain intact.
From an electronic point of view, the polymer should possess good charge injection
and charge transport to produce efficient PLEDs. Dependent upon the color needed,
the bandgap of the polymer needs to be tuned. High bandgap polymers can produce
136 | Part I Functional materials: Synthesis and applications

blue emission, while low bandgap polymers can produce a red emission. To obtain
a clear and clean color, the emission spectra of the polymer should be as narrow as
possible. In terms of performance, the device must present high electroluminescence
quantum efficiency (φEL ). That quantum efficiency is defined by the ratio of the num-
ber of photons emitted over the number of charges injected. This quantum efficiency
characterizes the electronic processes in the device but does not measure the light pro-
duced. Different approaches have been used to characterize the emitted light. First,
there are two units to characterize the light intensity: lumen (lm) and candela (cd).
Lumen corresponds to the total quantity of light emitted in all directions, whereas can-
dela represents the quantity of light emitted at a certain angle. In the literature, there
are three main parameters to measure in order to fully characterize the light emitted
by a LED. These are luminance (Lmax ), which is reported in candela per square me-
ter (cd/m2 ); luminous efficiency (LEmax ), reported in candela per ampere (cd/A), and
power efficiency (PE) in lumen per watt (lm/W). The latter is the figure of merit for
characterizing LEDs and is widely used by the lighting industry. Traditional lighting
technologies mostly rely on incandescent, fluorescent or halogen lamps. While light
bulbs show the lowest efficiency (10–35 lm/W), fluorescent or halogen lighting devices
are more efficient with power efficiency (PE) ranging from 50 to 100 lm/W. With perfor-
mance of about 15 lm/W, the best polymer-based light-emitting diodes (PLEDs) are not
yet competing with the best lighting sources [38, 39]. Future investigations should find
strategies to enhance (by about one order of magnitude) the luminosity of such de-
vices, probably by developing new multi-layered configurations and phosphorescent
polymeric materials. However, it is important to repeat here that vacuum-processed
OLEDs have already achieved the required electronic and optical properties to make
commercial devices.

6.6 Summary

Conjugated polymers are an important class of electroactive and photoactive mate-

rials. In the last 10–15 years, this research field has literally exploded owing to their
applications in energy, micro-electronics, and lighting, to name a few. However, in
order to obtain the desired performance with polymeric semiconductors, different pa-
rameters at the molecular and supramolecular levels (such as electronic structure,
molecular weight, regioregularity, solubility, morphology, etc.) must be carefully con-
trolled. These parameters are different for each application and should be considered
during the design of polymer molecular structures, the choice of monomers, and the
polymerization reactions.
With respect to applications, it is quite evident that conjugated polymers are
promised a bright future. Research is not only being conducted in academic institu-
tions, but many companies of all sizes are now involved in R&D projects devoted to
semiconducting polymers. On the basis of this tremendous momentum, it is firmly
6 Conjugated polymers for organic electronics | 137

believed that over the next few years conjugated polymers will continue to address
important problems of society, such as the production of renewable energy with min-
imum detrimental effects on the environment or printed electronic sensors for smart
packaging applications. As Benjamin Braddock (Dustin Hoffman) is told in The Gradu-
ate: “there is a great future in plastics. . . [or should we now say plastic electronics?] . . .
think about it” [40].


[1] Shirakawa H, Louis EJ, MacDiarmid AG, Chiang CK, Heeger AJ. Synthesis of electrically con-
ducting organic polymers : halogen derivatives of polyacetylene, (CHx ). J. Chem. Soc., Chem.
Commun. (1977), 578–580.
[2] MacDiarmid AG. Synthetic metals: a novel role for organic polymers (Nobel Lecture). Angew.
Chem. Int. Ed. 40 (2001), 2581–2590.
[3] Shirakawa H. The discovery of polyacetylene film: the dawning of an era of conducting poly-
mers (Nobel Lecture). Angew. Chem. Int. Ed. 40 (2001), 2574–2580.
[4] Heeger AJ. Semiconducting and metallic polymers: the fourth generation of polymeric materi-
als (Nobel Lecture). Angew. Chem. Int. Ed. 40 (2001), 2591–2611.
[5] Tourillon G, Garnier F. New electrochemically generated organic conducting polymers.
J. Electroanal. Chem. 135 (1982),173–178.
[6] Diaz AF, Kanazawa KK, Gardini GP. Electrochemical polymeriation of pyrrole. J. Chem. Soc.,
Chem. Commun. (1979), 635–636.
[7] Diaz AF, Logan JA. Electroactive polyaniline films. J. Electroanal. Chem. 111 (1980), 111–114.
[8] Edwards JH, Feast WJ. A new synthesis of poly(acetylene). Polymer 21 (1980), 595–596.
[9] Ballard DGH, Courtis A, Shirley IM, Taylor SC. A biotech route to polyphenylene. J. Chem. Soc.,
Chem. Commun. (1983), 954–955.
[10] McCullough RD. The chemistry of conducting polythiophenes. Adv. Mater. 10 (1998), 93–116.
[11] Leclerc M, Diaz FM, Wegner G. Structural analysis of poly(3-alkylthiophene)s. Makromol.
Chem. 190 (1989), 3105–3116.
[12] McCullough RD, Lowe RD. Enhanced electrical conductivity in regioselectively synthesized
poly(3-alkylthiophenes). J. Chem. Soc., Chem. Commun. (1992), 70–72.
[13] Chen TA, Rieke RD. The first regioregular head-to-tail poly(3-hexylthiophene-2,5-diyl) and a
regiorandom isopolymer: nickel versus palladium catalysis of 2(5)-bromo-5(2)-(bromozincio)-
3-hexylthiophene polymerization. J. Am. Chem. Soc. 114 (1992), 10087–10088.
[14] Loewe RS, Khersonsky SM, McCullough RD. A simple method to prepare head-to-tail coupled,
regioregular poly(3-alkylthiophenes) using Grignard metathesis. Adv. Mater. 11 (1999),
[15] Kirchmeyer S, Reuter K. Scientific importance, properties and growing applications of
poly(3,4-ethylenedioxythiophene). J. Mater. Chem. 15 (2005), 2077–2088.
[16] Roncali J. Molecular engineering of the band gap of π-conjugated systems: facing technologi-
cal applications. Macromol. Rapid Commun. 28, (2007), 1761–1775.
[17] Heeger AJ. Semiconducting polymers: the third generation. Chem. Soc. Rev. 39 (2010),
[18] Sakamoto J, Rehahn M, Wegner G, Schlüter AD. Suzuki polycondensation: polyarylenes à la
carte. Macromol. Rapid Commun. 30 (2009), 653–687.
138 | Part I Functional materials: Synthesis and applications

[19] Carsten B, He F, Son HJ, Xu T, Yu L. Stille polycondensation for synthesis of functional materi-
als. Chem. Rev. 111 (2011), 1493–1528.
[20] Mercier LG, Leclerc M. Direct (hetero)arylation: a new tool for polymer chemists. Acc. Chem.
Res. 46 (2013), 1597–1605.
[21] Ho H-A, Najari A, Leclerc M. Optical detection of DNA and proteins with cationic polythio-
phenes. Acc. Chem. Res. 41 (2008), 168–178.
[22] Rochat S, Swager TM. Conjugated amplifying polymers for optical sensing applications. ACS
Applied Materials & Interfaces 5 (2013), 4488–4502.
[23] He Z, Zhong C, Su S, Xu M, Wu H, Cao Y. Enhanced power-conversion efficiency in polymer solar
cells using an inverted device structure. Nat. Photon 6 (2012), 593–597.
[24] Park SH, Roy A, Beaupré S, et al. Bulk heterojunction solar cells with internal quantum effi-
ciency approaching 100 %. Nat. Photon 3 (2009), 297–302.
[25] Peters CH, Sachs-Quintana IT, Kastrop JP, Beaupré S, Leclerc M, McGehee MD. High efficiency
polymer solar cells with long operating lifetimes. Adv. Energy Mater. 1 (2011), 491–494.
[26] Zhao G, He Y, Li Y. 6.5 % efficiency of polymer solar cells based on poly(3-hexylthiophene) and
indene-C60 bisadduct by device optimization. Adv. Mater. 22 (2010), 4355–4358.
[27] Zhang Y, Zhou H, Seifter J, et al. Molecular doping enhances photoconductivity in polymer bulk
heterojunction solar cells. Adv. Mater. 25 (2013), 7038–7044.
[28] Shaw JM, Seidler PF. Organic electronics: introduction. IBM J. Res. & Dev. 45 (2001), 3–9.
[29] Namdas EB, Saricifti NS, Heeger AJ. Semiconducting and metallic polymers. Oxford University
Press, New York, 2010.
[30] Sirringhaus H, Brown PJ, Friend RH, et al. Two-dimensional charge transport in self-organized,
high-mobility conjugated polymers. Nature 401 (1999), 685–688.
[31] Li J, Zhao Y, Tan HS, et al. A stable solution-processed polymer semiconductor with record
high-mobility for printed transistors. Sci. Rep. 2 (2012), 754.
[32] Wang S, Kappl M, Liebewirth I, et al. Organic field-effect transistors based on highly ordered
single polymer fibers. Adv. Mater. 24 (2012), 417–420.
[33] Park JH, Jung EH, Jung JW, Jo WH. A fluorinated phenylene unit as a building block for high-
performance n-type semiconducting polymer. Adv. Mater. 25 (2013), 2583–2588.
[34] Holonyak N, Bevacqua SF. Coherent (visible) light emission from Ga(As1–xPx) junctions.
Appl. Phys. Lett. 1 (1962), 82–83.
[35] Pope M, Kallmann HP, Magnante P. Electroluminescence in organic crystals. J. Chem. Phys. 38
(1963), 2042–2043.
[36] Tang CW, Van Slyke SA. Organic electroluminescent diodes. Appl. Phys. Lett. 51 (1987),
[37] Burroughes JH, Bradley DDC, Brown AR, et al. Light-emitting diodes based on conjugated poly-
mers. Nature 347 (1990), 539–541.
[38] Huang J, Hou W-J, Li J-H, Li G, Yang Y. Improving the power efficiency of white light-emitting
diode by doping electron transport material. Appl. Phys. Lett. 89 (2006), 133509.
[39] Beaupré S, Boudreault P-LT, Leclerc M. Solar-energy production and energy-efficient lighting:
photovoltaic devices and white-light-emitting diodes using poly(2,7-fluorene),
poly(2,7-carbazole), and poly(2,7-dibenzosilole) derivatives. Adv. Mater. 22 (2010), E6–E27.
[40] (Accessed January 28, 2014)
A. Soldera
7 Theoretical tools for designing microscopic to
macroscopic properties of functional materials
The very essence of science is to unveil the huge palette of surrounding mysteries.
Deciphering the forces that sustain an object in water, or understanding how an ap-
ple falls from a tree are some well-known findings which have made a huge step to-
wards comprehending our world for humanity. Our contemplation of the world thus
changes according to the current advancements of science. Thanks to the extensive
progress made with experimental techniques and computers, it has become possible
to reveal some aspects of the behavior of the microscopic domain. From a computa-
tional viewpoint, approaches in solving present-day problems are certainly different
than those used 50 years ago. In the realm of functional materials, have we actually
reached what Dirac desired in 1929 in the Proceedings of the Royal Society of London.
Series A, Containing Papers of a Mathematical and Physical Character, Vol. 123, No. 792
(6 April 1929):

The underlying physical laws necessary for the mathematical theory of a large part of physics
and the whole of chemistry are thus completely known, and the difficulty is only that the exact
application of these laws leads to equations much too complicated to be soluble. It therefore
becomes desirable that approximate practical methods of applying quantum mechanics should
be developed, which can lead to an explanation of the main features of complex atomic systems
without too much computation.

This chapter intends to address the issue raised by Dirac to a certain extent. More
specifically, we will explore whether codes and computers are efficient enough to re-
veal the behavior of the complex atomic systems which ultimately lead to the world we
know. This would mean that, by improving computer performance, such as reaching
exascale computing and beyond, solving the “equation of everything” [1] i.e., the fa-
mous Schrödinger’s equation, the whole mechanism of the microscopic world would
be revealed. At the risk of disappointing the reader, the answer seems to be “No, it
cannot”, as claimed by Philip W. Anderson (Nobel prize in Physics in 1977) [2]. Viscos-
ity is an illustrative example. It can be roughly explained by the presence of dissipa-
tive forces which cannot be deduced as a result of solving the Schrödinger’s equation.
The answer actually divides two wholly opposed viewpoints: reductionism and emer-
gence [3, 4]. Reductionism envisions understanding a system through its “elementary”
constituents and their interactions. Emergence can be summarized by the vision of
Plato: the whole is greater than and different to the sum of its parts. Between these
two opposing perspectives, obtaining a clear response is not straightforward. Molec-
ular simulation is in some sense to illuminate this antagonism. Although this chapter
is not intended to discuss this aspect in detail, such a reflection should nevertheless
be kept in mind.
140 | Part I Functional materials: Synthesis and applications

The first part of the chapter is devoted to those issues ongoing from microscopic
to macroscopic scales, and thus the importance of possessing notions in statistical
thermodynamics to make meaningful molecular modeling. Examples of this state-
ment are presented in the second part of the chapter, and aim to illustrate the different
uses of simulation. These two examples originate from collaborations with my distin-
guished colleagues from the Quebec Centre for Functional Materials (CQMF), Profes-
sors Claverie and Morin, who are also authors of Chapters 2 and 3 in this book.

7.1 Methods

7.1.1 The link between microscopic and macroscopic scales

At the heart of molecular simulation it is necessary to establish the link between micro
and macro scales. Statistical physics is the branch of science which provides the tools
to build this bridge. However, relationships are first established for systems with no or
very weak interactions. Molecular simulation can be seen as an experiment performed
with a computer using a code [5]. It thus provides the means to apply such tools in com-
plex systems such as functional materials. However, special care is needed with regard
to applying these tools. The following example illustrates the nature of the problem
when deducing macroscopic properties from the microscopic realm. Without a doubt,
it takes some liberties from a rigorous approach which would necessitate a greater dis-
cussion of statistical physics. It is actually intended to give a simple overview of the
general issues required to relate the micro and the macro states and their properties, as
schematically summarized in Fig. 7.1. Consider a glass of water in a room whose ambi-
ent temperature is 10°C. Now, a Maxwell demon [6], instead of opening a gate to select
molecules with specific velocity, measures the speed of the water molecules inside the
glass of water by observing each individual water molecule. Moreover, he clearly ig-
nores any uncertainty principle raised by Heisenberg (see equation (7.2)). Accordingly,
he is able to provide the speed (vo ) of one molecule of water. The equipartition theo-
rem indicates that the square of speed is related to the temperature¹. He thus finds that
the computed temperature is very high, strictly higher than the ambient temperature.
Despite his demon character, he is astonished and says “Hey, there is some nonsense
going on here”.
Back to reality, the actual incongruity comes from applying the equipartition the-
orem to the speed of only one molecule. The real explanation is summarized in Fig. 7.2.
Distributions need to be considered, and the Maxwell–Boltzmann distribution must
be applied in this case in order to compute an average value. By integrating the sta-

1 More rigorously, each degree of freedom described by the square of any coordinate corresponds to
1/2 kB T, where kB is the Boltzmann constant.
7 Theoretical tools for designing functional materials | 141

5 0
4 0 νo
3 0
2 0
1 0
0 0
1 0
2 0

ν2 α T To >> 10°C Fig. 7.1. Principle of inconsistency in measur-

ing only properties of a single molecule.

tistical weights over the whole domain, or, in the case of discreteness, by adding up
all the possibilities, the result is one. Distributions are thus required to handle mi-
croscopic properties. Only then can the comparison between speeds measured at the
microscopic level and the value of the macroscopic temperature be undertaken in a
meaningful fashion. Applying this rule appropriately is at the heart of molecular sim-
ulation. Instead of speed, you can think of any other microscopic property, be it a
scalar, a vector or a tensor. Of course, it is not applied when you are only interested in
the configuration exhibiting the minimum energy, as in the crystal.

Micro Macro
ν2 α T

Distribution ρ(ν)
(Maxwell-boltzmann) Fig. 7.2. Distribution is required
when microscopic properties are
compared to macroscopic values;
vo is the result of the fishing in
V Fig. 7.1.

In Gibbs’ approach to statistical thermodynamics, the calculation of the average of a

distribution depends on the statistical ensemble of systems being applied. Consider a
distribution of systems in which the probability is described by pi for the appearance
of the ith system. Then ∑i pi = 1. For the sake of clarity, the integral form is not in-
troduced here. To compute the system ensemble average of a property (A), one needs
to compute, ⟨A⟩ = ∑i Ai pi where Ai is the value of the property A for the ith system.
At this point, the ergodic hypothesis must be introduced briefly. It specifies that the
value of A over a huge number of systems at a specific time, ⟨A⟩ equals the average
stemming from considering the distribution of the property of a single system over
a “long” period of time, A.̄ When Ā = ⟨A⟩ the computed property can be compared
to the experimental one. For a more detailed description of the ergodic hypothesis,
interested readers are referred to the seminal book of Allen and Tildesley [7]. How-
ever, computing Ā is time-consuming and it can be difficult to obtain by simulation.
A most accurate representation of ⟨A⟩ is thus mandatory. Experimentally, ergodicity
142 | Part I Functional materials: Synthesis and applications

is not attained in certain systems, such as molecular glasses and polymers. However,
conceptually, a simulation code must be ergodic.
The different scales have thus been exposed. The intent of this chapter is not to
provide the theoretical development underlying each scale. A list of books focusing on
this topic is available at the end of the chapter. This chapter aims to reveal the “very
substance”, as Gargantua said to Pantagruel, of molecular simulation, and to empha-
size the main difficulties most students will encounter during this course. The bridge
between scales will also be discussed.

7.1.2 Ab initio methods

“In the beginning was” Schrödinger’s equation, whose expression when it is not time-
dependent is:
Hψ̂ (r)⃗ = Eψ (r)⃗ (7.1)
This equation indicates that the Hamiltonian operator H,̂ acts on a wave function, ψ,
which depends on coordinates of electrons r ⃗ (considering only electronic contribu-
tion). This operation leads to the energy (eigenvalue) multiplied by the same wave
function which is then known as an eigenfunction. In bounded states² the resulting
energies are quantified. A difference in energy between two levels is then related to
a frequency which can be measured experimentally. Introducing the Dirac notation,
the ensuing equation (Fig. 7.3) can be represented similar to the one shown in Fig. 7.2,
and indicates that the resulting energy corresponds to an average. However, there is a
clear distinction to be made between these two averages. For the quantum average in
Fig. 7.3, the microscopic state, described by ψ, is known; in this case, the ensemble is
constituted only of systems in the ψ state, which is not necessarily an eigenstate of the
Hamiltonian. In statistical thermodynamics, in addition to the quantum uncertainty,
there is uncertainty regarding the state of the system; thus, the ensemble average of
Fig. 7.2 is over a distribution of systems in different states, each with its own statistical

ψ Ĥ ψ E
Distribution Fig. 7.3. Schrödinger’s equation in Dirac’s notation.

In the kingdom of electrons, it might seem amazing that no classic apparatus is able to
reveal one electron specifically. What is the actual problem? In a first step, we consider
our classical world, and try to use our understanding of it to disclose the microscopic

2 The electron confined to the field exerted by the nuclear charge is an example of confined state.
7 Theoretical tools for designing functional materials | 143

environment. The use of the adjective ‘classic’ indicates that the world is governed by
the classical laws of physics, such as the Newtonian laws of mechanics, or Maxwell
equations in optics. So, if we want to find our way in a room in the middle of the night
and avoid any untimely meetings with chairs, we need a flashlight. Does this light
change our environment? The answer is clearly no. Walls do not move, nor does their
shape change as a result of the beam of our flashlight. Can we use such an experiment
at the electron scale? Before answering this question, we have to answer to at least
two other ones.
First question: is the electron a particle, or a wave? The answer depends on the
kind of experiment carried out to reveal the presence of the electron. To symbolize this
problem, we use the image of the 2D world proposed by Carl Sagan in his famous book,
Cosmos, to represent the fourth dimension [8]. A cylinder is a 3D object. Its base lies on
the (x, y) plane. The demon (taking a little break from fishing for water molecules) has
only two possibilities for examining it: through the x or y axis, and along the z axis.
He actually changes color according to the direction he looks. In the first case, he ob-
serves a rectangle while in the second situation, a circle is revealed to him (Fig. 7.4).
Accordingly, dependant on the experience of looking at the cylinder from two possi-
ble observation points, the cylinder is either a rectangle or a circle. In the same way,
the electron is either a particle or a wave, depending on the type of instrument used
to observe it. Let us take a brief detour from this particular wave-particle duality of
electrons. In the late nineteenth century, the electron was recognized as carrying a
negative charge instead of being considered to be embedded in ether. The existence of
subatomic corpuscles was evidenced by J.J. Thomson in 1897 [9]. However, De Broglie
showed in his Ph.D. thesis published in 1924 that particles, e.g., electrons, behave as
waves. This was experimentally confirmed the following year [10].

Fig. 7.4. In a 2D world, is a cylinder
X a rectangle or a circle?

The second question concerns the fact that the position and momentum (velocity
times weight) of the electron cannot be determined at precisely the same time, i.e.
Heisenberg’s uncertainty principle:
σpx σx ≥ , (7.2)

144 | Part I Functional materials: Synthesis and applications

where σpx and σx are the standard mean deviation of (respectively) momentum and
position along the x axis, and h is the Planck constant.
The intrinsic behavior of electrons is thus barely grasped, is even impossible to
grasp, by classical methods. Using an analytic tool, i.e. the equivalent of the flash-
light but for electrons, the state of the electron will automatically change. How then is
the configuration of electrons uncovered? By using mathematics! Very unpopular an-
swer, as I usually hear from the students in my quantum chemistry class! The reader
can however be reassured: the purpose of this overview of simulation applied to func-
tional materials is not to develop mathematical models. It is aimed at explaining how
mathematics is used to grasp details that govern the microscopic world. Mathemat-
ical description can be retrieved from those textbooks suggested at the end of this
The equation in Fig. 7.3 indicates that the state of the system, and accordingly of
the electrons, can be described by a mathematical function (ψ), called the wave func-
tion. When it is squared, this equation leads to a physically relevant result: the prob-
ability of finding an electron somewhere in the space (Born interpretation). However,
it can only be solved analytically for one electron. Indeed, adding another electron,
combined with the fact that electrons are indistinguishable particles (being fermions),
as well as the presence of a spin and Heisenberg’s uncertainty principle make analyt-
ical solution of Schrödinger’s equation impossible. Computing Coulomb interactions
between these two negative charges becomes thus unfeasible, since their distance of
separation remains unknown. Accordingly, approximations are required. The prereq-
uisite in manipulating approximations to solve the Schrödinger equation gives rise to
different approaches. They can be roughly classified into three major types: ab initio
(Hartree–Fock method and derivatives), density functional theory (DFT), and semi-
empirical methods. Ab initio is a Latin phrase which means “from the beginning”. In
this approach, only fundamental constants are used and no experimental data are
introduced to compensate for the use of approximations. These approximations thus
lead to numerous commercial or freeware codes such as Gaussian [11, 12], Gamess [13],
DMol [3] Accelrys [14], and others. The Hartree–Fock method is the most common ab
initio approximation employed by scientists. The usual approximation employed by
chemists is to handle hydrogen atomic orbitals³. The intrinsic reason for employing
hydrogenoid atomic orbitals comes from the fact that Coulombic interactions between
electrons can be reduced to a central core problem. This makes it much easier to cal-
culate, through the introduction of the mean-field concept: each electron undergoes
interactions with all other electrons. In order to get the final wave function, a self-
consistency procedure is then applied. Nevertheless, initial wave functions are nec-
essary. Hydrogenoid wave functions are chosen since they correspond to the exact
solutions of the Schrödinger’s equation applied to the hydrogen atom. For practical

3 Physicists usually prefer the plane waves approach.

7 Theoretical tools for designing functional materials | 145

reasons, these wave functions are replaced by linear combinations of Gaussian type
orbitals (GTO), or Slater orbitals. A notation for GTO was introduced by Pople⁴ et al. to
indicate the number of Gaussian functions used to describe the different hydrogenoid
orbitals forming a basis set of equations. The essence of the basis set consists of func-
tions which fit the hydrogenoid atomic orbitals, such as 6-31g. 6-31g actually means
that 6 Gaussian functions depict the core electrons and the valence electrons are repre-
sented by a combination of two functions (Double Zeta), represented by 3 and 1 Gaus-
sian functions respectively; g stands for Gaussian orbitals. Specific functions can then
be added, such as polarization functions (d, p) to enhance directional effects, or dif-
fuse functions (+), to take long-range effects due to the presence of electronic delocal-
ization in the studied system. Post-Hartree–Fock methods, such as the Møller–Plesset
perturbation theory (MP), are aimed at refining this approach, but require more Cen-
tral Processing Units (CPU) and are therefore time-consuming.
In order to improve computer efficiency, the Density Functional Theory (DFT)
method offers a very interesting alternative. Within this approach, total energy is ex-
pressed in terms of total electron density, instead of the wave function which depends
on atomic position. This functional, i.e., the function of a function is inferred from
computations or experimental measurements. The PBE and hybrid B3LYP functionals
are the most commonly used, hybrid meaning that it also contains a Hartree–Fock
component. However, the choice of functional must be carried out carefully, since
a universal functional is not yet available. To further increase computer efficiency,
CPU and time consuming methods such as multiplication of overlap matrices can be
removed. These omitted data must then be balanced by experimental data such as the
heat of formation. The most-employed semi-empirical codes are the Austin model 1
(AM1) and Parameterized Model (PM3).
The main drawback in using quantum calculations is the prohibitive demand on
CPU. This problem increases with the number of atoms. As a typical example, the
computer time for the Hartree–Fock approach is proportional to N4 (which is the total
number of integrals to be evaluated), where N is the number of basis functions used to
described the studied system. For DFT, the computer time increases to N3 . DFT is gen-
erally employed for large systems. In practice, optimization (i.e. finding the structure
of lowest energy), can be carried out using a low level basis set. It can then be fol-
lowed by calculations carried out at a higher level of theory without the optimization
step, leading to single point energy. Alternatives also exist to address problems stem-
ming from larger systems such as polymers, where long range and large time scale
are relevant factors in computing properties. Thus, instead of considering electrons
as wave functions, the atomistic approach depicts atoms as particles with electronic
interactions embedded in a force field.

4 Nobel Prize in chemistry in 1998 “for his development of computational methods in quantum chem-
istry", shared with Kohn “for his development of the density-functional theory”.
146 | Part I Functional materials: Synthesis and applications

The Schrödinger’s equation and the different approximations are used when spe-
cific interactions involving electrons are of importance: electronic conjugation, proton
dissociation, infrared and NMR spectra, hyperpolarizabilities and others. However, it
is not always necessary to consider specifically all electrons during calculations. As
an atom moves in a medium, interactions with other atoms can be approximated in
many cases. This link to another level is discussed in the following section.

7.1.3 Bridging the gap between ab initio and atomistic levels

The graph shown in Fig. 7.2 corresponds to the Maxwell–Boltzmann distribution of

gas molecules. It may be used to reveal the bridge between the electronic and atomic
levels [15]. Rather than debating this 3D distribution, we focus on the distribution of
momentums in one direction only (x axis), which has the following form:

1 1/2
f(px ) = ( ) exp (− ), (7.3)
2πmkB T 2mkB T

where m is the weight of the particle, kB is the Boltzmann constant, and T is the tem-
perature. Again, this equation might seem complex, but it is essential to further under-
stand this approach. The shape of this distribution corresponds to a Gaussian curve
that presents the following generic form:

1 x2
g(x) = exp (− 2 ) , (7.4)
σ√2π 2σ

where σ is the standard mean deviation. This Gaussian curve is found in many systems
containing purely random events: tossing coins, dice etc. . . For example, let’s consider
tossing coins, for which the result can be either heads or tails. One event can be tossing
a coin 30 times. What is then the probability that in these 30 flips, the result is 15 times
heads, and thus 15 times tails? This event is actually x in equation (7.4), and the prob-
ability that this event occurs is g(x). However, in order for the discreteness⁵ (flipping
the coins, or the microscopic level) to reach continuity (Gaussian curve, and thus the
macro level), the number of events must be important. In fact, there is a theorem,
the central limit theorem, which actually specifies that the standard mean deviation
varies as 1/√N, where N is the number of events.
By comparing equations (7.3) and (7.4), it can be found that σpx = √mkB T. Inserting
this value into equation (7.2) leads to the conclusion that uncertainty in the x position
of a particle of weight m must be greater than h/(4π√mkB T). Thus at 300 K, the pre-
cision in the position of the electron is of the order of 15 Å, while for the hydrogen

5 This discretization of the distribution must be put in parallel with the measuring of the velocity of
the water molecules one by one as the demon did.
7 Theoretical tools for designing functional materials | 147

(van der Waals radius rH = 1.2 Å) and carbon atoms (van der Waals radius rC = 1.7 Å),
it is 0.35 and 0.1 Å respectively. This confirms that electrons are quantum particles,
since we have no idea where they can be located. Moreover, heavy atoms such as car-
bon atoms can be considered to be governed by classical mechanics. However, the
classical character of the hydrogen atom is difficult to establish clearly, although it
can generally be considered to be a particle, and can then be described by classical

7.1.4 Atomistic simulation

At the heart of any atomistic simulation we find the force field. Originally stemming
from vibrational spectroscopy, it was applied to depict interactions between all atoms,
including non-bonding terms. For example, consider an atom in a specific environ-
ment, for instance two linked sp3 carbon atoms. This environment can be defined by
a particular potential energy surface (PES). Such a PES does not change greatly if the
same environment is found in another system or molecule. Following the previous ex-
ample, the number of molecules possessing this C–C link is unlimited. A force field
is then aimed at representing the PES as accurately as possible. To achieve this goal,
a force field in fact consists of (1) equations in which terms tend to grasp specific in-
teractions, and (2) parameters which stem from the fit of these equations to the dif-
ferent PES. Experiments or quantum calculations are carried out to accurately depict
the PES. Returning to the C–C link, we will show how PES associated with interactions
stemming from this molecular characteristic is computed, and subsequently reveal the
great diversity in force fields. We actually describe the link between two carbon atoms
in a simple ethane molecule.

Eh = k (d–d0 )2 Harmonic function
EM = Ediss. [1–exp(–α (d–d0 ))]2
Energy (kcal/mol)

Morse function
60 4
40 2
20 0 Fig. 7.5. Potential energy with respect
1.4 1.5 1.6 1.7
0 to the C–C distance in an ethane
0 1 2 3 4 5 6 molecule. The dotted curve repre-
C-C distance (A) sents the actual energy behavior.
148 | Part I Functional materials: Synthesis and applications

The PES associated with the C–C elongation can be computed using quantum
methods. It is in fact reduced to a simple curve (Fig. 7.5). The potential energy is thus
computed, while the distance between two carbon atoms in an ethane molecule is var-
ied. A dotted line, joining the different data, is displayed. The Morse function is usu-
ally employed to fit these data: it is displayed in gray in Fig. 7.5. As can be seen, three
parameters are required to fit the curve: the dissociation energy, Ediss. , a parameter
which describes the width of the potential well, do , which corresponds to the distance
of minimum in potential energy, the actual equilibrium distance between the two car-
bon atoms, and α which defines the well width. Since a covalent bond is exposed to
small fluctuations around do most of the time, the Morse function can be expanded
in a Taylor development at d = do until the third term. This leads to a harmonic func-
tion (in black in Fig. 7.5), revealing in fact an elastic behavior (Hooke’s law). The three
parameters of the Morse function used to fit the data are then replaced by only two
fitting parameters: the spring constant, k = Ediss. α2 (stemming from the Taylor devel-
opment), and the equilibrium distance, do . The link between two atoms can thus be
described by either of these two functions. The harmonic function is usually preferred
to the Morse function since it needs two fitting parameters instead of three, and more
importantly it does not use an exponential function which consumes more CPU time.
Nevertheless, as can be observed in the inset of Fig. 7.5, a quadratic function is not
sufficient to describe interactions between two atoms properly. Computational data
are thus fitted by a 4th or 6th order polynomial. The higher the order, the better the
fit, and thus less important errors leading to better transferability (i.e., the availability
of a force field for the study of a great number of systems). Depending on the kind of
calculations to be carried out, several functions are thus available. In fact, there are
other parameters which also differentiate the force fields:
– Parameterization: what is the chosen value for do , 1.53 or 1.54 Å, or is greater pre-
cision needed? Actually, the equilibrium value depends on the kinds of systems
investigated and the type of calculation carried out.
– Environment: how many different environments of an atom are needed? Consider
the sp2 carbon atom. If only one series of fitting parameters is selected, one will
certainly encounter problems when willing to differentiate between alkene and
ketone bonding. In the case of second generation force field, carbon atoms can
have 20 different environments, making such a force field more transferable.

In summary, interactions between two linked atoms require a function and parame-
ters specific to each atom in order to be represented. However, the potential energy to
describe a molecule in a medium needs additional terms. They are characterized in
Fig. 7.6, where 4 atoms are schematically displayed.
The force field equation is in fact separated into two terms: bonding and non-
bonding equations.
7 Theoretical tools for designing functional materials | 149

φ rij

Fig. 7.6. Representation of the variables discussed in equations (7.5–7.9).

Bonding interaction terms

The bonding interaction terms depend on the internal coordinates of the molecule,
i.e., bonds, valence angles, and dihedral angles. Accordingly, its simplest expression
consists of the sum of all the terms that are functions of stretching, bending, and tor-
sion found in a molecule. Thus, in addition to the stretching term previously defined,
a simplified force field, usually classified as a class I generation force field, possesses
valence and dihedral angle terms to describe the bonding interactions.
– Stretching term:
Estretching (d) = ks (d − do )2 (7.5)
where d is the atomic distance at which the stretching potential energy is com-
puted, ks and do are the spring constant and the value of atomic distance for which
the stretching potential energy is zero respectively. This actually corresponds to
the black curve in Fig. 7.5.
– Bending term:
EBending (θ) = kθ (θ − θo )2 (7.6)
where θ is the valence angle at which the bending potential energy is computed,
kθ and θo are the spring constant and the valence angle for which the bending
potential energy is zero respectively.
– Dihedral term:
Ediheadral (φ) = ∑ Vn [1 − cos(nφ − φ0n )] (7.7)

where φ is the torsion angle at which the dihedral potential energy is computed,
Vn corresponds to the potential energy barrier, and φ0n the phase.

Other forms of bonding terms exist in a class I force field, such as the out-of-plane
term, due to variations in the planarity of a group of atoms as in the ester group. More-
over, due to increasing computational efficiency, bonding interaction cross terms, as
well as higher terms in Taylor development, can be added to better depict the PES,
leading to second generation force fields which cover a greater number of compounds
and thus have greater transferability. Accordingly, a great variety of force fields exists:
pcff, COMPASS, OPLS, MM2, MM4, UFF, AMBER, ESFF, CVFF, etc. [16–18]. Although
not explored in this chapter, interested readers are invited to read about the work of
Sun [16].
150 | Part I Functional materials: Synthesis and applications

Non-bonding interaction terms

Non-bonding terms express the interactions between linked atoms separated by a cer-
tain number of atoms (2 in Fig. 7.6), as well as between atoms belonging to different
molecules. They generally consist of two terms.
– van der Waals interactions
Aij Bij
EijLJ (rij ) = 12 − 6 (7.8)
rij rij
where Aij and Bij are fitting parameters, and rij is the actual distance between
atoms i and j. The equation displayed in (7.8) corresponds to the Lennard-Jones
potential: it is formed by a repulsive (in r−12 −6
ij ) and a dispersive (in rij ) term. While
the latter term results from the London–Keesom–Debye interactions [19], the for-
mer term can have different forms. The r−12 ij behavior can be substituted by an
exponential decay, exp(−ρrij ) to yield the Buckingham potential, or a less abrupt
function, r−9
ij [20].
– Electrostatic, or Coulomb potential energy:
qi qj
EijCoulomb (rij ) = (7.9)
4πεo rij

where qi , qj , and εo are respectively the partial charges of atoms i, j, and the di-
electric constant. The partial charges reproduce the first nonzero multipolar com-
ponent. They stem from the difference in electronegativity between atoms sharing
a bond.

Different equations are thus available to portray the whole PES. The approach used
to establish the PES from experiments or calculations can lead to different fitting pa-
rameters and also contribute to the great diversity of available force fields. The Monte-
Carlo method [21] and molecular mechanics consist of direct use of a force field. Static
properties such as energies [22], infrared spectra [23], and mechanical properties [24]
can be computed with the latter approach. Molecular dynamics (MD) is required to
reveal dynamic properties, such as autocorrelation function [25, 26] and relaxation
times [27]. It consists of integrating equations of motions, usually the second law of
Newton, to give mobility to atoms considered to be simple particles with interactions,
as previously stated. Most useful information is obtained by using MD. However, the
main problem with MD simulation is that the time period covered is usually small. In
order to integrate Newton’s equations, a time step (δt) is required for which various
types of integrators could be used. For example, a peculiarity of the Verlet algorithm
and its derivatives is the use of an “important” time step: δt = 1 fs, or 10−15 s. This
means that to produce a trajectory, i.e., a series of conformations during 1 ns, solv-
ing the equations would necessitate 1 000 000 integrations! Each integration consists
of computing the velocity and position of all atoms, according to the selected force
field (second law of Newton). So why would it be necessary to use such a low integra-
tion step? The main reason stems from the fact that all of the motions involving atoms
7 Theoretical tools for designing functional materials | 151

must be correctly reproduced. The most rapid motion existing among atoms is the C–
H stretching vibration, of the order of 10 fs. Thus, an integration step of 1 fs using the
Verlet algorithm makes the use of MD available to accurately describe atoms’ motions.
To increase the duration of MD, it is necessary to diminish the number of integration
calculations and/or increase the integration step. One method is to replace C–H with
a bigger “C” atom whose force field parameters will definitely be different to the whole
atom description. Such a procedure is called coarse-grained simulation and it is at the
beginning of the mesoscale level simulation.

7.1.5 Bridging the gap between atomistic and mesoscale levels

As the duration of MD increases, dissipative effects emerge. The Newton equations

of motion occur in systems with constant energy. In order to introduce a temperature
term into a system, various algorithms can be used. However, dissipative effects lead-
ing to an exchange of energy cannot be easily inserted into such systems, since all
these effects take place at the mesoscale level. The purpose is thus to reproduce the
general behavior in a more efficient way. The actual problems stemming from atom-
istic simulation are the large number of degrees of freedom. We must therefore distin-
guish between those that are of interest for the simulation and those that must reach
equilibrium rapidly. The general behavior of big molecular systems must be correctly
simulated, while specific details can be neglected at the mesoscale level. Let us con-
sider a polymer mixed in a solvent, where the polymer behavior is the main interest.
Details of the solvent molecules can actually be ignored in the simulation of such a
system. Similarly, if this polymer corresponds to a very long chain, details of atoms
or valence angles and dihedral angles are not of importance in describing how the
polymer behaves. On the other hand, the center of mass of a group of atoms whose
length can be Kuhn’s length is more important. Dissipative particle dynamics simula-
tion uses such approximations, and is very promising. We recently published a study
where we showed that water channels can be formed in fuel cell membranes by ap-
plying shearing. They are observed only for long chains and low water content [28].
Despite growing interest, mesoscale simulation is not discussed in this chapter.
The focus is on scales where chemical functionality is explicitly pictured, i.e., the
quantum computational and atomistic domains. Examples of the applications of these
two domains of approximation will be the focus of the next section.

7.2 Examples

Molecular simulation is a powerful tool but it must be manipulated with great care.
Calculations can be run easily (push-button temptation stemming from the simple
handling of commercial codes), providing some data. Conclusions can then be drawn
152 | Part I Functional materials: Synthesis and applications

rapidly, but how effective are they exactly? Many questions must be answered in or-
der to run a simulation efficiently. For instance, is the pertinent portion of the phase
space correctly explored? Have the right calculation characteristics been chosen, such
as the functional in DFT or the force field in atomistic simulation? Instead of exposing
a list of criteria which must be fulfilled before extracting any meaningful conclusion
from calculations, this section will emphasize the procedure that must be employed.
Since molecular simulation is based on approximations, correlation of results stem-
ming from calculations with experimental data must be carried out first. Corrections
can then be brought to the type of calculation or to the protocol used. Examples will
be discussed in order to illustrate this procedure.
The very purpose of simulations performed in the lab is to guide the synthesis
or formulation of new compounds through a better understanding of the molecular
interactions that give rise to the macroscopic properties. Functional materials partic-
ularly suit this approach, since small changes in the molecule can lead to important
changes in the final properties. Accordingly, if molecular simulation can deal with
such changes, new molecules with enhanced properties can be proposed. Molecular
simulation is thus particularly aimed at supporting experimentalists in their studies.

7.2.1 Quantum studies

Quantum calculations are of particular interest to guide experimentalists in the de-

velopment of materials and the optimization of their properties. Two illustrations of
this objective are reviewed. The first example is related to collaboration between our
laboratory and the laboratory of our colleague Professor Jean-François Morin, author
of Chapter 3 in the current book. The other example deals with a study carried out in
our laboratory.

Electronic properties [29]

Professor Morin and his group study organic and molecular electronics (Chapter 3).
Such compounds are very promising materials due to their great ability to tune
their properties by modifying their chemical structure. Among these compounds,
fullerene (C60 ) is a very interesting n-type candidate, since it can easily accommodate
negative charges in a reversible way. The reason for this peculiarity stems from the
presence of a triply degenerate lowest unoccupied molecular orbital (LUMO) level
which allows the addition of up to six electrons to a single fullerene cage. However,
the LUMO of C60 is 4.5 eV, which is too high for electronic applications. Variations of
this value can be addressed through the addition of specific groups. Molecular sim-
ulation then becomes of great interest, since the effect of this group on the value of
the LUMO can be rapidly revealed. Nevertheless, comparison with experimental data
must first be carried out. Koopmans theorem is used to compute approximate ioniza-
7 Theoretical tools for designing functional materials | 153

tion energy, considering that it corresponds to the negative of the highest occupied
molecular orbital (HOMO). Computing electron affinity, i.e., negative of LUMO, is less
straightforward, due mainly to the fact that it is a virtual molecular orbital.
A series of new molecules derived from molecules with an ethynyl-bridged func-
tion and linked to fullerene have been synthesized by Professor Morin’s group and
have also been characterized electrochemically. Selection of the appropriate func-
tional and basis set was first carried out. The ensuing values for the LUMO were
compared to computational data (Fig. 7.7, left). Despite an underestimation of the
LUMO, a linear relationship was obtained for almost all of the synthesized com-
pounds. However, the LUMO of the molecule possessing a penta-fluoro phenyl group
at the α position of the ethynyl-bridge in the fullerene (Fig. 7.7, right), was found out of
range, with a very low LUMO. The experimental LUMO was −3.84 eV. Using the PBE
functional, a value of −4.05 eV was computed. The chemical structure of the molecule
indicates that it mainly possesses one degree of freedom that is the torsion angle (φ)
associated with the bond linking the penta-fluoro phenyl group to the fullerene. This
dihedral angle was turned in 10° steps. At each step, the LUMO was calculated. The
resulting graph is shown in Fig. 7.7 on the right. As observed in this graph, the value
of the dihedral angle greatly influences the value of the LUMO.

–3.85 –3.70
–3.90 –3.75 OC8H17
Lumo simulated/eV

Lumo simulated/eV

–3.95 φ
–3.80 F

–4.00 F
5 BM

–3.85 F



–4.05 F



10 20 30 40 50 60 70 80










φ (deg.)
Lumo experimental /eV

Fig. 7.7. LUMO with respect to the dihedral φ.

When fluorine atoms are in an ortho position relative to the 1-ethynyl-4-(octyloxy)ben-

zene group, φ = 10° (position shown schematically in Fig. 7.7, right), the highest value
of LUMO (−3.71 eV) is obtained. Rotating the phenyl group until the fluorine atoms are
approximately at the same distance from the fullerene (φ = 80°), results in a much
lower LUMO energy level (−4.05 eV). This difference of 10 % between the two values
can be directly attributed to the δ− character of fluorine atoms, in agreement with the
electrochemical results. In fact, substituting the fluorine atoms with hydrogen atoms
in the phenyl group (leading to compound 3 in Fig. 7.7, left) does not have any impact
on the value of LUMO by rotating the dihedral angle. Accordingly, calculations show
154 | Part I Functional materials: Synthesis and applications

that LUMO can be tuned when a hexa-fluoro phenyl group is linked to C60 . Molecular
simulation thus acts as a guide to the synthesis of new compounds. More specifically, it
does not give the final molecule, but gives hints for the development of new functional
materials. In other words, molecular simulation can be used to guide the synthesis of
new and more efficient materials. A different approach, resulting in a proposition of
very new molecules based on simulations, will be developed in the next paragraph.

Design of new membranes [30, 31]

Polymer electrolyte membrane fuel cells (PEMFC) are the primary system being de-
veloped for use in automotive applications. The perfluorinated polymer Nafion® re-
mains extensively used as a PEM. Since it cannot be operated at low humidity and/or
at temperatures above 100°C, great efforts have been dedicated to proposing alterna-
tives to this copolymer. They can be roughly classified into two main approaches. The
first series of molecules is derived from the structure of Nafion® , i.e., they possess a
polytetrafluoroethylene (PTFE) backbone and randomly distributed fluorinated ether
side chains, each terminated with a sulfonic acid group. The second type of molecule
possesses the sulfonate acid part mentioned above, but the rest of the core is gen-
erally polyether ether ketone (PEEK)-like. The limited development of new classes of
membranes may stem from the lack of understanding of the factors contributing to
the morphology, conductivity, and hydration of Nafion® . Great effort has been made
to develop new membranes through trial and error and serendipity, although these
strategies are not discussed in this chapter.
Nafion® is a long polymer chain with a huge degree of freedom. Its polytetraflu-
oroethylene (PTFE) backbone is extremely hydrophobic, while the fluorinated side
chains have high electron affinity which promotes the dissociation of the proton of
the sulfonic acid, making it a superacid. Due to the electronic contribution, quantum
method is the appropriate approach for dealing with proton dissociation. Rather than
considering all the chains, our group performed calculations focused on trifluoro-
sulfonic acid (TfOH), CF3 SO3 H. Our methodology was then to find conformation of
minimum energy as a water molecule is added to the system. Infrared spectra were
computed at each step and compared to experimental ones, which allowed us to no-
tice that considering only one TfOH is not sufficient to accurately describe the infrared
spectra. Two acidic molecules were found to be sufficient. However, studies with three
TfOH molecules involve an important number of degrees of freedom (important con-
figurational space), making access to all potential energy wells hazardous. The agree-
ment between selected computed and experimental frequencies indicated that the
structural changes occurring during proton dissociation could be used. This sequence
of molecular systems can be seen as a flip book (i.e., a book with a series of pictures
that vary gradually from one page to the next, so that when the pages are turned
rapidly, the pictures appear to animate by simulating motion or some other change
(Wikipedia)). Each image of the sequence corresponds to the addition of one water
7 Theoretical tools for designing functional materials | 155

molecule. Initially, the two sulfonic acid groups form a complex through hydrogen
bonding. Analysis of the structural changes as a function of added water molecules
reveals that only one sulfonic group, the protogenic group, is affected by the inclusion
of water molecules before the first proton dissociation occurs. In fact, the distance be-
tween the two sulfur atoms remains relatively constant at 4.3 Å, as well as the HO–H
angle ∼ 178°, as shown in Fig. 7.8. It was then argued that new bis-sulfonic acid molec-
ular scaffolds possessing this optimal S–S distance and comparable behavior of the
̂ angle might perform as well as or better than Nafion® . This molecule is shown
in Fig. 7.8: the distance between the two sulfur atoms and the angle correspond to the
geometrical characteristics computed in trifluorosulfonic acids. By calculation, it is
demonstrated that they exhibit the same structural behavior as the bis-sulfonic sys-
tem when water molecules are added to the system. Moreover, the rigid structure that
supports the two sulfonic groups can possibly help reduce the water content necessary
for the first proton dissociation, and lead to a very low pKa. Ultimately, the mechani-
cal properties of the ensuing membrane may be improved. Synthesis of the proposed
molecules is currently underway. At the moment of editing this book, the molecule
had not yet been synthesized. Its synthesis is clearly not straightforward.

1st proton dissociation

0 1 2 3
6 4.3 Å


0 1 2 3
n/2 H2O

Fig. 7.8. Some architectural features of the two TfOH molecules and the new molecule are compara-

Molecular simulation is thus a complementary tool to experiment with in order to

guide the design of new molecules with optimal properties. It is the combination of
simulation and experiment that yields interesting conclusions, in agreement with the
principle of emergence. The next section focuses on the use of atomistic simulation to
provide insight into molecular interactions.
156 | Part I Functional materials: Synthesis and applications

7.2.2 Atomistic simulation

Thermal transitions of polymers such as melting points (Tm ) and glass transition tem-
peratures (Tg ) can be difficult to measure experimentally. For example, the thermal
transition of rigid rod polymers can be so high that the compound starts decomposing
when the transition temperature is reached. In other cases, the transition can be very
weak and hardly observable with the usual instrumental techniques, such as differen-
tial scanning calorimetry (DSC), or differential mechanical analysis (DMA). Recently,
nanomaterials have become more important due to their potential industrial applica-
tions. However, the measurement of thermal transitions of dispersed nanoparticles
in a continuous phase is often problematic, as the amount of material constituting
the nanoparticles can be very small. Nevertheless, the knowledge of their thermal
transitions is a key parameter for understanding the morphology of the nanoparticle,
as well as its ability to aggregate. In this regard, atomistic simulation is well-suited
for studying nanomaterials, as it probes comparable dimensions and thus provides a
valuable tool for the prediction of thermal properties in functional materials. It has
to be pointed out that, due to the great number of atoms required to represent a tran-
sition, quantum calculations cannot describe such phenomena. Measurement of Tm
and Tg is explained in the following section.

The melting transition [32]

Fundamental properties of crystals at the nanoscale are significantly different to those
of bulk crystals and melting and crystallization temperatures rarely correspond to
those of the bulk. Properties of nanocrystals are also strongly influenced by their shape
and size. An efficient way to clarify experimental observations relies on the use of ap-
propriate computational tools to probe the very nature of interactions at stake. The
study of Tm stems from a collaboration with Professor Jérôme Claverie, author of Chap-
ter 2 in the current book. Professor Claverie and his group are interested in functional
groups containing polyethylene (PE), such as the carboxylic acid group. Recently, they
demonstrated that nanoparticles of PE are thermoreversible [33]. Understanding this
phenomenon is not straightforward and the use of molecular simulation becomes a
reliable asset in unveiling the fundamental nature of this thermoreversibility.
Simulation of the melting point of functional materials conforms to this vision.
From a simulation viewpoint, the question was: how can interactions inside a PE
nanoparticle affect the melting point? A protocol had thus to be carefully designed
to model such a transition and consists of a simple but realistic representation of the
actual material. A nanocrystal constituted of alkane chains of fixed length was used
to mimic the PE nanocrystals. How to compute the melting point? It is known that it is
a first order transition (heat capacity tends to infinite) with latent heat. Moreover, in
the case of alkane chains, conformations change from all-trans to a Boltzmann distri-
bution of the rotameric states. These experimental facts are then considered in order
7 Theoretical tools for designing functional materials | 157

to envision by simulation the melting transition. The pcff force field was used, since
it was created specially to simulate polymers appropriately. First, calculations were
performed on a crystal formed with chains 24 atoms long, 4 chains in each x and y di-
rection, as shown in Fig. 7.9. The crystal is in the vacuum.

250 300 350 400
σe x % Trans 0.9
Cv (J/g-K)

l 7.8

z –160
y –320
x 250 300 350 400
Temp (K)

Fig. 7.9. Simulation of the melting point.

From the original nanocrystal, MD are run at different temperatures, each tempera-
ture leading to a trajectory. In fact, positions and velocities of each atom are recorded,
making visualization of the melting phenomenon possible, as well as the computa-
tion of many properties (temperature, pressure, local dynamics, and so on). In order to
observe the melting transition, three properties are reported with respect to tempera-
ture: van der Waals energy, heat capacity, and percentage of the trans-rotameric states
(Fig. 7.9). It has to be mentioned that the heat capacity is computed using the fluctua-
tion formula, (⟨E2 ⟩ − ⟨E⟩2 / kb T2 ), where E is the energy, kb is the Boltzmann constant,
and T is the temperature. The thermodynamic definition (𝜕⟨E⟩/ 𝜕T)V of heat capacity
cannot be used since the thermodynamic limit⁶ is not attained. As shown in Fig. 7.9, it
is observed that a transition occurs at the same temperature for the three properties
reported, which clearly corresponds to the melting temperature of the nanocrystal. All
details on the simulation can be found in Metatla et al., as we only present the main
conclusions of the study [32].
In Fig. 7.10, melting temperatures are reported with respect to the inverse of
the thickness of the crystals. The lines correspond to the linear fit using the Gibbs–

6 The thermodynamic limit is attained when N/V = constant with N → ∞ and V → ∞.

158 | Part I Functional materials: Synthesis and applications

Functionalized PE
0.90 experience


0.80 Alkanes
Alkanes experience
0.75 simulation


0.2 0.3 0.4 0.5 0.6
1/l (nm–1)

Fig. 7.10. Gibbs–Thomson equation representation reporting experimental (dark blue line) and sim-
ulated (blue line) Tm for alkane chains, and experimental Tm for functionalized PE (dotted line).

Thomson equation:
Tm 2σe 1
=1− (7.10)
Tom Δhm l
where Tom and Δhm are the melting temperature and the melting enthalpy per unit vol-
ume of bulk alkane chain respectively, σe is the interfacial tension of the crystal in
the plane normal to z (Fig. 7.9), and l is the crystal thickness in nm. There is excel-
lent agreement in the slope between experimental and simulated data. Such accuracy
reveals the precision of the procedure and the description of the σe /Δhm ratio.
If we compare simulated data with experimental values stemming from function-
alized PE, there is a clear discrepancy in the slope. This difference is in fact of great
interest and fostered this collaboration. Clearly, the dissimilarity in the slope relies
on the value of the interfacial tension, σe ⋅ Δhm can be considered comparable to
alkane and PE chains, while changes in the interfacial tension directly result from al-
terations at the surface. The occurrence of defects brought by chain loops or function-
alized groups modifies σe . The presence of these defects can be definitely simulated,
leading to a change in the slope (Fig. 7.10), approaching that displayed by functional-
ized PE. Such treatment is undoubtedly a molecular viewpoint of the surface problem.
Supported by experiments, this result should lead to interesting conclusions. Further
simulations are presently being carried out, and first data reveal that the presence of
defects on the surface lead to a change in Tm .
7 Theoretical tools for designing functional materials | 159

The glass transition [34–41]

The glass transition temperature (Tg ) has been known since ancient Egypt, as crafts-
men knew how to process a glassy material by heating it above a certain temperature
in order to shape it to a certain design. They may not have handled glass as glassblow-
ers from Murano Island do nowadays, but they had identified that glass is fragile and
is a very malleable material above it. More than three thousand years later, in 1995,
the Nobel Prize-winning physicist, Philip W. Anderson (already encountered at the
beginning of this chapter), wrote in the magazine Science: “The deepest and most in-
teresting unsolved problem in solid state theory is probably the theory of the nature
of glass and the glass transition.” It is peculiar that this unique phenomenon remains
an unsolved problem. This is due to the fact that the time and length domains it covers
are huge: from nanoseconds (10−9 sec) to years (108 sec), highlighting the problems
of polymer aging. It is thus very difficult for an experiment or a theory to uncover this
domain entirely. Can atomistic simulation possibly reveal such phenomenon? At first
glance, the response is clearly negative. But let’s have a look at what exactly glass tran-
sition is. Following that, the protocol and some results will be exposed, and it will be
left to the reader to decide whether atomistic simulation is capable of dealing with the
glass transition of polymers.
At high temperature, a polymer chain can reach many states. More particularly,
the different rotameric states⁷ associated with a bond along the backbone bonds are
reachable, as the chain is not restricted in its movements at all. As the temperature
is decreased, crossing the energetic barrier between rotameric bonds in order for the
bond to rotate occurs less often. The polymer thus reaches a state which makes it very
difficult for the backbone to move. Imagine the 15-block puzzle displayed in Fig. 7.11, in
which you want to move the G block. You need to move several blocks to make a space
available near G. One motion thus involves several motions. This observation incorpo-
rates the concept of cooperation, in fact. This important concept in the study of glass
transition corresponds to the fact that positions of neighboring entities (molecules or
bonds) are strongly correlated, leading to a modulation of density extending over a
certain molecular scale.
The glass transition phenomenon is thus a longer-range phenomenon than the
length scale associated with several molecules. In some ways, it confirms that atom-
istic simulation is not the best tool for dealing with this complex transition. However,
Roe and Rigby [42] first showed that it is possible to simulate the glass transition of
polymers by using atomistic simulated dilatometry. As in experiments, the dilatomet-
ric technique consists of reporting the specific volume i.e., the inverse density, with
respect to temperature. As the system cools down, the specific volume is reported for
different temperatures. The departure from a linear relationship between specific vol-
ume and temperature yields the value of Tg , which effectively corresponds to the tran-

7 A rotameric state is defined by a well in the potential energy as a bond is rotated.

160 | Part I Functional materials: Synthesis and applications

G I a s

s T r a

n s i t

i O n
Fig. 7.11. A 15-block puzzle representing a polymer chain
in the glassy state.

sition from the rubbery to the vitreous phases. Examples are shown in Fig. 7.12. In fact,
numerous atomic simulation studies can be found in the literature. The main draw-
back of this method is that a great disparity is found in the values of Tg higher than,
equal to or lower than the experimental values. Questions necessarily arise about the
pertinence of and the interest in using all-atomistic simulation to compute Tg . One of
the reasons is that atomistic MD simulations mainly inspect the region of space located
near the initial portion. Moreover, the length scale associated with the glass transi-
tion should exceed the domain available with atomistic simulations; coarse-grained
simulations addressing the underlying theories accountable for the occurrence of
the glass transition [43]. Nevertheless, meaningful information can be extracted from
such studies: the effect of side chain, tacticity, and so on. Atomistic simulation must
be considered a very complementary tool, as shown in the next paragraph.
Prior to any simulation, great care must be taken to select the appropriate force
field. It must be chosen properly in order to specifically address the problem it is re-
quired to solve. The glass transition temperature of polymers can be greatly affected
by the choice of the force field, as shown in Fig. 7.12. Simulated dilatometries of one
polymer, isotactic PMMA, using two force fields, AMBER (black circles) and AMBER-
OPLS (gray diamonds), are displayed in Fig. 7.12. The difference in the force fields stays
in computation of the partial charges in the Coulomb potential energy (equation (7.9)).
In the latter force field, partial charges have been computed in order to reproduce liq-
uid density, while in AMBER, charges have been computed using the ESPR method
based on quantum calculations. AMBER-OPLS was deliberately chosen, since a glass
transition is observed when it is used (black circles in Fig. 7.12).
Once a glass transition temperature simulation has been completed, two ques-
tions must be raised: what is the degree of trust attached to the conclusions stemming
from this analysis, and does this transition correspond to the experimentally observed
glass transition temperature? One way to answer these questions is to get reproducible
values of Tg and compare the stemming analysis to experimental data and theories.
The focus is thus to develop a procedure based on physics and statistical physics con-
cepts to compute Tg . The different steps are summarized in Fig. 7.13 and in the follow-
ing paragraph.
7 Theoretical tools for designing functional materials | 161

Specific volume (cm3.g–1 )



Fig. 7.12. Simulated dilatometric

curves for iso-PMMA according to
0 50 100 150 200 250 300 350 400 450 two first generation force fields:
Temperature (°C) AMBER (⧫), and AMBER-OPLS (∙).

Appropriate exploration of the configurational space is a key issue. It corresponds to

Step 1 in Fig. 7.13. The algorithm used for generating polymer chains (100) is very pow-
erful and was built by Theodorou and Suter [44], based on the Self-Avoiding Walk
protocol and the Rotational Isomeric States model [45, 46]. For each polymer chain,
the radius of gyration (Rg ) is then computed (Step 2). A Maxwell–Boltzmann-like dis-
tribution is obtained. Polymer chains which exhibit Rg with the most probable values
are selected. An additional criterion based on the lowest energy is then completed
to finally select ten configurations (Step 3). This number of configurations is consid-
ered a reasonable compromise between accurate representation of the configuration
(through Steps 1 to 3), and CPU capability. It is thus a way of circumventing the cen-
tral limit theorem. A heating-cooling process is then performed to relax the system
(Step 4). At this stage, Tg can be computed when the system has cooled down. How-
ever, it has been shown that this value is not reproducible. In fact, in order to effec-
tively characterize a small portion of the configurational space representative of the
entire domain, the simulated system must approach reality. The conclusion reached
was that the simulated polymers in their glassy state were not in mechanical equilib-
rium, leading to Step 5, which consists of getting 𝜕E/ 𝜕V = 0, where the internal pres-
sure and the external stress are made equal. Entropy and vibrational energy should
be introduced, but there are negligible quantities for this kind of simulation. When
this step has been completed, numerous properties such as mechanical properties
and values of Tg are found to be reproducible. This procedure was applied to get Tg
of organic glasses whose values are in very good agreement with experimental data.
We can thus suspect that the glass transition has been accurately reproduced, and that
local characteristics of this phenomenon can be disclosed. It has to be mentioned that
this procedure was transposed to simulate glass transition of molecular glasses [47]
and liquid crystalline phases [48] with the same reliable results.
As an illustrative example, Fig. 7.14 shows values of Tg for a series of vinyl polymers
computed following the procedure described in the previous paragraph with respect
162 | Part I Functional materials: Synthesis and applications

of 100
1 configurations


10 Computation
Number of

8 of the Rg
6 2
20 40 60 80 100 120 140 160

1. Higher probability 2 selection

2. Energetic criteria criteria

VSp/cm3 .g–1

1.04 Heating-Cooling

100 150 200 250 300

Temperature/°C Tg

–7660 1.08
Energy/kcal mol–1

VSp/cm3 .g–1

Mechanical 1.04

–7700 5 Equilibrium
–7710 1.00
27500 28000 28500 29000 29500 100 150 200 250 300

Volume/Å3 Temperature/°C

Fig. 7.13. Procedure for computing reproducible values of Tg .

to the experimental values. The linear relationship is excellent, which makes the iden-
tification of molecular events that lead to changes in Tg feasible. Thanks to the use of
MD, specific local dynamics can be revealed. Two generic vectors have been selected:
along C (backbone)-H, and originating from Cα (side chain), shown in Fig. 7.15. Anal-
ysis of the movement of these vectors according to temperature yields a correlation
time associated with each motion at a specific temperature. This correlation time can
be inserted into an Arrhenian diagram: the natural logarithm of this correlation time
is displayed with respect to the inverse of the temperature. Surprisingly, a non-linear
7 Theoretical tools for designing functional materials | 163

520 CH3 n
n n
500 N
480 X CH3
R n
460 O O N
Simulated Tg/K

440 n
340 n
320 O O
300 CH3
280 n









Experimental Tg/K

Fig. 7.14. Simulated and experimental Tg of a series of vinyl polymers and

linear fitting highlighting the correlation.

fit is observed. Arrhenian behavior, i.e., a linear relationship where the slope is re-
lated to the activation energy, is expected. The physical sense beneath the non-linear
fit is related to the occurrence of cooperation along the polymer chains, as schemati-
cally shown in Fig. 7.11. An equation which is better-suited to improving the fit of the
traditional Arrhenius equation is the Vogel–Fulcher–Tamman equation:

τ = τo exp ( ), (7.11)
T − To

where the three fitting parameters τo , To , and Eeff correspond respectively to the cor-
relation time for infinite temperature, Vogel temperature, and the effective activation
energy. The behavior of the effective activation energy with Tg for the different vinyl
polymers studied is of particular interest. Figure 7.15 shows the behavior of Eeff for the
two bonds. Values of Eeff related to the C–H bond (in blue) are relatively constant. This
means that the carbon backbone, to which C–H movement is related, possesses an in-
trinsic dynamic that does not influence Tg . Conversely, Eeff associated with the motion
of the side chain (dark blue) reveals a linear relationship with Tg . In fact, the dynamics
associated with this bond correspond to three motions: librational motion (not con-
sidered in the analysis), side chain rotation, and overall motion of the chain. The two
latter movements cannot be separated easily. As a consequence, they are studied si-
multaneously, leading to a molecular analysis of the coupling between the side chain
and the backbone. A linear relationship is used to fit the data. The Tg of a hypothetic
chain reduced to only a carbon backbone, and for which Eeff = 0 was found to be 250 K.
This is in fact close to the actual Tg of PE, which is of the order of 280 K. Thus, side
164 | Part I Functional materials: Synthesis and applications

Backbone 20

Effective Ea (kcal/mol)


X 12
chain coupling 8

250 300 350 400 450 500 550

Tg /K

Fig. 7.15. Local dynamics in simulated vinyl polymers.

chains greatly influence the backbone motions which ultimately yield differences in
the value of Tg through specific coupling.

7.3 Summary

Molecular simulation is becoming a very powerful tool in the lab and is not just for the-
oreticians. A corollary to this situation is that it has great difficulty finding its niche.
It is theory for experimentalists, and experiment for theoreticians. Moreover, extreme
caution should be exercised in interpreting ensuing results, as misleading interpreta-
tion can be very problematic. The examples presented in this chapter show that metic-
ulous protocols must be established. Clearly, simulations must be used as a tool to
complement experiment and theory. Consequently, these studies emphasize the great
strength of molecular simulation in unveiling molecular characteristics, thus acting
as a real guide in the design of new molecules with improved properties. Moreover,
codes and computers are now efficient enough to bring a new perspective to those
properties. Local phenomena can be scrutinized and molecular simulation acts as a
microscope. Can it solve the antagonism raised by the difference between the two vi-
sions of the world, reduction and emergence? We will see in the future!
In fact, increasing computer capacity and efficiency of codes, the molecular origin
of a property can be definitely disclosed at different length and time scales. The mul-
tiscale approach, as highlighted in this chapter, thus becomes more and more attrac-
tive. The very purpose is to make the most of conclusions stemming from one scale by
linking them to data extracted at another scale. A property can thus be comprehended
along the complete spectrum of interactions. This viewpoint has become so important
that Martin Karplus, Michael Levitt, and Arieh Warshel were awarded the 2013 Nobel
Prize in Chemistry “for the development of multiscale models for complex chemical
7 Theoretical tools for designing functional materials | 165

Molecular simulation
– Young DC. Computational chemistry: a practical guide for applying techniques to real world prob-
lems. Wiley, New York, 2001.
– Jensen F. Introduction to computational chemistry. 2nd ed. John Wiley & Sons, Chichester, UK,
Hoboken, USA, 2007.
– Leach AR. Molecular modelling: principles and applications. 2nd ed. Prentice Hall, Harlow, Eng-
land, New York, 2001.
– Cramer CJ. Essentials of computational chemistry: theories and models. 2nd ed. John Wiley &
Sons, Chichester, UK, Hoboken, USA, 2004.
– Ramachandran KI, Deepa G, Namboori K, SpringerLink (Online service). Computational chemistry
and molecular modeling principles and applications. Berlin: Springer; 2008.
– Hinchliffe A. Molecular modelling for beginners. 2nd ed. Wiley, Chichester, UK, Hoboken, USA,
– Höltje H-D. Molecular modeling: basic principles and applications. 3rd rev. and expanded ed.
Wiley-VCH, Weinheim, 2008.
– Leszczynski J. Handbook of computational chemistry. Springer reference. Springer Science+
Business Media B.V., Dordrecht, New York, 2012.
– Lewars E, Computational chemistry introduction to the theory and applications of molecular and
quantum mechanics. 2nd ed. Springer, Dordrecht, London, 2011.
– Reviews in computational chemistry. Editors: Lipkowitz KB, Boyd DB. 26 Volumes. Wiley-VCH,
New York.

Multiscale simulation
– Berendsen HJC. Simulating the physical world: hierarchical modeling from quantum mechanics
to fluid dynamics. Cambridge University Press, Cambridge, UK, 2007.

Quantum calculation
– Parr RG, Yang W. Density-functional theory of atoms and molecules. Oxford University Press,
New York, Oxford, 1994.
– Koch W, Holthausen MC, Wiley InterScience (Online service). A chemist’s guide to density func-
tional theory. Second ed. Wiley-VCH, Weinheim, New York, 2001.
– Sholl DS, Steckel JA. Density functional theory: a practical introduction. Wiley, Hoboken, N.J.,
– Hehre WJ. Ab initio molecular orbital theory. Wiley, New York, 1986.

Atomistic simulation
– Allen MP, Tildesley DJ. Computer simulation of liquids. Oxford University Press, Oxford, 1987.
– Frenkel D, Smit B. Understanding molecular simulation: from algorithms to applications. 2nd ed.
Academic Press, San Diego USA, London, UK, 2002.
– Kotelyanskii M, Theodorou DN. Simulation methods for polymers. Marcel Dekker, New York,
– Haile JM. Molecular Dynamics Simulation. John Wiley & Sons, New York, 1992.
166 | Part I Functional materials: Synthesis and applications


[1] Laughlin RB, Pines D. The theory of everything. Proceedings of the National Academy of
Sciences of the United States of America. 97 (2000) 28–31.
[2] Anderson PW. More Is Different. Science. 177 ( 1972) 393–6.
[3] Laughlin RB. A different universe: reinventing physics from the bottom down. Basic Books,
New York, 2005.
[4] Scerri E. Reduction and Emergence in Chemistry – Two Recent Approaches. Philosophy of
Science. 74 (2007) 920–31.
[5] Ramachandran KI, Deepa G, Namboori K, SpringerLink (Online service). Computational chem-
istry and molecular modeling principles and applications. xxi, 397 p. Springer, Berlin, 2008.
[6] Leff HS, Rex AF. Maxwell’s demon : entropy, information, computing. Princeton University
Press, Princeton, NJ, 1990.
[7] Allen MP, Tildesley DJ. Computer Simulation of Liquids. Clarendon Press, Oxford, 1987.
[8] Sagan C. Cosmos. 1st ed. Random House, New York, 1980.
[9] Thomson JJ. Cathode Rays. Phil Mag. 44 (1897) 293.
[10] Davisson C, Germer LH. Diffraction of Electrons by a Crystal of Nickel. Physical Review.
30 (1927) 705–40.
[11] Frisch MJ, Trucks GW, Schlegel HB, et al. Gaussian 09. Wallingford CT, Gaussian Inc, 2009.
[15] Berendsen HJC. Simulating the physical world: hierarchical modeling from quantum mechanics
to fluid dynamics. Cambridge University Press, Cambridge, 2007.
[16] Sun H. COMPASS: An ab initio force-field optimized for condensed-phase applications-
Overview with details on alkane and benzene compounds. J Phys Chem B. 102 (1998) 7338–64.
[17] Jorgensen WL, Maxwell DS, Tirado-rives J. Development and Testing of the OPLS All-Atom Force
Field on Conformational Energetics and Properties of Organic Liquids. J Am Chem Soc. 118
(1996) 11225–36.
[18] Cornell WD, Cieplak P, Bayly CI, et al. A second generation force field for the simulation of
proteins, nucleic acids, and organic molecules. J Am Chem Soc. 117 (1995) 5179–97.
[19] Levine IN. Physical chemistry. 6th ed. McGraw-Hill, Boston, 2009.
[20] Warshel A. Consistent force field calculations. II. Crystal structures, sublimation energies,
molecular and lattice vibrations, molecular conformations, and enthalpies of alkanes. J Chem
Phys.53 (1970) 582–94.
[21] Binder K, Heermann DW, SpringerLink (Service en ligne). Monte Carlo simulation in statistical
physics an introduction. 5th ed. Springer-Verlag, Heidelberg , New York, 2010.
[22] Soldera A. Energetic analysis of the two PMMA chain tacticities and PMA through molecular
dynamics simulations. Polymer. 43 (2002) 4269–75.
[23] Soldera A, Monterrat E. Mid-infrared optical properties of a polymer film: comparison between
classical molecular simulations: spectrometry, and ellipsometry techniques. Polymer. 43
(2002) 6027–35.
[24] Metatla N, Soldera A. Effect of the molar volume on the elastic properties of vinylic polymers:
A static molecular modeling approach. Macromol Theor Simul. 20 (2011) 266–74.
[25] McQuarrie DA. Statistical mechanics. Harper and Row, New York, 1976.
[26] Soldera A, Grohens Y. Local dynamics of stereoregular PMMAs using molecular simulation.
Macromolecules. 35 (2002) 722–6.
[27] Palato S, Metatla N, Soldera A. Temperature behavior of the Kohlrausch exponent for a series
of vinylic polymers modelled by an all-atomistic approach. Eur Phys J E. 34 (2011) 90.
7 Theoretical tools for designing functional materials | 167

[28] Metatla N, Palato S, Soldera A. Change in Morphology of Fuel Cell Membranes under Shearing.
Soft Matter. 9, (2013) 11093–7.
[29] Rondeau-Gagné S, Lafleur-Lambert A, Soldera A, Morin JF. Ethynyl-bridged fullerene deriva-
tives: Effect of the secondary group on electronic properties. New Journal of Chemistry.
35 (2011) 942–7.
[30] Laflamme P, Beaudoin A, Chapaton T, Spino C, Soldera A. Simulated infrared spectra of triflic
acid during proton dissociation. J Comput Chem. 33 (2012) 1190–6.
[31] Laflamme P, Beaudoin A, Chapaton T, Spino C, Soldera A. Molecular modeling assisted design
of new monomers utilized in fuel cell proton exchange membranes. J Membr Sci. 401–402
(2012) 56–60.
[32] Metatla N, Palato S, Commarieu B, Claverie JP, Soldera A. Melting of polymer nanocrystals:
a comparison between experiments and simulation. Soft Matter. 8 (2012) 347.
[33] Kryuchkov VA, Daigle J-C, Skupov KM, Claverie JP, Winnik FM. Amphiphilic Polyethylenes Lead-
ing to Surfactant-Free Thermoresponsive Nanoparticles. J Am Chem Soc. 132 (2010) 15573–9.
[34] Metatla N, Soldera A. Computation of densities, bulk moduli and glass transition temperatures
of vinylic polymers from atomistic simulation. Mol Simul. 32 (2006) 1187-93.
[35] Soldera A, Metatla N. Glass transition of polymers: Atomistic simulation versus experiments.
Phys Rev E. 74 (2006) 061803.
[36] Metatla N, Soldera A. The Vogel–Fulcher–Tamman equation investigated by atomistic simula-
tion with regard to the Adam-Gibbs model. Macromolecules. 40 (2007) 9680–5.
[37] Soldera A, Metatla N, Beaudoin A, Said S, Grohens Y. Heat capacities of both PMMA stere-
omers: Comparison between atomistic simulation and experimental data. Polymer. 51 (2010)
[38] Metatla N, Soldera A. Effect of the molar volume on the elastic properties of vinylic polymers:
A static molecular modeling approach. Macromolecular Theory and Simulations. 20 (2011)
[39] Palato S, Metatla N, Soldera A. Temperature behavior of the Kohlrausch exponent for a se-
ries of vinylic polymers modelled by an all-atomistic approach. European Physical Journal E.
34 (2011) 90.
[40] Soldera A. Atomistic simulations of vinyl polymers. Mol Simul. 38 (2012) 762–71.
[41] Plante A, Palato S, Lebel O, Soldera A. Functionalization of molecular glasses: effect on the
glass transition temperature. Journal of Materials Chemistry C. 1 (2013) 1037.
[42] Rigby D, Roe RJ. Molecular dynamics simulation of polymer liquid and glass. I. Glass transition.
J Chem Phys. 87(12), (1987) 7285–92.
[43] Baschnagel J, Binder K, Doruker P, et al. Bridging the gap between atomistic and coarse-
grained models of polymers: Status and perspectives. Advances in Polymer Science:
Viscoelasticity, Atomistic Models, Statistical Chemistry. 152 (2000) 41–156.
[44] Theodorou DN, Suter UW. Detailed molecular structure of a vinyl Polymer glass. Macro-
molecules. 18 (1985) 1467–78.
[45] Mattice WL, Suter UW. Conformational Theory of Large Molecules. The Rotational Isomeric
State Model in Macromolecular Systems. John Wiley & Sons, New York, 1994.
[46] Flory PJ. Statistical Mechanics of Chain Molecules. Hanser Publishers, New York, 1989.
[47] Porzio F, Levert E, Vadnais R, Soldera A. New Insights into the Thermal Stability of the Smectic
C Phase. Journal of Physical Chemistry B 118 (2014) 4037–43.
Part II: Development of new materials for energy
S. B. Schougaard and D. Bélanger
8 Electrochemical energy storage systems
8.1 Introduction

The ability to efficiently store and retrieve electrical energy is at the heart of the mo-
bile revolution which has swept through society since the late 1980s. The develop-
ment of green energies such as solar and wind, which could ultimately be used for
transportation in hybrid and electric cars, began during the same period. Making this
“green” transition will require an efficient, low cost energy storage system. Two of
the most promising candidates to meet the high power-high energy density storage
requirements are lithium-ion batteries and advanced electrochemical capacitors. Yet
the current state-of-the art performance requires improvements to meet customer ex-
pectations. To this end, the active materials still need to be perfected.
Materials used in energy storage systems which convert chemical energy into elec-
trical energy must have unique properties for this specific application. A wide variety
of materials such as metal, metal oxides, and even polymers have been used for this
purpose over the past two centuries. One of the first batteries developed which has
reached widespread commercial use is the lead-acid battery. This battery still occu-
pies a major segment of the energy storage market today, due to its use in cars and
other automotives. Other common battery technologies include alkaline and nickel-
cadmium cells. Materials used in current and future energy storage systems must sat-
isfy design criteria such as low cost, stability, low environmental impact, and toxicity,
and must be produced from widely available resources. This is obviously not the case
for all batteries thus far developed. Safety is also a major concern due to the quest for
higher energy and power density, which require more reactive materials.
The aim of this chapter is to present the fundamental concepts associated with
electrochemical energy storage systems in such a way that they will be accessible to a
broad audience working in the area of functional materials. Firstly, the metrics used to
evaluate the performance of batteries will be described. Secondly, the inner workings
of electrochemical energy storage will be explained using thermodynamic, kinetic,
structure, and mass transport properties of battery materials. The current state-of-the
art materials used in these electrochemical energy storage systems will also be pre-

8.2 Metrics and performance evaluation

The most fundamental characteristic of any energy storage system is how much en-
ergy per unit of mass can be stored (specific energy), and how fast this energy can be
retrieved and/or stored (power density). Alternatively, the storage capacity and power
172 | Part II Development of new materials for energy applications

density may be expressed in terms of volume, when space constraints are more rel-
evant. Electrochemical energy storage systems may be analyzed using several differ-
ent protocols, one of the most common being the constant current. Here, the current
is sourced (controlled) and the potential (voltage) is measured as a function of time.
Importantly, several different units and nomenclatures are used to express the con-
trolled current. Among the more common ones are A/g, A/cm2 and C-rate, where x C
represents full discharge in C/x hour, so that a current density of C/10 represents full
charge in 10 hours. Batteries and electrochemical capacitors differ in their fundamen-
tal charge storage mechanism; as such, it is not surprising that their electrochemical
responses are also different (Fig. 8.1).

10A/g 5A/g 1A/g

1.2 4.2

0.9 3.8
Potential (V)

Potential (V)
0.6 3.4

0.3 3.0 0.2A/g

0 2.6
0 40 80 120 160 0 40 80 120 160
Capacity (mA.h/g) Capacity (mA.h/g)
(a) (b)
Fig. 8.1. Potential vs. capacity curves. (a) electrochemical capacitor, and (b) battery. The stored
energy is proportional to the area under the curve when extended to 0 V.

From Fig. 8.1 it is evident that capacity is dependent on the current density used. To
put this into a meaningful metric, the power density can be defined as the product of
voltage and current density. As this product varies over the charge/discharge cycle,
an average is normally cited. This definition makes comparison of different devices
according to their power delivery ability and their total energy density in the Ragone
Plot possible (Fig. 8.2).
Figure 8.2 shows that electrochemical capacitors provide high power while batter-
ies provide high energy density. It is therefore important to understand the required
specifications for a specific application before choosing a battery or an electrochem-
ical capacitor. However, for an electric car, it is not a question of battery or electro-
chemical capacitor, since both serve distinct design parameters; i.e., electrochemical
capacitors are ideal for storing the power generated when the car decelerates rapidly
(e.g., regenerative braking), and for the high power needed when the car accelerates
from a standing position. However, the battery is essential to give the car a significant
range of autonomy due to its high energy density.
8 Electrochemical energy storage systems | 173

Specific power (W/kg)

105 engine


103 Electrochemical
100 Batteries Fuel

0.01 0.1 1 10 100 1000 Fig. 8.2. The Ragone Plot comparing
Specific energy (Wh/kg) different energy technologies.

For historical reasons, the energy storage capability of the electrochemical capacitor
is often expressed as capacitance (units: Farad (F)), using the idealized properties of
capacitors i.e., E = Q/C where Q is the stored charge, E is the potential, and C is the ca-
pacitance (see below). This relationship leads to W = 1/2 CE2 , so the stored energy (W)
is proportional to the capacitance.
Caution: Specific energy, specific power, and capacities are highly dependent on the
current density. As a rule of thumb, these values should not be extrapolated. Moreover,
care should be taken to identify the mass or volume on which the analysis is based (the
entire device including housing, the electrode, or the active material).

8.3 Models and theory of electrochemical charge storage

Electrochemical energy storage takes place by separating the connection between

two electrodes into two different charge transport paths, one for the ions (electrolyte)
which blocks the electrons, and one for the electrons (external circuit) which blocks
the ions (Fig. 8.3). The external circuit serves as the point where the electric energy is
injected or retrieved. It is this electrical energy that is converted into chemical energy
in the device.
The major difference between the electrochemical capacitor and the battery is the
processes which take place at the electrodes. In the battery it is the chemical process of
electron transfer, i.e., oxidation/reduction, while in the capacitor the energy is stored
electrostatically as “charges” on the electrode surfaces.
174 | Part II Development of new materials for energy applications

External circuit

Li+ + e–+ TiS2

Negative electrode

Positive electrode
Li+ + e–
Ions Electrons
Electrolyte Yes No


External No Yes

Fig. 8.3. Schematic of an electrochemical cell.

8.3.1 Battery operation – a Faradaic process

The separation of the charge transport allows the separation of the redox reaction (cell
reaction) into two Faradaic reactions (half cell), which occur simultaneously in two dif-
ferent locations. A Faradaic process involves a gain or loss of at least one electron. For
example, the relevant reactions during charging are given below for the first lithium
metal battery developed by Whittingham [1]:
Half cell (positive electrode): LiTiS2 → TiS2 + Li+ + e−
Half cell (negative electrode): Li+ + e− → Li (8.1)
Cell reaction: LiTiS2 → Li + TiS2
The operational principle of charging relies on electrons and lithium ions released
from the oxidation of LiTiS2 to TiS2 in the positive electrode being transported to the
negative electrode by the external circuit (e− ) and electrolyte (Li+ ) where they are used
to form Li(0). Therefore, the overall process does not generate or remove electrons or
ions; they are only transported by the energy source in the external circuit to a location
where they have higher potential energy. Thermodynamics tells us that the work done
by the electrons in the external circuit during discharge cannot exceed the chemical
energy stored in the chemical reaction of the battery. Moreover, the maximum work
possible for an electron in the external circuit is its potential energy, measured as the
voltage between the positive and the negative electrode when no current is flowing (E).
The maximum work that can be done by the chemical reaction (at constant pressure
and temperature) is its Gibbs free energy (Δ r G). This relationship leads to:

Δ r G = −nFE, (8.2)

where F is the Faraday constant 96 485 C/mol, and n is the number of electrons trans-
ferred (1 for the Li/TiS2 cell reaction). Δ r G can be determined from either thermo-
dynamic tables, calorimetry, first principles, or empirical considerations, which are
especially useful when designing new battery materials.
8 Electrochemical energy storage systems | 175

Armed with a method for approximating the potential of a battery (the y-axis of
Fig. 8.1(b)), another for predicting the maximum theoretical value for the x-axis would
be useful. The challenge is determining the charge transferred through the external
circuit by the reaction of one gram of reactant in (1). Thus:

theoretical capacity [in C/g] = nF/M, (8.3)

where M is the molecular mass of the reactant (g/mol), F is the Faraday constant, n is
the number of electrons transferred per formula unit. (Note: 1 C/g = 1/3.6 mAh/g).

8.3.2 Electrochemical capacitor operation – a non-Faradaic process

The mechanism behind the electrochemical capacitor differs fundamentally from the
battery, as no redox reaction occurs (e.g. no electron transfer). The classical electro-
chemical double layer capacitor comprises two porous carbon electrodes in the pres-
ence of an appropriate electrolyte, separated by a porous and inert separator. At the
electrode/electrolyte interface, opposite charges accumulate on the electrode surface
(within the porous structure) and in the electrolyte. At equilibrium, the electronic
charges in the electrode are counterbalanced by ionic charges from the electrolyte.
This leads to the formation of an electrochemical double layer, whose thickness is in
the nanometer range. The energy of an electrochemical double layer capacitor is de-
termined by the potential being developed and the capacitance of the electrodes, as
discussed previously in the metrics section.
The first description of a model for the double layer was proposed by Helmholtz,
and is illustrated in Fig. 8.4(a) [2]. In this simple model, which is similar to that of
a two-plate capacitor, two layers of opposite charge are present at the electrode/
electrolyte interface and are separated by a distance of molecular order. The Gouy–
Chapman model considers the fact that in contrast to the electrode, for which the
excess of charge is confined to the surface, the excess of charge required in the elec-
trolyte for charge compensation is spread over a thickness which depends on the con-
centration of the electrolyte and is referred to as a diffuse layer (Fig. 8.4(b)). A third
model, known as the Gouy–Chapman–Stern model, combines the first two models
and includes a first compact layer next to the electrode surface, followed by a diffusion
layer. In this model, the inner Helmholtz plane (IHP) refers to the plane that crosses
the species of the first layer, and the outer Helmholtz plane (OHP) passes through
the centres of solvated ions which are close to the electrode surface (Fig. 8.4(c)). The
effect of specific adsorption and the predominance of water molecules (in aqueous
electrolytes) for which the dipoles are oriented according to the electrode charge
was later proposed in a model by Grahame, Bockris, Devananthan, and Muller as
illustrated in Fig. 8.4(d) [3].
176 | Part II Development of new materials for energy applications

Diffusion Stern Diffusion Stern

layer layer layer layer
ψ0 ψ0 ψ0

Diffusion layer
Positively charged surface

Positively charged surface

Positively charged surface

Positively charged surface


ψ ψ
Primary Secondary
solvent solvent
plane Outer plane
layer layer
(a) (b) (c) (d)

Fig. 8.4. Schematic representation of the: (a) Helmholtz, (b) Gouy–Chapman, (c) Gouy–Chapman–
Stern, and (d) Bockris et al. models for the electrochemical double layer [2, 3].

The capacitance (C) of such an interface can be roughly estimated by analogy with a
double plate capacitor using the equation:

C = εr ε0 A/d, (8.4)

where εr is the dielectric constant of the electrolyte, ε0 is the dielectric constant of

vacuum, A is the area of the electrode, and d is the thickness of the double layer. From
this equation, a typical double layer capacitance of 10–15 μF cm−2 can be estimated
for metals in pure water by taking a value of 5 for the dielectric constant of a first layer
of water molecules and the radius of a water molecule [3].
Since the capacitance of an electrode material is related to its surface area, an
obvious approach to improvement has been to develop porous materials with a
large surface area. The pores of these materials are classified into four categories:
ultramicropores (< 1 nm), micropores (< 2 nm), mesopores (2–50 nm); and macro-
pores (> 50 nm). However, it is now clear that parts of the surface area are not electro-
chemically accessible when the pore size is smaller than the diameter of the electrolyte
ions. Nonetheless, recent studies have demonstrated a significant increase of capaci-
tance for electrodes presenting micropores (< 2 nm), which are smaller than the size
of hydrated ions [4].
Another important parameter of an electrochemical capacitor is the cell voltage.
Figure 8.5 shows a schematic representation of an ideal electrochemical double layer
capacitor, in which both electrode/electrolyte interfaces behave similarly. Initially, the
two electrodes have the same potential and therefore the potential of the cell is null.
When the electrochemical capacitor is charged, electrons are forced to flow through
an external circuit from one electrode to the other. The electrode which accumulates
electrons is termed the negative electrode, and cations from the electrolyte move to the
8 Electrochemical energy storage systems | 177

electrode surface under the influence of the electric field. Simultaneously, the second
electrode, from which electrons have been removed, becomes the positive electrode,
which is accompanied by displacement of anions of the electrolyte to the electrode
surface. Note that the terms negative and positive are used instead of anode and cath-
ode. This is because in contrast to the latter, which refers to the electrode where elec-
tron transfer occurs and involves a change in the oxidation state of the electrode or
solution species (see above), no electron transfer occurs between the electrodes and
the electrolyte in a classical electrochemical double layer. In these instances, the two
electrodes are said to be blocking electrodes.

Porous carbon Current collector

+ – – + + – + –
+ – + –
+ – – + charge + – + –
+ – + –
+ – – + + – + –
+– + – + –

Liquid double layer Liquid
(electrolyte) capacitor (electrolyte)

ψ0 ψ0

Discharged state

Charged state

Fig. 8.5. Schematic representation of the charges of each component (electrodes and electrolyte)
of an electrochemical capacitor at the electrode/electrolyte interface (top) and the potential profile
(bottom) for both its discharged and charged states.

Upon charging by application of a constant current, the double layer will become
charged to a level that will depend on the applied potential (Fig. 8.6). In the case of an
ideal electrochemical capacitor, the potential of each electrode varies linearly but in
opposite directions, with the charge passed during the charging process of the device.
The potential of each electrode can vary up to a limit, determined by the potential at
which oxidation or reduction of electrolytes commences.
178 | Part II Development of new materials for energy applications


Fig. 8.6. Variation of the potential of the positive (- - -) and

negative (– –) electrodes as a function during constant current
– charge/discharge together with the evolution of the voltage
Charge (—) of the electrochemical capacitor.

8.4 Electrolytes

From the operational principles of the battery and electrochemical capacitor, it is clear
that the energy required for transportation of charge through the electrolyte and the
potential limits outside which the electrolyte is oxidized or reduce, are key physico-
chemical design parameters.
Transport of species in solution (mass transport) is related to three phenomena:
diffusion, migration, and convection. Convection, which is forced movement of the liq-
uid including all its components, is only of importance in flow-batteries [5]. Diffusion,
which is important in batteries, is the process by which high concentration solutions
flow towards low concentration at a speed proportional to the concentration differ-
ence until equilibrium has been reached. Migration is the mechanism by which an
electric field moves charged species through a solution, and is crucial in both batter-
ies and electrochemical capacitors. Importantly, only migration can carry the current
required to close the circuit in Fig. 8.3. The drag force caused by moving ions through
the solution has two effects. First, there is a significant resistance to charge transport.
This leads to Ohm’s law behavior, where the current (I) is proportional to the poten-
tial difference between the electrodes (E), and conductivity (κ) is proportional to the
concentration of free ions (C+ , C− ) and their mobility (u+ , u− ), i.e.,

Iionic = E ∗ κ and κ ∝ u+ C+ + u− C− (8.5)

The energy lost to resistive heating in the electrolyte should be kept to a minimum,
which suggests that C and u should be maximized. Electrolytes should therefore be
highly concentrated, low viscosity solutions of free ions. This is a challenge, since as
the concentration of electrolyte increases, so does the tendency for ion-pairing. The
result is a viscous solution and a significant fraction of ions that are no longer free to
engage in migration. The second effect of the drag force is that it is different for ions
of different sizes. This leads to strong concentration differences during high current
operation, which the diffusion process is slow to counteract. In particular, the lithium
8 Electrochemical energy storage systems | 179

ion in the Li-battery electrolyte typically moves at only about 65 % of the speed of the
counter-ion. This leads to major problems during high power operation, since only
lithium ions can take part in the charge transfer at the electrodes.
The electrochemical potential range of stability of the electrolyte limits the max-
imum quantity of energy that can be stored per electron. For example, in aqueous
electrolytes used for electrochemical capacitors, this value is set at 1.23 V by the ther-
modynamic standard potentials of the H2 O/O2 and H2 O/H2 redox systems. However,
in practice a much lower value of about 1 V can be achieved. A typical nonaqueous-
based device using acetonitrile or propylene carbonate as solvent can, however, reach
close to 3 V. Even higher cell voltage could be obtained using materials with a wider
potential range of stability, such as room temperature ionic liquids, which are salts
that are liquid at room temperature. A list of common electrolyte systems is given in
Table 8.1.

Table 8.1. Various types of electrolytes and relevant properties.

Electrolyte type Example Practical Potential Comment

conductivity window
Aqueous solutions KCl (aq) ∼ 100 ∼ 1.23 V Cheap, nontoxic,
limited potential
Nonaqueous solutions LiPF6 in or- ∼ 10 ∼ 4.3 V Flammable, large
ganic carbon- potential window,
ates compatible with
low potential an-
Ionic liquids 1-ethyl- ∼ 0.1–30 > 5V Nonflammable,
3-methyl- large potential
imidazolium window.
Polymer Polyethylene ∼ 0.1 > 4V One of the two
oxide com- ions may be linked
bined with to the polymer to
LiPF6 . improve selective
ion transport.
Solid Lix POy Nz < 0.1 at room Highly de- Only high temper-
temperature. pendent on ature or thin film
system cho- application due to
sen. poor conductivity.
180 | Part II Development of new materials for energy applications

8.5 Electrode materials

8.5.1 Electrochemical capacitors

Carbon is the material of choice for the fabrication of electrode materials of electro-
chemical capacitor. This is primarily because its low cost, its good electrochemical
stability, high conductivity, and high specific surface area. A large variety of other
types of materials such as conducting polymers and metal nitrides have been used
for this application, but arguably the most promising materials are metal oxides such
as ruthenium dioxide and manganese oxide.
Carbon can be produced in various morphologies and shapes. Initial studies
mainly dealt with high specific surface area powders and high area fibers. Activated
carbons are the most widely investigated, but other forms such as templated car-
bons, aerogels, nanotubes, nano-onions, and graphene have also been examined for
electrochemical capacitor applications [6].
Carbon materials commonly used in electrochemical capacitors are characterized by
a high specific surface area which ranges from 1 000 to 2 500 m2 /g. Depending on the
production method, the porous texture of the carbon materials will display a variable
pore size distribution. As mentioned previously, it is now well-established that not
all of the porous surface area of the carbon material is active in the charge storage
process. Indeed, by taking a low value of 10 μF cm−2 for the double layer capacitance,
the expected specific capacitance of 250 F/g cannot be reached for the highest surface
area materials. Instead, the typical specific capacitance of standard activated carbons
in organic electrolyte is about 100 F/g [6]. It is possible to attain a higher value (150 F/g)
in aqueous electrolytes (e.g. aqueous KOH) due to the pseudocapacitive contribution
of oxygenated surface groups such as quinone. These surface functional groups will
also contribute to the wettability of carbon, thereby improving the capacitance of the
Several single metal and mixed metal oxides have been investigated for their ap-
plication as active electrode materials for electrochemical capacitors. One is ruthe-
nium dioxide (RuO2 ), which is characterized by pseudocapacitive behavior that, in
contrast to capacitive materials such as carbon, is Faradaic in origin. Thus, pseudo-
capacitive electrode materials undergo reversible redox reaction involving change of
oxidation states of the metallic species [7]. These redox reactions are distributed over
a large potential range, and are consequently characterized by a rectangular cyclic
voltammogram shape and a linear charge/discharge profile such as the one shown by
carbon electrodes (Fig. 8.7). On the other hand, most other metal oxides are charac-
terized by Faradaic redox processes (in fact they are battery material) which usually
encompass a relatively small potential window. Therefore, a symmetric electrochemi-
cal capacitor based on such material for both the positive and negative electrodes will
be characterized by a relatively small cell voltage, and would be almost useless unless
8 Electrochemical energy storage systems | 181

the Faradaic metal oxide is used with a complementary electrode material in a hybrid
electrochemical capacitor, as discussed below.
Another pseudocapacitive material that has attracted attention in the past decade
is MnO2 . It is obtained in various crystalline forms, including many tunnel and layer
structures which allow cation exchange. Amorphous or poorly crystallized MnO2 was
also widely investigated as an electrochemical capacitor material. These poorly crys-
tallized compounds consist of a random arrangement of MnO6 octahedra, which leads
to an intergrowth of different structures, alternating tunnels of different sizes along-
side water molecules and alkaline cations (e.g. Na+ , K+ ). The charge storage processes
for the MnO2 electrode can occur by two mechanisms depending on the nature of
the oxide, and involves cations (C+ = H+ , Li+ ) from the electrolyte in both cases, the
Mn3+ / Mn4+ redox interconversion (equation (8.6)) and the following electrochemical

MnOOC → MnO2 + C+ + e− (8.6)

(MnO2 )surface + C + e → (MnO−2 C+ )surface (8.7)

Spectroscopy investigations, typically using thin-film MnO2 electrodes, have con-

firmed that the Mn3+ / Mn4+ redox couple is primarily responsible for the observed
pseudocapacitance in mild aqueous electrolytes, as demonstrated for poorly crystal-
lized manganese dioxides using in-situ Mn K-Edge x-ray absorption and x-ray photo-
electron spectroscopy [8].

8.5.2 Hybrid electrochemical capacitors

A conventional electrochemical capacitor is usually comprised of two identical capac-

itive electrodes in a symmetrical configuration. On the other hand, in a hybrid electro-
chemical capacitor, one of the capacitive electrodes is replaced by a Faradaic elec-
trode. Several types of hybrid electrochemical capacitors have been developed; the
most common entails the positive electrode of a symmetric carbon/carbon electro-
chemical capacitor being replaced with an electrode characterized by a much higher
charge-storage capability. These include pseudocapacitive and Faradaic electrode ma-
terials, such as conducting polymers and metal oxides such as MnO2 , NiOOH and
PbO2 [8]. This type of electrochemical device, also called asymmetric capacitor, takes
advantage of the best performance characteristics of electrochemical capacitors and
batteries. This can be achieved by the fast charge storage process of the capacitive
negative electrode and by the increase of the specific capacity of the cell due to the
Faradaic charge storage mechanism of the positive electrode. In addition, by selecting
two electrodes characterized by different operating ranges it is possible to extend the
voltage of an aqueous hybrid electrochemical capacitor beyond the thermodynamic
limit of 1.23 V, as illustrated in Figs. 8.7(b) and (c).
182 | Part II Development of new materials for energy applications

I (mA cm–2)

Negative electrode 4 Positive electrode

(porous carbon) (porous carbon)

–1.0 –0.5 0 0.5 1.0 1.5 2.0

E (V) vs NHE

(a) –8

I (mA cm–2)

Negative electrode 4 Positive electrode

(porous carbon) (MnO2)

–1.0 –0.5 0 0.5 1.0 1.5 2.0

E (V) vs NHE –4

(b) –8
Positive electrode
8 (PbO2)
I (mA cm–2)

Negative electrode
(porous carbon)

–1.0 –0.5 0 0.5 1.0 1.5 2.0

E (V) vs NHE

(c) –8

Fig. 8.7. Schematic representation of cyclic voltammograms for three different configurations of
aqueous-based electrochemical capacitors (ECs), in which areas shaded in black and white repre-
sent the potential window of the positive and negative electrode, respectively for: (a) symmetric car-
bon//carbon device in 1 M H2 SO4 , (b) asymmetric activated carbon//MnO2 device in 0.5 M K2 SO4 ,
and (c) asymmetric activated carbon//PbO2 device in 1 M H2 SO4 . NHE, normal hydrogen electrode,
I, measured current, and E, electrode potential.
8 Electrochemical energy storage systems | 183

The operating range of an electrode material in a given electrolyte can be assessed

by cyclic voltammetry (Fig. 8.7). Figure 8.7(c) shows that the electrochemical potential
window of an activated carbon electrode ranges from −0.4 to 0.7 V in H2 SO4 (aque-
ous). Therefore, it can be used as a negative electrode of a hybrid device comprising
of a PbO2 positive electrode, with a range of electroactivity between 1.2 and 1.8 V, over
which the interconversion of PbO2 to PbSO4 is:

PbO2 + 4H+ + SO2−
4 + 2e → PbSO4 + 2H2 O (8.8)

Thus, Fig. 8.7(c) indicates that a carbon/PbO2 hybrid electrochemical capacitor can
deliver a cell voltage of about 2.3 V, in contrast to a carbon/carbon device, which will be
limited to 1.1 V at the most (Fig. 8.7(a)). In a symmetrical device, each carbon electrode
operates in a limited electrochemical window of about 0.55 V, which is only 50 % of
the full electrochemical potential window of carbon (Fig. 8.7(a), white or black areas).
This means that the resulting capacitance (F/g) of the symmetrical device is only 25 %
that of a single carbon electrode.

1/CEC = 1/C+ + 1/C− (8.9)

The use of a positive electrode with a larger capacitance (capacity) and a com-
plementary electrochemical window will result in full utilization of the carbon ma-
terial and will increase the energy density due to the capacitive and pseudocapac-
itive/Faradaic behavior of each electrode. Similarly, a carbon/MnO2 hybrid electro-
chemical capacitor can deliver a slightly larger cell voltage than a carbon/carbon de-
vice, since the electrochemical potential window of an MnO2 electrode extends to
slightly more positive value than a carbon electrode (Fig. 8.7(b)). The former is ar-
guably the most thoroughly investigated hybrid electrochemical capacitor due to the
fact that MnO2 is characterized by low cost, low toxicity, natural abundance, and en-
vironmentally friendly use in mild aqueous electrolytes.

8.5.3 Lithium battery¹ electrode materials

Charge storage requires that the charge/discharge cycle can be repeated thousands of
times. In batteries, this is a severe limitation for the type of chemical reactions which
can be employed since high yield is required (for example, a 99.99 % yield per cycle for
1 000 cycles leads to a 10 % total capacity loss). In fact, only three types of reactions
are currently being used in lithium batteries: insertion, alloying, and conversion.

1 Batteries where an intercalation electrode is used, both positive and negative electrodes, are known
as lithium-ion batteries.
184 | Part II Development of new materials for energy applications

8.5.4 Negative (anode) electrode materials

The simplest negative electrode is arguably metallic lithium. However, difficulties in

plating lithium uniformly lead to dendrite formation in liquid electrolytes. In turn,
this can lead to short-circuiting the electrodes, entailing thermal runaway and even
explosions. Graphitic carbon where Li(0) atoms are nominally intercalated between
graphene sheets is used to circumvent this problem (Fig. 8.8).


Fig. 8.8. Lithium (large circles) intercalation into graphite.

Current Graphite The reaction leads to a slight increase of the unit cell without
collector destroying the layered structure.

The limited capacity of the graphite lithium insertion reaction (339 mAh/g for LiC6 )
has led to the search for new materials for its replacement. A promising candidate is
silicon, which forms an alloy that can store 4.4 Li per Si atom (theoretical capacity
4 200 mAh/g). The accompanying 400 % volume change is however a major design
challenge, since it causes not only stress cracking of the silicon particle, but also in
the entire composite electrode. Another important property of all low potential nega-
tive electrodes is their lack of thermodynamic stability relative to the reduction of the
electrolyte. Fortunately, in organic carbonate electrolytes this reaction leads to the for-
mation of a thin layer at the solid electrolyte interface (SEI), which allows Li-ions to
pass and limits further reactions between electrode and electrolyte; see Fig. 8.9.


Li2CO3, alkoxides
Li2O, RCO2Li

Fig. 8.9. The solid electrolyte interface (SEI) layer
Current Graphite SEI Electrolyte formation (top) and ionic transport (bottom) at the
collector anode.
8 Electrochemical energy storage systems | 185

8.5.5 The positive (cathode) electrode

Most practical positive electrode materials are based on redox cycling of 4th row transi-
tion metals. These are combined with highly electronegative elements like oxygen and
fluorine to form solid-state structures, where elevated oxidation states of the metal are
stable. Unique to the structures selected for battery electrodes is that they are elec-
tron conductors and allow insertion and extraction of lithium ions to compensate the
change in metal ion oxidation state associated with the redox process.

b a

a b
c a

(a) (b) (c)

Fig. 8.10. Materials with (a) 3D, (b) 2D, and (c) one 1D lithium transport path, exemplified by a
spinel, layered α-NaFeO2 , and an olivine structure, respectively.

These intercalation compounds can be classified according to the dimensionality of

the lithium ions transport path, i.e. 3D, 2D, and 1D (Fig. 8.10). Arguably, the most well
known 3D material is the spinel LiMnTMO4 (TM: transition metal ion), which offers
“5V” vs. Li+ /Li. However, problems with Mn dissolution and electrolyte stability have
strongly limited its practical application. The 2D materials with α-NaFeO2 structure
have not suffered a similar fate. In fact, LiCoO2 , which was used in the first Li-ion bat-
tery launched by Sony in 1993, is still used in most cellphones today. More recently
Li(Ni0.8 Co0.15 Al0.05 )O2 , which has the same layered structure, has been targeted for
electric car batteries due to its vastly improved safety characteristics and lower cost.
Surprisingly, even a material with a 1D lithium transportation path, olivine LiFePO4 ,
has been commercialized. This material, once carbon coated, offers elevated storage
capacity, high power, and remarkable safety characteristics. However, its lithium in-
sertion/extraction mechanism is still a subject of intense debate. Yet it is clear that the
phase boundary between the lithium-poor heterosite FePO4 phase and the lithium-
rich olivine LiFePO4 phase formed during cycling plays a major role in reaction kinet-
ics. While insertion materials currently dominate the market, some transformation
186 | Part II Development of new materials for energy applications

materials have also been proposed. One notable example is nanosized FeF3 , which
is believed to convert reversibly into LiF and Fe, entailing a remarkable theoretical
capacity of 712 mAh/g.

8.5.6 Electrode production

The common positive electrode materials are hard ceramics with limited electronic
and ionic conductivity. They are therefore produced as small particles which are
mixed with conducting carbon particles and polymer binders and cast as thin films
(10–100 μm) onto metallic current collectors. This open composite structure allows
the electrolyte to penetrate the electrode. As such, it provides improved ionic and
electronic conduction compared to the active material itself. Unsurprisingly, the de-
tailed structure and composition of the electrode has a profound effect on the power
performance of the battery [9], and must consequently be optimized for the specific

8.6 Summary

Batteries and electrochemical capacitors are currently the technologies which allow
efficient electrical energy recovery. There are further great hopes both in industry
and in society as a whole that batteries may become the energy vector which makes
the use of solar, wind, and hydro energy in cars and trucks widely possible. Current
technology, however, does not fully reach the expectations of the average consumer.
Specifically, autonomy, i.e., driving distance per charge, appears to be a concern. The
challenge over the coming years will therefore be to improve the energy density of the
batteries or the charging times, to make them comparable to refuelling with gasoline
or diesel. Chemistries which have been identified as capable of yielding increased
energy density include lithium air and Li-sulphur, which pose unique challenges in
their own right. The alternative is to design batteries which allow for unprecedented
high charging rate, such as 70 % of a complete charge within five minutes. This rapid
charge will then be used for emergency charging in the few instances where the
average driver needs to travel beyond the range of a single charge within one day.
Obtaining such extreme redox kinetics will require that the charge compensating ions
involved in the redox process travel over very short distances inside the solid material.
As such, the distinction between the hybrid electrochemical capacitor and battery be-
comes less clear, since both rely on redox process in a thin layer of the solid. It will
therefore be fascinating to see if joint research in these fields will be able to combine
the high power of the electrochemical capacitor with the high energy density of the
battery in a single device.
8 Electrochemical energy storage systems | 187

Further reading
Readers interested in methods used for the production of various carbon materials will find valuable
information in the monographs by Kinoshita [10] and Conway [11]. Similarly, those interested in infor-
mation about the inner workings of batteries may consult references [12] and [13].


[1] Whittingham, M.S. Electrical Energy Storage and Intercalation Chemistry (1976) Science, 192,
[2] Bard, A.J., Faulkner, L.R. Electrochemical methods Fundamentals and applications. 2nd edition.
John Wiley & Sons, New York, 2001.
[3] Bockris, J.O’M., Khan, S.U.M. Surface electrochemistry A molecular level approach, Plenum
Press, New York, 1993.
[4] Simon, P., Gogotsi, Y. Materials for electrochemical capacitors (2008) Nature Materials, 7,
[5] Shin, S.-H., Yun, S.-H., Moon, S.-H. A review of current developments in non-aqueous re-
dox flow batteries: characterization of their membranes for design perspective. (2013)
RSC Advances. 3, 9095–9116.
[6] Frackowiak, E. Carbon materials for supercapacitor application (2007) Physical Chemistry
Chemical Physics, 9, 1774–1785.
[7] Conway, B.E. Transition from ’supercapacitor’ to ’battery’ behavior in electrochemical energy
storage (1991) Journal of the Electrochemical Society, 138, 1539–1548.
[8] Long, J.W., Bélanger, D., Brousse, T., Sugimoto, W., Sassin, M.B., Crosnier, O. Asymmetric
electrochemical capacitors-Stretching the limits of aqueous electrolytes (2011) MRS Bulletin,
36, 513–522.
[9] Yu, D.Y.W, Donoue, K., Inoue, T., Fujimoto. M., Fujitani, S. Effect of electrode parameters on
LiFePO4 cathodes. Journal of the Electrochemical Society (2006), 153, A835–A9.
[10] Kinoshita, K. Carbon: Electrochemical and physiochemical properties. John Wiley & Sons,
New York, 1988.
[11] Conway, B.E. Electrochemical Supercapacitors: Scientific Fundamentals and Technological
Applications, Kluwer Academic/Plenum Publishers, New York, 1999.
[12] Newman, J., Thomas-Alyea, K.E., Electrochemical Systems. 3rd edition. John Wiley & Sons,
New York, 2004.
[13] Reddy, T.B., Linden’s Handbook of Batteries. McGraw-Hill Companies. Inc., New York. 2011.
D. Rochefort
9 Functional ionic liquids electrolytes in
lithium-ion batteries
9.1 Introduction

When thinking about functional materials and their applications, advanced catalysts,
light harvesting materials, photovoltaic systems, biosensors and other systems usu-
ally come to mind, rather than electrolytes or ionic liquids. The reason is that we are
not in the habit of thinking about electrolytes as materials, and even less as materials
which can be functionalized. The origin of the interpretation of electrolytes as specta-
tor constituents of an electrochemical system can be traced back to our first encoun-
ters with electrochemical cells where the electrolyte, an aqueous solution of an inert
salt like KCl for instance, serves the purpose of balancing charges during the more im-
portant processes taking place at the electrode. While this might be true for the study
of redox behavior and the kinetics of dissolved species and other basic electrochem-
ical systems, the electrolyte plays a crucial role in the electrochemistry of advanced
systems such as energy storage devices. Charge is stored from ions forming the elec-
trical double-layer in electrochemical capacitors and carbonates are involved in the
formation of the solid electrolyte interface stabilising the anode of lithium-ion batter-
ies. In the latter, the design of the electrolyte will affect the transport of lithium ions
and the limits of the voltage of operation, thereby impacting the power and energy
density values of the system. These are only a few examples showing that develop-
ment of improved electrolytes must take place parallel with that of electrode material.
For reasons that will be explained, ionic liquids have the potential to play an important
role in the development of electrolytes which will increase the overall performance of
energy storage devices.
The aim of this contribution is to introduce the reader to the use of ionic liquids
as electrolytes in lithium-ion batteries. It will become clear that ionic liquids are still
far from being successfully widely applied in commercial batteries. This chapter will
serve the purpose of stimulating interest by demonstrating the progress and potential
developments which could stem from research into ionic liquid electrolytes, rather
than simply providing a list of articles on the subject. The more general topic of ionic
liquid-based electrolytes has already been the subject of several reviews and books
[1–6]. Contributions with a focus on electrolytes for energy storage devices and bat-
teries in particular [7–9] can be consulted as additional sources of information to this
chapter, which is dedicated to the discussion of functional ionic liquids (ILs).
The first section will cover general knowledge of ionic liquids which is required in
order for the reader unacquainted to ILs to understand the origin of the properties that
are responsible for their interest as electrolytes. Next, the use of typical ionic liquids
190 | Part II Development of new materials for energy applications

as battery electrolytes will be examined to show how their intrinsic properties, mostly
of high stability, can be an advantage in the development of safer and more efficient
lithium-ion batteries. Finally, the last section will deal with ionic liquids as organic
molecules which can be modified to bear a functional group. It will be demonstrated
that the modification, by known and accessible synthetic procedures, can provide ad-
ditional benefits to the use of ionic liquids as electrolytes. This represents the true
potential of the use of IL, which is to allow the development of customized solvents-
electrolytes which can perform tasks that no other conventional solvents can achieve.
Examples of ionic liquids designed to act as redox shuttles, to promote the forma-
tion of the solid electrolyte interphase, and to increase lithium ion transport will be
discussed. Before going into detail about advanced functional ionic liquid materials,
a quick glance at what exactly ionic liquids are is necessary.

9.1.1 Historical overview

The concept of ionic liquids is not a new one, or one that has spontaneously emerged
from a specific discovery. Their development stems from studies on molten salts
with a high melting point and, appropriate to this text, their electrochemical reac-
tions [10]. The most famous early example of the study of ionic liquid is attributed
to Paul Walden, who reported about salts such as ethylammonium nitrate which
“allow the reproduction at ordinary temperatures of the phenomena observed with
fused inorganic salts at higher temperature, and permit an approach to the condi-
tions prevailing in ordinary aqueous and non-aqueous solutions” in 1914 [11]. The
field of ionic liquids that we know now, a hundred years later, has been defined by
several key periods. The 1960s saw the beginning of the “chloroaluminate era”, dur-
ing which alkali-aluminum chloride systems were studied for the development of low
temperature molten salts. An important contribution to this era came from a program
in the U.S. Air Force aimed at developing electrolytes for thermal batteries [12]. Low
temperature electrolytes were obtained with chloroaluminate anions and organic
cations such as alkylpyridinium and dialkylimidazolium, which are now widely used
in ionic liquids. Such chloroaluminate ionic liquids are, however, reactive to water,
and while this does not represent a huge problem in the field of batteries studied
and produced in controlled environments, their reactivity hindered their use in many
other applications. The next major period occurred in the 1990s, when air and water
stable alkylimidazolium ionic liquids were demonstrated, using anions such as BF−4 ,
PF−6 , Br− , CN− , and later triflates, mesylates and bis(trifluoromethane)sulfonimide
(TFSI) [13–15]. The discovery of water stable ionic liquids resulted in a tremendous
increase in interest as depicted by the publication trend shown in Fig. 9.1. Ionic liquids
are now the focus of numerous research groups and a very broad range of applications
in the fields of electrochemistry [1, 2], catalysis [16–18], and analytical sciences [19,
20], to name but a few. Ionic liquids have also reached beyond the academic world and
9 Functional ionic liquids electrolytes in lithium-ion batteries | 191

found applications in industry [21]. The BASIL™ process, involving the ionic liquid
1-methylimidazolium chloride which is reused for the synthesis of alkoxyphenylphos-
phines, is the first and certainly the most famous example of a commercial process
using ionic liquids. While there are currently no ionic liquids commercialized as elec-
trolytes in batteries or supercapacitors, there is a huge research effort worldwide to
reach this goal. This is exemplified in Fig. 9.1, which shows that if the total amount
of publications on ionic liquids continues increasing over the years, the fraction of
those related to lithium-ion batteries will become larger, highlighting the relevance
of this chapter.

% of publications related to Li-ion batteries


Number of publications




0 0.0
1995 2000 2005 2010
Year of publication

Fig. 9.1. Analysis of the publication trends on the topic of ionic liquids. Only papers written in
English were selected for the compilation. The squares show the percentage of total publications
(gray bars) that were related to lithium-ion batteries.

9.1.2 What are ionic liquids?

Ionic liquids are broadly defined as salts, usually composed of a cation of organic
structure and an inorganic anion. When certain conditions relating to the chemical
structure are fulfilled (discussed below), such a combination of cation and anion can
melt at low temperatures, well below ambient in some cases. Provided with enough
thermal energy, any salt will reach a liquid phase composed exclusively of ions. NaCl
for instance melts at 800°C, and the resulting liquid phase has a viscosity of 1.03 cP
and a conductivity of 3.6 S cm−1 [22]. In order to make a clear distinction between these
high temperature molten salts and those able to form a liquid phase around room tem-
perature, a restrictive but generally accepted definition of ionic liquids is “salts with
a melting point below 100°C”. It should be noted that the set point of 100°C, used as
a comparison to limit the liquid phase of the ubiquitous liquid, water, is purely arbi-
trary and is not based on any particular property of ILs. However, this definition is very
192 | Part II Development of new materials for energy applications

helpful in setting a boundary between the molten salts which can be used in systems
operating at moderate temperatures and those requiring high temperatures. This is
especially true for electrolytes, such as those developed for Li-ion batteries and other
electrochemical energy storage devices which are meant to work at temperatures rang-
ing from −40 to +80°C. When used as electrolytes in their pure form, ionic liquids must
be liquid in this temperature range, so the general definition given above will not be
an issue in this chapter.
At the molecular level, what distinguishes typical ionic liquids (such as those
found in Fig. 9.2) from high temperature molten salts, NaCl for example, is a com-
bination of the low level of symmetry of the cation and charge delocalization on the
ions. This combination translates into systems for which ion packing in a solid state is
very difficult and results in a low lattice enthalpy [23]. In addition to the low symme-
try, longer flexible chains on the structure bring conformation flexibility, which fur-
ther reduces the melting point and renders crystallization more difficult. Ionic liquids
generally present a supercooling state in which crystallization does not occur even if
below the freezing point [24, 25]. The families of ionic liquids presented in Fig. 9.2 are
certainly the most studied, but this list is far from exhaustive. Considering the many
different possible substituents on the structures and different combinations of cations
and anions, billions of ionic liquid formulations could possibly exist.

9.1.3 Key properties as electrolytes

Several key properties of ionic liquids result from the nature of their components and
the strong ionic interactions existing between them. These interactions hold the con-
stituents strongly together, making the liquid very difficult to vaporize. Having an elec-
trolyte with a high boiling point and a very small (even immeasurable in some cases)
vapour pressure adds to the stability and safety of batteries. This is also a desirable fea-
ture which allows the operation of cells at higher temperatures than with conventional
organic solvents. However, due to these strong intermolecular interactions, ionic liq-
uids are often more viscous than solvent-solute systems, resulting in a lower conduc-
tivity. Since they are entirely composed of ions, the concentration of charge carriers is
very high, and ionic liquids should therefore be able to transport high currents. Their
conductivity is, however, limited by their viscosity and non-ideal ionicity, meaning
that the ions are screened by surrounding counter-ions and do not behave as freely
moving charges. Overall, their conductivity can be high enough for electrochemical
applications but remains in the same range as organic solvent-solute systems. The
excellent electrochemical stability is also a property which makes for a strong case in
the application of ionic liquids in batteries. Ionic liquids that begin oxidizing at poten-
tials above those of carbonates or acetonitrile can be useful for the application of high
voltage cathode materials. Imidazoliums, by far the most studied family of cations,
are reduced, however, at the anode at potentials above those of lithium deposition
9 Functional ionic liquids electrolytes in lithium-ion batteries | 193

R2 R1 R1
R1 R2
R1 R3 R4 N +
R2 R4 P +
R2 N+
R3 R3
Alkylimidazolium Tetraalkylammonium Tetraalkylphosphonium Dialkylpyrrolidinium

R1 R1 R1 R2 R1
R2 R2
N+ N+

O R3 R2
(R4)2N N(R3)2

Dialkylpiperidinium Dialkylmorpholinium Guanidinium Trialkylsulfonium

CI–, Br–, I– BF4– PF6– CH3SO4– CF3SO3–

Halides Tetrafluoroborate Hexafluorophosphate Methyl sulfate Trifluoromethane-
N– N– N–
N N F3 C CF3 F F
Dicyanamide (DCN) Bis(trifluoromethyl- Bis(fluorosulfonyl)imide
sulfonyl)imide (TFSI) (FSI)

Fig. 9.2. Cations and anions of common air and water stable organic room temperature ionic liquids.

and intercalation in some other materials. Various types of cations (phosphoniums

and pyrrolidiniums), which lack an aromatic system, can be used instead of imida-
zoliums, however, providing better cycling stability. Other properties such as tunable
water miscibility (due to the nature of the anion) and solvating power can be taken
advantage of for several applications.

9.2 Ionic liquids as Li and Lithium-ion battery electrolytes

Like all batteries and electrochemical cells, Li and Li-ion batteries are composed of
two electrodes (the positive cathode and negative anode) separated by an electrolyte.
Operation of Li-ion batteries is based on the reversible intercalation of Li-ions in ox-
ides (cathode) and graphite (anode) [26–28]. A Li battery is similar in construction,
but metallic lithium is used at the anode instead of graphite or another intercalating
material. While Li-ion batteries are secondary (i.e., rechargeable) batteries, lithium
batteries are primary devices which cannot usually be recharged, due to the forma-
tion of dendrites at the anode surface following the reduction of Li-ions. Controlling
194 | Part II Development of new materials for energy applications

Li deposition during recharging is an important issue due to the very high capacity of
Li(0) anodes. Ionic liquids have shown promising results in that field, but this topic
will not be discussed further here [29, 30]. For both types of battery, the electrolyte has
the important role of ionic conduction for the transportation of lithium ions between
the electrodes and balancing the charge passing through the external circuit as cur-
rent. The electrolyte should be sufficiently fluid to allow rapid ion diffusion, should be
good at solvating ions (high dielectric constant), and should be electrochemically and
thermally stable. The idea behind the use of ionic liquids as battery (and other elec-
trochemical energy storage devices) electrolyte is to take advantage of the intrinsic
properties of ILs which increase the thermal and electrochemical stability of the de-
vice. Replacing traditional volatile and flammable solvents with ionic liquids would
provide safer batteries, decreasing the risk of self-combustion and explosion in the
event of abuse or an accident of the system. The increased electrochemical stability
would also allow the use of high potential cathodes, thereby increasing the operating
voltage and the energy of the battery. On the downside, the significant disadvantages
of ionic liquids, mostly their high viscosity and low transport, will decrease the overall
cell performance, explaining in part why ionic liquid electrolytes have not yet made it
to commercial batteries, cost being another issue.
The most studied types of anode and cathode material for lithium and lithium-ion
batteries have been tested in cells with ionic liquid electrolytes. An extensive list can
be found in a review paper by A. Lewandowski [8]. The ionic liquids which have been
the most studied for batteries are based on the alkylimidazolium and pyrrolidinium
families. Ionic liquids based on imidazolium have the lowest viscosity, which is highly
important to favor charge transport by the Li-ions. Pyrrolidiniums are more stable
towards the reduction reaction than imidazoliums, which usually decompose at the
Li(0) and graphite anodes. The TFSI anion is the most reported due to the low viscos-
ity of the ionic liquids employing it, but tetrafluoroborate and hexafluorophosphate
have been studied extensively due to their wide application in conventional battery
electrolytes. There are currently no commercial lithium batteries based on ionic liquid
electrolytes, but research is progressing and studies on lab-scale stacked prototypes
are showing promise and helping to identify current technological challenges [31, 32].
In addition to the thermal stability they confer to electrochemical systems, ionic liq-
uids might very well make their place in batteries by achieving tasks which no other
electrolyte based on solvent-solute systems can do.

9.3 Functional ionic liquid electrolytes

The preceding section demonstrated the use of typical ionic liquids as Li-ion battery
electrolytes, taking advantage of their intrinsic thermal and electrochemical stability
to improve related properties in the battery. In this section, ionic liquids modified with
a functional group will be evaluated in the performance of certain tasks essential to
9 Functional ionic liquids electrolytes in lithium-ion batteries | 195

efficient battery operation. Such functionalization of ionic liquids is usually done on

organic cations (imidazoliums, phosphoniums or ammoniums) and uses well-known
chemistries which are generally accessible to research groups not specialized in or-
ganic molecule synthesis. The application of functional ionic liquids for generation
of the solid electrolyte interphase (SEI), to improve Li-ion transport and to prevent
over-oxidation of cathodes will be discussed.

9.3.1 Overview of functional ionic liquids

In a 2001 publication, R. Rogers (University of Alabama, U.S.A.) et al. demonstrated

the first example of an ionic liquid with a structure that was developed for a specific
purpose [33]. These “task-specific ionic liquids” or TSIL, were obtained by modify-
ing imidazoliums with a functional group based on the chemical structure of extrac-
tants used to extract heavy metal ions from non-aqueous solvents (Fig. 9.3). By com-
bining functionalized imidazoliums with the hydrophobic PF−6 anion, the ionic liq-
uids were non-miscible with water, allowing the extraction of Hg2+ and Cd2+ from
the ionic liquid from an aqueous phase. The distribution ratios obtained with vari-
ous TSIL based on the structures shown in Fig. 9.2 were above 300 for both M2+ ions
with pure TSIL. In comparison, the distribution ratio for both ions with the unmod-
ified 1-butyl-3-methylimidazolium: PF6 extracting phase were below unity, showing
the role of the functional group. This is a strong demonstration of how a properly de-
signed ionic liquid can perform the task of a solute-solvent system without requiring
dissolved additives. Several advantages are associated with this concept of adding a
functional group directly to the molecular structure of one of the constituents of a
liquid phase. The common benefits of ionic liquids such as low volatility, intrinsic
conductivity, and thermal stability are retained in TSIL and can facilitate the recycla-
bility of the liquid phase. These advantages motivated the development of numerous
functionalized ionic liquids over the years.

N + N N N
Fig. 9.3. Functional (task-specific) ionic liq-
S uids based on alkylimidazoliums modified with
R thiourea (top) and thioether (bottom) for the ex-
N + N R
traction of heavy metal ions. R represents a C3−6
PF6– alkyl chain.
196 | Part II Development of new materials for energy applications

Currently, TSIL are being developed for CO2 capture [34, 35], catalysis [36–39], and
extraction or solubilization of insoluble materials [40, 41]. There are only a few exam-
ples of functional ionic liquids being designed as electrolytes for energy storage and
conversion systems. Apart from specific examples of lithium and lithium-ion batteries
(discussed below), dye-sensitized solar cells [42], and lithium-air batteries [43], the
majority of papers on the functionalization of ionic liquids used as electrolytes are
based on ether modification to decrease viscosity and therefore address one of the
major limitations of ionic liquids [44]. In fact, most of the cations presented in Fig. 9.2
were functionalized with an ether group on one of their alkyl chains and their electro-
chemical behaviour was studied later. While modifying ionic liquids to reduce their
viscosity is highly important for their application as electrolytes, this chapter focuses
on ionic liquids functionalized with a reactive group designed to be involved in some
of the electrochemical reactions taking place during battery operation. The selected
examples presented below provide an overview of the possibilities offered by the use
of functional ionic liquid electrolytes in lithium and lithium-ion batteries.

9.3.2 Solid electrolyte interphase

During the first cycle of operation of Li-ion batteries, a passivation layer is formed at
the anode. This layer, called solid electrolyte interphase (SEI), is generated by the re-
duction of electrolyte (and additives if present), resulting in a complex mixture of inor-
ganic and organic materials with a thickness in the nanometer range [45]. The SEI pas-
sivates the anode which prevents further electrolyte decomposition, but allows Li-ions
to permeate and ensures good cyclability with sufficient cell performance. Common
solvents based on carbonates used in Li-ion battery electrolytes (ethylene, diethyl,
dimethyl, propylene carbonates) will typically decompose at the anode, generating
Li2 CO3 , polycarbonates, and several other species [45]. Additives like vinylene carbon-
ate (VC) are commonly added to the electrolyte to accelerate SEI formation and prevent
excessive generation of gases from the reduction of the carbonate solvents [46].
As it is usually done with electrolytes based on conventional solvents, VC and
other carbonates can be dissolved in ionic liquids (5–10 wt %) to accelerate SEI forma-
tion [47]. However, some issues related to the reduction of the ionic liquid itself (imida-
zolium) and limited Li+ intercalation might occur [8]. In the case of lithium batteries,
in which the anode is metallic lithium, and where Li deposition-dissolution processes
are involved rather than intercalation (as in graphite), the irreversible decomposition
of 1,3-substituted imidazolium cations essentially blocks the anode surface and de-
creases cell output. An approach taken by M. Egashira (Yamaguchi University, Japan)
to better control the SEI in Li batteries using ionic liquid electrolytes is to modify imi-
dazolium salts with cyano groups [48–50]. Examples of such functional ionic liquids
are presented in Fig. 9.4. The idea that motivated the design of these ionic liquids was
to deliberately incorporate a functionality of the cation that, once decomposed, will
9 Functional ionic liquids electrolytes in lithium-ion batteries | 197

enter the composition of the SEI to provide a stable film compatible with cell oper-
ation. The chemistry of the SEI is controlled by the solvent itself rather than via the
use of additives. The results obtained showed that the cyano group on the ionic liquid
accelerated the imidazolium ring opening reaction, leading to SEI formation [48]. The
SEI layer formed from CN-modified imidazolium on a stainless steel anode provided
a reversible lithium deposition and dissolution. There was no passivation on elec-
trodeposited Li with an electrolyte based on unmodified 1-ethyl-3-methylimidazolium
(EMIM) cation electrolyte. The capacity of a LiCoO2 cathode obtained in an electrolyte
composed of a mixture of the cyano-substituted quaternary ammonium TFSI (Fig. 9.4)
and EMIM TFSI was measured at 110 mAh/g, slightly lower than could be obtained in
conventional electrolytes [50]. Functional ionic liquids based on guanidinium cations
bearing ether groups have also been studied in lithium batteries to improve Li depo-
sition [44]. Interestingly, the authors reported the plating of Li on a Ni anode even if
the cathodic potential limit of their ionic liquids was higher than that of Li+ reduction.
This was attributed to the formation of the SEI, which was obtained by decomposition
of the functional ionic liquid (no SEI-forming additives were added). Their electrolyte
also provided good passivation in a Li/LiFePO4 cell, but the capacity obtained (110–
120 mAh/g) was lower than for conventional electrolytes, most likely due to the higher
viscosity of IL-based electrolytes.

N + N CN


Fig. 9.4. Imidazolium and ammonium ionic liquids functional-

N+ CN ized by cyano groups used to improve SEI formation in Li bat-

9.3.3 Transport of lithium ions

One of the major limitations arising from the use of ionic liquids as electrolytes for
Li-ion batteries is the low transport numbers for Li+ . For a battery to work at high
power rates during charging and discharging, the amount of ionic charge by the Li+
ions moving across the electrolyte must be as high as possible. The fraction of charge
transported by the lithium ions, represented by the transport number, is necessarily
lower in ionic liquids than in typical electrolytes due to the absence of solvent in the
former. This is clearly demonstrated by comparison of the equations for the transport
198 | Part II Development of new materials for energy applications

number of Li+ (tLi+ ) for the two systems:

Ionic liquid C+ A− + Li+ A− salt: tLi+ =
uLi+ + uC+ + 2uA−
Li+ A− salt in solvent: tLi+ =
uLi+ + uA−
In these equations, ui represents the mobility of the ion i. In other words, there is a
high concentration of cations other than Li+ in ionic liquids that participates in charge
transport, but these other ions will not intercalate and therefore won’t contribute to
energy storage. To address this issue, neutral lithium ion-complexing additives such
as glymes (oligoethers of CH3 (CH2 CH2 O)n CH3 composition) have been added to ionic
liquid electrolytes to reduce the interactions with the anion [51]. In this case, the ion-
icity of the ionic liquid – LiTFSI electrolyte was found to increase with the addition
of glyme. M. Watanabe (Yokohama University, Japan) demonstrated that by selecting
the appropriate anion, mixtures of Li salts and glymes formed a liquid phase on their
own with ionic liquid properties [52, 53]. Fig. 9.5 shows the structure of such an ionic
liquid in which the cation is composed of a Li+ solvated by a molecule of tetraglyme.
This family of ionic liquids are hence called solvate ionic liquids. Because of such sol-
vated complexes, the Li ions interact less with the TFSI anion, behave like an ion in a
neutral solvent and provide a high transport number. In the Li(tetraglyme)-TFSI, the
tLi+ was as high as 0.51 [52], representing a significant improvement to that of mixtures
of imidazolium ionic liquid with Li salts, which are found below 0.1 [54, 55].

O – O O O
S N S CF3 Li+
O O O O Fig. 9.5. Structure of a Li(glyme) ionic liquid based
on tetraglyme and the TFSI anion.

Lithium batteries employing LiFePO4 and LiNi1/3 Mn1/3 Co1/3 O2 as low (3V) and high
(4 V) voltage cathodes have been studied with the pure Li-(CH3 O(CH2 CH2 )3 CH3 )-TFSI
ionic liquid as the electrolyte [56]. The cell capacities as a function of cycle number
are shown in Fig. 9.6. Good stability was obtained for the LiFePO4 cells for 600 cycles,
showing stable electrochemical reactions at both positive and negative electrodes in
the ionic liquid. Decomposition of the electrolyte is likely to occur at higher potentials,
which could explain the important capacity loss in the 4 V cells (Fig. 9.6(b)).

9.3.4 Electroactive ionic liquids as redox shuttles

Redox shuttles are electroactive molecules added to the Li-ion battery electrolytes to
prevent oxidation of the cathode over the level required for charging (over-oxidation)
9 Functional ionic liquids electrolytes in lithium-ion batteries | 199

4.0 4.2

3.7 3.9
Cell Voltage/V

Cell Voltage/V
3.4 3.6

3.1 3.3
1st 300th 1st
2.8 50th 400th 3.0 10th 200th
100th 500th 50th 300th
200th 600th 100th 400th
2.5 2.7
0 50 100 150 200 0 25 50 75 100 125 150 175
Capacity/mAhg–1 Capacity/mAhg–1
(a) (b)

Fig. 9.6. Charge-discharge profiles of (a) Li/LiFePO4 , and (b) Li/LiNi1/3 Mn1/3 Co1/3 O2 cells using
Li(glyme)-TFSI ionic liquid electrolyte (C/8 rate at 30°C). Reproduced by permission from [56].

[57–59]. The operating principle is shown in Fig. 9.7. The over-oxidation of cathode
material leads not only to the irreversible degradation of the electrode, but will also
generate radicals, oxygen, and other reactive species which can cause a rapid ther-
mal runaway leading to explosion of the battery. Under normal operation, the redox
shuttle has no function and should not interfere with the processes involved in charg-
ing or discharging the battery. When current is passed through a fully charged (oxi-
dized) cathode, the redox shuttle (R-S[R]) will however be oxidized (to R-S[O]), thereby
preventing over-oxidation. The now oxidized shuttle will diffuse (D[O]) to the anode
to be reduced to complete the cycle and will become available to diffuse back to the
cathode (D[R] in Fig. 9.7).
The motivation behind the development of functional ionic liquids which can be
used as redox shuttle is two-fold. Firstly, there is the very appealing idea of combin-
ing the stability inherent to ionic liquids with the cathode protection offered by the
redox shuttle principle to develop a bi-functional electrolyte for safer batteries. Sec-
ondly, the modification of the ionic liquid could address the general solubility issues of
electroactive organic molecules in solvents and ionic liquids, leading to much higher
concentrations of redox moieties in the electrolyte. This higher concentration could
provide better protection, especially when high currents are applied to charge the bat-
tery in a shorter time. The following equation shows the direct relationship between
the maximum current (IMAX ) which can be carried out between two electrodes sepa-
rated by a distance (L) from the redox shuttle, and its concentration (C), in which n is
the number of charge per shuttle, F is the Faraday constant, A is the surface area of
the electrode, and D is the diffusion coefficient of the shuttle [60]:
IMAX = (9.1)
If a pure electroactive ionic liquid phase can be used (that is, if its viscosity is not too
high) the concentration of redox moiety in the liquid could exceed 2 M, which could
never be obtained by simply dissolving a redox molecule in an ionic liquid [61].
200 | Part II Development of new materials for energy applications


[O] [O]

[R] [R]

Anode Cathode

Fig. 9.7. Operating principle of a redox shuttle dissolved in a Li-ion battery electrolyte. R–S[O] and
R–S[R] are the redox shuttle in its oxidized and reduced states, respectively, D shows the diffusion
of the shuttle in the two states between the electrodes. Reproduced by permission from [65].

O N +
Link CF3 N
Side Chain O
S –
N + –N O O N O
Fe N O S S
n S F3C CF3

Fig. 9.8. Two types of electroactive ionic liquids studied as redox shuttle for Li-ion batteries,
obtained by the modification of imidazolium with ferrocene (left) and 2,5-di-tert-butylmethoxy-
benzyne (right).

The first demonstration of the use of electroactive ionic liquids as redox shuttle was
recently reported by our group [62]. In this report, an imidazolium-based ionic liquid
was functionalized with ferrocene as the redox moiety (Fig. 9.8).
The 1-ferrocenylmethyl-3-methylimidazolium bis(trifluoromethanesulfonyl)
amide (Fc-MIm TFSI; Fig. 9.8, left with n = 3) ionic liquid was dissolved in ethylene
carbonate–diethyl carbonate solvent (EC-DEC + 1.5 M LiTFSI) and incorporated into
a coin-type cell with a Li4 Ti5 O12 (LTO) cathode and a Li foil anode. LTO has a very
low charging potential (1.6 V vs. Li/Li+ ) and while not usable as commercial battery
material, it was used to demonstrate the concept. Figure 9.9 shows the specific ca-
pacity curves for the charge/discharge cycles of the test coin-type cell, starting with
a full charge, followed by a 100 % overcharge for a Li/Li4 Ti5 O12 coin cell at a C/10
rate. In this figure, the unmodified electrolyte (i.e., without electroactive ionic liquid)
9 Functional ionic liquids electrolytes in lithium-ion batteries | 201

is the solid line and the electrolyte containing 10 % of Fc-MIm TFSI is represented by
the broken lines. The overcharging situation appears very clearly for the cell without
redox shuttle added where the voltage increases sharply to a 4 V cut-off after the
charging plateau at 1.6 V for Li4 Ti5 O12 material. Adding the redox ionic liquid TO the
electrolyte prevents reaching the cut-off voltage of 4 V. A plateau is rather observed at
∼ 3.36 V, corresponding to the onset potential for oxidation of the ferrocene moiety
on the ionic liquid and demonstrating the protection against over-charging by the
shuttle mechanism.

No Fc-MIm TFSI, all cycles

4.0 10% Fc-MIm TFSI, 1st cycles
10% Fc-MIm TFSI, 2nd cycles
10% Fc-MIm TFSI, 3rd cycles
Cell voltage (V)






–50 0 50 100 150 200 250 300 350 400

Specific capacity (mAh/g)

Fig. 9.9. Capacity curves (C/10) for Li/Li4 Ti5 O12 cells using EC/DEC (1.5 M LiTFSI) electrolyte either
pure (solid line), or modified with 10 % Fc-MImTFSI (broken lines). The parameters were set to a full
charge followed by a 100 % overcharge and a cut-off at 4 V. Reproduced by permission from [62].

The use of Fc-MIm TFSI presents two important limitations. First, the electroactive
ionic liquid is very viscous due to a large cation and additional intermolecular interac-
tions between the ferrocene and imidazolium aromatic systems [63]. The high viscosity
prevented the use of an ionic liquid as an undiluted phase, but its dilution at 10 vol %
in an ethylene carbonate–diethyl carbonate solvent (EC-DEC + 1.5 M LiTFSI) provided
an electrolyte with acceptable viscosity (7.1 cP) and conductivity (5.5 mS/cm) for a use
in Li-ion cells [62]. The second challenge is related to the low oxidation potential of
ferrocene. The onset for the oxidation of Fc-MIm TFSI in the carbonate electrolyte (at
10 vol %) is 3.4 V vs. Li/Li+ , which is very close to the charging potential of LiFePO4 ,
a common Li-ion battery cathode material. In order to apply the electroactive ionic liq-
uid shuttle principle to this cathode material, a second series of ionic liquids was de-
signed, based on the successful commercial shuttle 2,5-di-tert-1,4-dimethoxybenzene
(DDB) [64]. These ionic liquids, dissolved in the EC-DEC + 1.5 M LiTFSI solution; start
being oxidized at potential values between 3.81 and 3.94 V vs. Li/Li+ due to the higher
202 | Part II Development of new materials for energy applications

redox potential of the DDB moiety (Fig. 9.8) [64]. Preliminary results in coin-type cells
using LiFePO4 cathodes demonstrated cathode protection capabilities against over-
charging (Fig. 9.10). Notably, the concentration in the redox shuttle was increased up
to 1 M due to the high solubility of imidazolium TFSI ionic liquid in carbonates. In
comparison, the poor solubility of DDB limits its working concentration in batteries
below 0.1 M.


Potential\V vs Li/Li+



0 50 100 150 200 250 300 350

Fig. 9.10. Voltage profile of a Li/LiFePO4 cell containing 1 M of the 2,5-di-tert-1,4-dimethoxybenzene-

based ionic liquid in 0.7 M LiTFSI EC/DEC (1 : 2 v/v).

9.3.5 Perspectives

Ionic liquids with various structures and properties have been applied and are still be-
ing studied to replace conventional electrolytes in energy storage devices. Advances
in understanding of how the structure of ionic liquid components affects its properties
made several of these discoveries possible. The application of functional ionic liquids
is likely to follow a similar pattern. The examples provided in this text demonstrate
the potential of task-specific ionic liquids to address specific issues of energy stor-
age devices although their application usually decreases the device performance. The
challenge is to determine if (and how) some of the most important viscosity, conduc-
tivity, and melting point limitations caused by the increase in ion size when adding a
large and complex (e.g., aromatic) functional group can be overcome. Therefore, more
work remains to be done to better understand the structure and the interactions taking
place between molecules in ionic liquids bearing a functional group.
9 Functional ionic liquids electrolytes in lithium-ion batteries | 203


[1] Armand M, Endres F, MacFarlane DR, Ohno H, Scrosati B. Ionic-liquid materials for the electro-
chemical challenges of the future. Nat Mater, 8 (2009) 621–9.
[2] Galinski M, Lewandowski A, Stepniak I. Ionic liquids as electrolytes. Electrochim Acta,
51 (2006) 5567–80.
[3] Macfarlane DR, Forsyth M, Howlett PC, et al. Ionic liquids in electrochemical devices and pro-
cesses: managing interfacial Electrochemistry. Acc Chem Res, 40 (2007)1165–73.
[4] Baldelli S. Surface structure at the ionic liquid-electrified metal interface. Acc Chem Res,
41 (2008) 421–31.
[5] Fernicola A, Scrosati B, Ohno H. Potentialities of ionic liquids as new electrolyte media in ad-
vanced electrochemical devices. Ionics, 12 (2006) 95–102.
[6] Ohno H, ed. Electrochemical Aspects of Ionic Liquids. 2nd ed. Hoboken, NJ: John Wiley & Sons;
[7] Garcia B, Lavallee S, Perron G, Michot C, Armand M. Room temperature molten salts as lithium
battery electrolyte. Electrochim Acta, 49 (2004) 4583–8.
[8] Lewandowski A, Swiderska-Mocek A. Ionic liquids as electrolytes for Li-ion batteries:
An overview of electrochemical studies. J Power Sources, 194 (2009) 601–9.
[9] Wishart JF. Energy applications of ionic liquids. Energy Environ Sci, 2 (2009) 956–61.
[10] Angell CA, Ansari Y, Zhao ZF. Ionic Liquids: Past, present and future. Faraday Discussions,
154 (2012) 9–27.
[11] Walden P. Molecular weights and electrical conductivity of several fused salts. Bulletin de
l’Académie Impériale des Sciences de St-Petersbourg 1914, 405–22.
[12] Wilkes JS. A short history of ionic liquids - from molten salts to neoteric solvents. Green
Chemistry, 4 (2002) 73–80.
[13] Fuller J, Carlin RT, Delong HC, Haworth D. Structure of 1-ethyl-3-methylimidazolium hexaflu-
orophosphate: model for room temperature molten salts . Journal of the Chemical Society-
Chemical Communications 1994, 299–300.
[14] Wilkes JS, Zaworotko MJ. Air and water stable 1-ethyl-3-methylimidazolium based ionic liquids.
Journal of the Chemical Society-Chemical Communications 1992, 965–7.
[15] Cooper EI, O’Sullivan EJ. New, stable, ambient-temperature molten salts. Proceedings of the
8th International Symposium on Molten Salts, 92-16, pp 386–96, 1992.
[16] Parvulescu VI, Hardacre C. Catalysis in ionic liquids. Chem Rev, 107 (2007) 2615–65.
[17] Sheldon R. Catalytic reactions in ionic liquids. Chem Comm 2001, 2399–407.
[18] Welton T. Room-temperature ionic liquids. Solvents for synthesis and catalysis. Chem Rev,
99 (1999) 2071–83.
[19] Ho TD, Zhang C, Hantao LW, Anderson JL. Ionic Liquids in Analytical Chemistry: Fundamentals,
Advances, and Perspectives. Anal Chem, 86 (2013) 262–85.
[20] Koel M, ed. Ionic Liquids in Chemical Analysis. Boca Raton, FL: CRC Press; 2012.
[21] Plechkova NV, Seddon KR. Applications of ionic liquids in the chemical industry. Chem Soc
Rev, 37 (2008) 123–50.
[22] Janz GJ. Molten salts data as reference standards for density, surface tension, viscosity and
electrical conductance - KNO3 and NaCl. J Phys Chem Ref Data, 9 (1980) 791–829.
[23] Krossing I, Slattery JM, Daguenet C, Dyson PJ, Oleinikova A, Weingartner H. Why are ionic liq-
uids liquid? A simple explanation based on lattice and solvation energies. J Am Chem Soc,
128 (2006) 13427–34.
[24] Faria LFO, Matos JR, Ribeiro MCC. Thermal Analysis and Raman Spectra of Different Phases of
the Ionic Liquid Butyltrimethylammonium Bis(trifluoromethylsulfonyl)imide. J Phys Chem B,
116 (2012) 9238–45.
204 | Part II Development of new materials for energy applications

[25] Ngo HL, LeCompte K, Hargens L, McEwen AB. Thermal properties of imidazolium ionic liquids.
Thermochimica Acta, 357 (2000) 97–102.
[26] Tarascon JM, Armand M. Issues and challenges facing rechargeable lithium batteries. Nature,
414 (2001) 359–67.
[27] Goodenough JB, Park K-S. The Li-Ion Rechargeable Battery: A Perspective. J Am Chem Soc,
135 (2013) 1167–76.
[28] Goodenough JB, Kim Y. Challenges for Rechargeable Li Batteries. Chem Mater, 22 (2010)
[29] Lane GH, Best AS, MacFarlane DR, Hollenkamp AF, Forsyth M. An Azo-Spiro Mixed Ionic Liquid
Electrolyte for Lithium Metal-LiFePO4 Batteries. J Electrochem Soc, 157 (2010) A876–A84.
[30] Schweikert N, Hofmann A, Schulz M, et al. Suppressed lithium dendrite growth in lithium
batteries using ionic liquid electrolytes: Investigation by electrochemical impedance spec-
troscopy, scanning electron microscopy and in situ 7Li nuclear magnetic resonance spectro-
scopy. J Power Sources, 228 (2013) 237–43.
[31] Balducci A, Jeong SS, Kim GT, et al. Development of safe, green and high performance ionic
liquids-based batteries (ILLIBATT project). J Power Sources, 196 (2011) 9719–30.
[32] Kim GT, Jeong SS, Xue MZ, et al. Development of ionic liquid-based lithium battery prototypes.
J Power Sources, 199 (2012) 239–46.
[33] Visser AE, Swatloski RP, Reichert WM, et al. Task-specific ionic liquids for the extraction of
metal ions from aqueous solutions. Chem Comm 2001, 135–6.
[34] Wang C-M, Mahurin SM, Luo H-M, Baker GA, Li H-R, Dai S. Reversible and robust CO2 capture
by equimolar task-specific ionic liquid-superbase mixtures. Green Chem, 12 (2010) 870–4.
[35] Zhang Z, Xie Y, Li W, et al. Hydrogenation of carbon dioxide is promoted by a task-specific ionic
liquid. Angew Chem, Int Ed, 47 (2008) 1127–9.
[36] Sun J, Cheng W, Fan W, Wang Y, Meng Z, Zhang S. Reusable and efficient polymer-supported
task-specific ionic liquid catalyst for cycloaddition of epoxide with CO2. Catal Today,
148 (2009) 361–7.
[37] Wang L, Li H, Li P. Task-specific ionic liquid as base, ligand and reaction medium for the
palladium-catalyzed Heck reaction. Tetrahedron, 65 (2008) 364–8.
[38] Xu J-M, Liu B-K, Wu W-B, Qian C, Wu Q, Lin X-F. Basic Ionic Liquid as Catalysis and Reaction
Medium: A Novel and Green Protocol for the Markovnikov Addition of N-Heterocycles to Vinyl
Esters, Using a Task-Specific Ionic Liquid, [bmIm]OH. J Org Chem, 71 (2006) 3991–3.
[39] Kamal A, Chouhan G. A task-specific ionic liquid [bmim]SCN for the conversion of alkyl halides
to alkyl thiocyanates at room temperature. Tetrahedron Lett, 46 (2005) 1489–91.
[40] Nockemann P, Thijs B, Pittois S, et al. Task-Specific Ionic Liquid for Solubilizing Metal Oxides.
J Phys Chem B, 110 (2006) 20978–92.
[41] Ohno H, Fukaya Y. Task specific ionic liquids for cellulose technology. Chem Lett, 38 (2009)
[42] Miao Q, Zhang S, Xu H, Zhang P, Li H. A novel ionic liquid-metal complex electrolyte for a
remarkable increase in the efficiency of dye-sensitized solar cells. Chem Comm, 49 (2013)
[43] Nakamoto H, Suzuki Y, Shiotsuki T, et al. Ether-functionalized ionic liquid electrolytes for
lithium-air batteries. J Power Sources, 243 (2013) 19–23.
[44] Jin YD, Fang SH, Yang L, Hirano SI, Tachibana K. Functionalized ionic liquids based on guani-
dinium cations with two ether groups as new electrolytes for lithium battery. J Power Sources,
196 (2011) 10658–66.
[45] Verma P, Maire P, Novak P. A review of the features and analyses of the solid electrolyte inter-
phase in Li-ion batteries. Electrochim Acta, 55 (2010) 6332–41.
9 Functional ionic liquids electrolytes in lithium-ion batteries | 205

[46] Ota H, Sakata Y, Inoue A, Yamaguchi S. Analysis of Vinylene Carbonate Derived SEI Layers on
Graphite Anode. J Electrochem Soc, 151 (2004) A1659–A69.
[47] Egashira M, Kiyabu T, Watanabe I, Okada S, Yamaki J. The effect of additives in room tempera-
ture molten salt - based lithium battery electrolytes. Electrochemistry, 71 (2003) 1114–6.
[48] Zhao L, Yamaki J-i, Egashira M. Analysis of SEI formed with cyano-containing imidazolium-
based ionic liquid electrolyte in lithium secondary batteries. J Power Sources, 174 (2007)
[49] Egashira M, Todo H, Yoshimoto N, Morita M, Yamaki J. Functionalized imidazolium ionic liquids
as electrolyte components of lithium batteries. J Power Sources, 174 (2007) 560–4.
[50] Egashira M, Tanaka-Nakagawa M, Watanabe I, Okada S, Yamaki JI. Charge-discharge and high
temperature reaction of LiCoO2 in ionic liquid electrolytes based on cyano-substituted quater-
nary ammonium cation. J Power Sources, 160 (2006) 1387–90.
[51] Bayley PM, Lane GH, Lyons LJ, MacFarlane DR, Forsyth M. Undoing Lithium Ion Association in
Ionic Liquids through the Complexation by Oligoethers†. J Phys Chem C, 114 (2010) 20569–76.
[52] Ueno K, Yoshida K, Tsuchiya M, Tachikawa N, Dokko K, Watanabe M. Glyme-Lithium Salt
Equimolar Molten Mixtures: Concentrated Solutions or Solvate Ionic Liquids? J Phys Chem B,
116 (2012) 11323–31.
[53] Tsuzuki S, Shinoda W, Seki S, et al. Intermolecular Interactions in Li+-glyme and Li+-glyme-
TFSA(-) Complexes: Relationship with Physicochemical Properties of Li(glyme) TFSA Ionic
Liquids. Chemphyschem, 14 (2013) 1993–2001.
[54] Wu TY, Hao L, Chen PR, Liao JW. Ionic Conductivity and Transporting Properties in LiTFSI-Doped
Bis(trifluoromethanesulfonyl)imide-Based Ionic Liquid Electrolyte. Int J Electrochem Sci,
8 (2013) 2606–24.
[55] Hayamizu K, Aihara Y, Nakagawa H, Nukuda T, Price WS. Ionic Conduction and Ion Diffusion in
Binary Room-Temperature Ionic Liquids Composed of [emim][BF4] and LiBF4. J Phys Chem B,
108 (2004) 19527–32.
[56] Seki S, Serizawa N, Takei K, Dokko K, Watanabe M. Charge/discharge performances of glyme–
lithium salt equimolar complex electrolyte for lithium secondary batteries. J Power Sources,
2 43 (2013) 323–7.
[57] Balakrishnan PG, Ramesh R, Kumar TP. Safety mechanisms in lithium-ion batteries. J Power
Sources, 155 (2006) 401–14.
[58] Chen ZH, Qin Y, Amine K. Redox shuttles for safer lithium-ion batteries. Electrochim Acta,
54 (2009) 5605–13.
[59] Chen J, Buhrmester C, Dahn JR. Chemical overcharge and overdischarge protection for lithium-
ion batteries. Electrochem Solid-State Lett, 8 (2005) A59–A62.
[60] Dahn JR, Jiang JW, Moshurchak LM, Fleischauer MD, Buhrmester C, Krause LJ. High-rate over-
charge protection of LiFePO4-based Li-ion cells using the redox shuttle additive
2,5-ditertbutyl-1,4-dimethoxybenzene. J Electrochem Soc, 152 (2005) A1283–A9.
[61] Balasubramanian R, Wang W, Murray RW. Redox Ionic Liquid Phases:? Ferrocenated Imida-
zoliums. J Am Chem Soc, 128 (2006) 9994–5.
[62] Forgie JC, El Khakani S, MacNeil DD, Rochefort D. Electrochemical characterisation of a lithium-
ion battery electrolyte based on mixtures of carbonates with a ferrocene-functionalised imida-
zolium electroactive ionic liquid. Phys Chem Chem Phys, 15 (2013) 7713–21.
[63] Fontaine O, Lagrost C, Ghilane J, et al. Mass transport and heterogeneous electron transfer of a
ferrocene derivative in a room-temperature ionic liquid. J Electroanal Chem, 632 (2009) 88–96.
[64] Forgie JC, Rochefort D. Electroactive imidazolium salts based on 1,4-dimethoxybenzene redox
groups: synthesis and electrochemical characterisation. RSC Adv, 3 (2013) 12035–8.
[65] Park J-K, ed. Principles and Applications of Lithium Secondary Batteries. Weinhem: Wiley-VCH;
C. de Bonis, A. D’Epifanio, B. Mecheri, S. Licoccia, and A. C. Tavares
10 Solid polymer proton conducting electrolytes
for fuel cells

10.1 Introduction

Numerous practical applications based on electrochemical cells use solid electrolytes.

These include batteries, fuel cells, sensors, electrolyzers, and also water purification,
electrodialysis, and seawater desalination [1–7].
An electrochemical cell is a device capable of producing electrical energy from
spontaneous chemical reactions (ΔGo < 0), or driving chemical reactions through an
external source of electrical energy (ΔGo > 0). Electrochemical cells contain two elec-
trodes (the anode and the cathode) with an electrolyte between them. Reactions at
the anode (oxidation) and at the cathode (reduction) are the driving forces of an elec-
trochemical cell. Oxidation refers to the loss of electrons by a chemical species (re-
ductant) and reduction to the gain of electrodes by a chemical species (oxidant). In
an electrochemical cell the electronic current flowing outside the cell equals the ionic
current flowing within the cell [8].
An electrolyte is a material able to conduct ions which is usually an electrical insu-
lator [8]. Electrolytes can be solutions (e.g., KOH, H2 SO4 ), molten salts (e.g. Li2 CO3 ),
solid ion conducting polymers (e.g., perfluorinated polymers bearing sulfonic acid
groups or benzyltrimethylammonium groups), and ionic crystals (e.g., ZrO2 :Y2 O3 ,
Na3 Zr2 PSiO12 ).
Solid electrolytes conducting O2− , H+ , Li+ , Na+ , Ag+ , F− , Cl− , OH− ions have been
reported for many years now. The conductivity range is typically 10−3 S/cm < σ <
10 S/cm depending on the material structure and operating temperature. For the pur-
pose of comparison, the conductivity of a solid electrolyte at room temperature is in-
ferior to that of a liquid electrolyte. For example, σ(KOH) = 0.6 S cm−1 (30%, 20°C),
σ(H2 SO4 ) = 0.82 S cm−1 (5.2 M, 20°C), and σ(Nafion® ) ≈ 0.07 S cm−1 (fully hydrated,
Solid electrolytes can be used in electrochemical cells as ion exchange mem-
branes to allow the passage of ionic current between the anode and the cathode
placed on opposite sides of the electrolyte, or in the electrodes when mixed (elec-
tronic and ionic) conductivity is needed [5, 9]. Ion conductivity σi is given by:
σi = zi ⋅ Ci ⋅ μi , (10.1)
where zi is the ion charge, Ci is the density of mobile ions, and μi the ion mobility. Thus,
a solid electrolyte has a large number of mobile ions. Ion conductivity is an activated
transport; therefore it increases exponentially as temperature increases:
σ = A exp (− ), (10.2)
208 | Part II Development of new materials for energy applications

where A is a proportional constant, Ea is the activation energy, R is the universal gas

constant (8.314 J mol−1 K−1 ), and T the temperature (in K).
There are two main classes of solid electrolytes: crystalline (or ionic) solids and
ion conducting polymers. In crystalline solid electrolytes, ion conductivity occurs by
means of ions hopping through energetically equivalent sites in the crystal structure.
High conductivity requires a large number of mobile ions, or in other words, a large
number of accessible empty sites, either vacancies or interstitial sites. A practical way
of increasing the density of mobile ions is by doping the crystalline solid with het-
erovalent ions forming solid solutions. For example, replacing three Li+ ions with one
Al3+ ion in Li4−3x Alx SiO4 to generate cation vacancies, or by replacing Zr4+ ions by Y3+
ions in yttria stabilized zirconia to generate anion vacancies. The activation energy
controls ion mobility. The empty and occupied sites should have similar potential en-
ergies with a low activation energy barrier for ion hopping between neighboring sites.
Ion mobility is thus related to the crystal structure. Ionic solids with a densely packed
crystal structure are characterized by large activation energy (1 eV or higher) and low
conductivity. Ionic solids such as α-AgI, RbAg4 I5, and Na β-Al2 O3 are known as fast
ion conductors, formed by solid frameworks with open conduction pathways, and are
characterized by low values of activation energy, for example 0.03 eV for AgI above
420 K. Practical applications of this class of compounds include, for example, ion se-
lective electrodes (Ag2 S for Ag+ and LaF3 for F− ), molten salts electrochemical cells
(Na β-alumina in ZEBRA batteries), and oxygen anion conductors (yttria stabilized zir-
conia) for solid oxide fuel cells and oxygen sensors.
Solid polymer electrolytes (SPE) consist of polymer backbones functionalized
with a high concentration of fixed ionic charges. The function of the SPE is deter-
mined by the charge of the ion exchange groups and the nature of the counter ions,
and can be classified as follows [10]:
(a) cation exchange membranes have anionic charged groups (-COO− , -SO−3 , etc.) and
cations can selectively permeate through them;
(b) anion exchange membranes have cationic charged groups (e.g. -NR+3 ) and anions
can selectively permeate through them;
(c) amphoteric ion exchange membranes contain randomly distributed cationic and
anionic functional groups;
(d) bipolar ion exchange membranes are bi-layer membranes with both a cation and
an anion exchange membrane layer;
(e) mosaic ion exchange membranes, which have separate domains with cationic and
anionic groups.
Types (a) and (b) are those used industrially.

Due to the presence of ionic groups, ion exchange membranes adsorb water molecules
to an extent dependent on the surrounding relative humidity. The electrical conduc-
tivity of ion exchange membranes depends on the concentration, size, and charge of
the ions, as well as on the water content, chemical structure, and morphology of the
10 Solid polymer proton conducting electrolytes for fuel cells | 209

membranes. In particular, ion mobility depends on its charge density and degree of
solvation [11].
Ion exchange membranes are used for the dehumidification of gases, in humidity
sensors, actuators, pervaporation, facilitated transport, and in electrochemical pro-
cesses such as electrodialysis, brine electrolysis, redox flow vanadium batteries, and
solid polymer electrolyte fuel cells [10]. A significant amount of the work on SPEs is
relevant to proton exchange membrane fuel cells, therefore this application has been
targeted for this chapter, with emphasis on SPEs functionalized with sulfonic acid

10.2 Proton exchange membranes

A fuel cell is an electrochemical device which converts the energy from a fuel into
electricity, and those using proton exchange membranes (PEM) are among the most
promising because of their potential application in portable electronics, stationary
and automotives. They operate between room temperature and 140°C and can use hy-
drogen, methanol or other liquid fuels. During operation the fuel is oxidized at the
anode, generating protons and electrons. The electrons flow from anode to cathode
through the external circuit, whereas the protons cross the electrolyte membrane to
reach the cathode. An oxygen reduction reaction takes place and water is produced:

Anode: H2 → 2 H+ + 2 e− (10.3)
+ −
Cathode: 1/2 O2 + 2 H + 2 e → H2 O (10.4)
Overall: H2 + 1/2 O2 → 2 H2 O (10.5)

The proton exchange membrane is the core component of PEM fuel cells. In order
to achieve high efficiency, the membrane must possess the following features: (a) high
proton conductivity to support high current with minimal resistive losses, (b) low per-
meability to reactants, (c) chemical and electrochemical stability under operating con-
ditions, (d) adequate mechanical strength and stability, and (e) production costs com-
patible with the intended application.
PEMs can be classified according to their polymer backbone as hydrocarbon mem-
branes, partially halogenated hydrocarbon membranes, and perfluorocarbon mem-
branes. The most common cation exchange group used for fuel cell applications is
the sulfonic acid group because it is a very strong acid (apparent pk −6 for –CF2 SO3 H
and 0–1 for alyl/alkyl –SO3 H), although phosphonic acid (pk1 = 2–3, pk2 = 7–8) and
imidazole are also protogenic groups of potential interest for operating temperatures
above 100°C and low relative humidity (RH) [10, 12].
PEMs rely on the mobility of protons in the aqueous network formed inside the
solid polymer [13]. Proton transport proceeds through the membrane following two
main mechanisms [14]. The first is the vehicle mechanism, where the proton diffuses
210 | Part II Development of new materials for energy applications

with the vehicle water. The counter diffusion of unprotonated water allows the net
transport of protons. Therefore, the observed conductivity depends on the rate of ve-
hicle diffusion and can be expressed as a function of the water self-diffusion coeffi-
cient (DH2 O ), which represents a measure of the average mobility of water in the mem-
brane. The other mechanism is known as the Grotthuss mechanism, “proton hopping”
or “structure diffusion”. In this process, the water molecules show pronounced local
dynamics but reside on their sites. The process consists of two steps: (1) proton transfer
from one water molecule to the other by hydrogen bonds, (2) consequent reorientation
of water dipoles resulting in the formation of an uninterrupted trajectory for proton
migration. A schematic description of two typical proton conduction mechanisms is
shown in Fig. 10.1 The prevalence of one mechanism or the other depends on the hy-
dration level of the membrane, and it has been suggested that proton hopping is more
significant for high water contents [15]. The activation energy for proton conduction
in SPEs depends on the water content, and typically decreases from 0.4–0.5 eV for dry
membranes (contain only residual water molecules) and 0.1 eV for fully hydrated and
swollen membranes [16].


Fig. 10.1. Simplified scheme of the proton transfer in

Nafion® by the Grotthuss mechanism (solid lines) and the
S S S S vehicle mechanism (dotted lines). Adapted with permission
from [17].

10.2.1 Nafion®

The key ionomer currently used in PEM fuel cell applications is Nafion® , which is
produced by DuPont. Nafion® is a perfluorosulfonic acid polymer (PFSA). It has a
polytetrafluoroethylene (PTFE) backbone, which confers high chemical inertness, and
side chains consisting of perfluorinated vinyl polyether ending in sulfonic acid groups
(–SO3 H), which give proton exchange capability to the polymer. The chemical struc-
ture is shown in Fig. 10.2, where the values of n, x, and y can be varied to produce
materials with different equivalent weights.
Nafion® shows excellent proton conductivity (0.09 to 0.12 S cm−1 at 80°C and RH be-
tween 34 and 100% RH [18]) and mechanical strength, as well as high thermal and
chemical stability. The structure of Nafion® as a function of its water content has
been the topic of many investigations and it has been investigated by swelling stud-
ies, infrared spectroscopy, small angle x-ray scattering, and transmission electron
10 Solid polymer proton conducting electrolytes for fuel cells | 211

CF y
CF2 x
CF z CF2 S
Fig. 10.2. Chemical structure of the
CF3 Nafion® ionomer.

microscopy to name a few [19]. These studies have shown that a hydrated membrane
contains two phases, an ionic phase which is associated with the hydrated sulfonic
acid groups, and a non-ionic phase which is the perfluorinated matrix. The actual
form of the phases depends on the water content. Since the early 1970s several mod-
els have been proposed for prediction of ionic transport properties of Nafion® , de-
scribing the way in which ionic groups aggregate. These models include the Mauritz–
Hopfinger Model [20], the Yeager Three Phase Model [21], and the Gierke Cluster
Network Model [22]. In the cluster network model proposed by Gierke and Hsu, the
structure is an inverted micelle in which the ion exchange sites are separated from the
fluorocarbon backbone, thus forming spherical clusters, connected by short narrow
channels; see Fig. 10.3.
Thus, with increasing water content, the clusters grow and form transitory in-
terconnections with one another. This network of collapsed channels leads to a
percolation-type phenomenon. Gierke and Hsu also used the percolation theory to
correlate electrical conductivity with the water content of the membrane, expressed
as λ, i.e., the number of water molecules per sulfonic group. According to this theory,
there is a critical amount of water available in the membrane, below which ion trans-
port is extremely difficult due to the absence of extended pathways. The percolation
threshold in Nafion® is around λ = 2 [20], as shown in Fig. 10.4, where conductivity
data is plotted against λ. At low hydration level, i.e., λ in the range of 1–2, it is rea-


SO–3 SO–3 SO–3 SO–3

SO–3 SO–3
SO–3 SO –
SO–3 3

SO–3 SO–3
4.0nm 1.0nm–
SO –
3 SO3 SO–3

SO–3 SO–3 SO–3 SO–3

– SO –
SO–3 –
SO 3 3

Fig. 10.3. Gierke’s cluster network model of Nafion® membranes.

Reprinted with permission from [22].
212 | Part II Development of new materials for energy applications

sonable to consider that all water molecules absorbed by the Nafion® membrane are
associated with the sulfonate heads because of the hydrophobic nature of the back-
bone and the hydrophilic nature of the sulfonic groups. Moreover, hydronium ions
will be localized on the sulfonate heads and conductivity will be extremely low, as
the amount of water absorbed is insufficient for the formation of a continuous water
phase [23]. For λ in the range of 3–5, the counterion clusters continue to grow and,
as λ approaches 5, the membrane will become more conductive since some counte-
rion clusters may connect. However, there is still insufficient water for all clusters to
coalesce. Molecular dynamics simulations indicate that 5 water molecules form the
primary hydration shell for the sulfonic groups and any additional water molecules
are not as strongly bound and thus form a free phase [24]. For λ ≥ 6, counterion clus-
ters coalesce to form larger clusters and eventually a continuous phase is formed and
the conductivity threshold is overcome.







0 5 10 15 20 25
Mole ratio H2O/SO3H

Fig. 10.4. Variation of the proton conductivity of Nafion® as a function of the water content in the
membrane. Reprinted with permission from [4].

Since proton mobility relies on the formation of a continuous aqueous network inside
the ionomer, proton conductivity shows a strong dependence on the hydration level of
the PEM, as shown in Fig. 10.4 for Nafion® . A humidification system is thus necessary
to keep the membrane hydrated during fuel cell operation, which represents a major
cost of the fuel cell system [25].
While water sorption improves proton conductivity, it also leads to morphological
instability and, at elevated water content, to membrane swelling [26]. The maximum
working temperature of all Nafion® -based fuel cells is limited to 80–90°C due to the
loss of mechanical strength of the membrane determined, at higher temperature and
10 Solid polymer proton conducting electrolytes for fuel cells | 213

large hydration, by the plasticizing effect of water. Moreover, under dynamic condi-
tions, swelling cycles contribute to mechanical fatigue.
In fact, one of the key challenges in the design of proton exchange membranes
is the retention of high conductivity at low water content [3], especially at high tem-
peratures [25]. Overall, operation at high temperature (> 100°C) is desirable to reduce
PEM fuel cell costs and to promote large-scale commercialization: it enhances reaction
kinetics at both electrodes thus reducing the catalyst loading on both electrodes, it al-
lows more efficient utilization of the waste heat, and simplifies the water and thermal
management systems [25].
Nafion® membranes possess an additional major hurdle which inhibits large-
scale commercialization of fuel cells operating with liquid fuels. The unique mi-
crostructure of the Nafion® ionomer results in a high crossover rate of liquid fuels
from the anode to the cathode through the membrane [27]. This not only lowers fuel
utilization at the anode but also increases the overpotential of the cathode, hence
lowering the cell performance.
All of these drawbacks essentially imply that the Nafion® membrane cannot be
used “as is” for fuel cell applications in a wide temperature range, relative humidity
(RH) and liquid fuels. Substantial effort is being made to develop membranes with
appropriate electrochemical and other physico-chemical properties under operating
conditions [28–30]. Different approaches are being pursued and include: (a) the devel-
opment of alternative ionomers based on non perfluorinated polymers [31, 32], on pol-
yarylene or on aliphatic main chains [33–38], (b) the modification of existing ionomer
membranes through the formation of blends and composites [29], and (c) the synthe-
sis of new hybrid systems [39, 40].

10.2.2 Alternative sulfonated ionomers and membranes

Among the alternative polymers to Nafion® , arylene main-chain polymers, such as

poly(ether ketone), poly(ether sulfone), poly(benzimidazole), and poly(phenylene
sulfone), as shown in Fig. 10.5, have been widely investigated [34–36, 41–44].
Such polymers are inexpensive and possess high chemical and mechanical sta-
bility at temperatures higher than 90–100°C [34–36]. Moreover, their aromatic struc-
ture offers the possibility of electrophilic and nucleophilic substitutions, to prepare
ionomers with the desired features for PEMFC and DMFC applications [45]. The most
important modification regards the introduction of sulfonic acid moieties to obtain
proton-conducting aromatic polymers. Several methods have been developed for the
preparation of proton-conducting electrolytes, including direct sulfonation of a poly-
mer backbone, total synthesis from monomer building blocks, and grafting of func-
tional groups onto a polymer main chain [46]. In general, the larger the number of sul-
fonic acid groups per structural unit, the larger the membranes’ ionic exchange capac-
ity and water uptake and the higher their conductivity. However, excessive swelling of
214 | Part II Development of new materials for energy applications


(a) n



(b) CH3 O n




Fig. 10.5. Chemical structure of (a) Polyetheretherketone; (b) Udel Polysulfone;

and (c) Poly(2,2󸀠 -m-(phenylene)-5,5󸀠 -bibenzimidazole).

the membranes could lead to dilution of the charge carriers and to lower proton con-
ductivity [33].
Some variations on Nafion® ’s hydration scheme are expected for sulfonated pol-
yarylene membranes. Sulfonated polyarylenes with sulfonic acid groups bound di-
rectly to the aromatic chain have less pronounced hydrophobic/hydrophilic separa-
tion with respect to Nafion® because their backbones are less hydrophobic and flex-
ible, and their sulfonic acid groups are less acidic and therefore also less polar. As a
consequence, narrower channels and a less-connected network of clusters are present
in the microstructure of sulfonated polyetherketones, resulting in a higher depen-
dence of the transport properties on water content due to percolation concepts [27].
A schematic representation of the microstructure of sulfonated polyetherketone com-
pared with that of Nafion® is shown in Fig. 10.6. As illustrated in Fig. 10.7, high con-
ductivity levels are achieved only with a high degree of sulfonation. This results in
low mechanical properties and a high rate of methanol crossover due to excessive
swelling [47].
Several strategies have been used to overcome the excessive swelling of highly sul-
fonated polyarylenes. These include the synthesis of aromatic polymer chains cross-
linked covalently by organic spacers such as α,ω-dihalogenoalkanes [48], the use of
partially fluorinated backbones [31], placing the protogenic groups on short pendant
side chains to increase the separation between the polymer main chains and the sul-
fonic acid groups [49], or building multi-block copolymers using coupling reactions
between hydrophilic and hydrophobic macromonomers [29]. Polymers with pendant
10 Solid polymer proton conducting electrolytes for fuel cells | 215

NAFION Sulfonated polyetherketone (PEEKK)

–(CF2–CF2)n–CF–CF2– O
O–(CF2–CF–O)m–CF2–CF2–SO3H O


1 nm

: –SO3–

: Protonic
: H2O

• Wide channels • Narrow channels

• More separated • Less separated
• Less branched • Highly branched
• Good connectivity • Dead-end channels
• Small –SO3–/–SO3– • Large –SO3–/–SO3–
separation separation
• pKa~–6 • pKa~–1

Fig. 10.6. Schematic representation of the microstructures of Nafion® and a sulfonated polyether-
ketone. Reprinted with permission from [27].

sulfonic acid groups in side chains are in general more stable against hydrolysis than
those with sulfonic acid groups attached directly to the polymer backbone. In addi-
tion, sulfonic acid groups on pendant side chains have a higher degree of freedom,
which results in better phase separation and higher proton conductivity with respect
to random sulfonated analogues [50, 51].
Hydrophilic-hydrophobic multiblock copolymers are considered an interesting
step forward in the rational design of PEMs. The ideal morphology has been pur-
sued by controlling the microphase separation in segmented block copolymers where
hydrophilic sulfonated polymer segments form an interconnected 3D network re-
sponsible for efficient proton transport especially at low relative humidity [52–54],
while a complementary network of hydrophobic non-sulfonated segments causes a
reinforcing effect, preventing excessive swelling in water and enhancing mechanical
216 | Part II Development of new materials for energy applications

Dow T = 300 K

0.10 625
σ/S cm–1


0.05 S-PEK
0 10 20 30 40 50
n = [H2O]/[–SO3H]

Fig. 10.7. Proton conductivity (measured at room-temperature) of two Dow membranes, Nafion® ,
two sulfonated poly(arylene ether ketone)s (SPEK and S-PEEKK), and sulfonated poly(phenoxy-
phosphazene) (S-POP) as a function of the degree of hydration n; the number below the compound
acronym/name indicates the equivalent weight of the ionomer. Reprinted with permission from [47].

properties [55, 56]. Proton and water transport increase significantly with increasing
block length because the longer block induces a more developed phase separation
[57]. However, their synthesis is often complex, thus increasing the materials’ cost.
Overall, synthetic approaches based on structure-property relationships of ionomers
represent a very promising method of obtaining more efficient proton-conducting
membranes with the desired features for fuel cell applications [31, 57, 58].
Anhydrous proton-conducting electrolytes consist of a more or less inert polymer
matrix which is swollen with an appropriate proton solvent, usually phosphoric acid.
These membranes are appealing for fuel cell operation at temperatures well above
100°C without the need for humidification. One of such membrane is poly(2,2󸀠 -m-
(phenylene)-5,5󸀠 -bibenzimidazole) (PBI) whose structure is shown in Fig. 10.8. Non-
modified PBI shows very low proton conductivity. It is therefore necessary to dope the
polymer with sulfuric or phosphoric acid to increase its proton conductivity [59, 60].
However, acid leaching from the membranes and corrosion of cell components are
two of the problems limiting the performance of fuel cell devices based on such
Alternative concepts use amphoteric heterocycles such as imidazole as a proton
conducting species “imbibed” in a polymer matrix. Proton transport occurs through
heterocyclic hydrogen-bonded networks under both anhydrous and low relative hu-
midity conditions. As for sulfonic acid-based ionomers, ion conductivity depends on
the local mobility of the heterocycles within the polymer films and on the effective
concentration of mobile protons in the membranes [12, 60–64]. The proton conductiv-
10 Solid polymer proton conducting electrolytes for fuel cells | 217

ity of these systems increases with the addition of strong acids due to the protonation
of some of the heterocycles within the polymer matrix [64, 65].
Recently, ionic liquids have also been proposed for high temperature proton
conductors, mainly due to their anhydrous high conductivity and good thermal sta-
bility. Nevertheless, the conductivity of ionic liquid-based composite membranes is
lower than that of the original ionic liquids. Therefore, only a few groups have re-
ported demonstrations of the ionic liquid-based solid membrane electrolytes in fuel
cells [66–68].
The composite strategy, where an inorganic phase is dispersed within the iono-
meric host has been demonstrated to be another effective way of improving the trans-
port and mechanical properties of ionomers. Several advantages can be obtained by
using composite membranes, such as: (a) improving the self-humidification of the
membrane at the anode side, (b) suppressing the fuel crossover, e.g., methanol in
DMFC, and (c) improving the mechanical strength of membranes without excessively
sacrificing proton conductivity [39, 40, 69]. Solid inorganic compounds can be classi-
fied as inert hygroscopic fillers, proton conductive fillers, and hydrophilic and proton
conductive fillers. They include: hygroscopic oxides (SiO2 , TiO2 , SnO2 ), clays, zeolites,
heteropoly acids, and zirconium phosphonates [39, 70–73]. For example, (a) hygro-
scopic fillers, e.g., SiO2 , TiO2 , SnO2 , and zeolites improve water retention and dimen-
sional stability of the membranes [74, 75], (b) the operation of fuel cells fed with liquid
fuels was sucessfully extended to high temperature [76, 77], and (c) beta and faujasite
zeolites improved proton conductivity and the DMFC performance of Nafion® [73, 78].
Composite membranes containing exfoliated layered compounds or 1D structures
such as nanotubes or nanorods as fillers are also an effective strategy for improving
relevant properties of electrolytes [76, 79]. The presence of one- and two-dimensional
nanomaterials, which have substantially different properties to those of nanometric
spherical particles, can enhance mechanical strength while acting as a physical bar-
rier to fuel crossover [76, 80, 81]. Performance of composite electrolytic membranes is
in fact strongly related to the polymer/inorganic phase interfacial properties. In detail,
the higher the interface interaction between the polymer and the dispersed particles,
the greater the filler’s influence on the original characteristics of the polymer [82, 83].
Composite membranes are generally prepared by casting the polymer solution
with an inorganic component. The main disadvantage of such composite systems is
related to the fact that it is very difficult to obtain homogenous systems, where the
inorganic particles are well dispersed in the polymeric matrix. Therefore, when appli-
cable, in-situ sol-gel synthesis of the inorganic filler in the hydrophilic clusters of the
PEM is a preferred alternative to the nanocasting method [73, 76].
According to recent reports [76, 83, 84], endeavours to identify the conditions un-
der which inorganic-organic membranes provide properties superior to those shown
by their polymer-only counterparts have been successful, and there is every rea-
son to be optimistic that MEAs based on nanocomposite membranes have a role to
play in liquid feed fuel cells, or in the highly strategic operating conditions of low
218 | Part II Development of new materials for energy applications

RH at 110–130°C. Current hurdles persist: membrane electrical resistance and long-

term durability under fuel cell operation. Surface functionalization of the inorganic
fillers with protogenic groups is being exploited to boost membrane conductivity
[73, 85–87]. However, in-depth studies of aging and degradation under realistic oper-
ating conditions are still needed to enable the synthesis of more advanced materials
and alignment with current targets.

10.3 Characterization of solid polymer electrolytes

The performance of H2 /O2 and liquid feed fuel cells is strongly influenced by the pro-
ton and water transport properties of the PEM. The fuel cell ohmic loss is proportional
to the ionic resistance of the PEM, and high conductivity is essential to assure the
required performance. Water molecules in the membrane increase proton mobility
according to the vehicle mechanism, but a high water uptake by the membrane de-
creases the density of sulfonic acid groups or charge carriers [88]. Therefore, changes
in water content and water mobility have an impact on the proton conductivity of
membranes [78, 88].
This section provides a short description and application of complementary char-
acterization tools (proton conductivity measurements, dynamic vapor sorption, and
differential scanning calorimetry) used to assess transport properties of PEMs. Al-
though these are the first properties to be considered when evaluating PEMs for po-
tential use in fuel cells, it should be stressed that other chemical, morphological,
mechanical, and thermal properties are also critical for the definition of the “ideal”
electrolyte for fuel cell applications and the study of structure–property relationships.
These properties can be studied by means of several characterization techniques in-
cluding bulk chemical analysis and ion exchange capacity, thermal gravimetric anal-
ysis, transmission electron microscopy, small-angle x-ray scattering, tensile tests, dy-
namic mechanical analysis, fuel cell life, and Fenton’s tests [28, 89–91].

10.3.1 Proton conductivity

The proton conductivity of a membrane is determined by measuring its resistance

against the flow of a direct current or an alternative current at controlled temperature
and hydration level. The conductivity σ is calculated by the equation:
10 Solid polymer proton conducting electrolytes for fuel cells | 219

σ= , (10.6)
where l is the distance between the two probe electrodes and S the cross-sectional area
of the membrane.
In the dc method, the potential difference across two probe electrodes in contact
with the membrane follows Ohm’s law over a wide range of current densities and the
resistance can be determined from the slope of the line ΔE vs j. In the ac method,
a periodic small-amplitude ac signal (voltage or current) is applied and the associ-
ated response (current or voltage) coming from the cell is measured [92]. The voltage
response to a sinusoidal current signal is a sinusoid, at the same frequency (ω) but
shifted in phase (φ):

it = i0 sin(ωt) (10.7)
Et = E0 sin(ωt + φ) (10.8)

The impedance, Z, is defined as the ratio of the voltage to the current at a given fre-
E E sin(ωt + φ)
Z= t = 0 (10.9)
it i0 sin(ωt)
By applying Euler’s relationship, the impedance can be expressed as a complex func-
tion with a real and an imaginary part:

Z = Z0 (cos φ + j sin φ) (10.10)

In an electrochemical system, slow kinetic reactions and diffusion of chemical

species can impede the flow of electrons. Electrochemical systems can thus be con-
sidered analogous to the resistors, capacitors, and inductors which hinder the flow
of electrons in an electrical circuit. In the case of a simple resistor, the phase shift is
zero degrees and the current is in phase with the voltage. Thus, according to equa-
tion (10.9), the impedance is purely real and independent of the frequency. For an
ohmic resistance Zt = R. Fig. 10.8 shows typical impedance data in the form of a Bode
plot obtained for a Nafion® 117 membrane using a four-probe cell (see below) at 100%
relative humidity. First, the frequency region over which the impedance has a con-
stant value is identified, and the impedance value taken to calculate the membrane’s
conductivity using equation (10.6) [93].
Electrochemical impedance spectroscopy is the most commonly used method for
measuring the membrane’s resistance and to determine its proton conductivity. It is a
rapid and accurate method and is quite suitable for dielectric materials such as PEMs.
The dc method has also been used [72, 94], the major advantage of this method being
the straightforward analysis of the E-j data.
Conductivity of the membrane can be measured perpendicular to the membrane’s
thickness (through-plane conductivity) or along the plane of the membrane (in-plane
conductivity). In addition, measurements can be made using either the four-probe or
220 | Part II Development of new materials for energy applications

103 8

Phase angle/degree

102 –8
0 1 2 3 4 5
Iog frequency

Fig. 10.8. Impedance data recorded for a Nafion® 117 membrane at 100% RH. Data was acquired in
the in-plane-plane direction using a four-probe cell (reprinted with permission from [93]).

the two-probe method. In fact, there is no standard method for measuring the proton
conductivity of ionomers, and each method/cell configuration has its own advantages
and disadvantages. Figure 10.9 illustrates some of the conductivity cells reported in the
During fuel cell operation, protons move through the cross-section of the mem-
brane. Thus, measurements made in this direction are more relevant for practical ap-
plications. However, in this configuration the area of the electrodes (≈ cm2 ) is much
larger than the distance between them (given by the membrane thickness, ≈ microme-
ters), so the cell constant (l/S) is small and the contribution from the interface formed
between the membrane and the electrodes is large. On the contrary, for in-plane mea-
surements the cell constant is larger since the distance between the electrodes is of the
order of mm to cm, and the section of interest is the cross-section of the membrane.
The bulk conductivity of the membrane is the dominant element contributing to the
measurement [95]. Conductivity measurements in both directions are nevertheless im-
portant to quantify the effect of morphological anisotropy of PEMs on their proton
conductivity [95–97].
In the two-probe method, the voltage drop is measured across the same two elec-
trodes where the current flows. Accordingly, the measured impedance (or resistance)
includes the contribution of all components on the current pathway. When determin-
ing, for example, the membrane resistance from the total cell impedance, all other
contributions such as electrode resistance, lead inductance, and membrane – elec-
trode contact resistance should be subtracted from the total cell impedance [95, 98].
This is done by recording the impedance of the short-circuited and open cells [95].
With the four-probe method, only the bulk membrane resistance is measured because
two distinct pairs of electrodes are used, and current flow and voltage sensing are
done independently. The current is imposed on the external pair and the voltage drop
10 Solid polymer proton conducting electrolytes for fuel cells | 221

Teflon block Teflon block

Contact electrode
Contact electrode PEM
PEM Contact electrode
Teflon block
Teflon block

Contact electrode
(a) (b)

Waveform generator


I High impedance I

Δ Eref
Cation exchange

8 cm

Saturated calomel
L reference electrode
Platinum gauze Luggin
counter electrode capillary Silicone-rubber
(1.5 cm2) gaskets

(d) 20 cm

Fig. 10.9. (a) In-plane and (b) through-plane two-probe conductivity cells; (c) In-plane and (d) four-
probe conductivity cells. (a) and (b) reprinted with permission from [94] and (d) reprinted with per-
mission from [95].

along the membrane sectional area is measured using the central pair of electrodes.
The effect of contact resistance is clearly seen in Fig. 10.10, which shows the variation
of the in-plane proton conductivity of a Nafion® 112 membrane exposed to 95% RH
and immersed in liquid water, as a function of the torque applied to the electrodes,
for a four-probe and a two-probe configuration. A lower interface resistance between
222 | Part II Development of new materials for energy applications

Four probe method at RH 95%
Two probe method at RH 95%
Four probe method in liquid water
1.5x10–1 Two probe method in liquid water
Proton conductivity (S/cm)



0.4 0.6 0.8 1.0 1.2 1.4 1.6

Torque (kgfcm)

Fig. 10.10. Effect of applied torque on the measured proton conductivity of Nafion® 112 by four-
probe (◼) and two-probe (∙) configurations at 95% RH and 60°C, and by four-probe (◻) and two-
probe (∘) methods in the liquid-water state at 60°C. Reprinted with permission from [98].

membrane and electrodes corresponds to a higher torque, and the influence of this ad-
ditional resistance is more important when the conductivity of the membrane is lower
(less hydrated membrane) [98].
Measurements of ion conductivity over a wide range of temperature and relative
humidity are important to determine the effect of composition and structure of the
new ionomers on the proton conduction and operational temperature. An example is
given in Fig. 10.11 which illustrates the effect of the filler content and composition on
the proton conductivity of Nafion® composite membranes containing TiO2 and propyl
sulfonic acid functionalized TiO2 nanoparticles [87]. The loading of an appropriate
amount of propylsulfonic-functionalized titania allows the preparation of Nafion® -
based composite membranes with higher conductivity and dimensional stability than
pristine Nafion® up to 140°C.

10.3.2 States of water and water mobility

Two critical parameters affecting the performance and proton conduction mechanism
of PEMs are their hydration level and water diffusion coefficient as a function of the
water content. It is therefore important to study the water sorption and diffusion be-
havior of electrolytes over a wide range of relative humidity.
10 Solid polymer proton conducting electrolytes for fuel cells | 223

140 120 100 80 60 40



log (σ/S cm–1)


–1.8 N_RC
–2.0 N_20TiO2-RSO3H

2.4 2.5 2.6 2.7 2.8 2.9 3.0 3.1 3.2 3.3
1000/T (K–1)

Fig. 10.11. Arrhenius plot of Nafion® , Nafion® -TiO2, and Nafion® propyl sulfonic acid functional-
ized TiO2 at 100% RH. The numbers in the legend indicate the wt% of filler with respect to Nafion® .
Reprinted with permission from [87].

Dynamic vapor sorption

Dynamic Vapor Sorption (DVS) is a gravimetric technique which allows fast and ac-
curate determination of vapor sorption isotherms and diffusion kinetics. A simplified
scheme of the DVS apparatus is shown in Fig. 10.12(a) the samples are placed in a
weighing pan and exposed to partial pressure- and temperature-controlled environ-
ment. In order to study water management in solid electrolytes, water is used as sor-
bate and the electrolyte as sorbent. The vapor partial pressure around the sample is
controlled by mixing saturated and dry carrier gas steams using electronic mass flow
controllers. A constant temperature is maintained by enclosing the entire system in a
temperature-controlled incubator. By measuring the change in mass as a function of
time to equilibrium, a typical diagram such as that shown in Fig. 10.12(b) is obtained.
The amount of water uptake (WU) by the sample exposed to a defined partial pres-
sure can thus be obtained using equation (10.11):

m(eq)aw − mdry
WUaw = , (10.11)

where m(eq)aw is the mass of the sample at equilibrium for a defined water activity
(aw ), and mdry is the dry mass of the sample.
A sorption isotherm is the graphic representation of WU values; it describes the
relationship between the water content of the electrolyte and water activity at con-
stant temperature. Water is a small molecule and a polar adsorptive, therefore its
adsorption mechanism is influenced by water affinity to the adsorbent surface. The
224 | Part II Development of new materials for energy applications

Temperature Microbalance
controlled module
Humidity regulated
sample module
Flow control

Dry gas

module Vapour
Sample reference

Camera option


Target aw 1.0

Water activity (aw=P/P0 )

7.00 Actual aw

6.50 Mass 0.6

6.25 0.4
0 500 1000 1500 2000
(b) t/min

Fig. 10.12. (a) Illustration of the DVS apparatus interfaced with a personal computer (reprinted with
permission from Surface Measurements Systems), and (b) kinetics of water adsorption of a typical
Nafion® membrane at 25°C and different partial pressures.

shape of the isotherm thus also reflects the hydrophilicity/hydrophobicity of the sur-
face. IUPAC proposed a classification for water sorption isotherms as illustrated in
Fig. 10.13 [99, 100].
Each isotherm shape is related to a material with specific hydrophilic characteris-
tics. Type I is characteristic of every hydrophilic material. Type II and type IV isotherms
are characteristic of moderate hydrophilic materials. Adsorbents showing a type IV
isotherm are hydrophilic as well. Adsorbents with low hydrophilicity will give rise to
type III and type V isotherms. Type VI is typical of a hydrophilic material with mul-
tiple sorbent–water interactions and stepwise sorption, while type VII isotherms are
characteristic of very hydrophobic materials.
10 Solid polymer proton conducting electrolytes for fuel cells | 225

Hydrophilic materials


0 P/PO 0 P/PO 0 P/PO 0 P/PO

Hydrophobic materials


0 P/PO 0 P/PO 0 P/PO

Fig. 10.13. IUPAC classification of adsorption isotherms for materials with different hydrophilicity.
Adapted with permission from [100].

Determination of the diffusion coefficient from DVS measurements

DVS measurements allow evaluation of the water diffusion coefficient through elec-
trolyte materials. Assuming that water sorption can be described by fickian behavior,
the water diffusion coefficient (D) can be calculated from the relationship between
mass variation and the time of water vapor exposure up to equilibrium [101]. This re-
lationship is obtained by combining Fick’s first law (10.12; describing the transfer of
solute atoms per unit area in a one-dimensional flow) and the conservation of mass
relationship (10.13) and expressed by Fick’s second law (10.14):
J = −D (10.12)
𝜕C 𝜕J
=− (10.13)
𝜕t 𝜕x
𝜕C 𝜕2 J
= −D 2 , (10.14)
𝜕t 𝜕x
where J is the amount of substance flowing per unit area as a function of time, C is
the concentration, and x is the position. Assuming constant diffusivity and that wa-
ter activity is constant across the membrane/vapor interface (c = c∞ at x ± d/2), solv-
ing (10.14) gives the normalized mass change as a function of the time:

Mt 4 D⋅t
= √ , (10.15)
M∞ d π
where Mt is the amount of water adsorbed at time t, M∞ is the amount of water ad-
sorbed at equilibrium, and d is the sample thickness.
D can be obtained by plotting Mt /M∞ for a sample exposed to a certain partial
pressure P/P0 (i.e., water activity) as a function of the square root of time (Fig. 10.14)
226 | Part II Development of new materials for energy applications


M(t)/M (∞)


Equation y = a + bx
Adj. R-Squar 0.9976
0.2 Value Standard error
Intercept –0.04147 0.01338
Slope 0.02505 0.00121

0 20 40 60 80 100 120 140

t1/2 / S1/2

Fig. 10.14. Typical plot of Mt /M∞ versus t1/2 at a given value of water activity (aw ). Adapted with
permission from [78].

and by fitting the curve to equation (10.15). This equation is valid for values of Mt /M∞ <
0.4, where the plot of Mt /M∞ against t1/2 is linear [102, 103].
Figure 10.15 shows the water diffusion coefficient of a Nafion® membrane as a
function of water activity, measured by DVS at 25°C. As shown, D increases with wa-
ter content in the membrane at low aw and reaches a maximum in the 0.3 to 0.4 aw
range. The increase in D in this aw range is due to the fact that water is less tightly as-
sociated with the sulfonic acid sites of Nafion® as water content increases. At higher
water activities, D decreases with increasing aw due to the water aggregation process
which occurs and provides kinetic limitations on the adsorption of water on the poly-
mer matrix [78].

Determination of the different states of water

It is also possible to obtain information about water mobility (and consequently
proton transport) from water sorption measurements by investigating the state of the
water in the electrolytes. In fact, specifically designed models can be applied to the
sorption isotherms in order to obtain insights into the water transport properties of
For instance, conventional dual mode sorption models (Langmuir-type) are effec-
tive in describing isotherms which are concave towards the activity axis, while the
engaged species induced clustering model (Flory-type) has been highly successful in
modeling isotherms in polymers which are convex to the aw axis [104]. Multimode
sorption models (Park-type) are particularly suited to sigmoidal isotherms, which are
the most common isotherm shapes among ionomers [105].
10 Solid polymer proton conducting electrolytes for fuel cells | 227








0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Fig. 10.15. Water diffusion coefficient (D) values of a recast Nafion® membrane as a function of water
activity at 25°C. Adapted with permission from [78].

By applying the multi-mode model proposed by Park to the sorption isotherms, the
presence of three different mechanisms in the sorption process can be hypothesized:
(a) specific adsorption at low water activity, described by the Langmuir model;
(b) nonspecific adsorption, described by Henry’s law;
(c) water clustering at high water activity.

All of these contributions can be formulated in the following equation:

aL KL aW
WU = + KH aw + nKA anW , (10.16)
1 + KL aW

where aL is the specific site capacity, KL is an affinity constant, KH is Henry’s law co-
efficient, KA is the aggregation equilibrium constant, and n is the aggregate size.
A distinct population of water adsorbed in the membrane can be associated to
each adsorption mechanism: specific adsorbed water (WSA ), nonspecific adsorbed wa-
ter (WNSA ), and clustered water (WC ). Each water population is described by the terms
constituting equation (10.16) as follows:
aL KL aW
WSA = (10.17)
1 + KL aW
WNSA = KH aw (10.18)
WC = nKA anW (10.19)

Figure 10.16 depicts the typical result of the curve fitting of a polymer membrane sorp-
tion isotherm, where the adsorbed water was separated into three contributions so
that the sum of WSA , WNSA , and WC matched the experimental isotherm data.
228 | Part II Development of new materials for energy applications

Taking into account that each type of adsorbed water is characterized by different
mobility, we correlated the different water population to the water mobility degree in
the membrane.
Being strongly bound to specific sites, the specific adsorbed water is character-
ized by low mobility, whereas the dissolved water molecules (Henry population) have
higher mobility. The growth of water clusters then reduces the mobility of the water
aggregates. As a consequence, among the three types of water population, the non-
specific adsorbed water is characterized by the highest mobility.
The amount of each type of adsorbed water was normalized to the total water con-
tent in the membranes as follows:
θW [SA] = × 100 (10.20)

θW [NSA] = × 100 (10.21)
θW [C] = × 100 (10.22)
As θ parameters were defined, θW [SA], θW [NSA], and θW [C] represent the “specific ad-
sorbed water degree”, “nonspecific adsorbed water degree”, and “clustered water de-
gree”, respectively. As shown in Fig. 10.16(b), specific adsorbed water dominates at
low relative humidity, nonspecific adsorbed water at intermediate values of RH, and
clustered water dominates at high relative humidity. These variations are consistent
with those found for D and shown in Fig. 10.15. The θW [NSA] parameter thus represents
the “water mobility degree” and allows comparison of different electrolytes in terms
of water mobility: the higher θW [NSA], the greater the water mobility in the electrolyte
is expected to be [72, 78].
As already mentioned, the analysis of water sorption isotherms of ionomers is
of paramount importance for the final fuel cell performance and scientific litera-
ture in this field is mainly based on adsorption properties of water on perfluorinated
polymers, in particular Nafion® , which is the state-of-the-art material [106]. Sorp-
tion isotherms of most common perfluorinated ionomers can be indeed described
by Park’s model. However, the involvement of five adjustable parameters (see equa-
tion (10.16)) makes the chemico-physical interpretation unclear for ionomers whose
microstructure is considerably different from that of Nafion® , which is the case for
polyaromatic polymers (see Fig. 10.6). The less pronounced hydrophobic/hydrophilic
separation in polyaromatic polymers compared to Nafion® makes the distinction be-
tween specific adsorbed water, nonspecific adsorbed water and clustered water (WC )
quite difficult.
As an alternative to the multimode Park’s model, the sorption behavior of the
membranes can be analyzed and interpreted on the basis of the dual mode sorption
model proposed by Feng [107]. The model is based on the Guggenheim–Anderson–
10 Solid polymer proton conducting electrolytes for fuel cells | 229

Equation: Adj. R-Square: 0.99828
Park model Red. Chi-Square: 0.0614
Value Standard error
Sample N_0_HP AL 2.89 0.418
KL 5.85 0.423
KH 9.38 0.729

15 n 10.1 0.87
KA 1.43 0.125

10 Experimental data
Curve fitting


0.0 0.2 0.4 0.6 0.8 1.0

(a) aw

70 θNSA






0.0 0.2 0.4 0.6 0.8 1.0

(b) aw

Fig. 10.16. (a) Typical curve fitting (Park’s model) of experimental sorption isotherm data (Nafion®
membrane at T = 25°C) and the corresponding fitting parameters, (b) variation of the three types of
water population in the membrane with the water activity. Adapted with permission from [78].

de Boer (GAB) multilayer sorption theory [108–110] and, at variance with the GAB
model which considers all sorption sites equivalent, the Feng model is based on the
assumption that the sorption sites can be divided into two different types, one be-
ing the polymer matrix region and the other the microvoid region (specific sorption
230 | Part II Development of new materials for energy applications

According to this model, the water content in the membranes can be described
using equation (10.23):

k󸀠 aw (A󸀠 − 1)k󸀠 aw
WU = Cp 󸀠
+ Cp , (10.23)
1 − k aw 1 + (A󸀠 − 1)k󸀠 aw

where Cp is the weighted mean value of the polymer sorption capacity, k󸀠 and A󸀠 are
temperature-dependent constants, k󸀠 provides a measure of the interaction between
water and the polymer matrix. Values lower than 1 indicate very weak interactions be-
tween water and polymer matrix. The higher the k󸀠 value, the greater the hydrophilic-
ity of the polymer. A󸀠 represents the difference between the interaction of a microvoid
and its first molecule adsorbed, and that of a microvoid and the subsequent adsorbed
molecules. Thus it provides a measure of the affinity between water and the polymer
microvoid. A󸀠 values close to 1 correspond to a polymer in a rubbery state without
microvoids; the higher A󸀠 is, the greater the dependence of sorption on microvoids
and affinity of specific sites to water.
Feng’s model requires only three parameters: Cp, the weighted mean value of the
sorption capacity of the polymer to water, k󸀠 , the affinity between water and the poly-
mer matrix (hydrophobic region), and A󸀠 , the affinity between water and the polymer
microvoid (hydrophilic domains). Rather than discriminating between the differ-
ent states of water in the membrane, the values and comparison of Feng’s parame-
ters make deep insight into chemical nature, and also into polymer microstructure
Figure 10.17 shows the water adsorption isotherms of Nafion® and sulfonated
polysulfone (SPS) with the result of a typical curve fitting Feng’s model. The figure
shows the very good match between the experimental data and the fit curves, and the
corresponding fitting parameters of all samples are summarized in the inset table.
Cp, k󸀠 , and A󸀠 parameters of the unfilled Nafion® membranes are in good agreement
with values found in previous papers. Both polymers showed low k󸀠 values, indicating
that sorption in the polymer matrix region is negligible. Hence, microvoid sorption is
predominant, as expected in the case of ionomer systems in which water associates
through the sulfonic acid groups, and Cp represents the monolayer sorption capacity
in the microvoid region (specific adsorption) [83, 111].
The comparison between the fitting parameters of unfilled Nafion® and SPS indi-
cated that Cp was higher for SPS than for Nafion® , whereas the A󸀠 parameter shows
the opposite trend. Differences in Cp and A󸀠 for Nafion® and SPS are ascribed to dif-
ferences in the microstructures of the two ionomers.
Both Nafion® and SPS phases separate in hydrophilic and hydrophobic domains.
The hydrophobic domains consist of the perfluorinated and polyaromatic backbone
of Nafion® and SPS respectively. The hydrophilic domains arise from the sulfonic acid
groups (–SO3 H) which are responsible for bonding with water molecules. Hydropho-
bic/hydrophilic separation is more pronounced in the case of Nafion® , as depicted
in Fig. 10.6. The greater tortuosity of the hydrophilic domains in SPS than in Nafion®
10 Solid polymer proton conducting electrolytes for fuel cells | 231

Sample Parameters Value SE
Nafion Cp 3.57 0.07
30 SPS
Reduced Chi-Sqr k' 0.88 0.03
25 Adj. R-Square A' 15.8 1.8
S Cp 7.46 0.37
Reduced Chi-Sqr k' 0.83 0.01

15 Adj. R-Square A' 5.32 0.42

0.0 0.2 0.4 0.6 0.8 1.0


Fig. 10.17. Curve fitting of experimental adsorption isotherm data of the unfilled Nafion® and SPS
membranes at T = 25°C. SE = Standard Error. Adapted with permission from [83].

makes the water phase in SPS lower interconnected than in Nafion® , thus explaining
the higher Cp and lower A󸀠 values of SPS compared to those of Nafion® .

Differential scanning calorimetry

Differential Scanning Calorimetry (DSC) is a thermoanalytical technique which mon-
itors heat effects associated with phase transitions and chemical reactions as a func-
tion of temperature [112–114]. It consists of measuring the difference in heat flow be-
tween the sample and a reference at the same temperature, the temperature of both
sample and reference being increased at a constant rate. The heat flow difference be-
tween the sample and the reference can be either positive or negative, depending on
whether the process is endothermic or exothermic. The result of a DSC experiment is
a curve of heat flux versus temperature or time. The area enclosed between the trend
line and the base line is a direct measurement of the amount of heat, ΔH, needed for
transformation. Useful information can be obtained by DSC analysis of polymer sam-
ples, such as degree of crystallinity (from the ratio of the heat of fusion of a polymer
sample and the enthalpy of a 100% crystalline sample), specific heat, the purity of the
polymer and occurrence of oxidation, cross-linking, and chain breakage.
As far as water management of electrolytes is concerned, DSC provides informa-
tion on states of water and water mobility through the electrolyte material. Focusing
on ionomer electrolytes, three different categories of water can be discerned by record-
ing DSC thermograms at subzero temperatures:
232 | Part II Development of new materials for energy applications

(1) nonfreezable bound water (WNF ), strongly bound to the ionic groups present in the
polymer. This type of water is characterized by the fact that it does not crystallize
even when the swollen sample is cooled down to −100°C. These water molecules
are in close proximity to an ionic group as in hydration shells, are highly polarized,
and are unable to crystallize. WNF does not yield characteristic thermal transition
in DSC analysis.
(2) Freezable bound water, weakly polarized. This type of water crystallizes at tem-
peratures below than 0°C.
(3) Freezable unbound water, which crystallizes at 0°C.

Freezable water (WF ), being more loosely bound, has higher mobility than nonfreez-
able water and is expected to give a more significant contribution to the proton trans-
port mechanism.
By performing DSC analysis in the range between −50°C and 10°C, freezable water
can be quantified from the endothermic peak below 0°C. An example of DSC thermo-
grams obtained from two different polymer electrolyte membranes showing an en-
dothermic peak ascribed to the melting of freezable water is given in Fig. 10.18.


Heat flow/Wg-1

–0.6 (b)


ENDO Fig. 10.18. DSC thermogram of
(a) (a) an unfilled Nafion® membrane,
–0.0 and (b) a composite Nafion® /
–40 –35 –30 –25 –20 –15 –10 –5 0 5 10 zeolite membrane. Reprinted with
T/°C permission from [78].

The percentage of freezable water in the sample can obtained from the following for-
A 1
WF (%) = ( ) × 100, (10.24)
ΔHW mdry
where A is the area of the endothermic peak, ΔHw is the enthalpy of melting for bulk
water (333 J g−1 ), and mdry is the mass of the dried sample.
The degree of freezable water, θF , can be defined by normalizing the freezable
water content to the total WU, which can be measured gravimetrically (for instance
by DVS).
θF = F × 100 (10.25)
10 Solid polymer proton conducting electrolytes for fuel cells | 233

As previously mentioned, a higher degree of mobile water corresponds to higher

proton conductivity, and since WF yields to thermal transitions similar to bulk water,
its content in the membrane can be discerned from total WU by using DSC [72, 78].
Figure 10.19 shows the variation of θF with the filler content for Nafion® –zeolite
composite membranes. Zeolites are aluminosilicates with cations relatively free to
move along the cavities of the framework. Moreover, they have a very high specific
surface area which results in a high water sorption capacity, further facilitating the ion
transport. As the zeolite content increases, qF values increase up to a maximum value
and then decrease at highest zeolite content. These findings indicate that the zeolite
likely contributes to the enhancement of the water mobility degree in the composite
membrane, which is related to its high water sorption capacity and to the introduction
of porosities at the polymer/filler interface. However, the reduction of this effect over
ca. 4 wt% zeolite content, suggests the formation of dead-end porosities which hinder
water mobility [78].



θF /%





0 5 10 15 20
Zeolite content/wt.%

Fig. 10.19. Variation of θF as a function of the zeolite content for Nafion® –Faujasite composite mem-
branes. Adapted with permission from [78].

10.4 Summary

Solid electrolytes are materials capable of conducting ions. They are used in many
electrochemical devices including batteries, sensors, electrolyzers, and fuel cells.
Proton exchange membrane fuel cells are considered attractive power sources for
portable applications, in-situ power generation, and for automotives. Nevertheless,
these systems still suffer from limitations which need to be addressed before they
can compete with batteries, fossil fuels, and internal combustion engines. Polymer
electrolyte membranes are one of the limiting elements of this technology. Nafion® ,
a perfluorinated sulphonic acid ionomer, is the most widely used electrolyte for both
234 | Part II Development of new materials for energy applications

hydrogen and liquid-fed proton exchange membrane fuel cells due to its high pro-
ton conductivity, chemical and mechanical stability. A unique feature of Nafion® is
the microphase separation between the hydrophobic backbone and the hydrated sul-
fonic acid domains, resulting in the formation of wide and well-separated water chan-
nels for proton transport. Nafion® membranes show a strong dependence of proton
conductivity on the membrane’s hydration level and are permeable to liquid fuels. In
the first case, the fuel cell system needs an expensive humidification auxiliary system
to keep the membranes hydrated. In the second, fuel cell efficiency is dramatically
reduced. Therefore the development of alternative polymer electrolyte membranes
with high proton conductivity in a wide range of temperature and hydration condi-
tions, which exhibit mechanical robustness, chemical and electrochemical stability,
low cost, and low fuel permeability remains a critical challenge for advancing fuel cell
Hydrocarbon membranes are potential candidates to replace Nafion® . Significant
efforts are being made to develop novel ionomers consisting of hydrocarbon back-
bones and pending side chains with terminal sulfonic acid groups to mimic Nafion® ’s
unique morphology. Structure–properties relationships are fundamental to learn
about the dependence of the transport properties on the membranes’ composition,
morphology, and water content, and to design better electrolytes.
Proton conductivity is a fundamental property of a proton exchange membrane.
When evaluating potential electrolytes for fuel cells, their proton conductivity is usu-
ally measured under controlled temperature and relative humidity. There is not yet
a standard method for measuring the membranes’ proton conductivity, but measure-
ments on both directions of the membrane (in-plane and through-plane) could pro-
vide valuable information on the membranes’ anisotropy. Proton conductivity of elec-
trolytes depends on their hydration level; hence it is important to study water sorption
and water diffusion over a wide range of relative humidity. The water states and water
diffusion in electrolytes can be assessed by dynamic vapor sorption and differential
scanning calorimetry. Excellent correlation has been found between proton conduc-
tivity and degree of mobile water determined by the two methods mentioned.
Advances in proton exchange membranes for fuel cells will likely contribute to the
development of other related fields including electrodialysis (water purification and
treatment) and redox flow batteries for energy conversion.


The authors would like to thank Maria J.V.R. Paulo (INRS-EMT) for helping with the
editing of this chapter.
10 Solid polymer proton conducting electrolytes for fuel cells | 235


[1] Hayashi A, Noi K, Sakuda A, Tatsumisago M. Superionic glass-ceramic electrolytes for room-
temperature rechargeable sodium batteries. Nat Commun. 3 (2012) 856.
[2] Tarascon JM, Armand M. Issues and challenges facing rechargeable lithium batteries. Nature.
414 (2001) 359–67.
[3] Hickner MA. Ion-containing polymers: new energy &amp; clean water. Materials Today. 13
(2010) 34–41.
[4] Carrette L, Friedrich KA, Stimming U. Fuel Cells – Fundamentals and Applications. Fuel Cells.
1 (2001) 5–39.
[5] Chroneos A, Yildiz B, Tarancon A, Parfitt D, Kilner JA. Oxygen diffusion in solid oxide fuel cell
cathode and electrolyte materials: mechanistic insights from atomistic simulations. Energy &
Environmental Science. 4 (2011) 2774–89.
[6] Brett DJL, Atkinson A, Brandon NP, Skinner SJ. Intermediate temperature solid oxide fuel cells.
Chemical Society Reviews. 37 (2008) 1568–78.
[7] Arico AS, Bruce P, Scrosati B, Tarascon J-M, van Schalkwijk W. Nanostructured materials for
advanced energy conversion and storage devices. Nat Mater. 4 (2005) 366–77.
[8] Goodenough JB. Ceramic solid electrolytes. Solid State Ionics. 94 (1997) 17–25.
[9] Passalacqua E, Lufrano F, Squadrito G, Patti A, Giorgi L. Nafion® content in the catalyst layer
of polymer electrolyte fuel cells: effects on structure and performance. Electrochimica Acta.
46 (2001) 799–805.
[10] Sata T. Ion Exchange Membranes: Preparation, Characterization, Modification and Applica-
tion, Royal Society of Chemistry; Oxford, UK, 2004.
[11] Hickner MA. Water-Mediated Transport in Ion-Containing Polymers. Journal of Polymer
Science Part B: Polymer Physics. 50 (2011) 9–20.
[12] Schuster M, Rager T, Noda A, Kreuer KD, Maier J. About the Choice of the Protogenic Group in
PEM Separator Materials for Intermediate Temperature, Low Humidity Operation: A Critical
Comparison of Sulfonic Acid, Phosphonic Acid and Imidazole Functionalized Model Com-
pounds. Fuel Cells. 5 (2005) 355–65.
[13] Eikerling M, Kornyshev A, Spohr E. Proton-Conducting Polymer Electrolyte Membranes: Water
and Structure in Charge. Advances in Polymer Science. 215 (2008) 15–54.
[14] Kreuer K-D. Proton Conductivity: Materials and Applications. Chemistry of Materials. 8 (1996)
[15] Zawodzinski TA, Neeman M, Sillerud LO, Gottesfeld S. Determination of water diffusion
coefficients in perfluorosulfonate ionomeric membranes. The Journal of Physical Chemistry.
95 (1991) 6040–4.
[16] Eikerling M, Kornyshev AA. Proton transfer in a single pore of a polymer electrolyte mem-
brane. Journal of Electroanalytical Chemistry. 502 (2001) 1–14.
[17] Choi P, Jalani NH, Datta R. Thermodynamics and Proton Transport in Nafion® : II. Proton Dif-
fusion Mechanisms and Conductivity. Journal of The Electrochemical Society. 152 (2005)
[18] Alberti G, Casciola M. Solid state protonic conductors, present main applications and future
prospects. Solid State Ionics. 145 (2001) 3–16.
[19] McLean RS, Doyle M, Sauer BB. High-Resolution Imaging of Ionic Domains and Crystal Mor-
phology in Ionomers Using AFM Techniques. Macromolecules.33 (2000) 6541–50.
[20] Mauritz K A, Hora C J, Hopfinger A J. Theoretical Model for the Structures of Ionomers: Appli-
cation to Nafion® Materials. Ions in Polymers: American Chemical Society. 1980: 123–44.
[21] Yeager HL, Steck A. Cation and Water Diffusion in Nafion® Ion Exchange Membranes: Influ-
ence of Polymer Structure. Journal of The Electrochemical Society. 128 (1981) 1880–4.
236 | Part II Development of new materials for energy applications

[22] Hsu WY, Gierke TD. Elastic theory for ionic clustering in perfluorinated ionomers. Macro-
molecules. 15 (1982) 101–5.
[23] Fimrite J, Struchtrup H, Djilali N. Transport Phenomena in Polymer Electrolyte Membranes: I.
Modeling Framework. Journal of The Electrochemical Society. 152 (2005) A1804–A14.
[24] Elliott JA, Hanna S, Elliott AMS, Cooley GE. Atomistic simulation and molecular dynamics
of model systems for perfluorinated ionomer membranes. Physical Chemistry Chemical
Physics.1 (1999) 4855–63.
[25] Li Q, He R, Jensen JO, Bjerrum NJ. Approaches and Recent Development of Polymer Elec-
trolyte Membranes for Fuel Cells Operating above 100°C. Chemistry of Materials. 15 (2003)
[26] Casciola M, Alberti G, Sganappa M, Narducci R. On the decay of Nafion® proton conductivity
at high temperature and relative humidity. Journal of Power Sources. 162 (2006) 141–5.
[27] Kreuer KD. On the development of proton conducting polymer membranes for hydrogen and
methanol fuel cells. Journal of Membrane Science. 185 (2001) 29–39.
[28] Peighambardoust SJ, Rowshanzamir S, Amjadi M. Review of the proton exchange membranes
for fuel cell applications. International Journal of Hydrogen Energy. 35 (2010) 9349–84.
[29] Zhang H, Shen PK. Recent Development of Polymer Electrolyte Membranes for Fuel Cells.
Chemical Reviews. 112 (2012) 2780–832.
[30] Di Noto V, Zawodzinski TA, Herring AM, Giffin GA, Negro E, Lavina S. Polymer electrolytes for a
hydrogen economy. International Journal of Hydrogen Energy. 37 (2012) 6120–31.
[31] Peckham TJ, Holdcroft S. Structure-Morphology-Property Relationships of Non-Perfluorinated
Proton-Conducting Membranes. Advanced Materials. 22 (2010) 4667–90.
[32] Alkan Gürsel S, Gubler L, Gupta B, Scherer G. Radiation Grafted Membranes. Advances in
Polymer Science. 215 (2008) 157–217.
[33] Yang Y, Siu A, Peckham T, Holdcroft S. Structural and Morphological Features of Acid-Bearing
Polymers for PEM Fuel Cells. Advances in Polymer Science. 215 (2008) 55–126.
[34] Wang F, Hickner M, Kim YS, Zawodzinski TA, McGrath JE. Direct polymerization of sulfonated
poly(arylene ether sulfone) random (statistical) copolymers: candidates for new proton ex-
change membranes. Journal of Membrane Science. 197 (2002) 231–42.
[35] Jannasch P. Recent developments in high-temperature proton conducting polymer electrolyte
membranes. Current Opinion in Colloid & Interface Science. 8 (2003) 96–102.
[36] Harrison WL, Hickner MA, Kim YS, McGrath JE. Poly(Arylene Ether Sulfone) Copolymers and
Related Systems from Disulfonated Monomer Building Blocks: Synthesis, Characterization,
and Performance – A Topical Review. Fuel Cells. 5 (2005) 201–12.
[37] Daigle JC, Dube-Savoie V, Tavares AC, Claverie JP. Copolymers of ethylene and sulfonated
norbornene for proton exchange membranes. Journal of Polymer Science Part A: Polymer
Chemistry. 51 (2013) 2669–76.
[38] Zaopo A, Lopes Correia Tavares AB, Ballabio O, inventors; PIRELLI & C SPA, Italy, assignee.
Fuel Cell and Polymer Electrolyte Membrane ; (2007) WO2007128330 A1; Application date:
2006-05-05; EP 2025031 B1 (2011).
[39] Herring AM. Inorganic–Polymer Composite Membranes for Proton Exchange Membrane Fuel
Cells. Journal of Macromolecular Science, Part C. 46 (2006) 245–96.
[40] Laberty-Robert C, Valle K, Pereira F, Sanchez C. Design and properties of functional hybrid
organic-inorganic membranes for fuel cells. Chemical Society Reviews. 40 (2011) 961–1005.
[41] Fu T, Cui Z, Zhong S, et al. Sulfonated poly(ether ether ketone)/clay-SO3H hybrid proton ex-
change membranes for direct methanol fuel cells. Journal of Power Sources. 185 (2008) 32–9.
[42] Kerres J. Covalent-Ionically Cross-linked Poly(Etheretherketone)-Basic Polysulfone Blend
Ionomer Membranes. Fuel Cells. 6 (2006) 251–60.
10 Solid polymer proton conducting electrolytes for fuel cells | 237

[43] Lafitte B, Karlsson LE, Jannasch P. Sulfophenylation of Polysulfones for Proton-Conducting

Fuel Cell Membranes. Macromolecular Rapid Communications. 23 (2002) 896–900.
[44] Tavares AC, Pedicini R, Gatto I, Dubitsky YA, Zaopo A, Passalacqua E. New Sulfonated Poly-
sulfone Co-Polymer Membrane for Low Temperature Fuel Cells. Journal of New Materials for
Electrochemical Systems. 6 (2003) 211–6.
[45] de Bonis C, D’Epifanio A, Di Vona ML, et al. Proton Conducting Hybrid Membranes Based on
Aromatic Polymers Blends for Direct Methanol Fuel Cell Applications. Fuel Cells. 9 (2009)
[46] Rozière J, Jones DJ. Non-fluorinated polymer materials for proton exchange membrane fuel
cells. Annual Review of Materials Research. 33 (2003) 503–55.
[47] Kreuer K-D, Paddison SJ, Spohr E, Schuster M. Transport in Proton Conductors for Fuel-Cell
Applications: Simulations, Elementary Reactions, and Phenomenology. Chemical Reviews.
104 (2004) 4637–78.
[48] Zhang W, Gogel V, Friedrich KA, Kerres J. Novel covalently cross-linked
poly(etheretherketone) ionomer membranes. Journal of Power Sources. 155 (2006) 3–12.
[49] Karlsson LE, Jannasch P. Polysulfone ionomers for proton-conducting fuel cell membranes:
2. Sulfophenylated polysulfones and polyphenylsulfones. Electrochimica Acta. 50 (2005)
[50] Pang J, Zhang H, Li X, Jiang Z. Novel Wholly Aromatic Sulfonated Poly(arylene ether) Copoly-
mers Containing Sulfonic Acid Groups on the Pendants for Proton Exchange Membrane Mate-
rials. Macromolecules. 40 (2007) 9435–42.
[51] Lafitte B, Jannasch P. Proton-Conducting Aromatic Polymers Carrying Hypersulfonated Side
Chains for Fuel Cell Applications. Advanced Functional Materials. 17 (2007) 2823–34.
[52] Markova D, Kumar A, Klapper M, Müllen K. Phosphonic acid-containing homo-, AB and BAB
block copolymers via ATRP designed for fuel cell applications. Polymer. 50 (2009) 3411–21.
[53] Lee H-S, Badami AS, Roy A, McGrath JE. Segmented sulfonated poly(arylene ether sulfone)-
b-polyimide copolymers for proton exchange membrane fuel cells. I. Copolymer synthesis
and fundamental properties. Journal of Polymer Science Part A: Polymer Chemistry. 45 (2007)
[54] Nakabayashi K, Matsumoto K, Ueda M. Synthesis and properties of sulfonated multiblock
copoly(ether sulfone)s by a chain extender. Journal of Polymer Science Part A: Polymer Chem-
istry. 46 (2008) 3947–57.
[55] Hickner MA, Ghassemi H, Kim YS, Einsla BR, McGrath JE. Alternative Polymer Systems for
Proton Exchange Membranes (PEMs). Chemical Reviews. 104 (2004) 4587–612.
[56] Ghassemi H, McGrath JE, Zawodzinski Jr TA. Multiblock sulfonated–fluorinated poly(arylene
ether)s for a proton exchange membrane fuel cell. Polymer. 47 (2006) 4132–9.
[57] Park MJ, Kim S, Minor AM, Hexemer A, Balsara NP. Control of Domain Orientation in Block
Copolymer Electrolyte Membranes at the Interface with Humid Air. Advanced Materials.
21 (2009) 203–8.
[58] Ulbricht M. Advanced functional polymer membranes. Polymer. 47 (2006) 2217–62.
[59] Mustarelli P, Quartarone E, Grandi S, Carollo A, Magistris A. Polybenzimidazole-Based Mem-
branes as a Real Alternative to Nafion® for Fuel Cells Operating at Low Temperature. Ad-
vanced Materials. 20 (2008) 1339–43.
[60] Schuster MFH, Meyer WH. Anhydrous Proton-Conducting Polymers. Annual Review of
Materials Research 33 (2003) 233–61.
[61] Herz HG, Kreuer KD, Maier J, Scharfenberger G, Schuster MFH, Meyer WH. New fully polymeric
proton solvents with high proton mobility. Electrochimica Acta. 48( 2003) 2165–71.
238 | Part II Development of new materials for energy applications

[62] Goward GR, Schuster MFH, Sebastiani D, Schnell I, Spiess HW. High-Resolution Solid-State
NMR Studies of Imidazole-Based Proton Conductors: Structure Motifs and Chemical Ex-
change from 1H NMR. The Journal of Physical Chemistry B. 106 (2002) 9322–34.
[63] Persson JC, Jannasch P. Intrinsically Proton-Conducting Benzimidazole Units Tethered to
Polysiloxanes. Macromolecules. 38 (2005) 3283–9.
[64] Schuster MFH, Meyer WH, Schuster M, Kreuer KD. Toward a New Type of Anhydrous Or-
ganic Proton Conductor Based on Immobilized Imidazole. Chemistry of Materials. 16 (2003)
[65] Granados-Focil S, Woudenberg RC, Yavuzcetin O, Tuominen MT, Coughlin EB. Water-Free
Proton-Conducting Polysiloxanes: A Study on the Effect of Heterocycle Structure. Macro-
molecules. 40 (2007) 8708–13.
[66] Che Q, He R, Yang J, Feng L, Savinell RF. Phosphoric acid doped high temperature proton
exchange membranes based on sulfonated polyetheretherketone incorporated with ionic
liquids. Electrochemistry Communications. 12 (2010) 647–9.
[67] Fernicola A, Navarra M, Panero S. Aprotic ionic liquids as electrolyte components in protonic
membranes. J Appl Electrochem. 38 (2008) 993–6.
[68] Sekhon SS, Krishnan P, Singh B, Yamada K, Kim CS. Proton conducting membrane containing
room temperature ionic liquid. Electrochimica Acta. 52 (2006) 1639–44.
[69] Alberti G, Casciola M. Composite Membranes for Medium-Temperature PEM Fuel Cells Annual
Review of Materials Research. 33 (2003) 129–54.
[70] Jones D, Rozière J. Advances in the Development of Inorganic–Organic Membranes for Fuel
Cell Applications. Advances in Polymer Science. 215 (2008) 219–64.
[71] de Bonis C, D’Epifanio A, Mecheri B, et al. Layered tetratitanate intercalating sulfanilic acid
for organic/inorganic proton conductors. Solid State Ionics. 227 (2012) 73–9.
[72] Zhang Z, Désilets F, Felice V, Mecheri B, Licoccia S, Tavares AC. On the proton conductivity
of Nafion® –Faujasite composite membranes for low temperature direct methanol fuel cells.
Journal of Power Sources. 196 (2011) 9176–87.
[73] Chen Z, Holmberg B, Li W, et al. Nafion® /Zeolite Nanocomposite Membrane by in Situ Crys-
tallization for a Direct Methanol Fuel Cell. Chemistry of Materials. 18 (2006) 5669–75.
[74] Wang Y, Chen KS, Mishler J, Cho SC, Adroher XC. A review of polymer electrolyte membrane
fuel cells: Technology, applications, and needs on fundamental research. Applied Energy. 88
(2011) 981–1007.
[75] Chen F, Mecheri B, D’Epifanio A, Traversa E, Licoccia S. Development of Nafion® /Tin Oxide
Composite MEA for DMFC Applications. Fuel Cells. 10 (2010) 790–7.
[76] Matos BR, Isidoro RA, Santiago EI, et al. In Situ Fabrication of Nafion –Titanate Hybrid Elec-
trolytes for High-Temperature Direct Ethanol Fuel Cell. The Journal of Physical Chemistry C.
117 (2013) 16863–70.
[77] Aricò AS, Baglio V, Di Blasi A, Creti P, Antonucci PL, Antonucci V. Influence of the acid–base
characteristics of inorganic fillers on the high temperature performance of composite mem-
branes in direct methanol fuel cells. Solid State Ionics. 161 (2003) 251–65.
[78] Mecheri B, Felice V, Zhang Z, D’Epifanio A, Licoccia S, Tavares AC. DSC and DVS Investigation
of Water Mobility in Nafion® /Zeolite Composite Membranes for Fuel Cell Applications. The
Journal of Physical Chemistry C. 116 (2012) 20820–9.
[79] Alberti G, Casciola M, Pica M, Tarpanelli T, Sganappa M. New Preparation Methods for Com-
posite Membranes for Medium Temperature Fuel Cells Based on Precursor Solutions of Insol-
uble Inorganic Compounds. Fuel Cells. 5 (2005) 366–74.
[80] Rhee CH, Kim Y, Lee JS, Kim HK, Chang H. Nanocomposite membranes of surface-sulfonated
titanate and Nafion® for direct methanol fuel cells. Journal of Power Sources 159 (2006)
10 Solid polymer proton conducting electrolytes for fuel cells | 239

[81] Thomassin J-M, Pagnoulle C, Bizzari D, Caldarella G, Germain A, Jérôme R. Improvement of

the barrier properties of Nafion® by fluoro-modified montmorillonite. Solid State Ionics 177
(2006) 1137–44.
[82] Marani D, D’Epifanio A, Traversa E, Miyayama M, Licoccia S. Titania Nanosheets
(TNS)/Sulfonated Poly Ether Ether Ketone (SPEEK) Nanocomposite Proton Exchange Mem-
branes for Fuel Cells. Chemistry of Materials 22 (2009) 1126–33.
[83] Mecheri B, Felice V, D’Epifanio A, Tavares AC, Licoccia S. Composite Polymer Electrolytes
for Fuel Cell Applications: Filler-Induced Effect on Water Sorption and Transport Properties.
ChemPhysChem. 14 (2013) 3814–21.
[84] Mecheri B, D’Epifanio A, Pisani L, et al. Effect of a Proton Conducting Filler on the Physico-
Chemical Properties of SPEEK-Based Membranes. Fuel Cells. 9 (2009) 372–80.
[85] Hamoudi S, Royer S, Kaliaguine S. Propyl- and arene-sulfonic acid functionalized periodic
mesoporous organosilicas. Microporous and Mesoporous Materials. 71 (2004) 17–25.
[86] Felice V, Ntais S, Tavares AC. Propyl sulfonic acid functionalization of faujasite-type zeolites:
Effect on water and methanol sorption and on proton conductivity. Microporous and Meso-
porous Materials. 169 (2013) 128–36.
[87] Cozzi D, de Bonis C, D’Epifanio A, Mecheri B, Tavares AC, Licoccia S. Organically functional-
ized titanium oxide/Nafion® composite proton exchange membranes for fuel cells applica-
tions. Journal of Power Sources. 248 (2014) 1127–32.
[88] Peckham TJ, Schmeisser J, Rodgers M, Holdcroft S. Main-chain, statistically sulfonated proton
exchange membranes: the relationships of acid concentration and proton mobility to water
content and their effect upon proton conductivity. Journal of Materials Chemistry. 17 (2007)
[89] Dupuis A-C. Proton exchange membranes for fuel cells operated at medium temperatures:
Materials and experimental techniques. Progress in Materials Science. 56 (2011) 289-327.
[90] Yang Y, Holdcroft S. Synthetic Strategies for Controlling the Morphology of Proton Conducting
Polymer Membranes. Fuel Cells. 5 (2005) 171–86.
[91] Hou H, Di Vona ML, Knauth P. Durability of Sulfonated Aromatic Polymers for Proton-
Exchange-Membrane Fuel Cells. Chem Sus Chem. 4 (2011) 1526–36.
[92] Barsoukov E, Macdonald JR. Impedance Spectroscopy. Theory Experiment and Applications
John Wiley & Sons, Hoboken NJ, 2005.
[93] Sone Y, Ekdunge P, Simonsson D. Proton Conductivity of Nafion® 117 as Measured by a Four-
Electrode AC Impedance Method. Journal of The Electrochemical Society. 143 (1996) 1254–9.
[94] Slade S, Campbell SA, Ralph TR, Walsh FC. Ionic Conductivity of an Extruded Nafion® 1100 EW
Series of Membranes. Journal of The Electrochemical Society. 149 (2002) A1556–A64.
[95] Soboleva T, Xie Z, Shi Z, Tsang E, Navessin T, Holdcroft S. Investigation of the through-plane
impedance technique for evaluation of anisotropy of proton conducting polymer membranes.
Journal of Electroanalytical Chemistry. 622 (2008) 145–52.
[96] Cooper K. Characterizing Through-Plane and In-Plane Ionic Conductivity of Polymer Elec-
trolyte Membranes. ECS Transactions. 41 (2011) 1371–80.
[97] Shuhua Ma ZS, Hirokazu Tanakaa. Anisotropic Conductivity Over In-Plane and Thickness
Directions in Nafion® -117. Journal of The Electrochemical Society. 153(12), (2006) A2274–81 .
[98] Lee CH, Park HB, Lee YM, Lee RD. Importance of Proton Conductivity Measurement in Polymer
Electrolyte Membrane for Fuel Cell Application. Industrial & Engineering Chemistry Research.
44 (2005) 7617–26.
[99] Sing KSW, Everett DH, Haul RAW, et al. Reporting Physisorption Data for Gas-Solid systems.
Pure and Applied Chemistry. 57 (1985) 603–19.
[100] Ng E-P, Mintova S. Nanoporous materials with enhanced hydrophilicity and high water sorp-
tion capacity. Microporous and Mesoporous Materials. 114 (2008) 1–26.
240 | Part II Development of new materials for energy applications

[101] Crank J, Park GS, (Eds). Diffusion in polymers, Academic Press, London New York, 1968.
[102] Morris DR, Sun X. Water-sorption and transport properties of Nafion® 117 H. Journal of Ap-
plied Polymer Science. 50 (1993) 1445–52.
[103] Takamatsu T, Hashiyama M, Eisenberg A. Sorption phenomena in nafion membranes.
Journal of Applied Polymer Science. 24 (1979) 2199–220.
[104] Foo KY, Hameed BH. Insights into the modeling of adsorption isotherm systems. Chemical
Engineering Journal. 156 (2010) 2–10.
[105] Park GS. Synthetic Membranes: Science, Engineering and Applications. Dordrecht, The
Netherlands: Reidel Publishing Company; 1986.
[106] Takata H, Mizuno N, Nishikawa M, Fukada S, Yoshitake M. Adsorption properties of water
vapor on sulfonated perfluoropolymer membranes. International Journal of Hydrogen Energy.
32 (2007) 371–9.
[107] Feng H. Modeling of vapor sorption in glassy polymers using a new dual mode sorption
model based on multilayer sorption theory. Polymer. 48 (2007) 2988–3002.
[108] Guggenheim EA. Applications of Statistical Mechanics. Oxford, U.K: Clarendon Press; 1966.
[109] Anderson RB. Modifications of the Brunauer, Emmett and Teller Equation1. Journal of the
American Chemical Society. 68 (1946) 686–91.
[110] de Boer JH. The Dynamical Character of Adsorption. Oxford, U.K.: Clarendon Press; 1968.
[111] Li Y, Nguyen QT, Buquet CL, Langevin D, Legras M, Marais S. Water sorption in Nafion® mem-
branes analyzed with an improved dual-mode sorption model—Structure/property relation-
ships. Journal of Membrane Science. 439 (2013) 1–11.
[112] Osswald T, Menges G. Materials Science of Polymers for Engineers. 2nd ed. Cincinnati, USA:
Hanser Gardner Publications, Inc.; 2003.
[113] Höhne G, Hemminger WF, Flammersheim HJ. Differential Scanning Calorimetry. 2nd ed. Berlin,
Heidelberg Germany: Springer; 2003.
[114] Wunderlich B. Thermal Analysis of Polymeric Materials. Berlin, Germany: Springer; 2005.
P. Bénard, A.-M. Beaulieu, D. Durette, and R. Chahine
11 Supercritical adsorption of hydrogen on
microporous adsorbents
11.1 Introduction

Physical adsorption on microporous adsorbents is widely used in industrial physico-

chemical processes such as gas separation and purification [1, 2]. The adsorption phe-
nomena can be defined as the enrichment or depletion of a species of fluid in an inter-
facial layer between the fluid and a substrate [3]. The absorption process, in contrast,
occurs when the species is integrated into the structure of the substrate. Chemisorp-
tion typically involves both phenomena. Physisorption (or physical adsorption) is as-
sociated with weak and nonspecific interactions, typically van der Waals forces. Ph-
ysisorption does not involve the hybridization of electronic orbitals. The distinction
between physisorption and chemisorption can be subtle. Typically, adsorption pro-
cesses with a characteristic energy scale smaller than 15 kJ/mol are considered to be
physical. The adsorption process occurs within the porous structure of the adsorbent,
which is the volume of the adsorbent externally accessible to the adsorbate molecules.
The pores are formally classified by IUPAC as micropores (characteristic size < 2 nm),
mesopores (2–50 nm) and macropores (characteristic size > 50 nm) [4]. Gas storage
applications of the adsorption phenomena usually require maximizing the micropore
volume of an adsorbent.
The adsorption process can be used to store light gases at substantially lower pres-
sures, offering the possibility of safer storage conditions with respect to high pressure
compression. The storage density of gases such as hydrogen and methane can thus be
increased significantly through the interatomic and intermolecular forces between the
adsorbate molecules (adatoms) and a solid surface. Materials-based storage of gases
such as hydrogen can also be achieved through chemical binding to other elements
(chemical hydrides) or through absorption inside a solid metallic matrix (metal hy-
drides). For materials-based storage, the characteristic binding energy of hydrogen
to the material is a determining factor in the thermodynamic description of systems.
It sets the energy scale required for charging and discharging gases, regulates its ther-
mal behavior, and determines boil-off rates. Materials with binding energies smaller
than 10 kJ/mol require cryogenic operation to achieve acceptable storage densities.
Passive heat sources from the environment can then be used to thermally manage out-
gassing. Low binding energies occur when hydrogen is reversibly bound in molecular
form to a substrate through weak, nonspecific interactions. The hard-to-achieve range
of 10–20 kJ/mol is believed to be optimal for materials-based hydrogen storage appli-
cations closer to room temperature.
242 | Part II Development of new materials for energy applications

11.2 Fundamentals of supercritical adsorption

Weak interactions between a gas and a substrate usually result in the adsorption of
gas molecules (the adsorbate) onto the surface of the substrate (the adsorbent). The
adsorption effect corresponds to a local enhancement of the density of the adsor-
bate close to the adsorbent’s surface resulting from those interactions, in the region
where the adsorbate/adsorbent interaction potential is significant (Fig. 11.1). The to-
tal adsorbed density nt is defined as the number of adsorbate molecules present in
the porous structure of the adsorbate. The adsorbed density of the adsorbate is usu-
ally divided by the quantity of adsorbent, either volumetrically (quantity of adsorbate
molecules per unit volume of adsorbent) or gravimetrically (per unit mass of the ad-
sorbent). Both are related to one another by the density of the adsorbent.
The density profile of the gas close to the surface can be separated into 3 regions:
(1) the solid phase, (2) the interface, and (3) the gas phase with volumes Vs , Va, and Vg ,
respectively. The volume of the solid phase (Vs ) is also called the volume of the skele-
ton, as it excludes the contribution of the pores. The total volume of the adsorbate is


Interface Gas


(b) Fig. 11.1. (a) Interaction potential between an ad-
sorbate molecule and the surface of an adsorbent.
ntot (b) Typical density profile of a gas close to an ad-
sorbing surface. The enhanced density in the vicinity
of the adsorbent is caused by gas-surface interac-

Gibbs surface tions. The boundary indicated on the figure depends

on the definition of a cut-off for the interaction po-
ng tential. (c) Gibbs dividing surface: the adsorbed
Gas atoms are smeared onto the solid. The dividing sur-
0 face is an ideal construct separating the adsorbent
X0 d
(c) and the adsorbate gas.
11 Supercritical adsorption of hydrogen on microporous adsorbents | 243

the sum of the three contributions. The adsorbate is usually described as belonging to
one of two phases: the gaseous phase (where the interaction potential is negligible)
or the adsorbed phase located in the interface.
The excess adsorbed density (na ) is defined as the number of molecules in a
porous structure in excess of the number that would be present in the same volume
under the same thermodynamic conditions in the absence of adsorbate-adsorbent
interactions. It can be defined as follows:

n a = n t − n g Vp .

The pore volume (Vp ) is defined as the sum of the volume of the gas phase in the ad-
sorbent and of the volume of the interface.
It is clear from Fig. 11.1 that the distinction between adsorbed molecules and the
bulk gas is ambiguous for long-range interaction potentials, which do not have a clear
cut-off. The boundary between the phases can be conventionally defined using a ref-
erence gas that is assumed to be nonadsorptive under ambient conditions. In Gibb’s
formulation of adsorption, this is done through the Gibbs dividing surface. The Gibbs
surface is chosen in such a way that the excess density of the reference gas (typically
helium) is zero under ambient conditions:

na (T0 , P0 ) = nt − ng Vp = 0.

This equation defines Vp and the location of the dividing surface. The pore volume
(Vp ) thus formally depends on the reference gas and the thermodynamic conditions
under which it is calculated.
In the Gibbs approach, the gas phase density is assumed to be constant and equal
to the bulk phase up to the Gibbs dividing surface [5]. The adsorbed gas molecules
(sometimes referred to as the adatoms) are assumed to be located on the surface. The
adsorbate phase has no volume (Va = 0). When considered as an interface problem,
the thermodynamic description of adsorption requires the introduction of the surface
area of the interface as a thermodynamic variable [6], coupled to a conjugate force
called the spreading pressure. It is coupled to the interfacial surface area in the free
energy equations of the system. The spreading pressure is not directly accessible and
acts as an intermediate calculation variable.
For highly microporous adsorbents (∼ 1.5 nm), the pore size is usually such that
from a microscopic point of view, an adatom is affected by adsorbent-adsorbate inter-
actions throughout the pore structure of the adsorbent. As a result, the local adsorbate
density remains larger than the free gas value in the pore volume. A macroscopic de-
scription of the adsorption process based on interface properties then becomes coun-
terintuitive, as a zero-volume interface has no simple microscopic correspondence.
From a macroscopic point of view, however, a consistent thermodynamic description
of the adsorption phenomena in microporous adsorbents based on interface thermo-
dynamic variables can still be used, as the latter are intermediate calculation variables
not directly accessible experimentally [6].
244 | Part II Development of new materials for energy applications

The adsorption isotherm (na (T, P)) expresses the relationship between the excess
density defined in the Gibbs sense and the pressure of the bulk gas phase with which
it is in equilibrium for a given temperature. It represents the difference between the
mass of adsorbate in the porous structure of the adsorbate and the amount that would
be present in the same volume in the absence of the adsorbate. The Gibbs adsorp-
tion isotherm is also called the excess adsorption isotherm and can be seen as the
gain in density over compression at the equilibrium pressure and temperature. The
absolute adsorption isotherm refers to the total amount of hydrogen present within
the potential field of the adsorbent. Obtaining the absolute adsorption isotherm from
the excess requires knowledge of the volume of the adsorbed phase, which is not di-
rectly accessible experimentally and depends very much on the definition of a cut-off
of the adsorbate/adsorbent interaction potential. For highly microporous adsorbents,
in which most of the volume of the pores is subject to the interaction potential, the
absolute adsorption isotherm corresponds to the total uptake of adsorbate inside the
pores. However, if macropores and mesopores are also present, adsorbate molecules
far from their surfaces should be excluded. An unambiguous definition of the abso-
lute adsorption isotherm can therefore be difficult, particularly for adsorbents with a
complex pore size distribution. The excess adsorption isotherm represents, however,
a clear and experimentally accessible concept.
The classification of adsorption isotherms by the International Union of Pure
and Applied Chemistry (IUPAC) consists of 6 categories [7]. Types I–III are reversible
isotherms which do not exhibit hysteresis. The Type I isotherm is concave with respect
to the pressure axis. It is usually associated with microporous solids (e.g., activated
carbons, molecular sieve zeolites, and certain porous oxides). The Type II isotherm
represents unrestricted monolayer-multilayer adsorption occurring in a macroporous
adsorbent. Point B in Fig. 11.2 indicates where multilayer adsorption begins to take
place. Type III isotherms are uncommon and characterized by a convex curve with re-
spect to the pressure axis. Type IV and V isotherms exhibit a hysteresis loop. Type IV
isotherms are associated with capillary condensation taking place in mesopores.
Type V isotherms are related to Type III, with a hysteresis loop as an additional
feature. Type VI isotherms are associated with stepwise multilayer adsorption on a
uniform nonporous surface. The height of each step is associated with the monolayer
capacity for each adsorbed layer.
Only adsorption processes associated with Type I isotherms will be addressed in
this chapter. We will only consider supercritical adsorption, which occurs above the
critical point of the adsorbate. We will thus limit ourselves to physical adsorption (or
physisorption) processes, associated with relatively weak adsorbate-adsorbent inter-
actions, and to the study of single component adsorbates, as the focus of this paper is
gas storage.
As discussed in the introduction, the adsorption phenomena, through gas sur-
face interactions, offers the possibility of substantially lowering the storage pressure
of gases [1, 2]. The presence of an interaction potential with a characteristic binding
11 Supercritical adsorption of hydrogen on microporous adsorbents | 245

energy (EA ) permeating narrow pores can substantially enhance the equilibrium den-
sities of the adsorbed gas inside the pore, compared to the far-field bulk gas, resulting
in very high local pressures in the pores of the adsorbing nanostructure. The resulting
fluid densities of the adsorbate in the pores can become comparable and sometimes
even exceed its liquid density, depending on the pressure and temperature of its bulk
phase. In addition to the characteristic energy (EA ), the surface area available for ad-
sorption processes and the bulk density of the adsorbent are important parameters for
adsorption-based applications. The specific surface is an important contributor to the
saturation density of an adsorbent, whereas the bulk density determines the weight of
the storage unit. Both parameters can be combined into the surface available to the ad-
sorbate per unit volume of the adsorbent, which should be maximized to optimize the
storage density. This quantity can be expressed in terms of adsorbate mass/adsorbent
volume, or adsorbate volume/adsorbent volume, the latter yielding the former when
multiplied by the bulk gas density at 298 K and 1 atmosphere of the adsorbate.


Amount odsorbed


Fig. 11.2. Types of adsorption isotherms.

Relative pressure Reprinted with permission from [7].
246 | Part II Development of new materials for energy applications

11.3 Supercritical adsorption isotherms

The adsorption isotherm is the equation of the state of the adsorption process and as
such, plays a key role in its thermodynamic description. Predictive strategies involve
first principles calculations using computer simulations and numerical analysis of its
properties in specific limits. Such approaches include ab initio quantum chemistry
calculations, grand canonical Monte Carlo simulations (with or without quantum cor-
rections), the virial expansion, and to some extent classical density functional theory.
Descriptive approaches aim to describe the isotherms using adsorption models (ana-
lytic or not) appropriate for the adsorbent structure and the adsorbate-adsorbent in-
teractions. These rely on approximations which invariably limit their applicability to
specific ranges of pressure and temperature.

11.3.1 Virial expansion of the excess density in terms of pressure

In the low pressure limit, the excess adsorbed density can be expressed as a virial
series expansion in terms of pressure [8]:

p p2 p3 p4
na (p, T) = BAS + BAAS 2
+ CAAS 3
+ DAAS + ..., (11.1)
kT (kT) (kT) (kT)4
where p is the pressure of the gas phase. At the lowest order in pressure, the exces-
sadsorbed density varies linearly with pressure, which corresponds to Henry’s law as
applied to the adsorption phenomena.
The coefficient BAS , generally called the second virial coefficient, is fully deter-
mined by the interaction U (r)⃗ between a single molecule and the surface of the adsor-
1 U (r)⃗
BAS (T) = ∫ (exp (− ) − 1) dr,⃗ (11.2)
M kT
where M is the mass of the adsorbent. The virial coefficient BAS can be obtained ex-
perimentally by finding the intercept of a plot of na (p, T) kT/p as a function of the
pressure (p). The interaction potential V(r)⃗ between an adsorbate molecule and the
adsorbent atoms can be written as:

U(r)⃗ = ∑ Uαi (r ⃗ − R⃗ i ), (11.3)


where r ⃗ and R⃗ i are the positions of the adsorbate molecule and the ith atom of the
adsorbent, and where αi refers to the atomic species of the atom at R⃗ i . The interaction
potential U(r)⃗ is often modeled using the Lennard-Jones potential:
6 12
Uαi (r)⃗ = 4ε [(σ/r) − (σ/r) ] , (11.4)

where σ is the distance where V(r)⃗ = 0 and 𝜖 is the minimum interaction energy.
11 Supercritical adsorption of hydrogen on microporous adsorbents | 247

Effective adsorbate-adsorbent potentials based on the Lennard-Jones potential

have been developed to model the interaction of a crystalline adsorbent surface with
adsorbate molecules. The planar Lennard-Jones potential, for instance, represents an
averaged interaction potential resulting from smearing the atoms of a corrugated pla-
nar crystal into a uniform, continuous distribution onto an infinite two-dimensional
plane. The 10-4 Lennard-Jones planar potential is given by the:

2 1 1
V (y) = 2εs ( − ), (11.5)
5 y10 y4

where εs = πθεσ2 . The parameter θ is the surface density of a graphene plane (0.38
atoms per Å [2]). The variable (y) is dimensionless and rescaled to σ (y = z/σ) and rep-
resents the distance (z) along the normal to the plane. The parameters ε = 30.5 K and
σ = 3.19 Å represent, respectively, the Lennard-Jones energy and distance parameters
between carbon atoms and hydrogen.
The pores of an activated carbon are typically modeled using the so-called slit
pore approach, in which they consist of two parallel smeared infinite graphene lay-
ers separated by the characteristic pore size (d). The resulting interaction potential
is shown in Fig. 11.3. As a function of decreasing distance between the graphene lay-
ers, the well depth of the potential gets deeper as the overlap of the contributions
from each plane increases. Ultimately, the two minima overlap completely, resulting
in a single well. Further decreasing the distance between the planes will decrease the
depth of the well, until the interaction potential in the pores becomes positive every-
where, quenching the adsorption process in the pore.



–2 –1 1 2 σ



Fig. 11.3. The slit pore potential for activated carbons. A pore is modeled by two infinite parallel
graphene layers smeared into planes, separated by a distance assumed to be representative of
the size of the pore resulting in the above adsorbate–adsorbent interaction potential within the
slit pore.
248 | Part II Development of new materials for energy applications

Assuming an average noncorrugated planar interaction potential between the ad-

sorbent planes and an adsorbate molecule, the second virial coefficient is then given
d∗ /2
= 2 ∫ [e kT [V(y+d /2)+V(y−d /2)] − 1] dy,
−1 ∗ ∗
SA σ

where d∗ = d/σ, y is a dimensonless integration variable, and SA is the specific surface,

which is the area of the adsorbent accessible to adsorbate molecules per unit mass of
the adsorbent. The second virial coefficient BAS for the activated carbon AX-21 is shown
in Fig. 11.4.



In (B AS/αZo )

0 2 4 6 8 10 12 14
X = ɛs/kT

Fig. 11.4. Normalized second virial coefficient as a function of the rescaled inverse temperature. The
line shows the theoretical curve. The black and white dots show experimental data points from the
adsorption of methane and hydrogen (respectively) on the activated carbon AX-21. The normalized
curve is expected to be universal for nonpolar classical gases. Reprinted with permission from [12].

The distance (d) and the specific surface of the activated carbon are treated as ad-
justable parameters, the values d = 8.1 Å for the interlayer distance and 2 900 m2 /Å
for the specific surface area for hydrogen sorption on the activated carbon AX-21 are
obtained. Although the estimated specific surface is close to the experimentally de-
termined BET value (2 800 m2 /Å), DFT analysis of the adsorption isotherm suggests a
pore size distribution peaked around 12.5 Å.
The second virial coefficient of single-wall nanotubes has also been studied using
a cylindrical potential in the continuum approximation [9]:

21 1 10 1 4
V∗ (r∗ , R∗ ) = 3 [ ( ∗ ) M11 (r∗ /R∗ ) − ( ∗ ) M5 (r∗ /R∗ )] , (11.7)
32 R R
11 Supercritical adsorption of hydrogen on microporous adsorbents | 249

where r∗ = r/σ, R∗ = R/σ, and V∗ (r∗ , R∗ ) = V(r∗ σ, R∗ σ)/εs are respectively the reduced
distance from the axis, the reduced radius of the SWNT, and the reduced potential,

Mn (x) = ∫ . (11.8)
(1 + x 2 − 2x cos φ)n/2
The interaction potential inside a nanotube is shown in Fig. 11.5.





0.0 0.2 0.4 0.6 0.8

Fig. 11.5. Potential of Stan and Cole for SWNT of various diameters R and helium, a function of
the distance from the center, divided by R. When the radius R is smaller than a critical value
Rc = 1.212 σ, the minimum is at the center. Further decreasing the radius leads to a shallower po-
tential well, and eventually a positive repulsive core [9].

Stan and Cole [9, 10] performed a study of the virial coefficient BAS for the adsorption
of rare gases on single wall nanotubes as a function of the radius using the cylin-
drical Lennard-Jones potential. The adsorbed density of Ne atoms on an SWNT was
compared to that obtained using a planar graphene sheet with the same surface area.
They found a significantly larger adsorbed density inside the SWNT compared to the
graphene sheet due to the curvature of the nanotubes. An estimate of the maximum
binding energy of adsorbate molecules in the cylindrical pore can be determined with
the cylindrical potential. The value of the potential at the centre of the tube is found
by setting normalized distance r∗ to zero in equation:
21 σ 10 σ 4
V (x = 0, R) = 3π2 θεσ2 ( ( ) − ( ) ). (11.9)
32 R R
From this expression, it can be determined that the maximum well depth is Vo =
V(r = 0, R) = −12.77 θε (or 12.5 kJ/mol for hydrogen) for a radius R0 = 1.086 σ, or
R0 = 3.43 Å for hydrogen.
250 | Part II Development of new materials for energy applications

The second virial coefficient, for uncapped SWNTs of length L, is given by the
following expression:
BAS = ∫ dr ⃗ (exp (− V (r,⃗ R)) − 1) . (11.10)

Since the interaction potential is only a function of the normalized radius, the volume
integral can be performed using cylindrical coordinates:
2π L/2 ∞
BAS = ∫ dθ ∫ dz ∫ rdr (exp (− V (r∗ , R∗ )) − 1), (11.11)
0 −L/2 0

which leads to

= 2π ∫ r∗ dr∗ (exp (− ∗ V∗ (r∗ , R∗ )) − 1). (11.12)
Lσ2 T

Capped nanotubes do not adsorb internally, so the volume integral must exclude the
internal volume of the nanotube:

= 2π ∫ rdr (exp (− V (r∗ , R∗ )) − 1). (11.13)
Lσ2 ∗

The behavior of the second virial coefficient as a function of the radius of the
nanotube is quite complex, and summarized in Fig. 11.6 below for hydrogen. At 77 K,
the coefficient BAS reaches its maximum value when the radius of the nanotube cor-
responds to the maximum well depth of the adsorption potential in the nanotube
(full line). The peak value of BAS is 3 orders of magnitude greater than the equiva-
lent graphene sheet. This feature, associated with a strong peak, is absent when the
nanotube is closed. The peak is quenched very quickly as a function of temperature.
At larger radii, the curves converge towards the value of an equivalent graphene sheet.
It is worthwhile to note that the smallest possible SWNT has a radius equal to that of
a C20 . These have been observed inside multi-walled carbon nanotubes as their inner-
most shells. The smallest experimentally accessible SWNTs have radii comparable to
C60 (3.5 Å). The quenching of the BAS of open SWNTs observed at higher temperatures
can be linked to the presence of repulsive and attractive regions inside the nanotube,
which at low temperature usually lead to a positive and large value of the integrant in
the expression for BAS , but which can become negative at lower temperature because
the exponential, which is sensitive to temperature, is subtracted by a temperature-
independent constant. It is then possible for the virial coefficient to become negative
and impair the adsorption process to first order in pressure. This occurs at tempera-
tures above the Boyle temperature.
11 Supercritical adsorption of hydrogen on microporous adsorbents | 251






10–3 C20 C60

0 1 2 3 4

Opened SWNT, T* = 0.2089 (H2: 77.4 K)

Closed SWNT, T* = 0.2089
Opened SWNT, T* = 0.8097 (H2: 300 K)
Closed SWNT, T* = 0.8097
Opened SWNT, T* = 1.000 (H2: 371 K)
Closed SWNT, T* = 1.000

Fig. 11.6. Behavior of the second virial coefficient of the adsorbed density of hydrogen on iso-
lated SWNTs as a function of its radius. The SWNT reduced radii corresponding to the C20 and
C60 fullerene cages are indicated by the vertical lines. Reprinted with permission from [11].

The behavior of the virial coefficient in bundles of SWNTs is rich, associated with the
grooves of the bundle (peripheral sites in Fig. 11.7(a)) and the interstices between the
nanotubes in the bundles (interstitial sites in Fig. 11.7(a)), which lead to additional
structures associated with the presence of additional extrema in the interaction po-
tential of the bundles (Fig. 11.7(b)).
The virial coefficient BAS is a readily physically accessible quantity which contains
a wealth of information about the energy scales of the adsorption process and in some
cases, the very structure of an adsorbent, particularly in ordered monoatomic solids
such as single wall nanotubes or other nanocrystalline carbon structures. It can be
determined experimentally from the slope of the excess adsorption isotherm at low
pressure, where the isotherm behaves according to Henry’s Law, from two pressure
measurements at a given temperature. The energy scales that can be extracted from
BAS require, however, the data from several low pressure isotherms. They can be ap-
proximately determined from a semi-logarithmic fit of BAS as a function of inverse tem-
252 | Part II Development of new materials for energy applications

The virial coefficient, however, does not contain much information on the high
pressure behavior of the adsorption isotherm, as it is determined in a régime where
intermolecular interactions between adatoms have not yet impacted the adsorption
phenomena, and provides little information on the eventual saturation of the adsorp-
tion isotherm. It can thus be used to determine the temperature scale of the adsorption
process, but not its saturation properties.

Tube site Peripheral site




Interstitial site
–15 x
(a) –15 –10 –5 0 5 10 15


x 0


–2 Fig. 11.7. (a) Adsorption sites and
–4 (b) interaction potential of a bundle
with 19 SWNTs with R∗ = 2.038 and
with a normalized lattice parame-
0 ter d∗ = 5.235 relative to hydrogen.
(b) 10 Reprinted with permission from [11].

11.3.2 Basic analytic models of the adsorption isotherm

The Langmuir model

The Langmuir isotherm is a local model of the absolute adsorption based on mono-
layer filling of noninteracting molecules. As such, it represents a model of supercritical
adsorption on surfaces. It predicts a monotonically increasing function of pressure,
which saturates asymptotically as a function of pressure to a value nm :
KL (T) P
n = nm , (11.14)
1 + KL (T) P
11 Supercritical adsorption of hydrogen on microporous adsorbents | 253

exp (ΔS0 /R) exp (−ΔH0 /RT),
KL (T) = (11.15)
where ΔS0 and ΔH0 represent the entropy and enthalpy variations associated with the
adsorption process, and where P is pressure [12].
There are several strategies for deriving the Langmuir isotherm. It can be obtained
from either kinetic theory or a lattice gas approach.
The Langmuir isotherm can be put into the following form:
P 1 P
= + , (11.16)
n KL (T)nm nm
from which the parameters nm and KL (T) can be determined from experimental
isotherms. The saturation density (nm ) and the coefficient (C) can be obtained from
a simple linear fit of the experimental data. Figure 11.8 shows that even excess
adsorption data (shown as points) represented in this way yields relatively linear
curves. Note, however, that the Langmuir isotherm is not a Gibbs (excess) adsorption
isotherm. A more detailed examination of the data represented in Fig. 11.8 would
show substantial deviations from linear behavior between the low and high pressure
The excess adsorption version of the Langmuir isotherm requires knowledge of
the pore volume:
KL (T)P󸀠 P󸀠
na (P󸀠 , T) = n(P󸀠 , T) − Vp ng (P󸀠 , T) = nm 󸀠
− Vp . (11.17)
1 + KL (T)P RT

T=273 K

P/n (Mpa gr/mmol)




T=77 K
0 1 2 3 4 5 6
Pressure (Mpa)

Fig. 11.8. Expression of the excess adsorption isotherms of hydrogen on the activated carbon AX-21
as the pressure over excess density as a function of pressure. Reprinted with permission from [12].
254 | Part II Development of new materials for energy applications

The validity of the Langmuir model is typically limited to low pressure, high tempera-
ture adsorption of gases in the supercritical regime. As such, it is of limited usefulness
to the description of adsorption over the wide temperature and pressure ranges rele-
vant to storage applications.

The Dubinin isotherm

The adsorption isotherm of microporous adsorbents has often been modeled by the
Dubinin–Astakhov model. In this approach, the adsorption process is viewed as a vol-
umetric process in which a pore of an adsorbent is progressively filled by a subcritical
gas (T < Tc ) whose vapor phase can be described by the ideal gas law. If the con-
ditions inside a pore are such that the local pressure is greater than the saturation,
then the adsorbate is assumed to condense to a liquid state. The characteristic energy
of adsorption of the pores is also assumed to be distributed according to a Weibull
distribution. These assumptions lead to an isotherm where total adsorbed density is
expressed as:
n = n0 exp(−(A/E)m ), (11.18)
where n0 is the saturation density of the adsorbate in the pores, m is a structural het-
erogeneity parameter, E is a characteristic energy of the average gas-solid interaction
potential in the pores [13], and A is the adsorption potential:
A = RT ln ( ), (11.19)
where R is the universal gas constant, and Ps the saturation pressure. The parame-
ter m is fitted in such a way that the characteristic curves of nA as functions of A are
independent of temperature. The limit m = 2 reduce to the Dubinin–Raduskevitch
isotherm [14, 15].
The excess density is obtained by considering the adsorption volume Va as a fitting
nex = na − Vp ρg , (11.20)
where ρg is the bulk gas density of hydrogen.
Surprisingly, this isotherm has been found to accurately describe the adsorption
process even in the supercritical regime. The pressure parameter Ps cannot in this
case be associated with the saturation pressure, which is undefined when T > Tc .
Amankwah and Schwarz have proposed the following equation [16] to determine Ps :
PS = (T/TC ) PC , (11.21)

where Tc and Pc are the critical temperature and pressure of the adsorbate. This pres-
sure, however, should be interpreted as the pressure at which the adsorbed phase be-
haves like an almost incompressible fluid in the supercritical regime, leading to a very
large value of Ps when the isotherm is fitted to experimental data over wide pressure
11 Supercritical adsorption of hydrogen on microporous adsorbents | 255

range. Formally, at P = PS , the isotherm reaches a maximum value as a function of

If the adsorbed phase behaves like a quasi-incompressible fluid, it can be ex-
pected that the characteristic energy also depends on temperature to reflect a limited
dependence on temperature that should be expected when P is such that the pore is
basically saturated. An analysis of the best fit of experimental isotherms of nitrogen
on activated carbon over wide ranges of pressure and temperature suggest a linear
dependence of the characteristic energy of adsorption as a function of temperature. A
modification of the Dubinin isotherm was then proposed in which the characteristic
energy of adsorption of the original model is replaced by a temperature-dependent
expression E = α + βT.

50 30 K
35 K
Excess Adsorption (mol/kg)

40 K
45 K
60 K
77 K
93 K
113 K
153 K
213 K
298 K
0 1 2 3 4 5 6
Pressure (Mpa)

Fig. 11.9. Modified D–A isotherm (solid lines) parameterized on the experimental excess ad-
sorption isotherms of hydrogen (shown as points) [19]. The optimal isotherm parameters are
n0 = 71.6 mol kg−1 , α = 3.08 kJ mol−1 , β = 18.9 J mol−1 K−1 , P0 = 1 470 MPa, and Vp = 1.43 l kg−1 .
The large values of P0 may be seen as coherent with the high density limit of the adsorbed gas.
Reprinted with permission from [17]

The model successfully describes the adsorption of hydrogen on several microporous

systems [17, 18], such as the activated carbon AX-21™ over the range 0–6 MPa and 30–
298 K, as shown in Fig. 11.9. The limiting density associated with complete filling of the
micropores can be estimated by dividing nmax by Vpore . A value of 50.2 mol/l is obtained
for hydrogen, which is a value close to the density of solid hydrogen (43.7 mol/l).
The model has also been used to describe the adsorption isotherms of hydrogen
on the activated carbon CNS-201™ and the metal-organic framework Cu3 (BTC)2 over
the range (0–6 MPa and 77–296 K), as well as the adsorption isotherms of nitrogen and
methane on single wall nanotubes over wide temperature and pressure ranges [19].
256 | Part II Development of new materials for energy applications

The Dubinin model has some issues, such as the absence of a proper Henry’s law
régime as a function of pressure. This approach should be viewed as a high density
approximation. This can lead to problematic results at low pressure, such as nega-
tive excess adsorption isotherms, for the adsorption of hydrogen on the metal organic
framework MOF 5. Its validitiy is also limited at high temperature by the fact that the
argument of the isotherm depends very little on temperature at large temperatures.

11.3.3 Self-consistent approaches

In addition to the examples of analytic isotherms described in the preceding section,

nonanalytic isotherm models have been proposed for supercritical gases over wide
pressure and temperature ranges.

Lattice gas isotherm

Aranovitch et al. proposed a set of self-consistent equations describing the adsorp-
tion process on carbon slit pores based on a lattice-gas approach [12, 20], applica-
ble in principle to the supercritical and subcritical regimes. This approach, which
can be seen as a lattice-gas version of the Langmuir isotherm, calculated assuming
a slit pore geometry, while accounting for proximity interactions between adsorbate
molecules in a molecular field approximation. It correctly yields Henry’s law in the
low pressure limit but does not, however, describe the porous structure of carbon.
The model assumes that the adsorbate molecules form N discrete layers sandwiched
by two graphene planes representing a slit pore. Adjacent molecules interact with an
Ising site potential E. Molecules adjacent to the two graphene planes are subjected
to a uniform potential EA . The coverage of the ith layer, defined by the relative den-
sity xi = ni /ns (where ns is the molar density of a completely filled adsorption layer)
is obtained from self-consistent equations with the boundary condition x1 = xN . The
excess adsorption isotherm is defined as:
nex = A ∑ xi − ρg Vp . (11.22)

The model has been successfully fitted to the experimental data shown in Fig. 11.10
for hydrogen on AX-21 with the following model parameters: Ea = −3.871 kJ/mol,
E = 0.519 kJ/mol and ns = 74.0 mol/l. The saturation constant (A) was fitted in the
following way:
A (T) = A0 − A1 exp (−B/T) , (11.23)
with A0 = 30.426 mol/l, A1 = 12.623 mol/l and B = 71.042 K. This isotherm leads to
a pore volume Vpore of 1.62 liter/kg [21].
11 Supercritical adsorption of hydrogen on microporous adsorbents | 257

30 K
35 K
Excess adsorbed amount [mol/kg]

45 K
40 60 K
77 K
93 K
30 113 K
133 K
153 K
20 173 K
193 K
213 K
10 233 K
253 K
273 K
295 K
0 1 2 3 4 5 6
Pressure [Mpa]

Fig. 11.10. Excess adsorption isotherms of hydrogen on the activated carbon AX-21. The full lines
represent a fit to the Ono-Kondo adsorption isotherm. Reprinted with permission from [21].

11.4 The thermodynamics of adsorption

Myers et al. [6] proposed a thermodynamic formulation of the adsorption process in

which the pore structure is considered to be a bulk property of the adsorbent (assum-
ing with Gibbs a zero volume adsorbed phase). This approach has the advantage of
eliminating the need for additional thermodynamic variables and their conjugated
The adsorption properties are generally referred to per unit quantity of adsorbent
(e.g., mass or volume). The adsorbed density thus represents a quantity of adsorbate
(mass, volume, moles or direct number of particles) per unit quantity of the adsorbent.
In their approach, all thermodynamic properties (Z) of the total system are con-
sidered to consist of three contributions: the adsorbate (a), the adsorbent (s) and the
gas phase (g):
Zt = Za + Zg + Zs , (11.24)
Zg = ng Vg zg . (11.25)
The density of the free gas (in the absence of the solid) is ng . The intensive property zg
is a thermodynamic property of the free gas which depends on temperature T, pres-
sure P and, in the case of a mixture, the set of mass fractions (yi ). The volume of the
gas phase Vg corresponds to the pore volume Vg . The properties of the solid phase
are determined through thermodynamic measurements of the solid in the absence of
the adsorbate. The properties of the adsorbed phase, on the other hand, intrinsically
depend on the interaction between the solid and the adsorbate and are obtained by
258 | Part II Development of new materials for energy applications

taking the difference between the thermodynamic properties of the total system and
the properties of the free gas and the solid adsorbent:

Za = Zt − Zg − Zs , (11.26)

which leads to the following expression for the adsorbed density:

na = n t − n g . (11.27)

This expression is also consistent with the excess adsorbed density defined earlier
if the system does not contain a free volume external to the porous structure of the
adsorbent, in which case Vg = Vp :

na = nt − ngaz = nt − ρg Vg = nt − ρg Vp , (11.28)

since the volume of the system is the sum of the volume of the solid and the volume
occupied by the gas in the adsorbent. At a given pressure and temperature, the ther-
modynamic properties of the adsorbed phase can be obtained by subtracting the con-
tribution of the gas phase and the adsorbent:

U a = Ut − Ug − Us = U − U s , (11.29)
Sa = St − Sg − Ss = S − Ss , (11.30)
Va = Vt − Vg − Vs = Vt − (Vg + Vs ) = 0. (11.31)

The internal energy of adsorption can be written:

Ua = U − Us = TSa + μg na + (μ − μs ) (11.32)

Ua = TSa + μg na + Φ, (11.33)
Φ = μ − μs . (11.34)
The potential Φ is called the surface potential, defined as the difference between the
chemical potential of the adsorbent interacting with the adsorbate and its value in the
absence of such interactions, but still subjected to hydrostatic pressure P. A descrip-
tion based on the surface potential considers the adsorption process basically as a
bulk property. It is as such appropriate from a conceptual point of view to the descrip-
tion of highly microporous adsorbents such as metal organic frameworks, in which a
clear interface cannot easily be defined.
The Gibbs potential can be found using the standard Legendre transform:

Ga = Ua + PVa − TSa = Ua − TSa = μg na + Φ. (11.35)

The free energies of the adsorbed phase thus depend on the adsorption isotherms (na )
and the surface potential.
11 Supercritical adsorption of hydrogen on microporous adsorbents | 259

The differential forms of the thermodynamic potentials can be written:

dUa = TdSa − μg dna , (11.36)

dGa = Sa dT − μg dna , (11.37)
dΦ = −Sa dT − na dμg . (11.38)

The adsorption isotherm and the entropy of the adsorbed phase can be obtained from:
𝜕Φ 󵄨󵄨󵄨󵄨
na = − μ 󵄨 (11.39)
𝜕 g 󵄨󵄨󵄨T
𝜕Φ 󵄨󵄨󵄨󵄨
Sa = − 󵄨 . (11.40)
𝜕T 󵄨󵄨󵄨μg
Using the Gibbs potential it can be shown that thermodynamic equilibrium between
the adsorbed molecules and their bulk gas phase is expressed through the equality of
their temperatures and their chemical potentials:

μa = μgaz . (11.41)

11.4.1 Properties of surface potential

At constant temperature, the difference equation for surface potential becomes:

dΦ|T = −na dμg . (11.42)

Expressing chemical potential in terms of the fugacity (f):

f = f0 exp . (11.43)
Fugacity converges to the pressure (P) as P→0. We obtain the constant temperature
contribution to the surface potential:
f df
dΦ|T = −RTna d (ln ) = −RTna . (11.44)
f0 f
In terms of the fugacity, the differential of the surface potential becomes:
dΦ = −Sa dT − na df. (11.45)
Note that the integrated surface potential should have the following form:

Φ(f, T) = u(f) + v(T) + w(f, T). (11.46)

When f → P → 0, however, we expect that Φ(f, T) = μ − μs = 0 because the chemi-

cal potential of the system becomes equal to the chemical potential of the adsorbent
(μ → μs ). This implies that v(T) = 0, and thus Φ(f, T) = Φ|T .
260 | Part II Development of new materials for energy applications

Low pressure limit

At low pressures, f ∝ P (this is exact for an ideal gas). Integrating from P󸀠 = 0 to P we
find that:
Φ = −RT ∫ na (P󸀠 , T) . (11.47)

In the low pressure limit, Henry’s law states that the total adsorbed density is propor-
tional to pressure:
n(P󸀠 , T) = K(T)P󸀠 , (11.48)
where K (T) is Henry’s coefficient. The excess adsorbed density in this limit is:

na (P󸀠 , T) = K(T)P󸀠 − Vp ng = K(T)P󸀠 − Vp . (11.49)
In the low pressure limit, the surface potential behaves as follows:

Φ = − RTK (T) P + Vp P = −RTna (P, T) + Vp P. (11.50)

The Langmuir isotherm

As an illustrative example, we consider the excess Langmuir adsorption isotherm,
which can be written in the following general form, assuming an ideal gas:

KL (T)P󸀠 P󸀠
na (P󸀠 , T) = n(P󸀠 , T) − Vp ng (P󸀠 , T) = nm 󸀠
− Vp , (11.51)
1 + KL (T)P RT

KL (T) = exp (ΔS0 /R) exp(−ΔH0 /RT), (11.52)
where ΔS0 is the entropy difference relative to the reference pressure P0 (typically one
atmosphere), and ΔH0 is the enthalpy difference associated with the adsorption pro-
cess. The coefficient KL (T) corresponds to Henry’s coefficient. The corresponding sur-
face potential, for an ideal gas, is thus given by:

Φ = −RTnm ln(KL (T)P + 1) + Vp P. (11.53)

The chemical potential of the ideal gas is given by: