You are on page 1of 299

Springer Laboratory

Springer Laboratory Manuals in Polymer Science

Editor

Prof. Harald Pasch Prof. Harald Pasch


Chair of Polymer Characterization Deutsches Kunststoff-Institut
Department of Chemistry and Polymer Science Abt. Analytik
University of Stellenbosch Schlogartenstr. 6
Private Bag X1 64289 Darmstadt
7602 Matieland Germany
South Africa e-mail: hpasch@dki.tu-darmstadt.de
e-mail: hpasch@sun.ac.za

Editorial Board

PD Dr. Ingo Alig


Deutsches Kunststoff-Institut
Abt. Physik
Schlogartenstr. 6
64289 Darmstadt
Germany
e-mail: ialig@dki.tu-darmstadt.de
Prof. Josef Janca
Universit de La Rochelle
Pole Sciences et Technologie
Avenue Michel Crpeau
17042 La Rochelle Cedex 01
France
e-mail: jjanca@univ-lr.fr
Prof. W.-M. Kulicke
Inst. f. Technische u. Makromol. Chemie
Universitt Hamburg
Bundesstr. 45
20146 Hamburg
Germany
e-mail: kulicke@chemie.uni-hamburg.de

For other titles published in this series, go to


www.springer.com/series/3721
Joseph L. Keddie Alexander F. Routh

Fundamentals of Latex Film


Formation
Processes and Properties
Joseph L. Keddie Alexander F. Routh
Department of Physics & Department of Chemical Engineering
Surrey Materials Institute and Biotechnology & BP Institute
University of Surrey University of Cambridge
Guildford Cambridge
UK UK

Published by Springer,
P.O. Box 17, 3300 AA Dordrecht, The Netherlands
In association with
Canopus Academic Publishing Limited,
15 Nelson Parade, Bedminster, Bristol, BS3 4HY, UK

www.springer.com and www.canopusbooks.com

ISBN 978-90-481-2844-0 e-ISBN 978-90-481-2845-7


Springer Dordrecht Heidelberg London New York

Library of Congress Control Number: 2009940449

Canopus Academic Publishing Limited 2010


No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any
means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written per-
mission from the Publisher, with the exception of any material supplied specifically for the purpose of
being entered and executed on a computer system, for exclusive use by the purchaser of the work.

Printed on acid-free paper

Springer is part of Springer Science+Business Media (www.springer.com)


Preface

This book has emerged out of our long-time research interests on the topic of latex
film formation. Over the years we have built up a repertoire of slides used in
conference presentations, short courses and tutorials on the topic. The story
presented in this book has thereby taken shape as it has been told and re-told to a
mix of academic and industrial audiences.
The book presents a wide body of work accumulated by the polymer colloids
community over the past five decades, but the selection of examples has been
flavoured by our particular experimental interests and development of mathematical
models. We intend the book to be a starting point for academic and industrial
scientists beginning research on latex film formation. The emphasis is on fundamen-
tal mechanisms, however, and not on applications nor on specific effects of formula-
tions. We hope that the book consolidates the understanding that has been achieved
to-date in the literature in a more comprehensive way than is possible in a review
article. We trust that the reader will appreciate the fascination of the topic.
We are very grateful to the many students and post-docs whom we have had the
privilege of supervising, and from whom we have learned much about latex film
formation. Several of them have provided artwork for this book, as is noted in the
figure captions. Post-docs, in a rough chronological order of their time at the
University of Surrey, include Jacky Mallgol, Jean-Philippe Gorce, Chun-Hong Lei,
Diana Andrei, Yong Zhao, Elisabetta Canetta, and Carolina de las Heras Alarcn.
Post-docs at Cambridge include Venkata Gundabala, Grace Yow and Milan Patel.
PhD students and visitors in Surrey, who have studied latex film formation, are
Katerina Tzitzinou, Elisabetta Ciampi, Juan Salamanca, Peter Doughty, Nicki Kessel,
Philippe Vandervorst, Tao Wang, Tecla Weerakkody, Alexander Knig, Argyrios
Georgiadis, and Andr Utgenannt. A special mention goes to Tao Wang who contrib-
uted to the writing of Chapter 7. Students at Sheffield and Cambridge are Wai Peng Lee
and Venkata Gundabala and at Cambridge are Richard Trueman and Merlin Etzold.
Athene Donald and Richard Jones first introduced JLK to the topic of latex film
formation and provided guidance in his early studies. Paul Meredith and Ruth
Cameron were his first academic collaborators in studies of film formation at the
Cavendish Lab in Cambridge. The Soft Matter Group members at Surrey, includ-
vi Preface

ing Peter McDonald, Richard Sear, and Alan Dalton, and formerly Michele
Sferrazza and Paul Glover, have shared their expertise related to experimental
techniques and polymer colloids. AFR is immensely grateful to Bill Russel and
Brian Vincent for providing an education in colloid science and to Bill Russel for
the introduction to film formation.
We have both learned much from our industrial collaborators and benefited
from their insight. They include Panos Sakellariou, David Taylor, Peter Palasz,
Olivier Dupont, Keltoum Ouzineb, Peter Mills, Guru Satguru, Jrgen Scheerder,
Derek Illsley, Martin Murray, Simon Emmett, Philip Beharrell, John Jennings,
Tom Annable, Steve Yeates, Ad Overbeek, Malcolm Chainey, Ann-Charlotte
Hellgren, Peter Weisenborn, Ismo Pietari, Tuija Heijen, Andrew Howe and Stuart
Lascelles among others. Martin Murray, Bob Groves and Peter Mills have each
provided particular help in the research for this book.
For over a decade, we have enjoyed the camaraderie and scientific excellence
of the UK Polymer Colloid Forum meetings. We have had the privilege of useful
discussions with some of the world leaders in polymer colloids including Mitch
Winnik, Mohamed El-Aasser, and Pete Lovell.
Both of us have benefited from taking part in the European Commission
Framework 6 project, NAPOLEON. We have gained much from our academic
collaborators including the film formation team of Yves Holl, Diethelm
Johannsmann, Catherine Gauthier, Laurent Chazeau, Jrg Adams, and Anders
Larsson. We have had helpful input from other academics in the project, including
Txema Asua, Costantino Creton, Mariaje Barandiaran, Maria Paulis, Elodie
Bourgeat-Lami, Tim McKenna, and Katharina Landfester, Christopher Plummer,
and Richard Guy, among others. Industrial partners in the NAPOLEON project
include Simon Dennington, Bas Lohmeijer, Wolf-Dieter Hergeth, Rob Adolph,
and Dirk Mestach. These lists are not exhaustive. We apologise for any omissions.
This book would not be possible without the co-operation of numerous authors
and journals in providing permission to reprint materials here. We are grateful for
their generosity. The staff in the George Edwards Library at the University of
Surrey were very helpful in helping to track down a few obscure references.
We thank our valiant referee who offered us encouragement in the writing
process combined with insightful comments and many highly apt recommenda-
tions. We appreciate the support of our publishers at Canopus Academic Publish-
ing (Robin Rees and Tom Spicer) in not losing patience with us and waiting more
than three years for the manuscript.
On a personal note, JLK was given encouragement in writing this book by his
grandmother, Jean DiMatteo, who did not live to see its completion. JLK has
received moral support from many people throughout the years of writing this
book and throughout his career, especially from Adam Towner and from his
parents, Linda and Dave. AFR would not have been able to write this book
without the support of Ruth.

Joseph L. Keddie, Guildford


Alexander F. Routh, Cambridge
Contents

1 AN INTRODUCTION TO LATEX AND THE PRINCIPLES OF


COLLOIDAL STABILITY.......................................................................... 1
1.1 What is Latex? ..................................................................................... 1
1.2 Latex Synthesis and Uses..................................................................... 2
1.3 Historical Context and Economic Importance ..................................... 8
1.4 Overview of the Film Formation Process .......................................... 10
1.5 Environmental Legislation ................................................................. 15
1.6 Relevant Colloid Science ................................................................... 17
1.6.1 Interaction Potentials ........................................................... 17
1.6.2 Fluid Motion........................................................................ 22
References ..................................................................................................... 24

2 ESTABLISHED AND EMERGING TECHNIQUES OF STUDYING


LATEX FILM FORMATION.................................................................... 27
2.1 Techniques to Study Latex in the Presence of Water (Wet
and Damp Films)................................................................................ 28
2.1.1 Physical Probes of Drying ................................................... 29
2.1.2 Specialist Electron Microscopies......................................... 36
2.1.3 Scattering Techniques ......................................................... 42
2.1.4 Profiling Water and Particles with Spectroscopies.............. 52
2.1.5 Probe Techniques for the Aqueous Environment ................ 58
2.2 Techniques to Study Particle Packing and Deformation
in Dry Films ....................................................................................... 61
2.2.1 Scanning Probe Microscopies ............................................. 61
2.2.2 Scanning Near-Field Optical Microscopy (SNOM)
and Shear Force Microscopy ............................................... 70
2.2.3 Electron Microscopies ......................................................... 71
2.3 Techniques to Study Film Crosslinking ............................................. 73
2.3.1 Ultrasonic Reflection and QCM .......................................... 73
2.3.2 Spectroscopic Techniques ................................................... 73
viii Contents

2.4 Techniques to Study Interdiffusion and Coalescence ........................ 74


2.4.1 Small Angle Neutron Scattering (SANS) ............................ 75
2.4.2 Fluorescence Resonance Energy Transfer (FRET) ............. 76
2.4.3 Transmission Spectrophotometry ........................................ 83
2.5 Concluding Remarks.......................................................................... 83
References ..................................................................................................... 83

3 DRYING OF LATEX FILMS.................................................................... 95


3.1 Humidity and Evaporation ................................................................. 95
3.1.1 Background ......................................................................... 95
3.2 Evaporation Rate from Pure Water .................................................... 96
3.3 Evaporation Rate from Latex Dispersions ......................................... 98
3.4 Vertical Drying Profiles ................................................................... 99
3.4.1 Scaling Argument.............................................................. 101
3.4.2 Governing Equations ......................................................... 102
3.4.3 Experimental Studies......................................................... 104
3.4.4 Consequence of Inhomogeneous Vertical Drying:
Skin Formation .................................................................. 107
3.5 Horizontal Packing and Drying Fronts............................................. 107
3.5.1 Model for Horizontal Drying Fronts ................................. 110
3.5.2 Lapping Time and Open Time........................................... 111
3.6 Colloidal Stability ............................................................................ 114
3.7 Film Cracking .................................................................................. 116
3.7.1 Do the Cracks Follow the Drying Front or Propagate
Quickly Over the Entire Film? .......................................... 116
3.7.2 What Sets the Crack Spacing?........................................... 117
References ................................................................................................... 117

4 PARTICLE DEFORMATION................................................................. 121


4.1 Introduction...................................................................................... 121
4.2 Driving Forces for Particle Deformation ......................................... 122
4.2.1 Wet Sintering..................................................................... 123
4.2.2 Dry Sintering ..................................................................... 123
4.2.3 Capillary Deformation....................................................... 124
4.2.4 Capillary Rings.................................................................. 126
4.2.5 Sheetz Deformation ........................................................... 126
4.3 Particle Deformations ...................................................................... 127
4.3.1 Hertz Theory Elastic Spheres with an
Applied Load ..................................................................... 127
4.3.2 JKR Theory Elastic Spheres with an
Applied Load and Surface Tension ................................... 127
4.3.3 Frenkel Theory Viscous Spheres with Surface
Tension .............................................................................. 128
4.3.4 Viscoelastic Particles......................................................... 130
Contents ix

4.4 The Problem with ParticleParticle Approach ................................. 130


4.4.1 Routh and Russel Film Deformation Model...................... 130
4.5 Deformation Maps ........................................................................... 133
4.5.1 Wet Sintering..................................................................... 133
4.5.2 Capillary Deformation....................................................... 133
4.5.3 Dry Sintering ..................................................................... 133
4.5.4 Receding Water Front........................................................ 133
4.5.5 Use of the Deformation Maps ........................................... 134
4.6 Dimensional Argument for Figure 4.6 ............................................. 135
4.6.1 Wet Sintering..................................................................... 135
4.6.2 Capillary Deformation....................................................... 135
4.6.3 Dry Sintering ..................................................................... 136
4.6.4 Sheetz Deformation ........................................................... 136
4.7 Effect of Temperature ...................................................................... 137
4.8 Effect of Particle Size....................................................................... 139
4.9 Experimental Evidence for Deformation Mechanisms .................... 140
4.9.1 Inferring Deformation Mechanisms from Water
Distributions ...................................................................... 140
4.9.2 Determination of Deformation Mechanisms Using
an MFFT Bar and Optical Techniques .............................. 143
4.9.3 Microscopy of Particle Deformation ................................. 143
4.9.4 Scattering Techniques ....................................................... 146
4.9.5 Detection of Skin Formation ............................................. 146
References ................................................................................................... 146

5 MOLECULAR DIFFUSION ACROSS PARTICLE


BOUNDARIES .......................................................................................... 151
5.1 Essential Polymer Physics................................................................ 153
5.1.1 Interface Width at Polymer-Polymer Interfaces ................ 153
5.1.2 Polymer Reptation ............................................................. 154
5.2 Development of Mechanical Strength and Toughness ..................... 158
5.2.1 Dependence on the Density of Chains Crossing
the Interface....................................................................... 162
5.2.2 Dependence on Interdiffusion Distance, ........................ 162
5.3 Factors that Influence Diffusivity .................................................... 164
5.3.1 Molecular Weight and Chain Branching ........................... 164
5.3.2 Temperature Dependence .................................................. 165
5.3.3 Influence of Hard Particles ................................................ 168
5.3.4 Latex Particle Size............................................................. 172
5.3.5 Particle Structure and Hydrophilic Membranes................. 172
5.4 Faster Diffusion with Coalescing Aids ............................................ 174
5.5 Simultaneous Crosslinking and Diffusion: Competing
Effects .............................................................................................. 175
References ................................................................................................... 179
x Contents

6 SURFACTANT DISTRIBUTION IN LATEX FILMS.......................... 185


6.1 Introduction...................................................................................... 185
6.1.1 Where Can Surfactant Go in a Dried Film?....................... 186
6.1.2 Effect of Non-Uniform Surfactant Distributions ............... 188
6.1.3 Mechanisms of Surfactant Transport................................. 191
6.2 Adsorption Isotherms ....................................................................... 192
6.3 Modelling of Surfactant Distribution during the Drying Stage........ 194
6.4 Effect of Surfactants Vertical Distribution on Film Topography ... 199
6.5 Experimental Evidence for Surfactant Locations............................. 201
6.5.1 Interfaces with Air and Substrates..................................... 201
6.5.2 Surfactant in the Bulk of the Film ..................................... 202
6.5.3 Depth Profiling and Mapping ............................................ 202
6.6 Reactive Surfactants......................................................................... 204
6.6.1 Reactive Surfactant Chemistry .......................................... 205
6.6.2 Effect of Surfmers on Film Properties............................... 205
6.7 Summary .......................................................................................... 207
References ................................................................................................... 207

7 NANOCOMPOSITE LATEX FILMS AND CONTROL OF


THEIR PROPERTIES.............................................................................. 213
7.1 Introduction...................................................................................... 213
7.1.1 Properties of Nanocomposites ........................................... 214
7.1.2 Applications of Colloidal Nanocomposites ....................... 216
7.2 Types of Hybrid Particles................................................................. 217
7.2.1 Polymer-Polymer Hybrid Particles.................................... 217
7.2.2 Inorganic and Polymer Nanocomposite Particles.............. 219
7.2.3 Self-Assembly of Nanocomposite Particles by
Precipitation or Flocculation of Pre-Formed
Nanoparticles..................................................................... 223
7.3 Colloidal Particle Deposition and Assembly Methods..................... 225
7.3.1 Deposition Methods........................................................... 227
7.3.2 Vertical Deposition............................................................ 229
7.3.3 Surface Pattern-Assisted Deposition ................................. 230
7.3.4 Long-Range Order from Self-Assembled Core-
Shell Particles .................................................................... 232
7.4 Colloidal Nanocomposites from Particle Blends ............................. 233
7.4.1 Advantages of Particle Blends........................................... 233
7.4.2 Dispersion of Nanoparticles .............................................. 233
7.4.3 Long-Range Order in Particle Blends ............................... 235
7.5 Three Lessons about the Properties of Waterborne
Nanocomposite Films ...................................................................... 238
7.5.1 Lesson One ........................................................................ 238
7.5.2 Lesson Two ....................................................................... 244
7.5.3 Lesson Three ..................................................................... 245
References ................................................................................................... 249
Contents xi

8 FUTURE DIRECTIONS AND CHALLENGES .................................... 261


8.1 Film Formation from Anisotropic Particles ..................................... 261
8.2 Assembly of Particles over Large Length Scales ............................. 263
8.3 Technique Development .................................................................. 265
8.4 Nanocomposite Structure and Property Correlations ....................... 265
8.5 Interdiffusion of Polymers in Multiphase Particles.......................... 267
8.6 Templating Film Topography .......................................................... 268
8.7 Resolving the Film Formation Dilemma.......................................... 269
References ................................................................................................... 272

APPENDICES

A Derivation of Creeping Flow and the Result for Low


Reynolds Number Flow Around a Sphere............................................... 275
A.1 Derivation of Creeping Flow ........................................................... 275
A.2 Scaling of the Navier-Stokes Equation ............................................ 276
A.3 Stokes Flow...................................................................................... 277
A.4 Sedimentation................................................................................... 277

B GARField Profiling Techniques and Experimental


Parameters ................................................................................................. 279
References ................................................................................................... 281

C Terminology of Humidity and an Expression for


Evaporation Rate....................................................................................... 283
C.1 Humidity .......................................................................................... 283
C.2 Relative Humidity ............................................................................ 283
C.3 Dry Bulb Temperature ..................................................................... 284
C.4 Wet Bulb Temperature..................................................................... 284
C.5 Specific Volume............................................................................... 284
C.6 Enthalpy of Air................................................................................. 285
C.7 Psychrometric Chart......................................................................... 285
C.8 Dew Point......................................................................................... 286
C.9 Relating Humidity to Partial Pressure .............................................. 286
Example 1.................................................................................................... 286
Example 2.................................................................................................... 287
Example 3.................................................................................................... 288
Example 4.................................................................................................... 290
Example 5.................................................................................................... 291
C.10 Evaporation Rate.............................................................................. 292
References ................................................................................................... 294
xii Contents

D Fracture Mechanics: Terminology and Tests ......................................... 295


D.1 Fracture Toughness, KIC ................................................................... 295
D.2 Plastic Zone Size at the Crack Tip, ry .............................................. 297
D.3 Critical Energy Release Rate, Gc ..................................................... 298
D.4 Fracture Strength.............................................................................. 298
D.5 Fracture Energy................................................................................ 299
References ................................................................................................... 299

INDEX................................................................................................................ 301
Symbols

Generic symbols used throughout this book

Symbol SI units Meaning


E ms -1
Evaporation rate (units of velocity)
H m Film thickness (wet)
R m Particle radius
P N m-2 Pressure
T K Temperature
Tg K Glass Transition Temperature
N s m-2 Dispersion viscosity
N s m-2 Solvent viscosity
Particle volume fraction
m Close packed volume fraction

Chapter 1

A Nm = J Hamaker Constant
D0 m2 s-1 Stokes-Einstein Diffusion coefficient
kT Nm = J Thermal energy
R m Particle radius
r m Spacing between particle centers
U Nm = J Interaction potential
Permittivity
0 A2 s4 kg-1 m-3 Permittivity of free space
1 m Debye length
V Electrostatic potential on particle surface
xiv Symbols

Chapter 2

A0, Asp m Tapping amplitudes for AFM


B T ( or V s m-2) Magnetic field strength
dI m Interpenetration distance
D0 m2 s-1 Stokes-Einstein diffusion coefficient for a particle
Dapp m2 s-1 Apparent molecular self-diffusion coefficient
dp m Path length for light through sample
db m Beam thickness
ep m Average spacing between particles
Ed Nm = J Energy dissipation in tapping mode AFM
Ep N m-2 Storage modulus (obtained from DWS)
f s-1 frequency
g2(t) Correlation function for light scattering
G T m-1 (V s m-3) Gradient in the magnetic field
G' G'' N m-2 Shear moduli (storage and loss)
hunit m Height of fcc/hcp unit cell
Hdry m Average thickness of dry film
I Intensity of radiation
I0 Intensity of incident radiation
IB Background scattering intensity
KP Porod constant
KB Constant of proportionality
L m Length of beam
L', L'' N m-2 Longitudinal moduli
m Ratio of refractive indices
n Refractive index
np Refractive index of particle
Q m-1 Scattering wave vector
q Quality factor for AFM cantilever
<r2> m2 Mean-squared displacement
Rf m Frster radius
Rg m Radius of gyration (particle or molecule)
s s-1 Speckle rate
t s time
tD s Fluorescence decay time
t0 s Initial time
Tf - Transmitted fraction of radiation
w s Interfacial width between two phases
Y N m-2 Youngs modulus (of beam)
z m Void radius
Constant of proportionality (for DWS)
Exponent in stretched exponential
s-1 T-1 Magnetogyric ratio
m Wavelength of radiation
min m Wavelength of minimum transmitted radiation
Poissons ratio
Scattering angle
f N m-2 Change in film stress
s Auto-correlation time
Symbols xv

Angle of deflection of beam


s-1 Angular frequency

Chapter 3

D m2 s-1 Diffusion coefficient


E m s-1 Evaporation rate (units of velocity)
H m Film thickness (wet)
Hdry m Film thickness (dry)
km m s-1 Mass transfer coefficient
kp m-2 Permeability of particle bed
K() Sedimentation Coefficient
L m Capillary length
Lb m Boundary layer thickness
P N m-2 Pressure
P* N m-2 Characteristic pressure
Pcap Dimensionless capillary pressure
Pe Peclet number
tevap s Time for evaporation
tdiff s Time for diffusion
y m Vertical distance
Z() Compressibility
Particle volume fraction
N m-1 Surface tension

Chapter 4

a0 m Radius of contact between two particles


E N m-2 Youngs modulus
F N Force pushing particles together
G N m-2 Shear modulus
m Approach of particle centers
R Strain along particle centers
Strain
wa N m-1 Water air interfacial tension
pw N m-1 Polymer water interfacial tension
pa N m-1 Polymer air interfacial tension
0 N s m-2 Polymer low shear viscosity
G Dimensionless group comparing polymer relaxation time
to evaporation time
Dimensionless group controlling particle deformation
Poisons ratio
t Dimensionless stress at top of film
Angle of contact between sintering spheres
xvi Symbols

Chapter 5

aT Time-temperature superposition shift factor


b m Kuhn length for polymer segment
C1, C2 -, K WLF equation factors
Db m-2 s-1 Fickian diffusion coefficient
Ea J mol-1 Activation energy
Gc J m-2 = N m-1 Critical energy release rate
KIC N m-3/2 Fracture toughness
M Polymer molecular weight
Me Entanglement molecular weight
N Degree of polymerisation
m Interpenetration distance
f N m-2 Failure stress
m-2 Density of polymer chains crossing interface
e s Rouse entanglement time
R s Rouse relaxation time
d s Reputation time
XL s Time for cross linking reaction
wI m Interfacial width for interdiffusion
Flory Huggins interaction parameter

Chapter 6

mol m-3 Critical concentration in Langmuir isotherm


Cs mol m-3 Surfactant concentration in solution
Dp m2 s-1 Particle diffusion coefficient
Ds m2 s-1 Surfactant diffusion coefficient
k, k' s-1 Reaction rate constants for surfactant adsorption and
desorption
mol m-2 Adsorbed amount in Langmuir isotherm
mol m-2 Maximum adsorbed amount
Pep Particle Peclet number
Pes Surfactant Peclet number

Chapter 7

D m Filler diameter
Lc m Critical length of inclusion
f N m-2 Fracture strength
N m-2 Interfacial shear strength
Symbols xvii

Appendix A

Fdrag N Drag force


G m s-2 Acceleration due to gravity
P N m-2 Pressure
P N m-2 Characteristic pressure
N m-2 Difference in density
kg m-3 Fluid density
U m s-1 Velocity
U* m s-1 Characteristic velocity

Appendix B

s-1 Pulse bandwidth


y m Pixel resolution in a profile
s-1 T-1 Magnetogyric ratio
Gy T m-1 Magnetic field gradient
Number of echoes in a train
Nacq Number of points acquired in an echo
NS Numbers of scans
D s Time interval between points in an echo
tpd s Time duration of an RF excitation pulse
s Pulse gap (delay time between RF pulses)
R s Repetition delay time (before sequence repeats)
T1 s Spin-lattice relaxation time
T2 s Spin-spin relaxation time

Appendix C

kg/kg Humidity
Mw kg mol-1 Molar mass
Pw N m-2 Partial pressure of water
P N m-2 Pressure
R J mol-1 K-1 Gas constant
V m3 Volume
xviii Symbols

Appendix D

a m Crack length
m Thickness of sample
E N m-2 Elastic (Youngs) modulus
Gc J m-2 = N m-1 Energy release rate
h, b m Specimen dimensions
K1 N m-3/2 Stress intensity factor
KIC N m-3/2 Fracture toughness (critical KI)
P N Applied load
ry m Radius of crack tip
N m-2 Applied stress
WB J Fracture energy
Y Geometric factor
Chapter 1
1 An Introduction to Latex and the Principles of
Colloidal Stability

1.1 What is Latex?

Latex is an example of a colloidal dispersion. It consists of polymeric particles,


which are usually a few hundred nanometres in diameter, dispersed in water. The
particles typically comprise about 50 percent by weight of the dispersion. Depend-
ing on the particular application, there will also be a complex mixture of pigments,
surfactants, plasticising aids and rheological modifiers. The whole dispersion is
colloidally stable, meaning that it can sit on a shelf for years and remain dispersed,
without sedimentation of particles making sludge at the bottom. In this book, the
word latex will be used as shorthand for a wet dispersion. Sometimes, however,
latex is used as an adjective, as in latex film. The plural of latex is latices, not
to be confused with lattices! (An alternative is to say latexes, but latices will
be used through this book.)
Although colloids have been used since ancient times in Egypt and Greece,
colloidal dispersions have been studied scientifically only since the nineteenth
century when Thomas Graham first coined the term colloid. It is derived from
the Greek work coll, meaning glue. A major advance in colloid science was
made by Robert Brown, who observed pollen particles dancing randomly when
dispersed in a fluid. This phenomenon is now termed Brownian motion and is a
defining characteristic of a colloidal dispersion. The effect is the result of an
imbalance of forces between the colloidal particles and the surrounding molecules
of the solvent, and it provides evidence for the existence of molecules. A disper-
sion in which Brownian motion is present is described as colloidal. An alternative
definition, from the International Union of Pure and Applied Chemistry (IUPAC),
is that colloidal particles have at least one dimension between 1 nm and 1 m
(IUPAC 2009).
2 1 An Introduction to Latex and the Principles of Colloidal Stability

Fig. 1.1 An emulsion polymerisation reactor. The structure spans an entire building and covers two
floors. The top of the reactor houses the stirring motor. (Photographs courtesy of AkzoNobel)

1.2 Latex Synthesis and Uses

Industrially and in the laboratory, latex is most often made by a reaction called
emulsion polymerisation. The process can be scaled up to large quantities, such
as is demonstrated by the large reactors in Fig. 1.1. The scale of the structure,
which spans two storeys, is indicated by the top two photographs. The top of the
reactor and the stirring motor are shown in the bottom photo. The emulsion
polymerisation process and newer developments, such as miniemulsion poly-
merisation, are described in numerous sources elsewhere (Lovell and El-Aasser
1997). A discussion of the polymerisation process is beyond our scope here.
Latex may be synthesised from a range of monomers, the typical ones being
acrylates (methyl methacrylate, butyl acrylate, ethylhexyl acrylate), styrene, vinyl
acetate and butadiene. Copolymers are used extensively. The molar ratio of the
monomers in the copolymer determines its glass transition temperature (Tg), the
point at which a solid polymer changes into a liquid-like polymer, because long
1.2 Latex Synthesis and Uses 3

range motion of the molecules backbone is enabled. A polymers viscosity falls


sub-exponentially as the temperature increases above the Tg. As we shall see in
our discussions in Chapter 5, the temperature at which a latex film is formed
should be greater than the (co)polymers Tg.
Anyone who coats a house with a modern, waterborne paint uses latex to their
benefit. The largest household use of latex is for architectural or decorative paints,
which typically comprise acrylic or styrene-acrylic copolymers. Most readers will
be familiar with commercial latex paints, such as those shown in Fig. 1.2.

Fig. 1.2 Latex paint is marketed in the UK and other countries under the trade name of Dulux, as
seen in this advertising material, showing a well-recognised mascot. (Photographs courtesy of
AkzoNobel)
4 1 An Introduction to Latex and the Principles of Colloidal Stability

Table 1.1 Typical formulations for gloss, silk and matte paints.
Gloss Paint
Silk Paint Matte Paint
(Schuler et
Component Function (Murray (Murray
al. 2000)
2009) (wt.%) 2009) (wt.%)
(wt.%)
Polymer particles Acts as a binder for the
26.5 20 15
(latex) pigment
Pigment to create
Titanium dioxide 23.0 16 10
opacity
Prevents bubble
Defoamer 0.15
formation during use
Thickener Adjusts viscosity 12.0
Stabilizes the pigment
Pigment dispersant
in the water phase 1.0
3
Lowers the film 3
(total for all
formation temperature (total for all
Coalescent 2.35 four)
and increases the four)
interdiffusion rate
Inexpensive filler,
Extender typically minerals such 2 32
as calcium carbonate
Water Carrier for the wet paint 35.0 59 40

Household paints and varnishes are usually applied over relatively small areas
and allowed to dry. Viscosity control is particularly important for paints. Under a
low shear rate, such as when coated vertically on a wall, the paint should not flow or
drip: the viscosity should be high. When applied with a brush at a higher shear rate,
the viscosity should be low to allow even spreading. Hence, rheology modifiers are
a key ingredient of commercial paints. Shear thinning behaviour, meaning that the
viscosity decreases as the shear rate increases, is desirable to ensure good spreading
with a brush or applicator. Film thicknesses for household paints are typically one
hundred micrometers. Typical formulations for three types of commercial emulsion
(i.e., latex) paint are shown in Table 1.1, and other recipes are found in literature
(Heldmann et al. 1999). Gloss paints are highly reflective of light, matte paints are
poorly reflective, and silk paints have an intermediate reflectivity.
You might be surprised to see that the latex polymer makes up less than 30% by
weight in the paint formulations. Note the presence of rheology modifiers and
defoamers. Wetting agents are also often added to ensure full coverage of hydro-
phobic surfaces. Extenders, which are inexpensive minerals used as fillers, are added
to matte paints to extend how far the more expensive polymers can go. Extenders
introduce light scattering and so are not added to gloss paints.
Increasingly, latex is being used as a binder in industrial coatings (applied in
factory production) and various types of protective coatings for automobiles, aircraft,
ships, appliances and more. The demands on industrial and protective coatings are
high and go beyond improving the appearance of an object. The coatings need to
provide resistance against corrosion, abrasion, wear, fire and chemical attack.
1.2 Latex Synthesis and Uses 5

Beyond paints and coatings, the uses for latex include pressure-sensitive adhe-
sives, sealants, carpet backings, construction additives (such as in cements and
mortars), paper coatings, inks, and latex gloves and condoms. Each will be
considered briefly here.
Pressure-sensitive adhesives (PSAs) are permanently tacky at room tempera-
ture, and provide permanent and nearly instant adhesion to almost any surface.
They are well-known in their application as tapes and sticky labels. Pressure-
sensitive adhesives have many specific applications in the household (Fig. 1.3)
and in cars (Fig. 1.4). They typically have a glass transition temperature that is
around 50oC, and hence are made from monomers such as butyl acrylate or 2-
ethyl hexyl acrylate to reduce the copolymers Tg.
Industrially, films are deposited onto long rolls of substrate material running at
high speeds through industrial coaters. Most often the transfer coating process is
used for labels. A sandwich structure, as in Fig. 1.5a, is made. The latex is spread
on a release liner with a low-surface-energy surface, such as a silicone polymer,
and then film-formed. The rate of drying is increased by drawing the film on the
backing layer through a long oven or by blowing hot air across it. The dry film
thickness for applications is typically 20 m. Next, a face material (or facestock)
is pressed into contact with the latex adhesive surface. The air surface of the
facestock is decorated appropriately for the adhesives application. To use the
adhesive, it is peeled off from the release liner and then pressed onto the surface of
choice, such as a beer bottle. The final film should be uniform and not have
defects, such as pinholes or surface ripples.

Fig. 1.3 Uses of pressure-sensitive adhesives in the home. (Image reproduced with permission of
Cytec Specialty Chemicals)
6 1 An Introduction to Latex and the Principles of Colloidal Stability

Fig. 1.4 Uses of pressure-sensitive adhesives in a car. (Image reproduced with permission of
Cytec Specialty Chemicals)

In a typical industrial coater, as in Fig. 1.5b, a steel roll is rotated through a reser-
voir of wet latex, picking it up as it rotates. The backing layer is then drawn at high
speed across the rotating wet roll to deposit a thick film. The excess latex in the film
is skimmed off by passing a grooved bar (called a Meyer rod) across it.
Another process, called gravure, uses a roll that has grooves engraved in it.
When the gravure roll is passed through a reservoir of latex, its grooves pick up
some of the latex. The roll is then pressed in contact with a backing layer onto
which it leaves a wet film. Higher coating speeds are obtained with the gravure
process compared to most other processes. Depending on the process, coating
speeds can range from 100 to over 1000 meters per minute.
Textiles and carpet backings are often made from acrylic polymers. The coat-
ing provides durability to the material surface as well as a resistance to a number
of solvents (Campos et al. 2006).
Construction materials often have hydrophobic polymer latices added to them.
These reduce the water permeability of the final material and will lead to a life
extension from reduced corrosion (Stancato et al. 2005).
Paper coatings use latex as a binder to enable the incorporation of clays and
inorganic fillers. Latex coatings on paper are cheaper than wood pulp and also
allow greater penetration of ink during the printing process (Zang and Aspler
1995, Alturaif et al. 1995).
Inks for printing on paper and card may use latex as a binder for the pigment or
dye particles (Hutton and Parker 2008). There is growing interest in water-based
inks to replace solvent-based materials. In inkjet printing, colloidal dispersions are
used in inks (Calvert 2001), and the use of latex particles has been demonstrated
in an inkjet printer (Wong et al. 1988).
1.2 Latex Synthesis and Uses 7

(a)

(b)

Fig. 1.5 Industrial coating of pressure-sensitive adhesive films. a In the transfer process, the latex
adhesive 3 is deposited onto a silicone release coating 2 on a release liner 1. Then the face
material 4 is laminated onto the adhesive surface. Image courtesy of Cytec Specialty Chemicals.
b An industrial coater. (Photograph used with permission of Cytec Specialty Chemicals)
8 1 An Introduction to Latex and the Principles of Colloidal Stability

Fig. 1.6 Latex gloves are being made by dipping moulds into a natural latex dispersion and
withdrawing. (Photograph used with the kind permission of Innovative Gloves Co. Ltd,
Songkhla, Thailand)

Latex gloves are made by dipping a hand-shaped, ceramic mould into a latex
dispersion and drying, such as shown in Fig. 1.6. Condoms are manufactured
similarly but of course with a different mould.

1.3 Historical Context and Economic Importance

There is a natural form of latex that is derived from a tree (Hevea brasiliensis),
which as its Latin name implies originated in Brazil (Fig. 1.7). Plantations
were developed in several Asian countries to support the growing rubber
industry in the twentieth century. Natural latex is essentially cis-isoprene
emulsified by a protein. During World War II, the supply of this latex from
Malaysia was disrupted, and the resulting shortages spurred the development of
synthetic latex made via emulsion polymerisation. The first commercially
available latex was introduced by the Glidden Company in 1948. Consequently,
a large synthetic latex industry grew through the remainder of the century. A
second important factor is that some persons have an allergy to the protein in
natural latex; about one in 10 persons show sensitisation (Turjanmaa et al.
1996, Liss et al. 1997). This drawback of natural latex has encouraged the
growth of the synthetic latex market as an alternative, especially for applica-
tions in gloves.
1.3 Historical Context and Economic Importance 9

Fig. 1.7 The white dripping liquid is the sap of the Hevea Brasiliensis tree, which provides the
natural form of latex. It is being collected in a bowl at the bottom of the tree for later use in the
manufacture of natural rubber objects. (iStockPhoto)

The market for emulsion polymers is large and growing, with worldwide sales
in 2007 of 17.9 billion US dollars. The vast range of industries that use polymeric
coatings makes the economic influence of the coatings industries particularly
immense. For example, the paper industry creates 91 million tonnes of paper in
the EU annually, using coatings to provide a brilliant white shine. The annual
market for paper coatings in the EU is 800 million Euros (Urban and Takamura
2002). Adhesives and sealants are used in more than 100 end-use product markets,
and they had a demand of four million tonnes in the European region in 2008. A
growth of 2.5% is predicted between 2008 and 2011. Of the entire European
market for adhesives, valued at 11.3 billion Euros (15.8 billion US dollars), a
share of 37 % was held by water-based adhesives, which are made from latex (von
Dungen 2009). The water-based sector is growing at the expense of other manu-
facturing processes that emit organic solvents.
In 2008, there were 135 billion latex gloves manufactured and the growth rate
is around 10% per annum (Tan 2008). In response to the allergic reaction of some
people to natural latex, the US market for nitrile latex gloves increased from
18.6% in 2006 to 26.8% in 2007, with the market share of natural rubber falling
from 48% to 41% (Tan 2008). The biggest exporter of rubber gloves is Malaysia.
In 2007, the total number of exported pairs of gloves was 41.7 billion pairs with a
sale value of 5.9 billion Malaysian dollars (1.72 billion US dollars). This corre-
sponds to approximately one percent of the Malaysian GDP (Department of
Statistics 2008). For textiles, the annual consumption in 2001 was 160,000 dry
10 1 An Introduction to Latex and the Principles of Colloidal Stability

tonnes of polymer, and for carpet backings it was 500,000 dry tonnes of latex in
1999 (Urban and Takamura 2002).
Latex is most associated with water-based paints used to paint the inside and
outside of buildings and houses, referred to as architectural paints. It is predicted
that as a result of tough environmental legislation 88% of all architectural
paints sold in Europe in the year 2011 will be water-based, using latex polymers
as their binder. Globally, 73% of the market share of architectural paints is
predicted to be water-based in that same year. The production of all architectural
paints is predicted to reach 21 million metric tonnes in 2011 and to have a value of
47 billion US dollars. Despite a harsh economic climate, an annual global growth
rate of nearly 4% is predicted for 20062011 (European Coatings Journal 2008).

1.4 Overview of the Film Formation Process

The process of transforming a stable dispersion of colloidal polymer particles into


a continuous film is called latex film formation. It involves many steps that span
from a dilute through to a concentrated dispersion, into a packed array of particles,
and eventually into a continuous polymer film. From a modelling perspective
(and for drawing cartoons!), it is conventional to split the process into three
sequential steps (drying, particle deformation, and diffusion). But, as will be
discussed in later chapters, the steps can overlap in time. There is a large literature
examining the mechanism of film formation with many excellent reviews and
commentaries (Steward et al. 2000, Winnik 1997, Dobler and Holl 1996, Winnik
1997b, Keddie 1997).
Crucially, the film formation process, sketched in Fig. 1.8, has a pronounced
influence on the final film properties. A recurring theme in this book is how the
process and properties are interrelated. When a stable dispersion (state 1) is
deposited on a surface and subject to evaporation, the particles consolidate into
some form of close packing (state 2). The latex dispersion typically does not dry
in a uniform manner across the film. As Fig. 1.9 illustrates, the edges often dry
first. In this illustration, light scattering by the particles in water makes the wet
regions turbid. There is a possibility of colloidal crystallisation (i.e. the formation
of an ordered array of particles) if the drying is slow enough. More likely, the
particles will collect in the form of a random close-packing with a volume fraction
in the region of 0.64 for mono-sized particles (Russel 1990, Routh and Russel
1998, 2001). This means that the spherical particles occupy 64% of the space and
the remainder is occupied by water. An image of an array of particles in close
packing is shown in Fig. 1.10a. Here, the individual particles make the topography
appear like a mountain range with the peak-to-valley distance corresponding to the
particle radius. The drying step is considered in detail in Chapter 3.
1.4 Overview of the Film Formation Process 11

Polymer-in-water dispersion
State 1
State 1
Close-packing of particles
State 2

1. Water loss

100 nm
2. Deformation
T> MFFT of particles

3. Interdiffusion Optical Clarity


and coalescence

T > Tg
Homogenous Film Dodecahedral structure
State 4 (honey-comb)
State 3

Fig. 1.8 Schematic of the process of film formation: a colloidal dispersions transition into a
continuous polymer film. (Drawing courtesy of Jacky Mallgol, University of Surrey)

0 min. 40 min. 90 min.


Fig. 1.9 Drying of 200 nm polystyrene particles in water. The particles pack and consolidate at
the edge, and these packed particle fronts propagate laterally across the film toward the centre.
(Photographs courtesy of Wai Peng Lee)

When the particles come into close contact, they will deform from their spheri-
cal shape to fill the void space around them. As the individual particles are
deformed, they can remain as distinct objects (state 3 in Fig. 1.8). With the loss of
the interparticle voids, the particle layer becomes optically transparent, because
light is no longer scattered by heterogeneities in the refractive index. The onset of
transparency is sometimes used to define the point of film formation.
12 1 An Introduction to Latex and the Principles of Colloidal Stability

(a)

(b)

Fig. 1.10 a AFM topographical image of a latex film surface in a three dimensional view. The
film was cast at room temperature, and since the polymers glass transition temperature is about
20 C, the particles do not flatten much. Scan area is 1 m x 1 m. b When viewed from above,
the high degree of order in the particle packing in a hexagonal array can be seen. (Images
courtesy of J. Mallgol, University of Surrey)

An atomic force microscope (AFM) image of deformed particles is shown in


Fig. 1.11a. The particles are no longer spherical, and only small voids are seen at
the particle boundaries. There are numerous possible driving forces for the particle
deformation, which are discussed at length in Chapter 4 (Brown 1956, Henson et
al. 1953, Vanderhoff et al. 1966, Sheetz 1965, Routh and Russel 1999, Dobler and
Holl 1996). The resistance to deformation comes from the particles themselves,
and hence the temperature, relative to the glass transition temperature, is a crucial
parameter in determining the extent of particle deformation.
If mono-sized particles are packed into a face-centred cubic array, each particle
(in the bulk of the film) will be in direct contact with twelve nearest neighbours.
When the contact regions between the particles flatten, the particles will each
create a twelve-sided geometric figure called a rhombic dodecahedron, drawn
in Fig. 1.11b. In a face-centred cubic crystal, each particle has six neighbours
1.4 Overview of the Film Formation Process 13

hexagonally arranged around it in the (111) plane (Fig. 1.11a). If a thick film of
deformed particles is sliced along a (111) plane, the particle cross-sections will be
hexagonal. The array will take on an appearance similar to natural honeycomb
(Fig. 1.11c), such as is shown schematically in State 3 in Fig. 1.8.

(a)

(b)

(c)

Fig. 1.11 a Example of flattening at particle/particle boundaries. Particle identity is still retained.
Size of AFM image is 1.5 m along each side. Image courtesy of A. Tzitzinou (University of
Surrey). b Two different views of a rhombic dodecahedron. There are six rhomboids on the
vertical faces. There are six square faces, three each on the top and bottom. c A cross-section
slice of the dodecahedral structure in a film shows a hexagonal array, which is reminiscent of
natural honeycomb. (iStockPhoto)
14 1 An Introduction to Latex and the Principles of Colloidal Stability

If the particles are too hard, then they will not be able to deform. The film will be
cloudy and cracked, as seen in Fig. 1.12a. It will be brittle and possibly powdery.
The polymer can be softened by heating to temperatures above its Tg. Then, the film
will be less brittle and will achieve optical transparency (Fig. 1.12b). The lowest
temperature at which optical transparency may be achieved is referred to as the
minimum film formation temperature or MFFT. It is an important characteristic of
latex formulation because it determines the conditions under which the latex can be
successfully applied. If hard particles are blended with softer particles, film forma-
tion may still occur (Fig. 1.12c). The softer particles will create the film, and the
harder particles will be dispersed throughout it.

(a) (b)

(c)

Fig. 1.12 a Latex Tg 80C; Film cast at room temperature; Film is opaque and has regular
cracking patterns. b Latex Tg 80C; Film formation at 150C for 10 min.; Film is cracked but is
now translucent, but note the haziness. c A blend of latex particles with Tg 80C (as in (a) and
(b)) and Tg 50C; Film formation at room temperature (21C). Film is smooth and transpar-
ent. (Photographs courtesy of Tao Wang, University of Surrey)

When the particles are deformed, their surfaces come into close contact over
large areas. At temperatures above the polymers Tg, the molecular chains within
the particles will move across the boundaries between particles. This diffusion
blurs the boundaries between individual particles and leads to a continuous film
with increased mechanical strength (State 4 in Fig. 1.8). The process by which
long polymer chains diffuse in a molten material is called reptation. It has been
shown that full mechanical strength is obtained once reptation has progressed by
the distance comparable to the polymers radius of gyration (Prager and Tirrell
1981, Richard and Maquet 1992). The diffusion process is considered in detail in
Chapter 5.
1.5 Environmental Legislation 15

While the different stages all occur sequentially for a local region of film, they
may all be occurring simultaneously in a whole film. For example, a film may have
regions that are still fluid (in State 1) and a further region that is fully dry and
interdiffused (State 4) with transitions between these two extremes (States 2 and 3).
Experimentally, drying fronts may be observed passing laterally across films thus
indicating a number of states (Routh et al. 2001). The physical reasons for drying
fronts and the consequences of non-uniform drying are explored in Chapter 3.

1.5 Environmental Legislation

In the past, polymers have been deposited from solution in organic solvents for
coatings applications as well as in various industrial processes. Small organic
molecules that vaporise and are emitted into the atmosphere are called volatile
organic compounds or VOCs. There are both environmental and health and safety
driving forces to reduce VOC levels in coatings and other products. Environmen-
tally, it has found that some VOCs damage the ozone layer, whereas others may
contribute to the greenhouse gases linked to global warming. Organic solvents
that have historically been used in coatings produce toxic oxidants that are
damaging to the atmosphere (Wicks et al. 1992). In workplaces that use solvent-
borne products, the odour may be unpleasant and long-term exposure to solvents
has raised health concerns for workers, especially professional painters (Bockel-
mann et al. 2002, 2003, 2004, Steinhauer et al. 2001, Morrow and Steinhauer
1995). Consumers increasingly prefer household products that do not emit VOCs.
Governments have responded by introducing regulations and pollution preven-
tion programmes (DeVito 1999). In Europe, the relevant legislation is the EU
Directive 2004/42/EC, and it has been implemented into British law with The
Volatile Organic Compounds in Paints, Varnishes and Vehicle Refinishing
Products Regulations 2005. For the purposes of the EU Directive, a VOC is
classified as an organic compound having a boiling point lower than or equal to
250C at an atmospheric pressure of 101.3 kPa. The specific limits on VOC level
depends on the application. The law comes fully into force in 2010, although the
VOC limits were reduced in 2007. For instance, for matt indoor paint, the VOC
limit for a water-based paint was 75 g per litre in 2007 but it will fall to 30 g per
litre in 2010.
Most coating manufacturers are striving to reduce their VOC levels below the
limits, not only to comply with the legislation, but also to achieve a competitive
advantage with the inevitable tightening of future legislation and to satisfy
consumer demands. Having a low VOC content can make a product more attrac-
tive to consumers and so add to its value (Dennington 2007). Many countries have
introduced special labelling schemes to identify environmentally sound products.
Workers in the manufacturing industries (such as in automotive finishing) are
protected by laws that set occupational exposure limits to restricted chemicals
(Jotischky 2000). These laws influence how chemicals are used in the workplace.
16 1 An Introduction to Latex and the Principles of Colloidal Stability

A key driver in the growth of the latex market has been the transformation of
solvent-cast polymer processing to waterborne processes. A few decades ago, it
was commonplace to use household paints that were solvent-based. Brushes were
cleaned with mineral spirits or turpentine. Now, most architectural coatings use
latex formulations, with only some types of varnish being solvent-based. Yet, even
latex formulations can contain a significant level of VOCs. Various small mole-
cules are employed to plasticise (i.e., soften) the polymer in the latex particles,
enabling film formation to occur at room temperature or even lower. After the
complete evaporation of the VOC, the polymer reverts to its hard, unplasticised
state leaving a high Tg polymer film. It is a major challenge to produce hard latex
coatings without VOC emission, and hence there is a continued need for the study
of latex film formation. This need adds a sense of urgency to scientific work and
has partly inspired the writing of this book.
These environmental and economic issues might also explain the growth of
interest in latex film formation in scientific literature. Fig. 1.13 gives an historical
perspective. Before 1990, there were only a handful of publications with a focus
on latex film formation. During the 1990s, there was a steady rise in the number of
publications each year, reaching a peak of 53 in the year 2000. Through the first
decade of the twenty-first century, the numbers have fluctuated around an average
of 41 per year.

Fig. 1.13 Number of publications, by year, on the topic of latex film formation according to a
keyword search on the ISI Web of KnowledgeSM database.
1.6 Relevant Colloid Science 17

1.6 Relevant Colloid Science

The physics of colloidal dispersions has received extensive experimental and


theoretical attention for hundreds of years. Here, we review the theories that are
directly relevant to latex. Interested readers may find additional information in the
many text books on colloid science (Hunter 2001, Russel et al. 1989).

1.6.1 Interaction Potentials

The stability of a colloidal dispersion, or the ability of particles to remain as


discrete objects, was first studied systematically by Schulz and Hardy (Hardy
1900), who examined the effect of electrolyte. A comprehensive theory, that is
still used extensively today, was derived in the 1940s independently by Derjaguin
and Landau in Russia and by Vervey and Overbeek in The Netherlands. It is now
known as the DLVO theory after the four scientists.
Colloidal stability may be understood by determining the energy of interaction
between two isolated particles. There are a number of separate constituent
energies and the DLVO theory assumes that they are additive. The two most
important interactions are the van der Waals attraction and an electrostatic
repulsion. In addition, depletion interactions are relevant to latex dispersions. Each
of these interaction energies will now be discussed individually.

1.6.1.1 Van der Waals Attraction

The electron clouds in the constituent atoms and molecules in a colloidal particle
fluctuate as a result of quantum effects. For an instantaneous moment when there
is a fluctuation, there will be an asymmetry of the electric charge, which produces
a dipole moment. Even non-polar, uncharged atoms and molecules are subject to
this effect. The instantaneous dipole, as it is sometimes called, will create an
electric field, which will then act upon the neighbouring molecules. The electric
field shifts the electron distributions in the neighbours and thereby polarises them.
The strength of their dipole is determined by their electron polarisability. These
neighbouring dipoles will interact with the instantaneous dipole, leading to an
attraction. It is somewhat surprising but true that all molecules (including
neutral, nonpolar molecules, such as the noble gases) feel a van der Waals
attraction to others. Van der Waals attraction explains why gases condense into
the liquid state as the thermal energy decreases.
The van der Waals attraction between two molecules is small, but when
summed up over two macroscopic bodies, the energy can be considerable. The
energy of interaction between two molecules is assumed to scale inversely with
the separation distance, r, to the sixth power. Upon summing up all the pair
18 1 An Introduction to Latex and the Principles of Colloidal Stability

interactions, the van der Waals attraction between two equal-sized spheres of
radius R with a centre-to-centre separation r is found to be

AR
U vdw = (1.1)
12 ( r 2 R )

where A is called the Hamaker constant and is an experimentally determined


quantity that depends on the dielectric constant and molecular densities of the
interacting particles and their dispersing medium. The negative sign indicates that
the interaction is attractive.
The van der Waals attraction is divergent, meaning that the attraction approaches
infinity at close contact (i.e. when r approaches 2R). Because of this, all colloidal
dispersions are inherently unstable. The lowest energy state is one where all the
particles are fused into a giant structure with minimum surface area. Therefore,
colloidal dispersions can only be kinetically stable, meaning that the rate of aggrega-
tion, towards the lowest energy state, is slow. The usual way to stabilise a colloidal
dispersion is to use an electrostatic repulsion between particles.

1.6.1.2 Electrostatic Repulsion Between Particles

Charged groups on the particle surfaces will repel other charges of the same sign,
and hence two particles with like charges experience a net repulsion. The magni-
tude of this repulsion is dependent on the charge, or potential, on the particle
surfaces and the medium they are interacting in. Using the same notation as in Eq.
(1.1), the electrostatic interaction energy between two like particles is given by

U elec = 2 0 R 2 exp ( ( r 2 R ) ) (1.2)

where and 0 are the permittivity of the medium and free space, respectively,
is the potential on the particle surfaces, and is the inverse of the Debye length.
The expression will be positive, reminding us that it is repulsive.
The Debye length, 1, is a measure of the range of distances over which the
electrostatic charges are significant. It is a function of the electrolyte concentration
because charged ions in solution will screen the effect of electrostatics and hence
reduce the repulsion between two charged particles. For a 1 mM solution of
sodium chloride in water, the Debye length is 9.6 nm, whereas for a 100 mM NaCl
solution, the Debye length reduces to 0.96 nm.
The average separation between particles in solution is surprisingly small. For
spherical particles at a volume fraction, , of 0.024, the average separation
between particle surfaces is equal to the particle diameter. As increases, the
average separation drops. For = 0.5, the separation between particle surfaces is
reduced to 9% of the particle radius. For illustration purposes, a 50% by volume
dispersion of 100 nm diameter particles has an average separation between
1.6 Relevant Colloid Science 19

particles of only 4.5 nm. Comparing this to the magnitude of the Debye length, it
may be deduced that if the particles are dispersed in 1 mM NaCl, the particle
electrostatic potentials will overlap to a considerable effect. This example shows
that the particles are experiencing an electrostatic repulsion from their neighbours.
The effect is seen in the dilute case of = 0.024 and becomes enhanced as the
volume fraction is increased and the average particle spacing decreases.

1.6.1.3 DLVO Theory

The concept behind DLVO theory is that the constituent interaction potentials are
additive. The original theory considered the sum of van der Waals attractions and
electrostatic repulsions. A typical result is shown in Fig. 1.14 where the divergent
van der Waals potential dominates at small separations. Because the electrostatic
repulsion typically acts over a longer range than the van der Waals attraction, the
total potential contains a repulsive part. If the maximum in this potential has a
significant value, in comparison to the thermal energy, then the dispersion is
kinetically stable. The addition of electrolyte causes the range of the electrostatic
repulsion to diminish and become overpowered by the van der Waals attraction.
After a long enough time, the kinetic stability will be lost, and the dispersion will
become unstable. The presence of electrolyte reduces the total interaction potential
as illustrated in Fig. 1.15. It is apparent from this illustration why the addition of
salt to a charge-stabilised colloidal dispersion will induce aggregation of particles.

Interaction Electrostatic repulsion. Range


energy is determined by the Debye length
which is a function of the electrolyte
concentration

Particle
Separation
Total potential from addition of
components: Dispersion stability is
Van der Waals attraction. Infinite
determined by the magnitude of the attraction at contact means all colloidal
maximum dispersions are thermodynamically
unstable

Fig. 1.14 The component potentials of electrostatic repulsion and van der Waals attraction
combine to give an overall interaction potential. The maximum in the total potential determines
whether the dispersion remains stable or aggregates.
20 1 An Introduction to Latex and the Principles of Colloidal Stability

Total
interaction
potential

Increasing
electrolyte
concentration

Separation

Fig. 1.15 Effect of increasing external electrolyte concentration. The van der Waals attraction
remains unaltered but the electrostatic repulsion is diminished. The resulting total potential
displays a diminishing maximum and hence the dispersion is destabilised.

1.6.1.4 Depletion Interactions

It is often observed that a stable colloidal dispersion, in the presence of a non-


adsorbing polymer, will experience aggregation1. The reason for this is the
depletion interaction, and its physical basis is sketched in Fig. 1.16. The non-
adsorbing polymer in the dispersing medium (water in the case of latex) will take
on the shape of an expanded random coil, if the medium is a good solvent. The
coil has a characteristic size given by its radius of gyration, Rg. As two colloidal
particles approach each other, the separation between them becomes less than the
size of the random coil, which is excluded from this overlap region. Hence, there
is a concentration difference between the bulk solution and the overlap region.
The result is an osmotic pressure on the outside region of the particles that is
larger than the osmotic pressure in the overlap region. This pressure imbalance
provides a force to push the particles together, causing them to aggregate.
An alternative way to view the same phenomena is to consider the entropy of
the polymer chains in the non-adsorbing polymer. If two particles, as a result of
their random thermal motion, approach each other, the exclusion regions will
1
In the colloid science literature, coagulation commonly refers to an irreversible aggregation,
and flocculation refers to a weak reversible aggregation. Here, aggregation is used to refer to
joining together of particles.
1.6 Relevant Colloid Science 21

overlap. Consequently, the total volume available to polymer chains will increase.
The greater available volume increases the entropy of the polymer chains and
reduces the Gibbs free energy of the overall system. Consequently, there is an
attraction between particles whose range extends over the size of the added
polymer chain. Its magnitude depends on the polymer concentration.

Overlap Region: The two


exclusion regions overlap

R
r

Rg

Exclusion zone: Polymer chains


Polymer chain: Size is set are excluded from approaching
by molecular weight and to a distance less than the
interaction between the radius of gyration
polymer and solvent

Energy of interaction is equal to the osmotic pressure of the dispersion, , multiplied


by the volume of the overlap region, Voverlap.

Udep = - Voverlap

For an ideal solution the osmotic pressure is simply related to the number
concentration of polymer chains, = n kT. The overlap region is determined from
geometry as


3



4 3r 1 r
Rg 3
Voverlap = 1 + R 3 1 +
3 Rg 16 R g R
4 R1 + R R1 +
R

Fig. 1.16 Depletion interactions. The exclusion of free polymer chains from the region between
two colloidal particles results in an attractive potential between them.
22 1 An Introduction to Latex and the Principles of Colloidal Stability

1.6.2 Fluid Motion

The motion of colloidal particles in a fluid was derived analytically by Stokes, and
the resulting flow pattern is known as Stokes flow. The effect of particles on
dispersion viscosity was also derived analytically, this time by Einstein in 1905.
He also built on the work of Brown by deriving the diffusion coefficient of
colloidal particles in a dilute dispersion, a result now known as the Stokes-
Einstein diffusion coefficient.
The flow of a Newtonian fluid around colloidal particles satisfies the Navier-
Stokes equation for momentum and mass conservation. Because of the small size
of colloidal particles, it is possible to greatly simplify the governing equations to
reach the so-called creeping flow limit, where fluid inertia is ignored. The
derivation of creeping flow is presented in Appendix A, wherein the solution to
the flow is shown in Fig. A.1. For creeping flow past a sphere, the solution is
referred to as Stokes flow. It tells us that the drag force acting on a sphere of
radius R that is travelling in a continuous medium of viscosity with a velocity of
U is given by 6RU. The drag coefficient of the particle is 6R.

1.6.2.1 Diffusion

Any colloidal particle is being continually subjected to fluid molecules colliding


with it. Because of the random nature of the molecule collisions, the average
displacement for a colloidal particle in one dimension is zero, although the
distribution of sampled positions increases with time. The random motion of the
particle is analysed as a diffusional process. In general, a diffusion coefficient is
given by the thermal energy, kT, which arises from the momentum of the fluid
molecules, divided by the drag coefficient, which describes the damping of the
particle motion. Here, k is the Boltzmann constant (1.38 x 1023 J K1) and T is the
absolute temperature in Kelvin. Following on from the expression for the Stokes
drag, the Stokes-Einstein diffusion coefficient is written as

kT
D0 = (1.3)
6 R

Equation (1.3) tells us that the diffusivity of a particle will be higher if its ra-
dius is smaller. On the other hand, a thickener that raises the viscosity will reduce
the diffusion coefficient. In Chapter 3, we shall see how particle diffusion is
important to take into account in the drying process.
1.6 Relevant Colloid Science 23

1.6.2.2 Low Shear Viscosity of Colloidal Dispersions

Adding colloidal particles to a fluid will increase the viscosity of the dispersion, .
That is, is a function of the volume fraction of the colloidal particles, . In the
dilute regime ( < 0.1), the dispersion viscosity can be predicted from the fluid
viscosity, , and the flow around an isolated particle. The result of the calculation
is a series expression in volume fraction where the higher order terms can be
safely neglected:


= 1 + + O ( 2 )
5
(1.4)
2

At higher particle concentrations, when reaches a value corresponding to


particle close packing, the dispersion forms a solid, and the viscosity diverges.
Accordingly, for higher volume fractions ( > 0.1), particle interactions must be
taken into account. In this regime, it is common to use the Krieger-Dougherty
expression
2
m
= (1.5)
m
where m is the volume fraction of particles at which the viscosity diverges. It
takes a value of about 0.64 for random close-packing of particles and 0.74 for a
face-centred cubic colloidal crystal (Russel et al. 1989). A graphical representation
of how the dispersion viscosity varies with particle volume fraction is shown in
Fig. 1.17. This viscosity dependence differs strongly from what is seen for a
solution of polymers dissolved in a solvent. In this latter case, the viscosity does
not diverge until it approaches a polymer fraction of 1. The use of a bimodal
distribution of particle sizes, i.e., a blend of large and small particles, enables a
higher value of m to be achieved.

= (1 + 2.5 )
1

m
Fig. 1.17 Effect of volume fraction on dispersion viscosity. At low volume fractions a linearisa-
tion applies and the viscosity diverges at a volume fraction of m.
24 1 An Introduction to Latex and the Principles of Colloidal Stability

References

Alturaif H., Unertl W.N., Lepoutre P. (1995) Effect of pigmentation on the


surface-chemistry and surface free-energy of paper coating binders. J Adhes
Sci Techn 9(7): 801-811
Bockelmann I., Darius S., McGauran N., Robra B.P., Peter B., Pfister E.A. (2002)
The psychological effects of exposure to mixed organic solvents on car paint-
ers. Disability and Rehabilitation 24: 455-461
Bockelmann I., Lindner H., Peters B., Pfister E.A. (2003) Influence of long term
occupational exposure to solvents on colour vision. Ophthalmologe 100: 133-
141
Bockelmann I., Pfister E.A., Peters B., Duchstein S. (2004) Psychological effects
of occupational exposure to organic solvent mixtures on printers. Disability and
Rehabilitation 26: 798-807
Brown G.L. (1956) Formation of films from polymer dispersions. J Polym Sci
22:423-434.
Calvert P. (2001) Inkjet printing for materials and devices. Chem Mater 13: 3299-
3305.
Campos G., Reyes Y., Soto N., Aremas J., Vasquez F. (2006) Development of
new carpet backings based on composite particles. J Reinforced Plast Compos-
ites 25(18): 1897-1901
Dennington S. (2007) personal communication
Department of Statistics (2008), Monthly Trade Statistics, Malaysia.
DeVito S.C. (1999) Present and future regulatory trends of the United States
Environmental Protection Agency. Prog Organ Coat 35: 55-61
Dobler F. and Holl Y. (1996) Mechanisms of Latex Film Formation, TRIP 4(5):
145-151
European Coatings Journal (2008) Global demand for architectural paint to rise.
Issue 4
Hardy W.B. (1900) A preliminary investigation of the conditions which determine
the stability of irreversible hydrosols, Proc Royal Soc 66: 110-125
Heldmann C., Cabrera R.I., Momper B., Kuropka R. and Zimmerschied K. (1999)
Influence of non-ionic emulsifiers on the properties of vinyl acetate/VeoVa10
and vinyl acetate/ethylene emulsions and paints Progress in Organic Coatings
35: 69-77.
Henson W.A., Taber D.A. and Bradford E.B. (1953) Mechanism of film formation
of latex paint. Indust Engin Chem, 45(4): 735-739
Hunter R.J. (2001) Foundations of Colloid Science, Oxford University Press, 2nd
Edition
Hutton B.H., Parker I.H. (2008) Immediate consolidation behaviour of aqueous
pigment coatings applied to porous substrates. Chemical Engineering Science
63: 3348-3357
IUPAC (2002) http://goldbook.iupac.org/C01172.html
References 25

Jotischky H. (2000) Coatings, regulations and the environment revisited. Surf.


Coatings Intern. Pt. B: Coatings Trans 84: 11-20
Keddie J.L. (1997) Film formation of latex. Mat Sci Eng R: Reports R21(3): 101-
169
Liss G.M., Sussman G.L., Deal K., Brown S., Cividino M., Siu S., Beezhold D.H.,
Smith G., Swanson M.C., Yunginger J., Douglas A., Holness D.L., Lebert P.,
Keith P., Wasserman S., Turianmaa (1997) Latex allergy: epidemiological
study of 1351 hospital workers. Occupational and Environmental Medicine 54:
335-342.
Lovell P.A., El-Aasser M.S. (1997) Emulsion Polymerisation and Emulsion
Polymers, John Wiley and Sons Ltd.
Morrow L.A. and Steinhauer S.R. (1995) Alterations in heart-rate and pupillary
response in persons with organic-solvent exposure. Biological Psychiatry 37:
721-730
Murray M. (2009) Personal communication.
Prager S. and Tirrell M. (1981) The healing process at polymer-polymer inter-
faces. J Chem Phys 75(10): 5194-5198
Richard J. and Maquet J. (1992) Dynamic micro-mechanical investigations into
particle/particle interfaces in latex films. Polymer 33(19): 4164-4173
Routh A.F. and Russel W.B. (1998) Horizontal drying fronts during solvent
evaporation from latex films. AIChE J 44 (9): 2088-2098.
Routh A.F. and Russel W.B. (1999) A process model for latex film formation:
limiting regimes for individual driving forces. Langmuir 15: 7762-7773
Routh A.F. and Russel W.B. (2001) Deformation Mechanisms During Latex Film
Formation: Experimental Evidence. Ind & Engin Chem Res 40(20): 4302-4308
Routh A.F., Russel W.B., Tang J. and El-Aasser M.S. (2001) A process model for
latex film formation: Optical drying fronts. J Coatings Techn 73 (916): 41-48
Russel W.B., Schowalter W.R. and Saville D.A. (1989) Colloidal Dispersions,
Cambridge University Press
Russel W.B. (1990) On the dynamics of the disorder-order transition. Phase
Transitions, 21: 127-137
Schuler B., Baumstarck R., Kirsch S., Pfau A., Sandor M., Zosel A. (2000)
Structure and properties of multiphase particles and their impact on the per-
formance of architectural coatings. Prog. Organ. Coatings 40: 139-150.
Sheetz D.P. (1965) Formation of films by drying of latex. J Appl Polym Sci,
9:3759-3773
Stancato A.C., Burke A.K. and Beraldo A.L. (2005) Mechanism of a vegetable
waste composite with a polymer-modified cement (VWCPMC), Cement and
Concrete Composites, 27(5): 599-603
Steinhauer S.R., Morrow L.A., Condray R., Scott A.J. (2001) Respiratory sinus
arrhythmia in persons with organic solvent exposure: Comparisons with anxi-
ety patients and controls. Archives of Environmental Health 56: 175-181
Steward P.A., Hearn J. and Wilkinson M.C. (2000) An overview of polymer latex
film formation and properties. Adv Coll Interf Sci 86(3):195-267
26 1 An Introduction to Latex and the Principles of Colloidal Stability

Tan, A. (2008) Its sunny days ahead for nitrile gloves, Rubber Asia Novem-
ber/December 139-140.
Turjanmaa K., Alenius H., Makinen-Kiljunen S., Reunala T., Palosuo T. (1996)
Natural rubber latex allergy. Allergy 51: 593-602.
Urban D. and Takamura K. editors (2002) Polymer Dispersions and Their
Industrial Applications, Wiley-V.C.H
Vanderhoff J.W., Tarkowski H.L., Jenkins M.C., Bradford E.B. (1966) Theoreti-
cal consideration of the interfacial forces involved in the coalescence of latex
particles. J Macromol Chem 1(2):361-397
Von Dungen M. (2009) Adhesives and sealants: the European market. European
Coatings Journal Issue 6: 12-15.
Wicks Z.W., Jones F.N., Peppas S.P. (1992) Organic coatings: Science and
technology. John Wiley & Sons, Chichester, p. 259
Winnik M.A. (1997) The formation and properties of latex films. Chapter 14 in
Emulsion Polymerisation and Emulsion Polymers, Edited by P.A. Lovell and
M.S. El-Aasser, John Wiley and Sons
Winnik M.A. (1997b) Latex Film Formation. Curr Opin Coll Interf Sci 2(2): 192-
199
Wong R., Hair M.L., Croucher M.D. (1988) Sterically stabilised polymer colloids
and their use as ink-jet inks. Journal of Imaging Technology 14(5): 129-131
Zang Y.H. and Aspler J.S. (1995) The influence of coating structure on the ink
receptivity and print gloss of model clay coatings. TAPPI Journal 78(1): 147-
154
Chapter 2
2 Established and Emerging Techniques of Studying
Latex Film Formation

The study of the processes of latex film formation presents many challenges to the
experimentalist. Each stage of film formation has specific requirements for any
analytical technique.
In the drying stage, there is a need to measure water concentration at various
positions laterally and through the depth of a wet film. Latex in the wet state
precludes the use of techniques that require the sample to be held in a high
vacuum, such as Auger spectroscopy or secondary ion mass spectrometry.
Electron microscopy conventionally has a high vacuum in the sample chamber,
and so it cannot be used in the standard way.
The particles are undergoing constant Brownian motion during the drying stage,
and any useful analytical technique must not perturb the particles while it probes
them. The ideal probe of particle motion will provide information on short time
scales, but a technique that can determine spatial profiles of particle concentration is
equally valuable. The drying process itself is discussed in Chapter 3.
In the particle deformation stage, water may still be present, and so there are
similar restrictions on the techniques. As the particles typically have a diameter
less than 300 nanometres, any imaging technique requires high resolution. There
is a need to know the particle structure through a film and not only at the interface
with air or substrate. Many microscopies, such as scanning probe techniques, may
be readily applied to the air interface. Information about the bulk of a film is
obtained after cutting cross-sections. There is always a concern that artefacts are
being introduced when preparing film cross-sections. Chapter 4 considers the
particle deformation mechanisms in detail.
Finally, in the interdiffusion stage, the needs are rather different. Here, the
molecular level is of greatest interest. Techniques are required to probe non-
invasively the individual molecules and the interfaces between particles. The
physical aspects of the interdiffusion stage are presented in Chapter 5.
This chapter will highlight techniques that are particularly well suited for ex-
amining latex film formation. Those techniques that study the polymer film itself
28 2 Established and Emerging Techniques of Studying Latex Film Formation

rather than its processing are not emphasised. Details of how the techniques
work may be found in the many cited sources. The emphasis here will be on the
basic principles of operation and on what the techniques can reveal. An historical
review of the techniques for studying film formation has been given elsewhere
(Keddie 1997) and will not be repeated here. Those techniques that can probe
latex in the presence of water (either liquid or solid) are considered separately
from those that analyse mainly, or exclusively, the dry film.
Through this chapter and the ones that follow, there will be a need to distin-
guish the plane of a film from the direction normal to it. In microscopy experi-
ments, a cross-sectional slice of a film is sometimes examined. The meanings of
these terms are clarified in Fig. 2.1.

Normal to film

Cross-sectional slice
Plane of film

substrate

Fig. 2.1 Illustration of the meaning of the direction normal to the film plane, which is also called
the vertical direction. The direction in the plane of the film is called the horizontal direction.
A cross-sectional slice is taken in a plane perpendicular to the plane of the film.

2.1 Techniques to Study Latex in the Presence of Water (Wet and


Damp Films)

The challenge of watching paint dry has attracted the application of numerous
sophisticated techniques. To interrogate latex that contains liquid water (or ice), a
probe must be used that does not perturb the material or the particle motion. The
probe may be in the form of electromagnetic waves (visible light, X-rays, radio
waves), ultrasound waves, or particles (electrons and neutrons). There is an
additional technique that uses neither waves nor particles, but instead it uses the
substrate as a probe of the mechanical state of the film. This technique, sometimes
referred to as beam-bending, will also be considered here. Yet another approach
is to impose an electric field across a film and to determine its electrical conduc-
tivity over time while drying.
2.1 Techniques to Study Latex in the Presence of Water (Wet and Damp Films) 29

2.1.1 Physical Probes of Drying

We will first consider those techniques that probe the physical state of wet films in
some way. These are all rather innovative and developed especially for latex film
studies. Optical probes will be considered separately.

2.1.1.1 MFFT Bar

Conceptually, the minimum film formation temperature (MFFT) bar is one of the
simplest techniques. Because of its simplicity, it is also one of the more common
approaches to the study of film formation.
First proposed by Protzman and Brown (1960), the MFFT bar consists of a
large platen that has a temperature gradient along its length. A film is cast on a
substrate that is put in contact with the platen such that one end of the film is held
at a temperature that is 20C or so higher than the other end. The atmosphere
above the film should be controlled. At low temperatures, below the dew point, a
dry atmosphere is required to prevent the condensation of water (Eckersley and
Rudin 1993). The method is defined by international standards (ASTM D2354-98
and ISO 2115:1996). An example of the apparatus may be seen in Fig. 2.2.
As its name implies, the MFFT bar finds the lowest temperature at which
a latex will form a film. The MFFT can be defined by one of three transitions in
the film. There is, however, some ambiguity in the physical meaning of these
definitions.
1. Cloudy-clear transition. This is essentially an indicator of the extent of particle
deformation. At temperatures below the MFFT, particles are packed into an
array, but they are not sufficiently deformed to be able to close up the void
space between them. Therefore, light is scattered from the dry film, and it ap-
pears turbid. At temperatures above the MFFT, the voids between particles
have closed sufficiently to prevent light scattering, and the film appears trans-
parent. The transition point is usually determined visually by the user and is
used to determine MFFT.
2. Crack point. This transition is determined visually as the point where cracks are
no longer observed in a dry latex film. When there is no interdiffusion between
particles in a latex film, it will be brittle. Zosel and Ley (1993) showed that a
brittle-to-ductile transition in a latex film is associated with polymer interdiffu-
sion. Hence, the crack point essentially determines the lowest temperature
where there is diffusion of polymer molecules across the particle interfaces.
3. Knife point. This transition corresponds to the minimum temperature at which
the film can resist mechanical shearing (Lee and Routh 2006).
The ASTM standard does not specify a single definition, but merely states that
the films discontinuity as evidenced by whitening or cracking or both should be
recorded. The ASTM standard says that the precision of the measurement has not
30 2 Established and Emerging Techniques of Studying Latex Film Formation

been experimentally obtained but that 2 C can be expected. Protzman and


Brown (1960) found that the MFFT is invariant with respect to the thickness of the
wet film and to the solids content of the latex.
The use of the MFFT bar requires a few words of caution. The ASTM standard
states that approximately one to two hours are required for the film to dry but it
does not specify a precise time for observation. This is a serious omission, which
could lead to a lack of reproducibility, because film formation is a dynamic
process. Particles can be slowly deformed under the action of surface energy over
long periods of time, so that their optical properties change (Keddie et al. 1996,
Lee and Routh 2006). It is understandable why Sperry et al. (1994) found that the
cloudy-clear MFFT moved to lower temperatures over time.
A second note of caution concerns the interpretation of the cloudy-clear point.
It is sometimes thought that it corresponds to good coalescence in a latex film,
which implies that all voids are eliminated and that there is extensive interdiffu-
sion between particles. In fact, electron microscopy of clear films has found that
they do, in fact, contain interparticle voids (Keddie et al. 1995).
Standard theories of optical transmission show us that optical clarity is a func-
tion of both the size and the number of air voids in a film (van Tent and te
Nijenhuis 2000). When very small particles are used in a latex, the size of the
interparticle voids will be reduced proportionally. This suggests that films made
from nanoparticles (with a radius of only tens of nm) could appear optically
transparent even when there is not much particle deformation. Simulations by van
Tent and te Nijenhuis (2000), shown in Fig. 2.3, offer great insight in this respect.
If particles are deformed such that the void volume fraction is only 0.03, larger
particles (with a radius of 400 nm) will naturally have larger voids compared to
smaller particles (radius of 50 nm). In the visible range of wavelengths, the film
from larger particles will appear turbid, whereas the film cast from the smaller
particles will appear transparent. To counter-act this problem, it has been proposed
that when comparing MFFT for latices with different particle sizes the ratio of
the wavelength of the light to the particle size should be kept constant (van Tent
and Nijenhuis 2000).
A third note of caution concerns the subjective nature of the measurement of
the transitions. In many laboratories, the same person is usually asked to make the
MFFT measurements to ensure consistency. Two different observers are likely to
define optically clear or not cracked in slightly different ways, leading to a
systematic shift in their MFFT readings. Multiple measurements should be made
and averaged to minimise errors. It is little wonder that MFFT has been called an
ill-defined concept (Dobler et al. 1992). Nevertheless, its measurement offers a
simple and fast check on the ability of a latex to form a film.
2.1 Techniques to Study Latex in the Presence of Water (Wet and Damp Films) 31

Fig. 2.2 Photograph of an MFFT bar showing two cast films. The cloudy-clear transition point is
indicated. (Photograph courtesy of P. Sperry and reproduced with permission from Sperry et al.
(1995); copyright 1995 American Chemical Society)

Fig. 2.3 Simulations of optical transmission under the assumption of a void volume fraction of
0.03 and with varying particle sizes, as indicated. The human eye is sensitive to wavelengths in
the range from 390 to 770 nm. (Reprinted from van Tent and te Nijenhuis (2000) with permis-
sion from Elsevier)
32 2 Established and Emerging Techniques of Studying Latex Film Formation

2.1.1.2 Film Scratching (Thin Film Analyser)

A specially designed instrument, named the thin film analyser by its inventors
(Bahra et al. 1992), provides a direct but invasive way of monitoring changes in
the rheology of films during the drying process. A wet film is spread on a tem-
perature-controlled platen set at a single value. A mechanical probe is scanned
back and forth in straight trajectories across the film surface (in the film plane),
moving to a new position after each stroke. The probe is held by an arm that is
joined to an angular displacement potentiometer. The amount of displacement can
be calibrated against known imposed forces.
What the technique measures depends largely on whether a spherical (ball) or
needle probe is used. A spherical or hemispherical probe will mainly impose a
shear stress on the film and hence a measure of the drag force is related to
viscosity. A drying film will have a complex dynamic modulus, but for Newtonian
oils it has been shown that the drag force on a hemispherical probe is linearly
related to the viscosity. When the particles in a colloidal dispersion hit the gel
point, associated with their close-packing, the instrument is sensitive to the sharp
rise in viscosity (Bahra et al. 1992). A needle probe, on the other hand, is capable
of scratching films and is thereby sensitive to the hardening and stiffening of a
film associated with its crosslinking. The marks left on a surface by a moving
needle are a complex function of hardness, friction and tear resistance (Schal-
lamach 1952).

2.1.1.3 Gravimetry

One of the simplest ways to study the drying of latex films is to record the mass as
a function of time. The solids fraction, , can then be followed through the film
formation process. A standard laboratory balance is often used for large-area
films, and there has been a report of the use of a sorption balance for small
droplets, up to 20 mg in mass (Erkselius et al. 2007). Of course, no direct informa-
tion is obtained about the distribution of water within the film from its mass.
Gravimetry is therefore most effective when combined with other techniques such
as optical microscopy and photography (Winnik and Feng 1996), IR spectroscopy
(Guigner et al. 2001), and magnetic resonance profiling (Mallgol et al. 2006).
Narita et al. (2005) have compared their data to a model of drying to determine
the importance of osmotic pressure. Erkselius et al. (2007) have pointed out the
effects of evaporative cooling on the rate of water evaporation from latex films.

2.1.1.4 Beam Bending (or Optical Cantilever) Technique

When two materials are coupled together in parallel in a bi-layer bar, and one of
the materials changes dimensions more than the other, the bar will curve. One of
the materials will be pulled in tension while the other is placed in compression. A
2.1 Techniques to Study Latex in the Presence of Water (Wet and Damp Films) 33

common example is when there is a mismatch in thermal expansivities in a bi-


metallic strip. For many years, a cantilever technique operating on this principle
was used to measure stress in paint and coatings (Corcoran 1969, Perera and Van
den Eynde 1981). But, it was not employed as an in situ probe of film formation
itself.
Petersen et al. (1999) provided the first major report of the optical lever (beam
bending) technique applied to the study of latex film formation. The basic
principle is that stress transfer from the film along the length of the beam will
cause curvature. If the film is pulling inward (in tension), then the beam will curve
upwards (with a concave curvature). If, however, the film is pushing outward
against the substrate (in compression), then the beam will curve downward (with a
convex curvature).
In the experiment of Petersen et al. (1999), a film is cast on a thin strip of metal
or other stiff material with a known elastic modulus. The strip is clamped in a
fixed position at one end, and the other end is free. Light from a laser source is
reflected off a mirror placed on the free end in order to monitor the displacement
of the strip. The movement of the reflected light beam is sensed with a quadrant
detector, which is sensitive to lateral movement of the light. Fig. 2.4 shows the
experimental set-up.
The technique requires that the Youngs modulus, Y, of the beam is known. A
beam with a length of L will be deflected from a straight position by an angle,
(in radians). For a beam of thickness db, the change in the stress of the film, f, is
given by the Stoney equation:

Ydb 2
f =
6 H L
(2.1)
dry

where Hdry represents the average thickness of the dry film.


Petersen et al. (1999) pointed out several caveats in using the beam-bending
method.
1. The technique measures the integral stress (force per unit width) and not the
local stress.
2. It is not sensitive to vertical stress or hydrostatic pressures just the lateral
stresses.
3. The zero position of stress cannot be determined, as there is no way of reaching
a stress-free state when the measurement has started (about a minute after film
casting). Instead, the zero point is defined as the starting point of data collec-
tion.
4. It must be assumed that there is no slip at the interface with the substrate a
sensible assumption for soft films with good adhesion.
They also noted that the loss of mass from water evaporation makes the exter-
nal torque on the beam change over time. The relative contribution of the film
weight scales with L2, so shorter beams result in reduced error. In the experiments
34 2 Established and Emerging Techniques of Studying Latex Film Formation

of Petersen et al., the effect of film weight on is 0.0007, which is not large in
comparison to the total deflection of about 0.1. The beam-bending technique has
also been used to investigate the causes of cracking in latex films (Tirumkudulu
and Russel 2005) and to determine the role of capillarity in generating film stress
during drying (Tirumkudulu and Russel 2004).
An interesting extension of the technique is to map the stress distribution in a
latex film in two dimensions in its plane by measuring the deformation when it is
cast on a thin membrane. This type of experiment was originally demonstrated
successfully by biophysicists as a means of examining the stresses generated by
stationary living cells (Schwarz et al. 2002). In the approach used by von der Ehe
and Johannsmann (2007), the back of the membrane has a mirrored surface. Stress
in a latex film induces distortions in the membrane, which are apparent when the
image of a grid is reflected off the back of it. The data analysis assumes that the
lateral tension in the membrane is the source of the deformation. The vertical
displacement of the membrane is related to the local stress in the plane of the film.
Variations of stress across a film can be mapped, and the stress development
leading up to cracking events can be followed in time.

Laser
beam
Quadrant detector

Mirror
Latex film

Clamp

Flexible beam

Fig. 2.4 Experimental set-up for a beam-bending experiment.

2.1.1.5 Ultrasonic Reflection

Acoustic waves are influenced by both polymer structure and molecular relaxa-
tion. The reflection of shear waves at ultrasonic frequencies (MHz range) is a well
established way of measuring the complex shear modulus (real and imaginary
components: G' and G") of a wide variety of materials. Alig et al. (1997) devel-
oped a fully automated set-up with digital signal analysis that enabled time
dependent studies. They showed the evolution of the shear moduli during the film
formation of an acrylic latex film. The technique is highly sensitive to the transi-
tion between liquid and solid. The shear modulus of the liquid is close to zero, but
2.1 Techniques to Study Latex in the Presence of Water (Wet and Damp Films) 35

it rises sharply when the particles pack together to make a more solid-like mate-
rial. There is also sensitivity to water loss in the later stages of drying.
A further improvement was made by the same group when they devised a
method and instrument to measure the complex longitudinal modulus (L' and L")
simultaneously with G' and G" over time during film formation (Lellinger et al.
2002). In an isotropic medium, these two moduli are related mathematically to the
Youngs modulus, the compression modulus, and Poissons ratio, . For instance,
can be calculated from this relation:

L' G'
= 2 (2.2)
L ' G '

Hence, a complete mechanical characterisation of a film, albeit at ultrasonic


(MHz) frequencies, can be obtained.
In the experimental set-up described by Lellinger et al. (2002), films are cast on
the end of a glass rod. A LiNbO3 transducer is attached to the other end, as shown
in Fig. 2.5. The transducer is used both to transmit and to receive ultrasonic
waves. This single transducer generated longitudinal waves of 8 MHz and shear
waves of 5 MHz. These two types of waves travel at different velocities in the
glass rod and then reflect from the film interface. The longitudinal wave arrives
back at the transducer in a shorter time than taken by the shear wave. Their
differing velocities thereby allow separation of the longitudinal and shear proper-
ties. The acoustic reflectivity depends on the complex dynamic shear impedance.
In turn, the shear impedance is a function of G' and G". There is an analogous
relationship between the longitudinal impedance and L' and L".
In a wet latex film, the compression modulus is sensitive to the fraction of
solids in water, as the two components have differing values. At the gel point, the
film can support a shear stress, and G' rises sharply from zero. On the other hand,
L' is approximately equal to the compression modulus and so it goes through a
more gradual transition, rising steadily with water loss.

Fig. 2.5 Experimental set-up of an ultrasonic reflection experiment in which both longitudinal
and shear waves are generated and detected. (Reprinted from Lellinger et al. (2002); copyright
Wiley-VCH Verlag GmbH & Co. KGaA; reproduced with permission)
36 2 Established and Emerging Techniques of Studying Latex Film Formation

2.1.1.6 Electrical Conductivity

Measurements of the electric conductivity of wet films are conceptually simple


and just require a sample cell in which electrical contact may be made by inserting
metallic wires into the wet film. In this design, a conductivity meter must be
sensitive enough to measure values on the order of S/cm (Bouchama et al. 2002).
Another approach is to attach electrodes to a substrate and then spread the film
across them. Electrical impedance can be recorded over time starting from the
deposition of a wet film. One difficulty with this method is that film cracking
leads to a sharp rise in the electrical resistivity of the film and interferes with the
measurement (Mulvihill et al. 1997).
In the later stages of drying an emulsion, a biliquid foam structure can develop
in which the particles (soft droplets) are deformed to fill space with a thin water
layer separating them. There is an established relationship between the conductiv-
ity of a biliquid foam and its water content, as derived by Lemlich (1978). It may
be used to determine whether it is consistent with the proposed foam structure.
Hence, information on the deformation of droplets (or soft particles) may be
extracted from an experiment. Conductivity measurements are particularly
effective in determining whether particle boundaries exist in a film that has dried
to a state of optical clarity. A drawback is that the measurement provides an
average reading for the entire film. In the methods used to date, there is no
information as a function of depth or lateral position.
Conductivity measurements are most informative when coupled with simulta-
neous visual observation using optical microscopy or measurements of water
content via gravimetry to aid analysis (Bouchama et al. 2002). This combined
analysis has been demonstrated for oil-in-water emulsions, but is less well
developed for the study of latex films.

2.1.2 Specialist Electron Microscopies

A conventional scanning electron microscope requires specimens that are electri-


cally conductive, in order to avoid negative charge build-up on the surface. A high
vacuum is required in the sample chamber and along the path down the micro-
scope column from the source of electrons. These two restrictions make the study
of wet latex films impossible within a conventional electron microscope. One way
around the problem is to operate at very low temperatures in a cryogenic electron
microscope so that the water is frozen (ideally in a glassy state). Another approach
is to use environmental scanning electron microscopy (ESEM), which enables a
low water vapour pressure in the sample chamber and thereby allows the analysis
of wet samples. More recently, environmental microscopes have been adapted to
allow scanning transmission electron microscopy (STEM) of wet films. Each of
these three will be discussed here.
2.1 Techniques to Study Latex in the Presence of Water (Wet and Damp Films) 37

2.1.2.1 Cryogenic Scanning Electron Microscopy

Ma et al. (2005) have shown that wet latex films may be successfully quenched to
cryogenic temperatures so that the water solidifies. Latex particles in water are
immobilised being literally frozen in place. The specimen can then be fractured
and loaded into a scanning electron microscope and examined while being
maintained at this temperature. Cryogenic microscopy has provided the first direct
visualisations of the three-dimensional distribution of latex particles during film
drying. The technique is highly effective, and stunning cryo-images are presented
in Chapter 3. Yet even so, Ma et al. (2005) outlined numerous artefacts of which
the microscopist must be aware. (An artefact is a feature that is created during the
analysis and that is not present in the original specimen.)
1. Freezing. For best results, ice formation must be avoided, because the volume
change associated with the freezing phase transition disrupts the local structure.
Instead, through superfast cooling, at least 105 C per second, glassy water can
be formed. High pressure is useful here, as it has proven much less disruptive
to film structure in comparison to plunge freezing.
2. Freeze-fracture. It is important to remember when viewing cryogenic SEM
images of film cross-sections that the surface has been created by fracture.
During the fracture process, particles can be elongated and plastically de-
formed. Surface roughness is the result of the fracture process and not necessar-
ily indicative of material structure.
3. Sublimation. Often in the frozen specimen, ice covers the material of interest.
Sublimation of ice on the surface enables underlying latex particles to be ob-
served. Furthermore, etching a small amount of the surrounding ice increases
the contrast with the particles. If the sublimation proceeds too far, however, the
particles are fully exposed and can rearrange their relative positions.
4. Cryo-transfer. After a specimen is frozen, it must be transferred to the cold
stage of the electron microscope. If a frozen latex film is not protected in a dry
atmosphere, such as the vapour of liquid nitrogen, or in vacuum, it is subject to
the formation of ice crystals from condensed atmospheric moisture.
5. Imaging. The electron beam may damage the specimen. Latex particles may
change shape, melt, and even move position under high electron doses. To
minimise these effects, low beam current and electron voltages are often used.
Furthermore, residual water vapour within the high vacuum microscope cham-
ber should be avoided, such as by using cold traps, to prevent condensation on
the surface (Sheehan et al. 1993)

2.1.2.2 Environmental Scanning Electron Microscopy (ESEM)

As indicated in the previous section, there are many potential artefacts in cryo-
genic electron microscopy. An alternative approach, ESEM, permits the observa-
tion of hydrated substances, including colloidal films, at temperatures above the
38 2 Established and Emerging Techniques of Studying Latex Film Formation

freezing point of water. ESEM has undoubtedly yielded many new insights into
the film formation process. Its primary attraction over cryogenic SEM is that
sample preparation is minimal. A film can be cast, placed into the microscope
chamber, and examined within a few minutes (Donald 2003). An image of
a partially wet latex film can be obtained (Fig. 2.6) at the point where particles
have packed together but where water still surrounds them. Under the right
conditions, latex particle aggregation can be observed at the surface of water
(Donald et al. 2000).
Many of the key insights leading to the development of ESEM were made by
Danilatos (1998). In particular, he developed the use of a microscope column
(along which the electron beam travels) with a differential pressure along it.
Pressure-limiting apertures maintain the pressure gradient that varies typically
from 107 Torr around the electron source (the gun filament) up to about 10 Torr in
the specimen chamber (Fig. 2.7). In a conventional electron microscope, by
comparison, the specimen chamber is maintained at a high vacuum along the
whole length of the microscope column. A variety of gases may fill the sample
chamber in an environmental electron microscope (Fletcher et al. 1997), but for
wet latex films, water vapour is the obvious choice.
In addition to the pressure gradient along the microscope column, there are two
other key requirements for ESEM, as explained by Donald et al. (2000). One
requirement is that there must not be too much scattering of electrons from the
gas in the chamber as they pass from the sample surface to the detector. The
scattering may be reduced by keeping the distance between sample surface and the
detector short, and by keeping the gas pressure in the sample low. A second
advantageous requirement is that the detector must operate in a gaseous envi-
ronment rather than a vacuum. The secondary and backscattered electrons emitted
from the sample collide with the chambers gas molecules to create daughter
electrons and positive ions. Because of this cascade process, many more electrons
are detected than are originally emitted from the sample. A positively biased
environmental (gaseous) detector is used in environmental microscopes to detect
the amplified signal.
An added benefit of having a low-pressure gas in the specimen chamber is
that the positive ions are drawn to the sample surface where they neutralise the
negative charge that builds from bombardment by the electron beam (Donald et
al. 2000). In conventional microscopy, a thin layer of a conducting material,
such as gold or carbon, is deposited on the sample surface to create a path for
electron charge transport. Without a conductive coating, fine surface features in
an environmental electron microscope will not be hidden or covered up. Fur-
thermore, the sample will not need to be exposed to the vacuum, and hence
dried, as is required for the deposition of metallic coatings by thermal evapora-
tion or sputtering.
2.1 Techniques to Study Latex in the Presence of Water (Wet and Damp Films) 39

Fig. 2.6 ESEM image of a latex film in its partially dry state. The water appears as the light grey
areas. The dark grey spots are the particles that are emerging from the wet surface. The scale bar
is 2 m. (Image courtesy of Paul Meredith, University of Cambridge)

Fig. 2.7 A diagram of an environmental scanning electron microscope (ESEM) showing the
pressure decrease along the electron beam path. (Reprinted from Keddie et al. 1995; copyright
1995 American Chemical Society)
40 2 Established and Emerging Techniques of Studying Latex Film Formation

To reduce electron scattering and thus to obtain a quality image in the envi-
ronmental microscope, the pressure must typically be about 6 Torr (or 6 Torr/760
Torr 0.008 of an atmosphere). At this pressure, water will boil at room tempera-
ture. Considering the temperature dependence of the vapour pressure of water
(Fig. 2.8), it is apparent that thermodynamics will favour the condensed state at a
pressure of 6 Torr when the temperature is held at 6C or below. Hence, there is a
range of pressure and temperature combinations under which evaporation is
prevented and a latex, or any other sample, can be kept wet but not frozen in
ice. The sample is loaded into the specimen chamber at ambient pressure, and a
careful pump-down procedure is required to prevent sample drying (Cameron and
Donald 1994).
One challenge in ESEM is the prevention of electron beam damage to the sam-
ple. As the electron beam is rastered back and forth across a surface, it dwells for a
longer time at the start of a line. Along the edges of a scanned region, where the
amount of irradiation is greater, visible changes may be made to the sample, such
as melting or particle rearrangement. When electrons interact with water in the
chamber, free radicals are created through the process of radiolysis. These free
radicals react with polymers, such as in a latex, and thereby alter the chemical
composition (Donald 2000). When polymer surfaces are coated with water, the
beam damage is even worse than when water is only in the gaseous state (Kitching
and Donald 1998).

Water
condenses

Water
evaporates

Fig. 2.8 Water vapour pressure as a function of temperature. At higher vapour pressures and
lower temperatures in the ESEM chamber, water will condense, and imaging of wet samples is
possible.
2.1 Techniques to Study Latex in the Presence of Water (Wet and Damp Films) 41

2.1.2.3 Wet STEM

Clearly, ESEM is a powerful tool for the visualisation of colloidal processes,


including the drying of films. As explained in the previous section, in ESEM an
electron beam is rastered across a surface and an image is created from the
secondary and backscattered electrons. A variation on the technique uses a
rastered electron beam that is transmitted through a thin wet film, as is illustrated
schematically in Fig. 2.9. Called wet STEM, this technique was pioneered by
Bogner et al. (2005), who demonstrated its utility for colloids dispersed in water.
In their first report of the technique, they pointed out its main advantages. Wet
STEM shares the attractions of ESEM in that it is suitable for wet samples without
the need for staining, coating or destructive sample preparation. A clear benefit of
wet STEM is that it provides information about the entire volume (i.e., depth) of
the wet sample and not just what emerges at the surface. In comparison to TEM,
wet STEM yields greater contrast in an image, as a result of the lower voltages
leading to a greater number of electron-sample interactions. Using microscopes
with a field emission electron source, the resolution can be as low as 1 nm at a
voltage of 20 kV, and a resolution of 5 nm was achieved in the study of gold
nanoparticles. Somewhat surprisingly, when the sample volumes are sufficiently
small, such as in the m-sized holes of a carbon grid, the images are not affected
by particle drift or diffusion.
To prepare a sample, a droplet of the colloidal dispersion of interest is placed
on a grid used for transmission electron microscopy. Surface tension ensures that
the liquid is self-supporting, so that the electron beam can pass through the
dispersion layer and to the detector, provided that the layer is sufficiently thin
(Bogner et al. 2005). The sample is loaded into the specimen chamber at ambient
pressure, and a careful pump-down procedure is required to prevent sample
drying, sample freezing, or unwanted condensation (Cameron and Donald 1994).
As in the case for ESEM, the specimen chamber must be evacuated without drying
the sample. Then, the pressure and temperature of the aqueous sample must be
carefully maintained to either induce slow evaporation of water or to maintain
a constant volume of water in the sample. The liquid state of water can be
maintained at 2C and a water vapour pressure of 5.3 Torr. Decreasing the
pressure or increasing the temperature controls the evaporation of the sample and
enables particles the deformation and the film forming process to take place
(Arnold et al. 2009).
Bogner et al. (2005) used annular dark-field conditions in which two semi-
annular solid-state detectors are used to collect the scattered beam but not the
transmitted one. This approach achieves better contrast than the use of the
transmitted beam in the bright field. The contrast in the image is influenced by the
distance between the sample and the detector, and it can be optimised through
systematic variation. An optimum distance was found at 7 mm.
42 2 Established and Emerging Techniques of Studying Latex Film Formation

Fig. 2.9 The arrangement used for wet scanning transmission electron microscopy. A Peltier stage a
controls the temperature. The incident beam i is transmitted through a thin, wet layer w supported
on a stand b within an environmental chamber. A backscattered-electron detector c is positioned
under the sample. (Reprinted from Bogner et al. (2005) with permission from Elsevier)

2.1.3 Scattering Techniques

The scattering of waves of particles (electrons or neutrons) or electromagnetic


radiation (e.g. visible light and X-rays) offers an ideal non-invasive method to
probe latex in its wet state.

2.1.3.1 Small Angle Neutron Scattering (SANS) and Small Angle X-Ray
Scattering (SAXS)

The scattering of neutrons and X-rays gives information on the radial structure of
latex particles (and hence its surfactant layer and membrane state) and on the
spatial correlation between particles. That is, scattering can tell us whether the
particles are spaced at equal distances. It also can tell us whether particles are
arranged on an ordered lattice or randomly distributed in space. As a consequence
2.1 Techniques to Study Latex in the Presence of Water (Wet and Damp Films) 43

of the power of these scattering techniques, they have often been used in the study
of colloids (Ballauff 2001).
The scattering of both X-rays and neutrons uses a common approach. A beam
of known intensity, I, and a wavelength, , is transmitted into the sample. The
scattered radiation is detected at a scattering angle with respect to the transmit-
ted beam, as is shown in Fig. 2.10. In a typical experiment, measurements of the
intensity of the scattered radiation are made over a range of or more often
over a range of . The latter is straightforward when using position-sensitive
detectors. In this way, the intensity of the scattered beam is measured as a function
of the wavevector, Q:

4
Q= sin (2.3)
2

and data are usually presented as a spectrum of I(Q). Depending on the system, the
scattering spectrum is interpreted in terms of an appropriate model. Q is inversely
related to the size of the feature that is probed. Hence, to study the relatively large
distances associated with colloids such as particle spacings and diameters low
values of Q must be obtained. Radiation scattered at low are thus the most
useful, leading to the suitability of small angle rather than wide neutron and X-
ray scattering (SANS and SAXS). The first conclusive demonstration of how
SANS can interrogate particle deformation in wet colloidal specimens was given
by Crowley et al. (1992).


I I(0)
latex

I(Q)
Fig. 2.10 Geometry of the arrangement for a small-angle neutron or small angle X-ray scattering
experiment.

As an example of experimental details for SANS, at the neutron source at the


National Institute of Standards and Technology in Gaithersburg, Maryland (USA),
Kim et al. (2000) used a neutron beam of = 20 scattered over a range of to
provide Q from 0.004 1 to 0.039 1. As a second example, Belaroui et al.
(2003) obtained even lower values of Q, from 0.0015 1 to 0.034 1, on the D11
and D22 diffractometers at the Institute Laue Langevin in Grenoble, France, using
two wavelengths of = 6 and = 10 .
SANS has provided evidence for surfactant desorption from colloidal polymer
particles in partially dried films (Belaroui et al. 2003). To achieve contrast
44 2 Established and Emerging Techniques of Studying Latex Film Formation

between the phases, selective deuterium labelling is used. (The scattering density
of deuterium differs strongly from hydrogen, and so it is an ideal label.) In the
experiments of Belaroui et al. (2003), the continuous serum phase contained 20%
D2O with the rest being H2O. Hence, the neutron scattering density of the particles
and the serum was matched, so that the dispersion was not expected to scatter
neutrons. When the surfactant (sodium dodecyl sulphate) was labelled with
deuterium, it had contrast with the particles and the serum, so that desorption
could be followed with SANS.
There are relatively few examples of the use of X-ray scattering to study latex
drying and particle packing. In the past, when Kratky cameras were used, the
range of Q was not sufficient to study larger distances in real space. With a
modified camera, Dingenouts and Ballauff (1998, 1999) obtained Q in the range
from 0.03 to 4 nm1 for use in studying the ordering of non-deformable particles in
dried latex films. Comparing two latices, they were able to relate greater ordering
of packed particles to a narrower polydispersity of particle sizes (Dingenouts and
Ballauff 1999). X-ray reflection measurements from a concentrated dispersion of
colloidal silica particles in water have probed the particle ordering near the
interface with air (Madsen et al. 2001). In this case, X-rays were reflected from
the wet surface, and the specular beam was detected. (A specular beam reflects
from the surface at the same angle as the incident beam strikes it.) To date, this
type of experiment has not been reported for latex films.

2.1.3.2 Photo Correlation Spectroscopy, Diffusing Wave Spectroscopy,


and Speckle Interferometry

Light scattering is a common technique for measuring particle size distributions in


dilute dispersions. In a latex dispersion, the particles are undergoing constant
random movement referred to as Brownian motion. As solvent molecules collide
from all sides of a colloidal particle, there is a slight imbalance of momentum
transfer, so that the particle is pushed a short distance in a certain direction. Over
longer times, the particle is seen to follow a random path.
The self-diffusion coefficient for a Brownian particle, D0, is defined for an
ensemble of particles through the mean of the square of the distance travelled,
<r2>, in a given time, t, as:

r2
D0 = (2.4)
6t

A well-known equation, derived from the work of Stokes and Einstein (SE) and
presented already in Section 1.6.2.1, shows us how the rate of self-diffusion for a
spherical particle of radius, R, is affected by the viscosity, , of the continuous
phase (e.g., the latex serum):
2.1 Techniques to Study Latex in the Presence of Water (Wet and Damp Films) 45

kT
D0 = (2.5)
6 R

where the product of the Boltzmann constant, k, and the absolute temperature, T,
is a measure of the thermal energy. The SE equation (2.5) applies in the dilute
limit where particle motion is not affected by interaction with neighbours.
We see in (2.5) that measurements of D0, such as by observing the paths of
diffusing latex particles with a microscope, provides a measure of R, provided that
the dispersion is dilute, the viscosity of the continuous phase is known, and
temperature is kept constant during the measurement. The same equation tells us
that measurements of D0 for a particle of known size can tell us the viscosity of the
continuous medium. In this way, a colloidal particle may be used as a probe of
in its local region. Equation (2.5) can be generalised to consider a viscoelastic
medium, rather than a simple viscous liquid, so that the complex dynamic
modulus can be extracted. This concept forms the basis of a microrheology
technique (Waigh 2005) in which colloidal particles are used as local probes of the
viscoelasticity of concentrated polymer solutions (Papagiannopoulos et al. 2005)
and gels (Djabourov et al. 1995).
In the type of dynamic light scattering known as photon correlation spectros-
copy (PCS), the detector is fixed at a certain angle , and the scattered intensity is
measured over time, I(t). Because colloidal particles in a fluid diffuse over time,
there will be fluctuations in the intensity of the scattered light such that I varies
with t. Thus at a t = 0, the scattered intensity I(0) will be slightly different than at a
later time t. An autocorrelation function, g2, is a measure of these fluctuations over
time (Maret and Wolf 1987):

I (0) I (t )
g 2 (t ) = (2.6)
I (0) 2

where the < > brackets indicate an ensemble average over time. Fig. 2.11 illus-
trates the physical meaning of the autocorrelation function and gives a typical
example. When g2(t) is found in an experiment on a dilute colloidal dispersion, it
follows an exponential dependence on time: g2(t) = g0 exp(t/), where is
called the autocorrelation time. It tells us how long it takes for g2(t) to decay to
1/e of its value at time 0. A is useful because it is related to the parameter of
interest, the Brownian diffusion coefficient, D0. PCS makes use of the autocorrela-
tion function to extract D0 and hence to determine the particle size.
There are two reasons why PCS cannot be used in the usual way for the study
of latex film formation. First, it applies when the incoming photons are scattered
only once. To achieve this condition a short optical path is created by using a
dilute dispersion. Concentrated latex dispersions cause multiple light scattering,
and PCS is not valid. A second reason stems from the fact that to get good results,
g2(t) must be measured over sufficiently long times such that t is about a thousand
46 2 Established and Emerging Techniques of Studying Latex Film Formation

times greater than A. In the later stages of drying, the latex concentration is high,
so that D0 is small and A is large.
A probe of the drying of latex films must be able to study multiple photon
scattering and must be relatively fast. The study of multiple scattering is enabled
by diffusing wave spectroscopy (DWS). The technique can be made faster by
averaging over a series of positions in space at one time rather than averaging over
time at one position.
The name diffusing wave spectroscopy was proposed by Pine et al. (1988)
who wrote out g2 for the case of multiple scattering. They considered the case of
light transmitted through a sample as well as the case of light backscattered from a
sample. Latex films are usually quite turbid, and so the backscattering geometry is
more suitable. For this geometry, when the film thickness is relatively large, Pine
et al. derived an expression for g2(t) in the limit where the thickness of the
scattering slab is far greater than the mean free path of the diffusing photons:


2 n
g 2 (t ) exp (2.7)
1
r 2 (t ) 2

Variables are as defined earlier; n represents the refractive index of the medium
(e.g., the latex serum) and 2 (Pine et al. 1988). Thus, measurements from
multiple light scattering give us g2(t) and from that the mean-squared displace-
ment of particles can be found. With a measurement of particle displacement, and
applying Eqs. (2.5) and (2.6), information about the other unknown, such as local
viscosity, particle size, or the effects of colloidal interactions, is gained.
The first two published reports of applying DWS to the drying of waterborne
colloidal films did not appear until 17 years after the techniques invention. In a
major advance, Breugem et al. (2005) used the Brownian diffusion of colloidal
particles as a probe of the storage modulus, Ep, and the viscosity of the film, ,
and how these parameters evolved with ongoing water evaporation. Both transmit-
ted and backscattered light was used. In these experiments, g2(t) was measured
experimentally and used to extract <r2>. Then, in turn, the Laplace transform was
used in the generalised Stokes-Einstein equation to find the Laplace-transformed
viscoelastic modulus (containing Ep and ).
Near the same time as the work of Breugem et al., a different research group
(Narita et al. 2004) employed multispeckle DWS to examine the dynamics of latex
particles (both deformable and non-deformable) through the drying process by
measuring the characteristic time of relaxation. Their approach was significantly
different than that of Breugem et al. They recorded the scattered light using a
CCD camera so that each pixel in the image recorded a scattering event in a
technique described thoroughly by Viasnoff et al. (2002). They then averaged over
an ensemble of positions (represented by each point on the detector) rather than
averaging over time. They specifically measured the characteristic time of the
2.1 Techniques to Study Latex in the Presence of Water (Wet and Damp Films) 47

relaxation, which represents the collective rearrangement of particles in a colloidal


gel (i.e., when the particles have packed together).
DWS has been combined with MR profiling to provide information on the
particle motion in a drying film along with simultaneous information on the
distribution of water (Knig et al. 2008). Fig. 2.12 shows the experimental set-up.
A typical autocorrelation function in these experiments, which is presented in Fig.
2.11, was fit to a stretched exponential function of the form:

t
g 2 (t ) 1 = A exp (2.8)

where the exponent provides qualitative information on the dynamical heteroge-


neity (coexistence of fast and slow dynamics). A value of the amplitude, A, that is
far less than unity provides evidence for the presence of static scatterers. It drops
sharply when the particles are elastically coupled together in a solid-like form
(Knig et al. 2008).

I(t)
I(t)

0.50

Time Time
g2(')1

0.25

0.00
1E-5 1E-4 1E-3 0.01 0.1
' [s]

Fig. 2.11 An autocorrelation function (g2()-1)obtained from scattering from a latex film. The
light intensity fluctuates over time. On the left side, there is a large overlap of the intensity signal
with itself when the time is shifted by a small time, . On the right, the time shift is greater, and
there is less overlap. The autocorrelation function is related to the amount of overlap at various
values. Mathematically the autocorrelation function is defined as

I (t ) I (t + )dt
g 2 ( ) = 0

I
2
(t ) dt
0

such that g2(0) = 1 and it decays monotonically with (equivalent to Eq. (2.6)). (Data courtesy of
Alexander Knig and Diethelm Johannsmann, Clausthal University of Technology)
48 2 Established and Emerging Techniques of Studying Latex Film Formation

Fig. 2.12 Experimental set-up for the combined use of GARField NMR profiling and diffusing
wave spectroscopy (DWS). The light in the laser beam is spread out and then is scattered from
the latex film. A photomultiplier tube detects the signal before the light is passed into an
autocorrelator. (Drawing courtesy of Alexander Knig, Clausthal University of Technology)

Prior to the development of DWS for the study of film formation, a closely
related technique known as dynamic speckle interferometry was reported
(Amalvy et al. 2001). Speckle refers to the pattern of light spots that are seen
when reflecting an intense coherent light source on a dynamic surface. Speckle is
caused by the interference of dephased coherent light after reflecting from a
surface having roughness and heterogeneities in its refractive index. In a system in
which there is particulate motion, such as a wet latex film, the speckle pattern will
fluctuate over time in a dynamic speckle pattern.
In an experiment, a laser is used to create a dynamic speckle pattern on a wet
film. A CCD camera is employed to capture 512 successive images of the speckle
pattern at short time intervals (e.g., 0.08 s). Each image consists of 512 x 512
pixels. The temporal history of the speckle pattern is then presented in an image
made up of a column from each of the 512 successive images. Analysis of the
image enables the calculation of a co-occurrence matrix and the moment of inertia
of the matrix (Almavy et al. 2001, Faccia et al. 2009). The moments of inertia
have been shown to correlate with the amount of water in a latex film. A slow-
down of water loss rate in a second stage of drying is observed and is comparable
to what is found with gravimetric analysis.
There have been recent technique developments to create speckle interferometry
images of surfaces to visualise heterogeneities (Faccia et al. 2009). An image is
created in which the grey scale indicates the activity, which is related to the
movement of scatterers and hence the extent of drying. Simultaneous monitoring of
2.1 Techniques to Study Latex in the Presence of Water (Wet and Damp Films) 49

zones with differing amounts of water from fully wet to partially dry is possible.
This non-invasive method is particularly well suited to probe lateral drying fronts
and particle packing fronts in latex films (see Fig. 3.9 and Section 3.5).
An instrument using speckle analysis has already been commercialised (Brun et
al. 2006, 2008). Using a proprietary data processing technique, the instrument
provides advanced analysis of the speckle pattern through all stages of drying. The
analysis does not find an autocorrelation time, A, but instead it measures the
speckle rate, s, which is the inverse of a characteristic time. As the latex dries,
particle motion slows down as a result of particle crowding. This effect is seen in
the dependence of D0 on the solids volume fraction, . When the particles create a
gel, one would expect an abrupt slowing down of particle dynamics and for s to
decrease. Experimentally for a latex, s is seen to decrease by up to five orders of
magnitude as the characteristic time goes from 0.1 s to 10,000 s. Research is in
progress to correlate the speckle rate with the physical state of a latex film and the
stages of film formation.

2.1.3.3 Evanescent Dynamic Light Scattering

Photon correlation spectroscopy may also be performed by using an evanescent


wave rather than through directly transmitted light (Lan et al. 1986). In this
experiment, light is brought in through a prism to reflect from the interface with a
cell containing a concentrated colloidal dispersion. As the refractive index of the
prism is greater than the wet latex film, the light is totally reflected from the
interface, and the film is probed with an evanescent wave. Autocorrelation spectra
can be obtained and used to study the Brownian motion of latex particles near the
interface with the prism.
Whereas the first evanescent light scattering experiments (Lan et al. 1986) were
performed on latex dispersions in a closed cell, Schmidt et al. (1997) extended the
technique to a latex undergoing film formation. They cast a latex film on a prism
and brought laser light through the prism to reflect from the interface with the
film. Their preliminary work used evanescent dynamic light scattering to observe
the slowing down of particle motion brought on by the increase of solids content
during drying. As with DWS, the technique measures the autocorrelation function
as it evolves with the film state.

2.1.3.4 Ultramicroscopy and Confocal Microscopy

The rather obvious obstacle in using optical microscopy to study latex film
formation is that standard optical microscopes do not have sufficient resolving
power to provide useful information on particles, sometimes with sizes less than
100 nm. The reason for the poor resolution stems from the so-called diffraction
limit that sets the resolution of optical microscopy at about /2 (Pohl and Van den
Eynde 1984).
50 2 Established and Emerging Techniques of Studying Latex Film Formation

Two different types of optical microscopy have been developed to overcome


this limit. One is known as ultramicroscopy, and it detects the scattered light
from moving particles. The other type is usually known as scanning near field
optical microscopy (SNOM), and it beats the diffraction limit by bringing the
light source close to the sample surface. SNOM will be discussed in Section 2.2.4,
as it is usually applied to dry films.
A rather old technique used for some of the early 20th century studies of
Brownian motion, ultramicroscopy has been rediscovered and applied to the
study of latex film formation (Keslarek and Galembeck 2003). A direct image of
sub- individual particles dispersed in liquid is not created. Instead, the scattered
light from the particles is detected, providing information on particle motion and
clustering but not yielding information on the structure of individual particles.
Experimentally, observations are carried out in the dark-field of an optical
microscope with illumination at 90 from the optical axis. Light is scattered by the
particles and collected in the objective lens. The light scattering is seen as flashes
of light in the dark field. Ultramicroscopy has the potential to be particularly
useful when information is required on particle dynamics.
Although most dispersions used for coatings and adhesives have particles that
are too small for optical microscopy, fundamental scientific studies use larger
particles with success. Confocal laser scanning microscopy (typically with = 633
or 488 nm) can be used for relatively large latex particles (R > 200 nm). For
instance, the technique has been used to show the formation of planar arrays of
large particles (R 650 nm) in a density matched medium under the application of
shear at a constant rate (Derks et al. 2004). The tracking of colloidal particles in
video analysis is made much easier when they contain a fluorescent core and a
non-fluorescent shell (Dullens et al. 2004). Confocal microscopy of this type of
core-shell particle composed of poly(methyl methacrylate) prepared by dispersion
polymerisation can provide insight into the latex drying process. A two-
dimensional radial distribution function, which reveals the long range ordering in
the packing in the film plane, has been obtained for the surface of a dried layer
(Dullens et al. 2003) and along a glass wall for a wet layer (Dullens et al. 2004).

2.1.3.5 Optical Techniques: Transmission Spectrophotometry and


Ellipsometry

The experimental arrangement for measurements of optical transmission is


straightforward. A latex film is cast on a transparent substrate. A light beam of a
known wavelength, , and with an intensity of I0 shines on the film perpendicular
to the surface. The intensity of the transmitted light, I, is recorded so that the
fraction transmitted, Tf, can be determined:

I
Tf = (2.9)
I0
2.1 Techniques to Study Latex in the Presence of Water (Wet and Damp Films) 51

Measurements are made over a range of , typically extending from the UV to the
near infrared.
A similar approach is to relate the optical transmission through a wet latex dis-
persion directly to the latex particle radius R and particle volume fraction, , via:
2
32 4 R3 d p m2 1
ln T f = 2 (2.10)
4 m + 2

where dp is the path-length through which the light is transmitted (i.e., sample
thickness) and m is the ratio of the refractive index of the polymer to the refractive
index of the solvent (e.g., the latex serum) (van Tent and te Nijenhuis 1992b). In
turn, some of these parameters can be related to the particle/particle spacing, ep, by
the relation:

R 3
= (2.11)
ep
C1 ( R + )3
2

where C1 is a geometric constant that depends on the particle arrangement in


space. Hence, measurements of Tf over time can be used to deduce particle
spacing during the drying process, provided that R and dp are known.
It is of particular interest to know how the particle spacing and geometric ar-
rangement evolve as a film dries. This information is provided in the signature of
interparticle interference. It is a fortunate coincidence that wavelengths of visible
light cover the relevant range of distances to probe the interparticle distances in
concentrated polymer colloids. A minimum in the transmission spectra will arise
from interparticle interference at a wavelength of min. If the particles with a
refractive index of np are arranged into a face-centred cubic (or hexagonal close-
packed) unit of height hunit, then min is simply hunitnp (van Tent and te Nijenhuis
1992b). Strong differences in the interparticle interference can be observed for
latices at temperatures above, below and near their MFFT (van Tent and te
Nijenhuis 2000). The interference diminishes at high temperatures where particles
deform to fill all available void space.
Once the film has almost completely dried, the amount of residual water can be
detected using a similar method. For isolated spheres of radius, z, in a continuous
medium, Rayleigh theory states that lnTf varies with z6 (van Tent and te Nijenhuis
1992). In the later stages of drying, interparticle pockets of water in a latex film
can be adequately modelled as spherical. In this case, the continuous medium is
the polymer phase and not the serum phase. Information on void closure can
thereby be obtained from measurements of optical transmission.
The Clausius-Mossotti equation tells us that the refractive index increases with
the density of a substance. For instance, if there is a phase change, the denser
phase will have a higher refractive index. Therefore, measurements of refractive
index can be used to estimate the volume fraction of air voids in a film. This
52 2 Established and Emerging Techniques of Studying Latex Film Formation

technique is best applied to dry films, when there is less ambiguity in interpreta-
tion of the data.
Tzitzinou et al. (2000) have shown how measurements of refractive index over
time can be performed using ellipsometry. The results were interpreted as the
closing of air voids as particles deform to fill space. Ellipsometry uses polarised
light that is reflected from a surface at a known angle of incidence. The change in
the state of the polarisation is characterised by the amplitude and phase of the
components of the light in the plane of the reflection and perpendicular to the
plane. Ellipsometry determines the change in the polarisation state upon reflection,
and the data analysis is used to extract useful parameters, such as the complex
refractive index.
Ellipsometry is applicable to all stages of latex film formation, starting from a
freshly cast film. The technique is particularly sensitive to the creation of air voids
in latex films in which the water level drops below the film surface (Keddie et al.
1995). In modelling ellipsometry data, a rough surface can be imagined to contain
air voids. Such a surface can be described as having a refractive index, n, that is
midway between the ambient air (n = 1) and the bulk material (for a polymer, n is
typically 1.5 1.6). Tzitzinou et al. (2000) showed that ellipsometry is sensitive to
the flattening of latex surfaces as particles deform and flow.

2.1.4 Profiling Water and Particles with Spectroscopies

During the drying of uniform films on large areas, the direction normal to the film is
of greatest interest. Several different spectroscopies have been adapted and im-
proved to provide profiles of water and/or polymer content in the vertical direction
with sufficient resolution. Techniques of infrared and Raman spectroscopy have
been well advanced to provide chemical and structural information as a function of
position within dynamic polymer systems (Koenig and Bobiak 2007). Infrared
radiation is absorbed by condensed matter at specific frequencies that correspond to
the energies of particular molecular vibrations, given by hc/, where h is Plancks
constant and c is the speed of light. Each bond and type of vibration has a character-
istic frequency. The molecular groups in large molecules can be identified through
determination of its IR absorption spectra. In Raman spectroscopy, light of a single
wavelength is inelastically scattered from the sample with a shift in frequency (and
energy) that is dependent on the particular molecular group.

2.1.4.1 Confocal Raman Microscopy

In the confocal arrangement, laser light is focused on a specified depth below the
surface to provide local information. The optics is designed so that only light from
the focal plane is collected in an image. The focus can be moved stepwise from
the substrate interface to the air interface. The technique can be properly quantita-
2.1 Techniques to Study Latex in the Presence of Water (Wet and Damp Films) 53

tive, because the Raman signal is linearly proportional to the concentration for the
particular Raman-active species.
The technique was first established as a way of providing information on film
composition as a function of depth from the surface. It thus provided insight into
the transport and distribution of surfactants in latex films in what is sometimes
called a post-mortem analysis, i.e., after drying was completed rather than as it
occurred (Belaroui et al. 2000).
Belaroui et al. (2001) provided a clear demonstration of the capability of the
technique for determining the water distribution in the vertical direction as latex
films dried. For water depth profiling, the water peak at 3500 cm1 can be followed
and compared to a peak from the polymer phase to determine the concentration. A
resolution in the vertical direction of 25 m has been reported (Belaroui et al.
2001), which is sufficient for profiling films that are a few hundred m thick. A
potential problem in the technique is fluorescence from chemical impurities or
from the substrate.
A recent development has been inverse-micro-Raman-spectroscopy, in which a
Raman spectrometer is coupled to an inverse microscope (Ludwig et al. 2007).
Laser light is brought into the specimen through a transparent plate that serves as the
substrate after being focused by an objective lens (Fig. 2.13a). As the light travels
through a wet latex film from the bottom, there is a decrease in the Raman signal
intensity, which is attributed to the effects of elastic light scattering. The loss in
intensity is not a function of the wavenumber (i.e., the frequency shift) but decreases
uniformly across the entire spectrum (Fig. 2.13b). Consequently, the ratio of the
polymer to water peak intensities can be used to measure the water content.
Measurements of water concentration can be made with inverse micro-Raman
spectroscopy both laterally in the plane of the film and as a function of the
direction normal to the film. A time resolution of about 1 second and an optical
resolution of 23 m have been reported. Inverse-micro-Raman spectroscopy can
therefore measure the progression of a drying front from the edge of a latex film.
It can also determine the ingress of water into a partially dry film in a re-
dispersion experiment in which a wet latex drop is applied to a partially dry film
(Ludwig et al. 2007).

2.1.4.2 IR Microscopy

In IR microscopy, a focused beam of infrared radiation is transmitted through, or


reflected from, a layer. A beam spot size of 50 m can be achieved. Guigner et al.
(2001) obtained water concentrations over distances of several mm. In their
experimental set-up, a rather unusual sample configuration was used. A 25 m
thick layer was sandwiched in a cell and evaporation was allowed only from one
side, rather than from the face of the film. Monitoring the change in intensity of
a water absorption band (such as the OH deformation band at 1645 cm1)
provides a quantitative measure of water content.
54 2 Established and Emerging Techniques of Studying Latex Film Formation

Fig. 2.13 a Experimental set-up for inverse confocal microscopy experiment. The light comes up
through the substrate (glass plate) and is focused at a desired lateral position. b Raman spectra
obtained at three different positions within a wet latex film. Peaks for the polymer phase and the
water phase are both observed. The ratio of the peaks remains the same, regardless of the
position, which is what is expected if there is a uniform lateral composition. (Reprinted with
permission from Ludwig et al. (2007))

2.1.4.3 NMR Profiling and Imaging

At the heart of nuclear magnetic resonance (NMR) techniques, is the fact that
spin- nuclei behave like tiny bar magnets and will align in a magnetic field in
one of two energy states, with the distribution between the states depending on the
temperature. If a pulse of electromagnetic radiation with the correct energy (in the
radio-frequency region of the spectrum) strikes the nuclei, they can be tipped from
the alignment direction. The nuclei precess along the direction of the field, just as
a childs toy top precesses around the direction of gravity, at the frequency of the
RF pulse. As they relax from an excited state, the nuclei emit energy in the RF
range, which is detected as an NMR signal.
NMR techniques rely on the principle that the resonant angular frequency, ,
of a spin- nucleus (e.g., 1H or 13C) is proportional to the local magnetic field
strength, B0. The two quantities are related through the magnetogyric ratio, , as
expressed by a simple equation:

= B0. (2.12)

The value of the magnetogyric ratio, , for 1H is given as /2 = 42.58 MHz T1


(Hore 1995). The frequency that satisfies (2.12) corresponds to electromagnetic
radiation in the radiofrequency (MHz) range. In NMR spectroscopy, there is a
small shift in (in parts per million) as a result of atomic nuclei feeling slightly
different B0 in different chemical environments because of the shielding effect of
electron clouds. Solid state NMR 1H spectroscopy can provide information on the
local environment of water in a latex film: whether free in the serum phase, in the
2.1 Techniques to Study Latex in the Presence of Water (Wet and Damp Films) 55

surfactant layer at the particle/water interfaces, or dissolved in a hydrophilic


polymer particle, such as poly(vinyl acetate) (Rottstegge et al. 2000).
When a system of nuclei is placed in a magnetic field gradient, defined as Gy =
dB/dy, then Eq. (2.12) must be modified. The expression for the resonant fre-
quency then becomes

(y) = (B0 + Gyy) (2.13)

At each position along the gradient direction, there will be a different as


illustrated in Fig. 2.14. Hence, the magnetic field gradient encodes spatial position
through the resonant frequency. This is the fundamental principle at the heart of
magnetic resonance imaging or MRI. Techniques of MRI are commonly em-
ployed in hospitals for medical diagnosis and follow up to treatment. In recogni-
tion of the medical importance of the technique, the 2003 Nobel Prize in Physiol-
ogy or Medicine was awarded to Paul Lauterbur and Sir Peter Mansfield for the
development of MRI.
In MRI scanners, the gradient in the magnetic field is created within a supercon-
ducting magnetic using electromagnetic gradient coils. The use of coils enables the
gradient to be turned on and off and to be varied in strength. Two-dimensional MR
images can be used to find the relative distribution of water laterally in drying latex
films (Salamanca et al. 2001), but it lacks sufficient resolution (ca. 25 m) to
provide much information about water in the vertical direction in films thinner than
a few hundred m. Nevertheless, MR imaging has sufficient resolution to yield
valuable information on the lateral flow of water in drying latex and emulsion films
(Ciampi et al. 2000) and enabled the successful validation of drying models (Sala-
manca et al. 2001). Furthermore, MRI can identify heterogeneous drying, with a
polymeric skin layer over a more dilute dispersion, in latex layers that are several
mm thick with an in-plane resolution of ca. 60 m (Rottstegge et al. 2003).
Standard MR images can be considered to be a map of the density of the nu-
clear spin of the particular nuclei, e.g., 1H, and hence are ideal for identifying
heterogeneities in water distribution. Images can also be weighted by the spin-spin
relaxation time, T2, which is correlated to the local molecular mobility. T2-
weighted images can be used to identify reduced polymer mobility in the skin
layers of drying latex films (Rottstegge et al. 2003).

Stray Field Imaging Techniques

To obtain high resolution images, strong and stable field gradients are required.
When imaging in a single direction, an attractive approach to achieving such a
field gradient is to do the experiment in the fringe or stray field of a supercon-
ducting magnet. At distances close to the magnet, the strength of the magnetic
field decreases with increasing distance, hence providing a stable, large G in a
field of high strength. Stray-field imaging (StraFI) (Newling and McDonald
1998), as the approach is called, has grown in popularity over the past few
56 2 Established and Emerging Techniques of Studying Latex Film Formation

decades. It has been applied to several problems in condensed matter and engi-
neering materials, such as water transport in cements or solvent ingress into
polymers (Mitchell et al. 2006). In these applications, information is provided over
distances of up to a few mm. The technique is suitable for the determination of
concentration profiles in the vertical direction in dispersions and emulsions
undergoing sedimentation or creaming (Newling et al. 1996).
StraFI is a relatively expensive technique because it requires a superconducting
magnet and the associated consumption of costly cryogens. Usually restricted to
small sample sizes, the profiling resolution can be detrimentally affected by the
meniscus of wet films. Although the resolution of stray field imaging is better than
that found in two and three-dimensional imaging techniques, it is not acceptable for
layers thinner than a few hundred m. Furthermore, as explained in the next
paragraph, the parallel arrangement of G with respect to the main field Bo is not
optimum as it precludes the use of a planar surface coil to create the RF field pulse.
Recognising these various limitations of conventional stray-field NMR for the
profiling of coatings, Glover et al. (1999) specially designed a magnet for this
application. The magnet is called GARField, which stands for Gradient At
Right-angles to the Field. In imaging the stray field of a high-field, vertical-bore
superconducting magnet, the static field with strength Bo is oriented parallel to the
magnetic field gradient, Gy, in the vertical direction. Therefore B1 from the RF coil
must be parallel to the sample plane. In the GARField geometry, on the other
hand, B0 is parallel to the sample plane, so that B1 is perpendicular to it. The
sample can thereby be positioned against a surface coil that produces the RF pulse,
and greater sensitivity can be achieved. Gy is perpendicular to B0 thus explaining
the name of the technique.
In order to image a thin slice in a film or coating, the B0 must be highly uni-
form in the plane of the film. Hence G must be highly homogeneous. In the
GARField design, the gradient in the magnetic field is produced by the introduc-
tion of a taper in the pole-pieces on a pair of permanent magnets (made from
NdFeB or similar material). The required shape to create the desired gradient field
was found analytically using a scalar potential method to solve the Laplace
equation (Glover et al. 1999). The design of the GARField magnet is shown in
Fig. 2.15. There is a curvature in the field of less than 5 m over a 5 mm x 5 mm
area. In the original design, samples are placed in a permanent magnetic field
strength of B0 = 0.7 T, and a magnetic field gradient strength of G = 17.5 T m1
was achieved. A newer magnet design has pole pieces shaped to provide two
different magnetic field gradients (G/Bo = 16.7 m1 and G/Bo = 33.3 m1) using a
single pair of permanent magnets providing the same static field (Bennet et al.
2003). The design has since been modified to accommodate samples with a
greater area. In the GARField design of Erich et al. (2005, 2005b), shaped pole
pieces are mounted on an electromagnet system, rather than on a permanent
magnet. A constant magnetic field of B0 = 1.4 T and a gradient of Gy of 34 T/m
are achieved. The GARField technique is closely related to the NMR MOUSE
(Eidmann et al. 1996), which provides one-dimensional depth profiles into a
sample by pressing a radiofrequency coil up against a free surface.
2.1 Techniques to Study Latex in the Presence of Water (Wet and Damp Films) 57

(a) (b)
B

Gy
B0

B y
(c)

y
Fig. 2.14 The basic principle of magnetic resonance profiling. a The resonant frequency, (also
known as the Larmor frequency) is related to the magnetic field strength through . b When a
magnetic field gradient, Gy, is applied, the magnetic field varies with position. c The position is
thereby encoded in the frequency, as is stated in Eq. (2.13).

(a) B1
y Gy

Latex film
Curved Curved
pole piece pole piece B0
RF coil

Permanent magnets

(b)

Fig. 2.15 a Illustration of the curved pole pieces in a GARField magnet, which create a magnetic
field gradient, Gy, in the vertical direction. The latex film rests on a radiofrequency coil in the
gradient field between the pole pieces of the magnets. b A photograph of the pole pieces in a
supporting frame, viewed from above. (Courtesy of the University of Surrey)
58 2 Established and Emerging Techniques of Studying Latex Film Formation

GARField Data Interpretation

Experimental details and a discussion of the factors that influence the resolution of
GARField profiles are found in Appendix B. Fig. 2.16 relates three different types
of vertical water distribution to the resulting GARField profile. In a profile, the
NMR signal intensity is plotted as a function of the vertical position (height)
within the film. The intensity is proportional to the density of the mobile 1H. The
1
H in glassy polymers is not sufficiently mobile to yield an NMR signal in this
experiment. Hence, the profiles are only sensitive to water in a hard latex film. A
signal is obtained, however, from viscous or rubbery polymers, such as in adhe-
sives and alkyd coatings, which are not crosslinked. The signal is then the sum of
contributions from the water and polymer. The water content can be deduced by
subtracting out the contribution to the signal from the polymer phase.
Drying uniformity in waterborne coatings in the vertical direction has been
probed with the GARField magnet. A typical result is presented in Fig. 2.17, where
the changes over time of the film thickness and water content are observed. Water
profiles can be used to test models of drying (Gorce et al. 2002, Ekanayake et al.
2009) and to surmise the extent and mechanisms of particle deformation (Mallgol
et al. 2001). The penetration of water from a latex film into a porous substrate can
also be examined (Bennet et al. 2003). Low water content near latex film surfaces
has confirmed the process of skin formation during drying (Mallgol et al. 2006).
Reduced polymer mobility resulting from crosslinking reactions can be followed as
a function of depth into waterborne films (Wallin et al. 2000) through measurements
of the T2 spin-spin relaxation time as a function of depth in a film. (Measurements T2
are obtained from an exponential fit to multi-echo profiles.) Drying uniformity can
be determined simultaneously with studies of crosslinking.

2.1.5 Probe Techniques for the Aqueous Environment

A new approach to the study of latex drying, as proposed by Baumgart et al.


(2000), is to use a variety of small molecules to probe the aqueous phase and
interfacial regions throughout the process. There are three main types of probe: (1)
spin probe molecules (with unpaired electrons) for electron paramagnetic reso-
nance (EPR); (2) photochromic dyes for forced Rayleigh scattering (FRS) and (3)
donor/acceptor fluorescent dye molecules for non-radiative energy transfer
(NRET) experiments. The value of probe molecules is that they are inside the wet
film and yield insight into the local environment, which is continuously changing
as water content decreases and particles come closer together. Each of these non-
invasive approaches will be considered separately, although their impact is
potentially greatest when used in parallel.
2.1 Techniques to Study Latex in the Presence of Water (Wet and Damp Films) 59

Fig. 2.16 Interpretation of GARField NMR profiles. The signal is proportional to the density of
mobile 1H. The horizontal axis shows the vertical position in the layer. a If the particle distribu-
tion is uniform, then the profiles are square. When water evaporates, the thickness decreases and
the water concentration decreases. b If there some particle accumulation near the film surface,
then the profile slopes downward. c If the particles at the film surface coalesce to create a skin
layer with water underneath, then there will be a step in the profile. (It is assumed here that there
is a signal from the 1H in the polymer.)
60 2 Established and Emerging Techniques of Studying Latex Film Formation

intensity
0.5

0.4
intensity
Time
signal
0.3
Relative

0.2
Relative

0.1

0.0
-50 0 50 100 150 200 250
Depth (m)
Height (m)
Fig. 2.17 A series of GARField MR profiles obtained over time from a drying latex film. The top
of the film is represented at the greater heights. The signal intensity is proportional to the water
concentration. With increasing drying time, each successive profile indicates a lower film
thickness and a lower water concentration.

EPR requires the introduction of a spin probe, which is a molecule with an


unpaired electron. Nearly spherical molecules are often used, such as TEMPO
(2,2,6,6-tetramethylpiperidine-1-oxyl), which has an unpaired electron in its
nitrogen atom. In wet latex, probes are introduced easily as small molecules or
surfactant probes in the serum phase at concentrations of 1 mmol L1 or less
(Cramer et al. 2002a). The interaction of the probe with its environment is found
by sweeping an external magnetic field while continuously irradiating with a fixed
microwave frequency, so as to induce transitions of the electron spin (Weil et al.
1994). EPR of drying latex has provided information on the local viscosity in the
vicinity of ionic and polar probes as a function of water content. It also measures
the mobility of molecular or surfactant probes in the changing local environment
within the latex. The orientation and position of surfactants in relation to the
particles and serum can also be determined. Micellisation and surfactant aggrega-
tion can be followed (Cramer et al. 2002b).
FRS is ideal for monitoring slow diffusion processes, ranging between 1013
and 1021 m2s1, and hence is well suited for studying the later stages of latex
drying (Veniaminov et al. 2002). The relative hydrophobicity of the dye deter-
mines what is probed in an experiment. A hydrophobic dye is sensitive to the
polymer and polymer-water interface. A hydrophilic dye can probe the serum
phase and surfactant clusters but also can be partitioned in the polymer phase and
interfacial regions (Veniaminov et al. 2003). Thus, depending on the probe, it is
possible to study topics ranging from the continuity of the serum phase and water
loss from various components to the possible break-up of particle boundaries and
the onset of coalescence.
2.2 Techniques to Study Particle Packing and Deformation in Dry Films 61

The main concept of FRS is to produce a period structure, or grating, in a mate-


rial using dye molecules and then to follow the loss of this grating as dye diffusion
proceeds (Eichler et al. 1986). Diffraction of light from the grating is sensitive to
the spatial modulation of the complex refractive index that is created by the dye
molecules. The relaxation of the diffraction over time provides a way to determine
the tracer diffusion coefficient of the dye. The spacing of the lines in the grating
can be adjusted to measure diffusion over various length scales and hence to
characterize the size of heterogeneities (Veniaminov et al. 2002).
NRET, known also as FRET (fluorescence resonance energy transfer), is well
established as a means of probing the interdiffusion of polymer molecules
between latex particles. Its principle of operation will be explained in Section 2.4.
In the work of Baumgart et al. (2000), the aim of NRET was to determine if, and
to what extent, the hydrophilic interface (or membrane) of latex particles can
restrict the transport of tracer molecules. Hence, the tracer molecules can be used
to probe the integrity of the particle interfaces.
In a typical NRET experiment, one phase contains a donor dye, such as phe-
nanthrene, and the other phase contains an acceptor dye, such as perylene or
anthracene. In latex experiments, particles containing donor dyes are blended with
particles containing acceptor molecules. The steady-state fluorescent intensity of
the acceptor molecules, IA, is affected by their distance from the donor molecules.
As their separation decreases, such as when the membranes separating the
particles break down, IA increases in a known way. The technique has been used
to study the early stages of latex film formation when particles first make contact
in a drying film (Turshatov et al. 2008). However, it has primarily been used to
study the interdiffusion of polymer chains across particle boundaries in dry films,
and it will be considered in detail in Section 2.4.

2.2 Techniques to Study Particle Packing and Deformation in Dry


Films

Many of the techniques that yield information on the particle distribution in wet
films can be extended to the case where particles start packing together and
deforming. SANS, SAXS and optical transmission all fall in this category. There
are other techniques that cannot, for various reasons, be applied to films in the wet
state. These will now be considered.

2.2.1 Scanning Probe Microscopies

Scanning probe microscopies consist of several techniques, each with a fundamen-


tal strategy in common. An ultra sharp probe tip on the end of the cantilever is
brought in close enough proximity with a surface to be able to feel attractive and
62 2 Established and Emerging Techniques of Studying Latex Film Formation

repulsive forces. Examples of cantilevers and tips are shown in Fig. 2.18. The tips
can have a radius of curvature as low as 10 nm. The various microscopies can then
be classified depending on the type of force that is detected: molecular, electro-
static charge, magnetic, shear, etc. Although there is at least one example of
scanning probe microscopy on a wet latex surface (Butt et al. 1994) and numerous
examples of microscopy in liquids, it is usually applied to dry latex surfaces or to
solid surfaces immersed in liquids.

(a) Tip (b) (c)

Cantilever

Fig. 2.18 SEM images of three different AFM cantilevers. a A cantilever with a triangular end
and a sharp conical tip that is angled away from the end of the cantilever. Scale bar is 40 m.
b A rectangular cantilever with a conical tip that is normal to the cantilever. Scale bar is 20 m.
c A cantilever with a triangular end and a pyramidal tip. Scale bar is 15 m. (Images courtesy of
Dr. Chun-Hong Lei, University of Surrey)

2.2.1.1 Contact Atomic Force Microscopy (AFM)

Just six years after the invention of the atomic force microscope was reported
(Binnig et al. 1986), it was used to look at close-packed arrays of latex particles
(Wang et al. 1992). In the first mode of AFM to be developed, known as the
contact mode, the tip on the cantilever is brought to the surface where it feels a
repulsive force and is deflected. The tip is moved across the surface, usually by
the x-y motion of a piezoelectric crystal on which the sample sits. A laser light
beam that is reflected from the top side of the cantilever tracks the cantilever
deflection in a position-sensitive detector. In the constant force mode of operation,
a feedback loop is used to adjust the vertical position of the sample while scanning
in order to maintain a constant deflection of the cantilever, and hence a constant
force. A record of the movements in the vertical direction is used to build a
topographic image of the surface (Wang et al. 1992).
When AFM is operated in the constant force mode, a relatively weak cantilever
is used to provide good sensitivity. The spring constant, which characterises the
stiffness, is as low as 0.02 N/m, and the imparted force is on the order of 10 nN
(Butt and Kuropka 1995).
In the 1990s, contact AFM provided many fundamental insights into film struc-
ture, including the effects of surfactant on particle packing (Juhu and Lang
1994a), the flattening of particles during annealing (Goudy et al. 1995, Goh et al.
1993), the role of capillary forces in particle deformation (Lin and Meier 1995),
and the exudation of surfactant (Juhu et al. 1995, Tzitzinou et al. 1999). In these
early studies, the radius of the AFM tip was as large as 40 nm and the angle of tip
2.2 Techniques to Study Particle Packing and Deformation in Dry Films 63

as large as 110 (Lin and Meier 1995). Hence, there were physical limitations to
the tip being able to sense the bottom of the valleys between convex particles
(Butt and Gerharz 1995). Researchers devised methods to measure the tip geome-
try (Odin et al. 1994) and to compensate for its effects on images through decon-
volution methods (Markiewicz and Goh 1995). Now, ultra sharp tips are commer-
cially available, and so tip deconvolution is not always needed.
When the patch of the reflected light beam is perpendicular to the direction of
the scan, the tip deflection is an indication mainly of surface topography. How-
ever, when the light path is in the scan direction, the cantilever deflection has a
contribution from variations in friction between the tip and sample. This enables
the creation of maps of the friction variations across a latex surface as a means of
distinguishing material components (Butt et al. 1994).
AFM is an attractive technique because there is no need for sample preparation
and it operates in air. As demonstrated in an early study of latex (Butt et al. 1994),
contact AFM with a constant force suffers from one limitation: even though the
force of the tip acting on the surface is small, it is still sufficient to damage the
surface or to scrape away molecular layers of surfactant.

2.2.1.2 Intermittent Contact AFM and Phase Imaging

Starting in the mid 1990s, another mode of AFM began to be used to determine
latex film structure. Called intermittent contact or by the trade name of Tap-
pingModeTM, this mode is potentially less damaging to soft surfaces, such as latex
(Sommer et al. 1995, Patel et al. 1996).
In intermittent contact AFM, the tip and the cantilever are made from the same
material, usually silicon. The cantilever is driven to oscillate near its natural
resonant frequency (in the kHz range) by driver electronics. When it is brought
into close proximity to a surface, the tip makes intermittent contact at the bottom
of the oscillation. Parameters needed to describe the tapping conditions are the
free amplitude Ao (corresponding to oscillation in air) and the set point amplitude
Asp (corresponding to the amplitude when the tip is in contact with the sample
surface) (Fig. 2.19a).
A freely oscillating cantilever has a resonant frequency f0, at which the tapping
amplitude will be at a maximum value of A0(f0). Imaging is performed at a
frequency f set slightly lower than f0 such that A0(f) is typically 5% less than the
peak value at A0(f0). It is important to note that the AFM tip does exert an average
force, Fav, on the sample surface, and this force is sufficient to indent into it. The
lower that the set point tapping amplitude Asp is in comparison to A0, the greater is
the value of Fav. This important parameter is given in a semi-empirical equation
(Yu et al. 2000) as:

1k Asp
Fav = 1 A0 ( f ) , (2.14)
2 q A0 ( f 0 )
64 2 Established and Emerging Techniques of Studying Latex Film Formation

with the off-resonance parameter given as

A0 ( f )
= , (2.15)
A0 ( f 0 )

and where q is the quality factor of the cantilever. Following from the above
statements, < 1 in all cases. Note the importance of the ratio of Asp to A0 in
determining Fav, rather than the individual value of Asp alone. For a particular
AFM tip, q is obtained from the ratio of f0 to the full bandwidth at 0.707 of the
maximum amplitude, f :

f0
q= . (2.16)
f

q is a function of the atmosphere around the cantilever. The narrowest resonance


is obtained in vacuum, and q is thus higher in vacuum compared to air or a liquid.
Fav can be calculated in experiments using Eqs. 2.14, 2.15 and 2.16.

Indentation Depth and Height Artefacts

It is, however, important to realise that an AFM tip does indent to a depth of zind
into a soft (low Tg) surface (Mallgol et al. 2001). When the tip of an oscillating
cantilever with a set point amplitude of Asp is in contact with a soft polymer
surface, the distance between the tip and sample, dsp, is always less than Asp. Then,
zind is found from the difference between these two values:

zind = Asp dsp . (2.17)

This relationship between the three parameters is illustrated in Fig. 2.19b.


A chief concern among AFM users is how to determine the true structure of a
soft surface. In providing topographic images of soft surfaces, the technique is
prone to height artefacts that may be misleading to the viewer. For several years, it
has been recognised that height images obtained with intermittent contact (also
called the tapping mode) do not necessarily indicate true surface topography
(Knoll et al. 2001). For instance, it has been shown that differences in the tip
indentation depth into hard and soft phases on a smooth triblock copolymer
surface create a false impression of surface topography (Kopp-Marsandon et al.
2000). That is, instead of representing the true surface topography, the image is a
map of indentation depths. In a different study of a triblock copolymer, an
inversion was observed between which block is considered to have the highest
topography. This inversion was explained by specific tip-sample interactions
dominating in different set point regimes (Wang et al. 2003). In similar work on
2.2 Techniques to Study Particle Packing and Deformation in Dry Films 65

blends of glassy and rubbery polymers (Raghavan et al. 2000), differences in the
stiffness of the two domains caused a reversal in which phase appeared to have a
greater height, when the tapping force was high. A second reversal was observed
at lower tapping forces and was attributed to a switch over between attractive and
repulsive regimes between the tip and sample. This type of artefact in colloidal
films has not been well documented or described in the literature in comparison to
the amount of attention devoted to diblock polymers and polymer blends. In soft
latex films, the apparent roughness and topography is highly sensitive to the
tapping force, as adjusted through the set point ratio (Mallgol et al. 2001).

2
(a)
Ao: free amplitude

Asp: setpoint amplitude

(b)

2Asp
dsp
Asp
zind
Soft surface

Fig. 2.19 a Illustration of an AFM cantilever oscillating in air with a free amplitude, Ao and in
contact with a surface at a set point amplitude, Asp. b Diagram showing the inter-relationship of
Asp, the tip-sample distance (dsp) and the indentation depth (zind). (Reprinted from Lei et al.
(2008) with permission from Elsevier)

The choice of AFM parameters affects whether the latex particles in a film
appear convex or concave (Lei et al. 2007). Fig. 2.20 shows images of the same
latex film surface using two different set point ratios. Tip indentation into the film
surface is variable, depending on the position, because the latex particles are
surrounded by a surfactant-rich component. In Fig. 2.20a, a relatively high Asp/A0
ratio was used, such that Fav is low (1.3 nN), whereas Fig. 2.20b is the result of a
relatively low Asp/A0 ratio, such that Fav is high (4.4 nN). In the height images, it is
notable that particles appear to be concave under a low Fav but convex under the
higher Fav. These observations stem from how the indentation force varies with
depth in the latex particles in comparison to the surfactant-rich component.
66 2 Established and Emerging Techniques of Studying Latex Film Formation

(a) (b)

Fig. 2.20 Topographic images of the same latex film surface. a With a high set point ratio, the
particles appear concave in shape. b With a low set point ratio, the particles appear to be convex
in shape. (Reprinted from Lei et al. (2008) with permission from Elsevier)

Phase-Contrast Imaging

In the intermittent contact mode of operation, a phase image may be obtained


simultaneously with the height image. The phase image presents the phase lag, ,
of the photodiode output signal in relation to the driving piezoelectric signal as a
function of position within the scan area. Changes in reflect variations in the
energy dissipation, ED, of the cantilever as its tip moves laterally across a surface.
is related mathematically to ED as (James et al. 2001, Anczykowski et al. 1999):

f Asp qE D
sin = +
. (2.18)
f 0 A0 kAsp A0

During a scan, all of the parameters in Eq. (2.18) except ED are fixed, so the
phase image provides a map of ED (Anczykowski et al. 1999). When the tip
interacts with a highly viscous region on a surface, or a viscoelastic region in
which the viscous component is dominant, more energy will be dissipated, and
therefore ED and will be greater. When the tip interacts with a viscoelastic
region in which the elastic component is dominant, it is however expected that less
energy will be dissipated, making smaller (Scott and Bhushan 2003). In a phase
image, the magnitude of at a particular position is represented by the relative
shade (i.e., darker or lighter). Two possible scenarios in which topographic and
phase-contrast images from the same surface differ are presented in Fig. 2.21. In
images obtained from latex films, the topographic and phase-contrast images do
indeed show different features (Fig. 2.22).
2.2 Techniques to Study Particle Packing and Deformation in Dry Films 67

The reasons for contrast in phase images have been a subject of frequent inves-
tigation (Anczykowski et al. 1999, Scott and Bhushan 2003). Phase contrast has
been shown to be independent of variations in the elastic moduli (i.e., stiffness) of
polymers (Tamayo and Garcia 1997, Bar et al. 1999) because the energy used in
surface deformation is elastically recovered. One exception is when the elastic
modulus of a region is low. In this case, the AFM tip will push more deeply into
the surface and thus have a greater contact area with the deformed surface. The
interaction energy will then be greater (Anczykowski et al. 1999). On the other
hand, when the tip interacts with a viscous material, energy is dissipated. There is
convincing evidence from experiments and modelling (Scott and Bhushan 2003,
Leclre et al. 2002) that the energy dissipation in tip-sample interactions is greater
in viscoelastic materials with a high viscosity. Variations in viscosity across a
surface therefore lead to contrast in phase images and have indeed been observed
over lateral distances of a few nm (Leclre et al. 2002). The contrast in both height
and phase images has been found to reverse in regimes where there is either a very
low or a very high Asp (Scott and Bhushan 2003, Raghavan et al. 2000). A bistable
regime for has also been reported (James et al. 2001).
The air surface of a latex film is most commonly studied with AFM, but the
film can be also delaminated from its substrate to examine this lower interface.
Furthermore, cross-sections can be cut in a microtome apparatus that uses an ultra
sharp blade to slowly cut a film, usually while keeping the sample at a low
temperature to prevent mechanical deformation. An example AFM imaging being
applied to a film cross-section is given in Fig. 2.23.

Fig. 2.21 Examples of the difference between topographic and phase-contrast images. a A surface
with depressions of different depths and with a uniform composition might not show any features in
the phase-contrast image. The shade in the topographic image is proportional to the depth of the
depression. b A surface that is smooth but that is chemically heterogeneous might show features in
the phase-contrast image. More viscous regions will result in greater energy dissipation when the tip
interacts with a surface, and hence there is a phase shift in the image. The chemical heterogeneity
will not necessarily be apparent in a topographic image. (The possible effect of greater indentation
into viscous regions is neglected here.)
68 2 Established and Emerging Techniques of Studying Latex Film Formation

(a) (b)

(c) (d)

Fig. 2.22 AFM images (10 m x 10 m) of a latex film. a Topographic image showing circular
regions with a shallow depression. b Phase-contrast image for the same film surface. There is no
evidence for the circular regions. Reprinted from Lei et al. 2008 with permission from Elsevier. c
In the AFM images of a latex film with a surfactant layer on its surface, the topographic image
shows a flower-like pattern. d Phase-contrast image shows a pattern that corresponds to the
topography but additionally shows evidence for surface layers that are not apparent in the
topographic image. (Images courtesy of Dr C. Lei, University of Surrey)

ca. 40 m
latex film
Supporting
polypropylene
sheets AFM cantilever

Fig. 2.23 Photograph showing a view from above while performing AFM imaging of a latex film
cross-section. The film is sandwiched between two polypropylene sheets to provide mechanical
support. (Photograph courtesy of Dr C. Lei, University of Surrey)

A particular problem in the AFM analysis of latex films is the contamination of


the tip. Latex films contain free surfactant and other water-soluble materials (e.g.,
oligomers and salts) which can adsorb onto the tip. The interaction forces with the
surface are modified when the tip is contaminated (Fig. 2.24). Its effective size
becomes larger with contamination, and hence the image resolution can be
decreased. When imaging hard surfaces or when applying a high force on a
cantilever, the sharp tip can become blunted, with a detrimental effect on image
quality and resolution.
2.2 Techniques to Study Particle Packing and Deformation in Dry Films 69

(a) (b) (c)

contamination

Fig. 2.24 SEM images at high magnification showing AFM tips of various condition. a A conical
tip that is sharp and uncontaminated. Scale bar is 375 nm. b A pyramidal tip that has been
damaged. This blunted tip will not yield images with a high resolution. Scale bar is 750 nm. c A
conical tip after being used to scan a latex film. Contamination, attributed to surfactant, is visible
on the surface. Scale bar is 375 nm. (Images courtesy of Dr. Chun-Hong Lei, University of
Surrey)

2.2.1.3 Electric Force Microscopy (EFM) and Scanning Electric Potential


Microscopy (SEPM)

Electric force microscopy, EFM, gives information on the electrostatic charge on a


surface as a function of position. It was first developed not only as a way of
imaging the charge distribution laterally across a surface, but also as a way to
deposit charge in selected areas at the nm length scale (Terris et al. 1990). EFM is
a so-called two pass technique, in that the creation of an image requires two
scans or passes. In the first scan, the topography is determined by non-contact
AFM. Then, in the second scan, the cantilever is driven at its resonant frequency
while grounded or biased by a DC voltage. The spatial derivative of the capacitive
tip-sample electric force leads to a shift in the resonant frequency. Accordingly,
the oscillation amplitude decreases and there is a shift in the phase of the oscilla-
tion. Images of the electric potential distribution can be created from the devia-
tions of either the amplitude or the phase. It was over a decade from its develop-
ment before reports of its application to latex films appeared. Keslarek et al.
(2002) found variations of 200 mV across the surface of dialysed latex film
surface compared to 80 mV variations in as-prepared material.
The technique of scanning electric potential microscopy (SEPM) has been used
extensively for the study of dry latex films. The technique is similar to other
scanning probe microscopies in that a tip in close proximity to the surface is
rastered back and forth. In this case, the scanning probe is a Kelvin-bridge
electrode, so that SEPM is also known as scanning Kelvin probe force micros-
copy (Nonnenmacher et al. 1991).
An SEPM image provides a map of the distribution of electric potential sensed
by the tip at a fixed distance from the surface. The tip is coated with Pt and fed
with an AC-signal that matches the resonant frequency of the cantilever-tip
system. Charges buried up to 100 nm below the surface can be detected, depend-
ing on the dielectric constant and charge distribution in the near-surface region.
However, there is greater sensitivity to charges at the surface. A charge buried 20
nm below the surface is nine times less effective than a surface charge (Keslarek
70 2 Established and Emerging Techniques of Studying Latex Film Formation

et al. 2001). As latex contains many charged species, such as salts and ionic
surfactants, it is well suited for the technique. Topography is independent of the
charge distribution; the two type of features can be studied independently (Galem-
beck et al. 2001). Remarkable electrical patterns in dried films have been observed
and correlated to the packing of particles into arrays (Santos et al. 2004).
Comparing the two related techniques, EFM provides sharper images in compari-
son to SEPM. But, SEPM provides absolute measurements of surface potential,
whereas EFM gives information only on its spatial variation (Keslarek et al. 2002).

2.2.2 Scanning Near-Field Optical Microscopy (SNOM) and Shear Force


Microscopy

When this technique was first developed, it was called by the descriptive name of
optical stethoscopy. Pohl et al. (1984) pointed out that a doctors stethoscope is
able to resolve the position of a beating heart by detecting very long sound waves
( ~100 m), and thus achieving resolution on the order of /1000. They drew an
analogy to the detection of light waves that are relatively long in comparison to a
sub-m object. The stethoscope achieves this good resolution by having a very
narrow aperture at its end and by bringing it very close relative to to the
object being studied, the heart. This approach is used in the SNOM technique on
shorter length scales. Light is brought in through a very narrow aperture, and the
distance between the aperture and object under study is kept smaller than . To
increase the resolution, the tip-surface distance must be made shorter, or the radius
of the tip must be made smaller. In the first demonstration of the technique (Pohl
et al. 1984), a resolution of /20 or 25 nm was achieved. The optical stetho-
scopy technique was later called near-field optical scanning or NFOS micros-
copy (Drig et al. 1986), but today SNOM is the standard name.
At the heart of the original SNOM design, is a tiny aperture in a conducting
metal screen. The aperture is on the apex of a transparent material that is illumi-
nated from behind. When light is forced through an aperture that is smaller than ,
evanescent wavelets are created that have a wavelength less than the size of the
aperture. The wavelets cannot travel through free space but wrap around the
aperture to create contours of constant energy. In this way, energy is radiated
across the aperture (Drig et al. 1986).
An important further development of SNOM came with the use of a transmissive
aperture at the apex of the tapered tip of a metal-coated optical fibre (Betzig et al.
1991). This very small aperture (< 20 nm) was brought in close proximity of the
surface (/50) to achieve a high resolution (12 nm) corresponding to about /43.
For an optical fibre tip to be scanned across a surface, while keeping the dis-
tance between the tip and sample at a fixed value, there must be a means of
regulating the tip vertical position in response to the changing surface topography.
A commonly used approach is to detect the shear forces between the near-field tip
2.2 Techniques to Study Particle Packing and Deformation in Dry Films 71

and the sample surface in a technique known as shear-force microscopy (Betzig


et al. 1992). One of the first demonstrated applications was the imaging of
polystyrene latex particles of diameters 230 nm and 19 nm (Betzig et al. 1992).
In contrast to conventional scanning probe microscopy, shear-force microscopy
uses cantilevers that have a high stiffness (indicated by its spring constant) normal
to the surface but a low stiffness in the horizontal plane. When shear-force
microscopy and SNOM are combined in use, the tapered optical fibre tip is used
as a mechanical probe of surface topography while also being the source of
evanescent light for optical image formation.
There have been just a few reports of SNOM used for the examination of struc-
ture in latex films. The serum solute (including water-soluble polymers and
surfactant) in sub-monolayers of latex particles has been observed by SNOM
(Teixeira-Neto et al. 2004). Variations in evanescent wave intensity found in
SNOM images of latex films have been attributed to minute differences in
chemical composition of individual particles (Teixeira-Neto et al. 2003).
Image contrast in SNOM analysis arises mainly from differences in the refrac-
tive index across a surface. Most polymers and organic materials have roughly
similar refractive indices, and so contrast is not strong. One way to improve
contrast is through the addition of fluorescent molecules. The emitted light can be
detected in reflection or transmission modes.

2.2.3 Electron Microscopies

2.2.3.1 Transmission Electron Microscopy (TEM)

As a way of determining the extent of particle deformation and packing, TEM has
been employed longer than all other techniques, with reports on the topic emerg-
ing in the early 1950s (Dillon et al. 1951, Bradford 1952) when the technique was
still new. (The worlds first design of an electron microscope was in 1933 by Ernst
Ruska, who shared a Nobel Prize for this work 53 years later.) There has been a
gradual evolution in the approach to sample preparation for TEM analysis. In the
earliest work (Bradford 1952, Bradford and Vanderhoff 1963), replicas of the
latex surface were made. The replicas are usually deposited onto the latex surface
from the vapour phase by sputtering or thermal evaporation. Next, the latex film is
dissolved in a good solvent for the latex to leave a free-standing representation of
the original surface. TEM is then performed on the replicas rather than on the
original sample. Replicas in the past were evaporated silicon monoxide or
collodion deposited from a solution (Bradford 1952, Bradford and Vanderhoff
1963), but carbon replicas were used in later work.
In the earlier work, replicas were only made of the interfaces with air or the
substrate. Microscopists had to ensure that the replicas faithfully represented the
surfaces without loss of resolution or the introduction of artefacts. In some work,
72 2 Established and Emerging Techniques of Studying Latex Film Formation

the silicon monoxide replicas were highly fragile, so that cracks can be seen in the
images (Vanderhoff 1970). A more important limitation is that no information
about the internal particle structure and arrangement is provided.
In the 1990s, there was an important technical development, learned from bi-
ologists, which overcame the limitations of previous work. The inside of dry
latex films was revealed for the first time by freezing the films, fracturing them in
a vacuum while maintaining temperatures ranging between 115C and 140C.
The fracture surface was then replicated for TEM analysis. More durable replicas
were created by depositing thin layers of Pt onto the surface of interest, and then
supporting the replica from the backside with a thicker layer of deposited carbon
(Roulstone et al. 1991, Wang et al. 1992). Freeze-fracture TEM (FFTEM), as the
technique is called, provided the first experimental evidence that latex particles
packed in a cubic array deformed to fill space and created rhombic dodecahedra
(Wang et al. 1992). One of their striking images appears in Chapter 4 as Fig. 4.1.
Point defects in colloidal crystal films, as also observed at the atomic level in ionic
crystals, have been identified using FFTEM (Sosnowski et al. 1994).
As an alternative to replication, ultrathin slices of dry films can be cut using a
microtome (Distler and Kanig 1980). The slices, which are supported on a fine
metallic or carbon grid, are thin enough so that the electron beam can pass through
it in TEM analysis. Microtoming on low Tg materials is often carried out at
cryogenic temperatures to avoid damage during cutting.
To achieve greater image contrast between polymer phases, selective staining is
often used. For instance, OsO4 is an effective stain for ester groups (Du Chesne et
al. 1997) as is uranyl acetate for acrylics and acrylic acid groups (Joanicot et al.
1993). In a core-shell latex film, RuO4 exposure was used to stain poly(butadiene)
and poly(butyl acrylate) in particle cores to create contrast with a poly(methyl
methacrylate) shell (Domingues dos Santos et al. 2000). The cellular structure of
latex films usually cannot be seen in TEM without the use of stains (Joanicot et al.
1993, Distler and Kanig 1980). An alternative to staining is the use of energy
filtering TEM with electron spectroscopic imaging (ESI). To obtain images of
surfactants that contain hetero-atoms not found in the continuous polymer matrix,
ESI and elemental mapping can be applied to microtomed film slices or on film
replicas (Du Chesne et al. 1997).

2.2.3.2 Scanning Electron Microscopy (SEM)

Since becoming available commercially in the 1960s, and since reports of its
application to latex film formation in the 1970s (El-Aasser and Robertson 1975),
SEM has become a relatively routine method for a wide range of applications.
Latex film structure and surfactant distribution can readily be visualised, even
when the surface is coated with a thin conducting film, such as Au (Eckersley and
Rudin 1993). Environmental SEM, which was described in Section 2.1.2.2, is
applied to dry films also, as it eliminates the need for an electrically conductive
coating and the associated imaging artefacts (Keddie et al. 1996). Microscopes
2.3 Techniques to Study Film Crosslinking 73

with a field emission source of electrons are often used to achieve higher resolu-
tion (Ming et al. 1995, Teixeira-Neto et al. 2003). Low voltage SEM reduces
polymer beam damage during the analysis. Consequently, there is greater time
available for focussing, so that images at higher magnifications can be obtained
(Gaillard et al. 2004). Images created from backscattered electrons provide
information about the distribution of elements across the surface, because the
backscattered energy is proportional to atomic number. Backscattered electron
images offer poorer resolution than images produced from secondary electrons,
but they are not prone to the effects of charge build-up (Teixeira-Neto et al. 2002).

2.3 Techniques to Study Film Crosslinking

A complete study of crosslinking reactions requires both chemical information (to


confirm the relevant reactions) and mechanical information (to correlate with
crosslink density). The ideal technique would be non-invasive, applicable in a range
of atmospheres, and provide data as a function of depth and lateral position in a film.

2.3.1 Ultrasonic Reflection and QCM

Ultrasonic reflection techniques, already described here, can provide measure-


ments of the dynamic moduli as they evolve in crosslinking materials, such as
epoxy resin (Lellinger et al. 2002). A related technique uses the shear wave
generated by a quartz crystal vibrating at its resonant frequency to probe the
viscoelastic properties of thin films. This technique, called quartz crystal micro-
balance (QCM) with dissipation, has shown some promise for characterising
crosslinking in waterborne alkyd emulsion films (Hellgren et al. 2001). A meas-
urement of the amount of energy dissipated during the shear oscillation indicates
the extent of crosslinking. A film with a high crosslink density will be stiffer and
dissipate less energy. A limitation of this approach is that it is only applicable to
films that have a thickness less than the wavelength of the ultrasound (< 2 m),
but is has the attractions of being non-invasive and applicable in any atmosphere.
Recent developments in the applications of QCM have been reviewed (Jo-
hannsmann 2008)

2.3.2 Spectroscopic Techniques

An enormous body of work in developing photoacoustic (PAS) FTIR spectros-


copy and other IR techniques was produced by the team led by Urban (1997).
Their work mainly concerns the bonding interactions and distribution of surfac-
74 2 Established and Emerging Techniques of Studying Latex Film Formation

tants in dry latex films. According to Belaroui et al. (2000), PAS is not properly
quantitative and depth profiling is only possible through analysis of layers of
increasing thickness ranging up to 20 m.
In NMR spectroscopy, the lifetime of a transient decay, known as the spin-
spin relaxation time, T2, is highly sensitive to motion at the molecular level. The
values of T2 have been shown to be correlated to the degree of crosslinking in
alkyd coating formulation, falling from a few ms to s as crosslinking proceeds
(Mallgol et al. 2002).
In an alkyd film undergoing autoxidative crosslinking, the progression of
crosslinking from the top film surface has been determined with GARField NMR
profiling (see Section 2.1.4.3). The signal intensity falls as the molecular mobility
decreases as the alkyd hardens (Mallgol et al. 2002, Erich et al. 2005). The
technique has been used to determine the effect of catalysts (known as driers) on
how the oxidative crosslinking extent varies with depth into the film, as a result of
oxygen ingress from the surface (Mallgol et al. 2002, Erich et al. 2006, 2006b).
NMR profiling has likewise been used to study how the extent of crosslinking
varies with depth into coatings undergoing a photo-initiated chemical reaction
(Wallin et al. 2000).
Complementary to NMR probes of molecular mobility, and hence viscosity,
confocal Raman microscopy detects the chemical state as a function of position.
An early report of confocal Raman microscopy to obtain crosslinking profiles in
coatings came from Schrof et al. (1999) who studied UV curing and the effect of
photostabilisers. More recently, Erich et al. (2005b) have used it to probe
crosslinking in alkyd coatings. The disappearance of the double bonds in the
alkyds fatty acid groups was determined as a function of depth. The front position
separating the crosslinked and non-crosslinked regions was well correlated with
the fatty acid chemical reaction (Erich et al. 2005b).

2.4 Techniques to Study Interdiffusion and Coalescence

Although the interdiffusion of polymer molecules between particles during the last
stage of film formation has a profound influence on properties, there are compara-
tively few techniques to probe this process at the molecular level. Commonly, film
properties are evaluated, especially mechanical properties, and the extent of
interdiffusion is deduced. Similarly, electron microscopy is used to assess the
amount and extent of particle coalescence by visualisation of the particle bounda-
ries. Microscopy will not be discussed further in this chapter, but we will turn our
attention to three specialist techniques for diffusion studies.
There is a similarity between the techniques of SANS and FRET. In each,
molecules in a fraction of the particle are labelled (either with deuterium or with a
fluorescent dye), and the smearing of the boundaries with the neighbouring
particles is detected.
2.4 Techniques to Study Interdiffusion and Coalescence 75

2.4.1 Small Angle Neutron Scattering (SANS)

SANS was discussed earlier in Section 2.1.3.1 in the context of film drying. This
technique has offered conclusive evidence for the interdiffusion of molecules in
the later stage of film formation. In a typical experiment, particles that are labelled
with deuterium are blended with normal hydrogenous particles. When the polymer
is at temperatures above its Tg, molecules will diffuse across the interface with
other particles and the effective size of the particle will be seen to grow. This fact
is exploited in SANS experiments. The scattering intensity falls as the radius of
gyration, Rg, of a particle increases, according to the relation:

Q 2 Rg 2
I (Q ) = I (0) exp . (2.19)
3

When the logarithm of I(Q) is plotted against Q2, in a representation called a


Guinier plot, the slope of the line can be used to find Rg. A Guinier plot is shown
schematically in Fig. 2.25. The radius of the particle is then found from

R = 5 Rg . (2.20)
3

As the molecules diffuse out of the confines of the particle at an initial time t0, the
size that each particle occupies will increase. The dependence of Rg on time t is
predicted to increase to a greater extent when the apparent molecular diffusion
coefficient, Dapp, is higher:

Rg 2 (t ) = Rg 2 (t0 )+ 6 Dapp (t t0 ) . (2.21)

Measurements of the evolution of Rg thereby provide a direct means to determine


Dapp. This approach, however, requires a measurement of Rg at t0.
As interdiffusion occurs, the interfaces between particles can be characterised
by an interpenetration distance, dI, or by an interfacial width, wI. The first of these,
dI, represents merely the difference between R at a given time and the radius at the
initial time, R0. The other measurement, wI, is obtained from an analysis using a
modification of the Porod law. Whereas the Porod law describes a two-phase
system with sharp boundaries, a modified expression is required when there is a
diffuse interface. If a Gaussian profile with a standard deviation wI is invoked to
describe the interface between deuterated and non-deuterated particles, then I(Q)
is given as:

KP
I (Q) = exp(Q 2 wI 2 ) + K B I B (Q) . (2.22)
Q4
76 2 Established and Emerging Techniques of Studying Latex Film Formation


Log (Intensity)

Short time




Longer time

Q2 (-2)
Fig. 2.25 A schematic illustration of Guinier plots for deuterium-labelled particles with little
interdiffusion (at a short time) and with more interdiffusion (at a longer time).

Here, Kp is the Porod constant. The second term contains the product of the
background scattering intensity, IB, and a constant of proportionality, KB. Kim et
al. (2000) found that that dI is 1.67 times greater than wI.

2.4.2 Fluorescence Resonance Energy Transfer (FRET)

In a major development, Peckan et al. (1988) first reported the use of non-
radiative energy transfer (NRET) for the study of latex particle structure. The
group of Winnik at the University of Toronto have gone on to apply the technique
to answer many questions about interdiffusion in latex films. The technique is now
usually called fluorescence resonance energy transfer (FRET).
The basic principle of FRET is that the fluorescence lifetime of fluorescent
donor molecules decreases when they are in close proximity to acceptor mole-
cules. The donor and acceptor molecules are grafted to polymer molecules, and
they are initially in separate particles. As is shown in Fig. 2.26, when there is
diffusion of molecules from one particle to another, the donor and acceptor pairs
come into closer contact. The area under a fluorescent decay curve is proportional
to the fraction of mixing between the particles (Wang and Winnik 1993).

2.4.2.1 Rate of Energy Transfer

Here we will follow the presentation of Farinha et al. (2000). The rate of fluores-
cence decay for an isolated donor dye molecule is the reciprocal of its lifetime, 1/D.
If the dipoles of all of the donor and acceptor pairs are separated by a distance, r,
then the decay rate will be faster, and is given by the Frster relation as:
2.4 Techniques to Study Interdiffusion and Coalescence 77

[Cdonor]
o o oo o o
o
oo o o o
o

o
o o
o
oo o o o o
o

o o
x
o o o o oo o

o o o o o o
o o o o o o
o
o o o oo o
x
o o o o
o o o
o

o o o oo o
o o
o
o o o o o o o
o
o
x x

Fig. 2.26 Schematic illustrations of a boundary between a phase labelled with donor molecules
(filled circles) and acceptor molecules (open circles). Initially there is no mixing between the
phases (top). As diffusion proceeds, the donor and acceptor molecules cross the boundary
(middle) until there is no longer a concentration gradient and mixing is complete (bottom). The
donor concentration profiles are sketch on the right side.

6
3 2 1 RF
W (r ) = (2.23)
2 D r

where RF is called the Frster radius. Its value depends on the particular donor
and acceptor pair, but it is typically between 2 and 7 nm. For the commonly used
pair of phenanthrene and anthracene, RF has been found to be 2.3 nm (Wang and
Winnik 1993). The 32/2 term, which is related to the orientation of donor and
acceptor dipoles, is typically close to unity and is often neglected in calculations.
In the Frster relations, W varies with r 6, thus showing great sensitivity to small
values of r, with RF setting the range of sensitivity. This gives the technique its
attractiveness and power.
In the case of isolated donor dye molecules, the fluorescence intensity, I, is
simply an exponential function of time, t.

t
I D (t ) = a1 exp (2.24)
D

where a1 is a constant. Ideally, a plot of the logarithm of I(t) against t (where t


usually extends from 0 up to a few hundred ns) will give a straight line with a
gradient equal to 1/D. When donors and acceptor pairs are all homogeneously
78 2 Established and Emerging Techniques of Studying Latex Film Formation

separated by r in an infinite volume, the decay of I(t) for the donor is faster as a
result of quenching by the acceptor. The equation is modified as

I DA (t ) = exp(t / D ) exp[W (r )t ] (2.25)

In this equation, we can see how the fluorescence decay contains information
about the donor and acceptor pairs and hence can serve as a measure of very short
distances between interdiffusing molecules.
In latex systems, there are finite volume effects, and there is not a homogene-
ous distribution of donors and acceptors but a distribution of r. The donors
fluorescent decay profile, however, can be fit to a phenomenological expression:

t t t
0.5
I DA (t ) = a1 exp + a2 exp exp , (2.26)
D D D

where a2 and are constants (Farinha et al. 2000). The first component in this
equation is sometimes attributed to the donors that have not yet mixed with
acceptor molecules, whereas the second component is attributed to those that have
mixed. Fig. 2.27 shows how the fluorescent decay curves are expected to change
when there is interdiffusion across a particle interface.

2.4.2.2 Quantum Efficiency and Fraction of Mixing

The quantum efficiency of energy transfer, ET, can be related to the fraction of
molecular mixing in a system of labelled particles (Farinha et al. 2000). ET is
defined in terms of the areas under the fluorescence decay curves, A, of a donor in
the presence of an acceptor, IDA(t) in comparison to the area under the curve of the
donor in isolation, ID(t). Of course, the areas can be obtained by integration over
time, such that

I DA (t )dt I DA (t )dt
(2.27)
ET = 1 0

= 1 0

D
I
0
D (t )dt

where we see that the donor lifetime, D, is found by integration of the decay
curve. In reality, in latex systems, there is always energy transfer at the start of the
experiment (t=0) such that

I
0
D (t ) dt <D . (2.28)
2.4 Techniques to Study Interdiffusion and Coalescence 79

In the study of latex, the fraction of mixing of molecules, fm, is of great interest
in assessing interdiffusion. It is related to ET in a rather complex way. We can
see in the defining equation for ET that it will increase over time as molecules
interdiffuse and donor and acceptor molecules come into greater proximity. A
valid and useful method for determining fm is by using a normalised ET that
compares A for the fluorescence decay profiles at a particular time t in comparison
to that for the fully mixed state achieved after a long time (t=) (Farinha et al.
1995). The areas are corrected for the contribution of mixing at the time when the
experiment starts, t = 0, so that fm is given as

A(t ) A(0)
fm = (2.29)
A( ) A(0)

Fig. 2.27 a Simulated distributions of donor-labelled polymer chains for a particle with an initial
radius of 60 nm for three interfacial widths of 1 0, 2 20, and 3 26 nm. b Simulated donor decay
functions corresponding to the donor distributions in a. The simulation assumes 10 wt.% donor
labelled spheres are mixing with 90 wt.% labelled material. (Reprinted from Farinha et al.
(2000); copyright 2000 American Chemical Society)
80 2 Established and Emerging Techniques of Studying Latex Film Formation

Fig. 2.28 The results of simulations, such as shown in Fig. 2.27, can be used to predict the
efficiency of the energy transfer as the interface widens as a result of interdiffusion. When more
donors and acceptors are mixed, the efficiency increases. (Reprinted from Farinha et al. (2000),
copyright (2000) American Chemical Society)

When fm is found at various times of interdiffusion (such as when annealing


films at a constant temperature), a value of the interdiffusion coefficient can be
obtained by analysis of the data with an appropriate diffusion model (Winnik et al.
1992). The quantum efficiency of energy transfer is predicted to increase (Fig.
2.28) as diffusion proceeds and the interface broadens.
More exact treatments require an equation to correlate particular donor and
acceptor profiles with the fluorescent decay. An analytic expression to relate
fluorescence intensity decay, ID(t) of donors in the presence of acceptors to
specific donor and acceptor profiles, has been derived for systems with spherical
symmetry (Yekta et al. 1997), which are relevant to latex diffusion experiments.
Fig. 2.29 shows typical fluorescence decay curves (IDA(t)) for differing extents of
mixing between the donor and acceptor-labelled polymers in a film made from a
blend of latex. Over the years, FRET has proven its usefulness by yielding insight
into the influence of a wide range of parameters on interdiffusion in latex films.
Among the topics studied include various effects, such as coalescing aids (Wang and
Winnik 1990, Winnik et al. 1992, Juhu et al. 1993); plasticising solvents (Juhu
and Lang 1994b); effects of water content on diffusion of hydrophobic and hydro-
philic polymers (Feng and Winnik 1997); chain entanglements (Odrobina and
Winnik 2001); molar mass (and gel content) (Oh et al. 2003), presence of crosslinks
(Tamai et al. 1999)), polymer Tg (Boczar et al. 1993), surface acid group presence
and neutralisation (Kim and Winnik 1995), core-shell structure (Juhu and Lang
1995), or the effect of free, low-molecular-weight polymers (Farinha et al. 1996).
In a recent technical development, a pulsed diode has been used as the excitation
source in FRET experiments, resulting in faster acquisition times in comparison to
what is obtained with a pulsed flash lamp. In the past, FRET was mainly used to
obtain an average value of the fraction of mixing of molecules within a latex film.
2.4 Techniques to Study Interdiffusion and Coalescence 81

The use of focusing optics enables the excitation beam to be reduced to a spot size
less than 1 mm. In an exciting innovation, it has been shown that interdiffusion can
be measured in wet latex films as a function of the lateral position (Haley et al.
2007). The lateral position in a drying film can be correlated with the time elapsed
since the point of particle-particle contact. Recent exciting experiments have
determined the influence of the presence of water on particle contact and on polymer
plasticisation throughout the film formation process (Haley et al. 2008).
FRET has also determined the influence of various types of polymer and inor-
ganic fillers: both deformable (i.e., soft) latex particles (Odrobina et al. 2001) and
non-deformable (i.e., hard) latex particles (Feng et al. 1998) in a film-forming
matrix, silica nanoparticles (Kobayashi et al. 2002), and pigment (CaCO3)
particles (Kobayashi et al. 2001).
In most cases, phenanthrene is used as the donor and anthracene as the acceptor
dye, but other pairs are possible, such as 1-napthylethyl methacrylate and 9-anthryl
methacrylate (Boczar et al. 1993). Alternative acceptors include perylene (Baumgert
et al. 2000) and the non-fluorescent dye [1-(4-nitrophenyl)-2-pyrrolidinmethyl]-
acrylate, which has an absorption spectrum with good spectral overlap with the
emission spectrum of phenanthrene (Turshatov and Adams 2007).

Fig. 2.29 Experimental fluorescence decay curves from a phenathrene (Phe) donor from four
different samples. a A Phe-labelled latex without an acceptor. b A blend of the Phe-labelled latex
and an acceptor-labelled latex after being dried at a temperature below the glass transition
temperature of the polymer. There is no donor-acceptor close contact and therefore the curve
overlays on a. c The same film as in b after annealing at 45C for 30 min. d A film with full
mixing between donors and acceptors obtained by dissolving the latex in an organic solvent.
(Reprinted from Oh et al. (2003b); copyright 2003 American Chemical Society)
82 2 Established and Emerging Techniques of Studying Latex Film Formation

Table 2.1. Specialist Texts on Characterisation Techniques for Latex Films and Film Formation
Specialist Texts
High Vacuum Surface Analysis (SIMS, XPS, Auger)
Sabbatini L, Zambonin PG (1993) Surface Characterisation of Advanced Polymers. VCH,
Weinheim
Vickerman JC, ed. (1999) Surface Analysis The Principal Techniques. John Wiley & Sons,
Chichester.
Electron Microscopy
Goldstein J et al. (2003) Scanning electron microscopy and X-ray microanalysis, 3rd ed.,
Springer, Berlin.
Goodhew PJ, Humphreys J, Beanland R (2001) Electron Microscopy and Analysis, 3rd ed.
Taylor & Francis, London
Reimer L (1995) Energy-filtering transmission electron microscopy, Springer, Berlin.
Reimer L (1998) Scanning electron microscopy: Physics of image formation and microanaly-
sis, 2nd ed., Springer, Berlin.
Williams DB, Carter CB (2009) Transmission Electron Microscopy: A Textbook for Materials
Science, 2nd ed., Springer, Berlin.
X-ray and Neutron Scattering
Feigin LA, Svergun DI (1987) Structure analysis by small-angle X-ray and neutron scattering.
Plenum Press, New York.
Glatter O, Kratky O (eds.) (1982) Small angle X-ray scattering. Academic Press, London.
Higgins JS, Benoit HC (1994) Polymers and neutron scattering. Clarendon Press, Oxford.
Lindner P, Zemb Th. (eds.) (1991) Neutron, X-ray and light scattering: Introduction to an
investigative tool for colloidal and polymeric systems. North-Holland, Amsterdam.
Dynamic Light Scattering
Berne BJ, Pecora R (1976) Dynamic light scattering with applications to chemistry, biology
and physics. Wiley, New York.
Ishimaru A (1978) Wave propagation and scattering in random media. Academic Press, New
York.
van de Hulst HC (1980) Multiple light scattering. Academic Press, Amsterdam.
Other Microscopies and Spectroscopies
Birks JB (1970) Photophysics of aromatic molecules. Wiley, London
Bonnell D (ed.) (2001) Scanning Probe Microscopy and Spectroscopy: Theory, Techniques,
and Applications, 2nd. Wiley-VCH: New York
Callaghan PT (1991) Principles of nuclear magnetic resonance microscopy. Clarendon Press,
Oxford
Sarid D (1994) Scanning Force Microscopy, Oxford University Press, New York.
Sawyer LC, Grubb DT (1987) Polymer Microscopy, Chapman and Hall, London.
Schrader B (1995) Infrared and Raman Spectroscopy. VCH, Weinheim.
Wilson T (1990) Confocal Microscopy. Academic Press, London.
Ultrasound
Mason WP (ed.) Physical Acoustics. Vol. 1. Academic Press, New York.
Probe Techniques
Eichler HJ, Gnther P, Pohl DW (1986) Laser-induced dynamic gratings. Springer, Berlin.
Weil JA, Bolton JE, Wertz JE (1994) Electron paramagnetic resonance. Wiley, New York.
2.5 Concluding Remarks 83

2.4.3 Transmission Spectrophotometry

An indirect probe of coalescence is to swell latex in water (or another solvent) and
to observe the turbidity. When the coalescence is complete, boundaries between
particles should be erased, and water will only be able to be absorbed at the
molecular level as a solvent in the polymer. But, prior to the completion
of coalescence, pockets of water can gather in interparticle void space or at
particle boundaries. These pockets will scatter light. Van Tent and te Nijenhuis
(1992a) used turbidity measurements to show that particle coalescence occurs
slowly over time.

2.5 Concluding Remarks

In this review of techniques employed in the study of latex film formation, there
have been numerous examples of techniques being applied to film formation soon
after their development. The story is similar for electron microscopies, scanning
probe microscopies, FRET, MR profiling and most recently diffusing wave
spectroscopy. This trend seems to indicate that the mysteries of latex film forma-
tion have attracted the interests of the leading experimentalists. It also suggests
that our understanding of latex film formation is well advanced because of the
application of so many state-of-the-art techniques. To learn more about specific
analytical techniques, the relevant books listed in Table 2.1 may be consulted.

References

Anczykowski B., Gotsmann B., Fuchs H., Cleveland J.P., Elings V.B. (1999) How
to measure energy dissipation in dynamic mode atomic force microscopy. Appl
Surf Sci 140: 376-382
Alig I., Lellinger D., Sulimma J., Tadjbakhsch S. (1997) Ultrasonic shear wave
reflection method for measurements of the viscoelastic properties of polymer
films. Rev Sci Instrum 68: 1536-1542
Amalvy J.I., Lasquibar C.A., Arizaga R., Rabal H. and Trivi M. (2001) Applica-
tion of dynamic speckle interferometry to the drying of coatings. Prog Organic
Coat 42: 89-99
Arnold C. et al. (2009) Mapping the route from wet to dry. European Coatings
Journal 11: 28-32
Bahra M., Elliott D., Reading M., Ryan R. (1992) A novel approach to thermal
analysis of thin films. J Therm Anal 38: 543-555
Ballauff M. (2001) SAXS. and SANS studies of polymer colloids. Curr Opin Coll
Interf Sci 6: 132-139
84 2 Established and Emerging Techniques of Studying Latex Film Formation

Bar G., Delineau L., Brandsch R., Bruch M. Whangbo (1999) Importance of the
indentation depth in tapping-mode atomic force microscopy study of compliant
materials. Appl. Phys. Lett. 75: 4198-4200.
Baumgart T., Cramer S., Jahr T., Veniaminov A., Adams J., Fuhrmann J., Jeschke
G., Wiesner U., Spiess H.W., Bartsch E., Sillescu H. (2000) Film-forming
colloidal dispersions studied by tracer methods. Macromol Symp 151: 451-457
Belaroui F., Grohens Y., Boyer H., Holl Y. (2000) Depth profiling of small
molecules in dry latex films by confocal Raman spectroscopy. Polymer 41:
7641-7645
Belaroui F., Cabane B., Dorget M., Grohens Y., Marie P., Holl Y. (2003) Small-
angle neutron scattering study of particle coalescence and SDS desorption
during film formation from carboxylated acrylic latices. J Coll Interf Sci 262:
409-417
Belaroui F., Hirn M.P., Grohens Y., Marie P., Holl Y. (2003) Distribution of
water-soluble and surface-active low-molecular-weight species in acrylic latex
films. J Coll Interf Sci 261: 336-348
Bennett G., Gorce J.-P., Keddie J.L., McDonald P.J., Berglind H. (2003) Magnetic
resonance profiling studies of the drying of film-forming aqueous dispersions
and glue layers. Mag Res Imaging 21: 235-41.
Betzig E., Trautman J.K., Harris T.D., Weiner J.S., Kostelak R.L. (1991) Breaking
the diffraction barrier Optical microscopy on a nanometric scale. Science
251: 1468-1470
Betzig E., Finn P.L., Weiner J.S. (1992) Combined shear force and near-field
scanning optical microscopy. Appl Phys Lett 60: 2484-2486
Binnig G., Quate C.F., Gerber Ch. (1986) Atomic force microscope. Phys Rev
Lett 56: 930-933
Boczar E.M., Dionne B.C., Fu Z., Kirk A.B., Lesko P.M., Koller A.D. (1993)
Spectroscopic studies of polymer interdiffusion during film formation. Macro-
molecules 26: 5772-5781
Bogner A., Thollet G., Basset D., Jouneau P.-H., Gauthier C. (2005) Wet STEM:
A new development in environmental SEM. for imaging nano-objects included
in a liquid phase. Ultramicroscopy 104: 290-301
Bradford E.B. (1952) Electron microscope study of plasticized latices. J Appl
Phys 23: 609-612
Bradford E.B., Vanderhoff J.W. (1963) The morphology of synthetic latexes. J
Polym Sci: Pt C. 3: 41-64
Breugem A.J., Bouchama F., Koper G.J.M. (2005) Diffusing wave spectroscopy:
A novel rheological method for drying paint films. Surf Coat Intern Pt B: Coat
Trans 88, B2: 83-156
Brun A., Brunel L., Snabre P. (2006) Adaptive speckle imaging interferometry
(ASII): New technology of advance drying analysis of coatings. Surface Coat-
ings International Pt B: Coatings Transactions 88: 251-254
Brun A., Dihang H., Brunel L. (2008) Film formation of coatings studied by
diffusing-wave spectroscopy. Prog Organic Coatings 61: 181-191
References 85

Butt H.-J., Kuropka R., Christensen B. (1994) Latex film formation studied with
the atomic force microscope: Influence of aging and annealing. Coll Polym Sci
272: 1218-1223
Butt H.-J., Gerharz B. (1995) Imaging homogeneous and composite latex particles
with an atomic force microscope. Langmuir 11: 4735-4741
Butt H.-J., Kuropka R. (1995) Surface structure of latex films, varnishes, and paint
films studied with an atomic force microscope. J Coatings Tech 67(848):
101-107
Cameron R.E., Donald A.M. (1994) Minimizing sample evaporation in the
environmental scanning electron microscope. J Microscopy Oxford 173: 227-
237
Ciampi E., Goerke U., Keddie J.L., McDonald P.J. (2000) Lateral transport of
water during drying of alkyd emulsions. Langmuir 16: 1057-1065
Corcoran E.M. (1969) Determining stresses in organic coatings using plate beam
deflection. J Paint Techn 41 (538): 635-40
Cramer S.E., Jeschke G., Spiess H.W. (2002a) EPR studies on film formation of
colloidal dispersions, 1. Site selectivity and techniques. Macromol Chem Phys
203: 182-191
Cramer S.E., Jeschke G., Spiess H.W. (2002b) EPR studies on film formation of
colloidal dispersions, 2. Drying, film annealing, and film rewetting. Macromol
Chem Phys 203: 192-198
Crowley T.L., Sanderson A.R., Morrison J.D., Barry M.D., Morton-Jones A.J.,
Rennie A.R. (1992) Formation of bilayers and Plateau borders during the dry-
ing of film-forming latexes as investigated by small-angle neutron scattering.
Langmuir 8: 2110-2123
Danilatos G.D. (1988) Foundations of environmental scanning electron micros-
copy. Adv Electronics & Electron Phys 71: 109-250
Derks D., Wisman H., van Blaaderen A., Imhof A. (2004) Confocal microscopy of
colloidal dispersions in shear flow using a counter-rotating cone-plate shear
cell. J Phys Condens Matt 16: S3917-S3927
Dillon R.E., Matheson L.A., Bradford E.B. (1951) Sintering of synthetic latex
particles. J Coll Sci 6: 108-117
Dingenouts N., Ballauff M. (1999) First stage of film formation by latexes
investigated by small-angle X-ray scattering. Langmuir 15: 3283-3288
Dingenouts N., Ballauff M. (1998) Assessment of spatial order in dried latexes by
small-angle X-ray scattering. Macromolecules 31: 7423-7429
Distler D., Kanig G. (1980) Morphology of (meth-)acrylic latex films. Org
Coating Plast Chem 43: 606-610
Djabourov M., Grillon Y., Leblond J. (1995) The sol-gel transition in gelatin
viewed by diffusing colloidal probes. Polym Gels Networks 3: 407-428
Dobler F., Pith T., Lambla M., Holl Y. (1992) Coalescence mechanisms of
polymer colloids. 2. Coalescence with evaporation of water. J Coll Interf Sci
152: 12-21.
Donald A.M., He C.B., Royall C.P., Sferrazza M., Stelmashenko N.A., Thiel B.L.
(2000) Applications of environmental scanning electron microscopy to colloi-
86 2 Established and Emerging Techniques of Studying Latex Film Formation

dal aggregation and film formation. Coll Surf A Physicochem Eng Asp 174:
37-53
Donald A.M. (2003) The use of environmental scanning electron microscopy for
imaging wet and insulating materials. Nature Materials 2: 511-516
Du Chesne A., Gerharz B., Lieser G. (1997) The segregation of surfactant upon
film formation of latex dispersions: an investigation by energy filtering trans-
mission electron microscopy. Polym Intern 43: 187-196
Dullens R.P.A., Claesson M., Derks D., van Blaaderen A., Kegel W.K. (2003)
Preparation and properties of cross-linked fluorescent poly(methyl methacry-
late) latex colloids. Langmuir 19: 5963-5966
Dullens R.P.A., Claesson E.M., Kegel W.K. (2004) Preparation and properties of
cross-linked fluorescent poly(methyl methacrylate) latex colloids. Langmuir
20: 658-664
Drig U., Pohl D.W., Rohner F. (1986) Near-field optical-scanning microscopy, J
Appl Phys 59: 3318-3327.
Eckersley, Rudin A. (1993) The effect of plasticization and pH. on film formation
of acrylic latexes. J Appl Polym Sci 48: 1369-1381
Eichler H.J., Gnther P., Pohl D.W. (1986) Laser-induced dynamic gratings.
Springer, Berlin.
Eidmann G., Savelberg R., Blmler P. (1996) The NMR MOUSE, a mobile
universal surface explorer. J Magn Res Ser A 122: 104-109
Ekanayake P., McDonald P.J., Keddie J.L. (2009) An experimental test of the
scaling prediction for the spatial distribution of water during the drying of
colloidal films. European Physical Journal Special Topics 166: 21-27.
El-Aasser M.S., Robertson A.A. (19975) Film formation of latexes. J Paint Techn
47(611): 50-53
Erich S.J.F., Laven J., Pel L., Huinink H.P., Kopinga K. (2005) Dynamics of cross
linking fronts in alkyd coatings. Appl Phys Lett 86: 134105 (3 pp)
Erich S.J.F., Laven J., Pel L., Huinink H.P., Kopinga K. (2005b) Comparison of
NMR and confocal Raman microscopy as coatings research tools. Prog Org
Coat 52: 210-216
Erich S.J.F., Laven J., Pel L., Huinink H.P., Kopinga K. (2006) NMR depth
profiling of curing alkyd coatings with different catalysts. Prog Org Coat 55:
105-111
Erich S.J.F., Laven J., Pel L., Huinink H.P., Kopinga K. (2006b) Influence of
catalyst type on the curing process and network structure of alkyd coatings.
Polymer 47: 1141-1149
Erkselius S., Wadsoe L., Karlsson O.J. (2007) A sorption balance-based method to
study the initial drying of dispersion droplets. Coll Polym Sci 285: 1707-1712
Faccia, P.A., Pardini O.R., Amalvy J.I., Cap N., Grumel E.E., Arizaga R., Trivi
M. (2009) Prog Organic Coatings 64: 350-355
Farinha J.P.S., Martinho J.M.G., Yekta A., Winnik M.A. (1995) Direct nonradia-
tive energy transfer in polymer interphases: fluorescence decay functions from
concentration profiles generated by Fickian diffusion. Macromolecules 28:
6084-6088
References 87

Farinha J.P.S., Martinho J.M.G., Kawaguchi S., Yekta A., Winnik M.A. (1996)
Latex film formation probed by nonradiative energy transfer: Effect of grafted
and free poly(ethylene oxide) on a poly(n-butyl methacrylate) latex. J Phys
Chem 100: 12552-12558
Farinha J.P.S., Vorobyova O., Winnik M.A. (2000) An energy transfer study of
the interface thickness in blends of poly(butyl methacrylate) and poly(2-
ethylhexyl methacrylate). Macromolecules 33: 5863-5873
Feng J.R., Winnik M.A. (1997) Effect of water on polymer diffusion in latex
films. Macromolecules 30: 4324-4331
Feng J.R., Odrobina E., Winnik M.A. (1998) Effect of hard polymer filler
particles on polymer diffusion in a low-Tg latex film. Macromolecules 31:
5290-5299
Fletcher A.L., Thiel B.L., Donald A.M. (1997) Amplification measurements of
alternative imaging gases in environmental SEM J Phys D. 30: 2249-2257
Fouchama F., Estramil G., Autin A.J.E., Koper G.J.M. (2002) Film formation
from concentrated emulsions studied by simultaneous conductometry and gra-
vimetry. Coll Surf A: Physicochem Eng Asp 210: 129-135
Gaillard C., Stadelmann P.A., Plummer C.J.G., Fuchs G. (2004) Practical method
for high-resolution imaging of polymers by low-voltage scanning electron
microscopy. Scanning 26: 122-130.
Galembeck A., Costa C.A.R., da Silva M.C.V.M., Souza E.F., Galembeck F.
(2001) Scanning electric potential microscopy imaging of polymers: electrical
charge distribution in dielectrics. Polymer 42: 4845-4851
Glover P.M., Aptaker P.S., Bowler J.R., Ciampi E., McDonald P.J. (1999) A
novel high-gradient permanent magnet for the profiling of planar films and
coatings. J Magn Reson 139: 90-97
Goh M.C., Juhu D., Leung O.-M., Wang Y., Winnik M.A. (1993) Annealing
effects on the surface structure of latex films studied by atomic force micros-
copy. Langmuir 9: 1319-1322
Gorce J.-P., McDonald P.J., Keddie J.L. (2002) Vertical water distribution during
the drying of polymer films formed from emulsions. Eur Phys J E 8: 421-429.
Goudy A., Gee M.L., Biggs S., Underwood S. (1995) Atomic force microscopy
study of polystyrene latex film morphology: Effects of aging and annealing.
Langmuir 11: 4454-4459
Guigner D., Fischer C., Holl (2001) Film formation from concentrated reactive
silicone emulsions. 1. Drying mechanism. Langmuir 17: 3598-3606
Haley J.C., Liu Y., Winnik M.A., Demmer D., Haslett T., Lau W. (2007) Tracking
polymer diffusion in a wet latex film with fluorescence resonance energy trans-
fer. Rev Sci Instrum 78: 084101
Haley J.C., Liu Y., Winnik M.A., Lau W. (2008) The onset of polymer diffusion
in a drying acrylate latex: how water initially retards coalescence but ultimately
enhances diffusion. J Coat Technol Res 5(2): 157-168
Hellgren A.-C., Wallin M., Weissenborn P., McDonald P.J., Glover P.M., Keddie
J.L. (2001) New techniques for determining the extent of crosslinking in coat-
ings. Prog Org Coatings 43: 85-98
88 2 Established and Emerging Techniques of Studying Latex Film Formation

Hore P.J. (1995) Nuclear Magnetic Resonance, Oxford University Press


James P.J., Antognozzi M., Tamayo J., McMaster T.J., Newton J.M., Miles M.J.
(2001) Langmuir 17: 349-360
Joanicot M., Wong K., Richard J., Maquet J., Cabane B. (1993) Ripening of
cellular latex films. Macromolecules 26: 3168-3175
Johannsmann D. (2008) Viscoelastic, mechanical and dielectric measurements on
complex samples with the quartz crystal microbalance. Phys Chem Chem Phys
10: 4516-4534
Juhu D., Wang Y., Winnik M.A. (1993) Influence of a coalescing aid on polymer
diffusion in poly(butyl methacrylate) latex films. Makromol Chem Rapid
Commun 14: 345-349
Juhu D., Lang J. (1994a) Latex film surface morphology studies by atomic force
microscopy: effect of a non-ionic surfactant post-added to latex dispersion. Coll
Surf A: Physicochem Eng Asp 87: 177-185
Juhu D., Lang J. (1994b) Latex film formation in the presence of organic
solvents. Macromolecules 27: 695-701
Juhu D., Lang J. (1995) Film formation from dispersions of core-shell latex
particles. Macromolecules 28: 1306-1308
Juhu D., Wang Y., Lang J., Leung O.-M., Goh M.C., Winnik M.A. (1995)
Surfactant exudation in the presence of a coalescing aid in latex films studied
by atomic force microscopy. J Polym Sci: Pt B: Pol Phys 33: 1123-1133
Keddie J.L. (1997) Film formation of latex. Mat Sci Eng R: Reports, R21: 101-
170.
Keddie J.L., Meredith P., Jones R.A.L., Donald A.M. (1995) Kinetics of film
formation in latex studied with ellipsometry and environmental-SEM Macro-
molecules 28: 2673-2682
Keddie J.L., Meredith P., Jones R.A.L., Donald A.M. (1996) Film formation of
acrylic latices with varying concentrations of non-film-forming latex particles.
Langmuir 12: 3793-3801.
Keslarek A.J., Costa C.A.R., Galembeck F. (2001) Electric charge clustering and
migration in latex films: A study by scanning electric potential microscopy.
Langmuir 17: 7886-7892
Keslarek A.J., Costa C.A.R., Galembeck F. (2002) Latex film morphology and
electrical potential pattern dependence on serum components: A scanning
probe microscopy study. J Coll Interf Sci 255:107-114
Keslarek A.J., Galembeck F. (2003) Effect of serum components on styrene/butyl
acrylate latex film formation: An in situ examination by ultramicroscopy. J
Appl Polym Sci 87: 159-167
Kim H.-B., Winnik M.A. (1995) Factors affecting interdiffusion rates in films
prepared from latex particles with a surface rich in acid groups and their salts.
Macromolecules 28: 2033-2041
Kim S.D., Klein A., Sperling L.H., Boczar E.M., Bauer B.J. (2000) SANS study
of sulfonate end group effect on polystyrene self-diffusion. Macromolecules
33: 8334-8343
References 89

Kitching S., Donald A.M. (1998) Beam damage of polypropylene in the environ-
mental scanning electron microscope: an FTIR study. J Microsc Oxford 190:
357-365
Knoll A., Magerle R., Krausch G. (2001) Tapping mode atomic force microscopy
on polymers: Where is the true sample surface? Macromolecules 34: 4159-
4165.
Kobayashi M., Rharbi Y., Winnik M.A. (2001) Effect of inorganic pigments on
polymer interdiffusion in a low Tg latex film. Macromolecules 34: 1855-1863
Kobayashi M., Rharbi Y., Brauge L., Cao L., Winnik M.A. (2002) Effect of silica
as fillers on polymer interdiffusion in poly(butyl methacrylate) latex films.
Macromolecules 35:7387-7399
Koenig J.L., Bobiak J.P. (2007) Raman and infrared imaging of dynamic polymer
systems. Macromol. Mater. Engin. 292: 801-806
Knig A.M., Weerakkody T.G., Keddie J.L., Johannsmann D. (2008) Heterogene-
ous drying of colloidal polymer films: Dependence on added salt. Langmuir 24:
75807589
Kopp-Marsaudon S., Leclre Ph, Dubourg F., Lazzaroni R., Aim J.P(2000)
Quantitative measurement of the mechanical contribution to tapping-mode
atomic force microscopy images of soft materials. Langmuir 16: 8432-8437.
Lan K.H., Ostrowsky N., Sornette D. (1986) Brownian dynamics close to a wall
studied by photon correlation spectroscopy from an evanescent wave. Phys Rev
Lett 57: 17-20
Leclre Ph, Dubourg F., Kopp-Marsaudon S., Brdas J.L., Lazzaroni R., Aim J.P.
(2002) Dynamic force microscopy analysis of block copolymers: beyond imag-
ing the morphology. Appl. Surf. Sci. 188: 524-533.
Lee W.P., Routh A.F. (2006) Time evolution of transition points in drying latex
films. J Coatings Techn. Research 3: 301-306
Lei C., Ouzineb K., Dupont O., Keddie J.L. (2007) Probing Particle Structure in
Waterborne Pressure-Sensitive Adhesives with Atomic Force Microscopy.
Journal of Colloid and Interface Science 307: 56-63.
Lellinger D., Tadjbach S., Alig I. (2002) Determination of the elastic moduli of
polymer films by a new ultrasonic reflection method. Macromol Symp 184:
203-213
Lemlich R. (1978) Theory for limiting conductivity of polyhedral foam at low
density. J Coll Interf Sci 64: 107-110
Lin F., Meier D.J. (1995) A study of latex film formation by atomic force micros-
copy. 1. A comparison of wet and dry conditions. Langmuir 11: 2726-2733
Ludwig I., Schabel W., Kind M., Castaing J.-C., Ferlin P. (2007) Drying and film
formation of industrial waterborne latices. AIChE Journal 53: 549-560
Ma Y., Davis H.T., Scriven L.E. (2005) Microstructure development in drying
latex coatings. Prog Org Coat 52: 46-62
Madsen A., Konovalov O., Robert A., Grubel G. (2001) Surface ordering in a
concentrated suspension of colloidal particles investigated by x-ray scattering
methods. Phys Rev E. 64: 061406
90 2 Established and Emerging Techniques of Studying Latex Film Formation

Mallgol J., Dupont O., Keddie J.L. (2001) Obtaining and interpreting images of
acrylic pressure-sensitive adhesives by tapping-mode atomic force microscopy.
Langmuir 17: 7022-7031.
Mallgol J., Barry A.M., Ciampi E., Glover P.M., McDonald P.J., Keddie J.L.,
Wallin M., Motiejauskaite A., Weissenborn P.K. (2002) Influence of drier
combination on through-drying in waterborne alkyd emulsion coatings ob-
served with magnetic resonance profiling. J Coatings Techn 74(933): 113-124
Mallgol J., Gorce J.-P., Dupont O., Jeynes C., McDonald P.J., Keddie J.L. (2002)
Origins and effects of a surfactant excess at the surface of waterborne acrylic
pressure-sensitive adhesives. Langmuir 18: 4478-4487.
Mallgol J., Dupont O., Keddie J.L. (2003) Morphology and elasticity of water-
borne acrylic pressure-sensitive adhesives investigated with atomic force mi-
croscopy. J Adhes Sci Techn 17: 243-259.
Mallgol J., Bennett G., Dupont O., McDonald P.J., Keddie J.L. (2006) Skin
development during the film formation of waterborne acrylic pressure-sensitive
adhesives containing tackifying resin. J Adhesion, 82: 217-238
Markiewicz P., Goh M.C. (1995) Atomic-force microscope tip deconvolution
using calibration arrays. Rev Sci Intrum 66: 3186-3190
McDonald P.J. and Newling B. (1998) Stray field magnetic resonance imaging.
Rep Prog Phys 61: 1441-1493
Ming Y., Takamura K., Davis H.T., Scriven L.E. (1995) Microstructure evolution
in latex coatings. Tappi J 78(11): 151-159
Mitchell J., Blmler P., McDonald P.J. (2006) Spatially resolved nuclear magnetic
resonance studies of planar samples. Progr Nucl Magn Reson Spectrosc 48:
161-181
Mulvihill J., Toussaint A., De Wilde M. (1997) Onset, follow up and assessment
of coalescence. Prog Org Coat 30: 127-139
Narita T., Beauvais C., Hbraud P., Lequeux F. (2004) Dynamics of concentrated
colloidal suspensions during drying aging, rejuvenation and overaging. Eur
Phys J E 14: 287-292
Narita T., Hbraud P., Lequeux F. (2005) Effects of the rate of evaporation and
film thickness on nonuniform drying of film-forming concentrated colloidal
suspensions. Eur Phys J E 17: 69-75
Newling B., Glover P.M., Keddie J.L., Lane D.M., McDonald P.J. (1997) Concen-
tration profiles in creaming oil-in-water emulsion layers determined with stray
field magnetic resonance imaging. Langmuir 13: 3621-3626
Nonnenmacher M., OBoyle M.P., Wickramasinghe H.K. (1991) Kelvin probe
force microscopy. App Phys Lett 58: 2921-2923
Odin C., Aim J.P., El Kaakour Z., Bouhacina T. (1994) Tip finite-size effects on
atomic-force microscopy in the contact mode Simple geometrical considera-
tions for rapid estimation of apex radius and tip angle base on the study of
polystyrene latex balls. Surf Sci 317: 321-340
Oh J.K., Yang J., Tomba J.P., Rademacher J., Farwaha R., Winnik M.A. (2003)
Molar mass effect on the rate of polymer diffusion in poly(vinyl acetate-co-
butyl acrylate) latex films. Macromolecules 36: 8836-8845
References 91

Oh J.K., Tomba P., Ye X., Eley R., Rademacher J., Farwaha R., Winnik M.A.
(2003b) Film formation and polymer diffusion in poly(vinyl acetate-co-butyl
acrylate). Temperature dependence. Macromolecules 36: 5804-5814.
Odrobina E., Winnik M.A. (2001) Influence of entanglements on the time
dependence of mixing in nonradiative energy transfer studies of polymer diffu-
sion in latex films. Macromolecules 34: 6029-6038
Odrobina E., Feng J.R., Pham H.H., Winnik M.A. (2001) Effect of soft filler
particles on polymer diffusion in poly(butyl methacrylate) latex films. Macro-
molecules 34: 6039-6051
Papagiannopoulos A., Fernyhough C.M., Waigh T.A. (2005) The microrheology
of polystyrene sulfonate combs in aqueous solution. J Chem Phys 123: 214904
Patel A.A., Feng J., Winnik M.A., Vansco G.J., Dittman McBain C.B. (1996)
Characterization of latex blend films by atomic force microscopy. Polymer 37:
5577-5582
Peckan O., Winnik M.A., Croucher M.D. (1988) Direct energy-transfer studies on
doped and labelled polymer latex particles. Phys Rev Lett 61: 641-644.
Perera D.Y., Van den Eynde D. (1981) Considerations on a cantilever (beam)
method for measuring the internal stress in organic coatings. J Coatings Tech
53 (677): 39-44
Petersen C., Heldmann C., Johannsmann D. (1999) Internal stresses during film
formation of polymer latices. Langmuir 15: 7745-7751
Pohl D.W., Denk W., Lanz M. (1984) Optical stethoscopy: Image recording with
resolution /20, Appl Phys Lett 44: 651-653
Protzman T.F., Brown G.L. (1960) An apparatus for the determination of the
minimum film temperature of polymer emulsions. J Appl Polym Sci 4: 81-85
Raghavan D., VanLandingham M., Gu X., Nguyen T. (2000) Characterization of
heterogeneous regions in polymer systems using tapping mode and force mode
atomic force microscopy. Langmuir 16: 9448-9459.
Rottstegge J., Landfester K., Wilhelm M., Spiess H.W., Heldmann C. (2000)
Different types of water in the film formation process of latex dispersions as
detected by solid-state nuclear magnetic resonance spectroscopy. Colloid Po-
lym. Sci. 278: 236-244.
Rottstegge J., Traub B., Wilhelm M., Landfester K., Heldmann C., Spiess H.W.
(2003) Investigations on the film-formation process of latex dispersions by
solid-state NMR spectroscopy. Macromol. Chem. Phys. 204: 787802
Roulstone B.J., Wilkinson M.C., Hearn J., Wilson A.J. (1991) Studies of polymer
latex films: 1. A study of latex film morphology. Polym Intern 24:87-94
Salamanca J.M., Faux D.A., Glover P.M., Keddie J.L., McDonald P.J., Routh A.F.
(2001) Lateral drying in thick films of waterborne colloidal particles. Langmuir
17:3202-3207
Santos J.P., Corpart P., Wong K., Galembeck F. (2004) Heterogeneity in styrene-
butadiene latex films. Langmuir 20: 10576-10582
Schmidt M., Krieger S., Johannsmann D. (1997) Film formation of latex disper-
sions observed with evanescent dynamic light scattering. Progr Coll Polym Sci
104: 191-194
92 2 Established and Emerging Techniques of Studying Latex Film Formation

Schrof W., Beck E., Kniger R., Reich W., Schwalm R. (1999) Depth profiling of
U.V. cured coatings containing photostabilisers by confocal Raman micros-
copy. Prog Org Coat 35: 197-204
Schallamach A. (1952) Abrasion of rubber by a needle. J Polym Sci 9:385-404
Schwarz U.S., Balaban N.O., Riveline D., Bershadsky A., Geiger B., Safran S.A.
(2002) Calculation of forces at focal adhesions from elastic substrate data: The
effect of localised force and the need for regularisation. Biophys J 83: 1380-
1394.
Scott W.W., Bhushan B. (2003) Use of phase imaging in atomic force microscopy
for measurement of viscoelastic contrast in polymer nanocomposites and
molecularly thick lubricant films. Ultramicroscopy 97:151-169
Sheehan J.G., Takamura K., Davis H.T., Scriven S.E. (1993) Microstucture
development in particulate coatings examined with high-resolution cryogenic
scanning electron microscopy. Tappi J 76(12):93-101
Sommer F., Duc T.M., Pirri R., Meunier G., Quet C. (1995) Surface morphology
of poly(butyl acrylate)/poly(methyl methacrylate) core shell latex by atomic
force microscopy. Langmuir 11:440-448
Tamai T., Pinenq P., Winnik M.A. (1999) Effect of crosslinking on polymer
diffusion in poly(butyl methacrylate-co-butyl acrylate) latex films. Macromole-
cules 32:6102-6110
Tamayo J., Garcia R. (1997) Effects of elastic and inelastic interactions on phase
contrast images in tapping-mode scanning force microscopy. Appl. Phys. Lett.
71: 2394-2396.
Teixeira-Neto E., Kaupp G., Galembeck F. (2003) Latex particle heterogeneity
and clustering in films. J Phys Chem B. 107: 14255-14260
Teixeira-Neto E., Kaupp G., Galembeck F. (2004) Spatial distribution of serum
solutes on dry latex sub-monolayers determined by SEPM, SNOM and SC
microscopy. Coll Surf A Physicochem Eng Asp 243: 79-87
Terris B.D., Stern J.E., Rugar D., Mamin H.J. (1990) Localised charge force
microscopy. J Vac Sci Technol A 8: 374-377
Tirumkudulu M.S., Russel W.B. (2004) Role of capillary stresses in film forma-
tion. Langmuir 20: 2947-2961
Tirumkudulu M.S., Russel W.B. (2005) Cracking in drying latex films. Langmuir,
21: 4938-4948
Turshatov A., Adams J. (2007) A new monomeric FRET-acceptor for polymer
interdiffusion experiments on polymer dispersions. Polymer 48: 7444-7448
Turshatov A., Adams J., Johannsmann D. (2008) Interparticle contact in drying
polymer dispersions probed by time resolved fluorescence. Macromolecules
41: 5365-5372
Tzitzinou A., Jenneson P.M., Clough A.S., Keddie J.L., Lu J.R., Zhdan P.,
Treacher K.E., Satguru R. (1999) Surfactant concentration and morphology at
the surfaces of acrylic latex films. Prog Org Coatings, 35:89-99
Tzitzinou A., Keddie J.L., Geurts J.M., Peters A.C.I.A., Satguru R. (2000) Film
formation of latex blends with bimodal particle size distributions: Considera-
References 93

tion of particle deformability and continuity of the dispersed phase. Macro-


molecules 33: 2695-2708
Urban M.W. (1997) Surfactants in latex films; interactions with latex components
and quantitative analysis using ATR and step-scan PAS FT-IR spectroscopy.
Prog Org Coatings 32: 215-229
van Tent A., te Nijenhuis K. (1992) Turbidity study of the process of film forma-
tion of thin films of aqueous acrylic dispersions. Prog Org Coat 20: 459-470
van Tent A., te Nijenhuis K. (1992b) Turbidity study of the process of film
formation of polymer particles in drying thin films of acrylic latices. J Coll
Interf Sci 150: 97-114
van Tent A., te Nijenhuis K. (2000) The film formation of polymer particles in
drying thin films of aqueous acrylic latices - II. Coalescence, studied with
transmission spectrophotometry. J Coll Interf Sci 232: 350-363
Veniaminov A., Jahr T., Sillescu H., Bartsch E. (2002) Length scale dependent
probe diffusion in drying acrylate latex films. Macromolecules 35: 808-819
Veniaminov A., Jahr T., Sillescu H., Bartsch E. (2003) Probing poly(n-butyl-
methacrylate) latex film via diffusion of hydrophilic and hydrophobic dye
molecules. Macromolecules 36: 4944-4953
Viasnoff V., Lequeux F., Pine D.J. (2002) Multispeckle diffusing-wave spectros-
copy: a tool to study slow relaxation and time-dependent dynamics. Rev Sci
Instrum 73: 2336-2344
von der Ehe K., Johannsmann D. (2007) Maps of the stress distributions in drying
latex films. Rev Sci Instrum 78: 113904 (5 pages)
Waigh T.A. (2005) Microrheology of complex fluids. Rep Prog Phys 68: 685-742
Wallin M., Glover P.M., Hellgren A.-C., Keddie J.L., McDonald P.J. (2000)
Depth profiles of polymer mobility during the film formation of a latex disper-
sion undergoing photo-initiated crosslinking. Macromolecules 33: 8443-8452.
Wang Y., Juhu D., Winnik M.A., Leung O.M., Goh M.C. (1992) Atomic force
microscopy study of latex film formation. Langmuir 8: 760-762
Wang Y., Winnik M.A. (1993) Polymer diffusion across interfaces in latex films.
J Phys Chem 97: 2507-2515
Wang Y., Song R., Li Y.S., Shen J.S. (2003) Understanding tapping-mode atomic
force microscopy data on the surface of soft block copolymers. Surf. Sci. 530:
136-148
Wang Y.C., Winnik M.A. (1990) Latex film formation. 3. Effect of coalescing aid
on polymer diffusion in latex films. Macromolecules 23: 4731-4732
Weil J.A., Bolton J.E., Wertz J.E. (1994) Electron paramagnetic resonance. Wiley,
New York.
Winnik M.A., Feng J. (1996) Latex blends: An approach to zero VOC coatings,
Journal of Coatings Technology 68(852): 39-50
Winnik M.A., Wang Y., Haley F. (1992) Latex film formation at the molecular
level: The effect of coalescing aids on polymer diffusion. J Coatings Tech
64(811): 51-61
Yekta A., Winnik M.A., Farinha J.P.S., Martinho J.M.G. (1997) Dipole-dipole
electronic energy transfer. Fluorescence decay functions for arbitrary distribu-
94 2 Established and Emerging Techniques of Studying Latex Film Formation

tions of donors and acceptors. II. Systems with spherical symmetry. J Phys
Chem A 101: 1787-1792
Yu M.-F., Kowalewski T., Ruoff R.S. (2000) Investigation of the radial deform-
ability of individual carbon nanotubes under controlled indentation force. Phys
Rev Lett 85: 1456-1459
Chapter 3
3 Drying of Latex Films

Drying is an important operation in a manufacturing process. An example is film


formation which requires the removal of water to allow the particles to pack and to
form a coherent film. This water removal is achieved by evaporation. In addition
to film formation, there are a number of other examples where it is desirable to
reduce the amount of water (or indeed any solvent) from a product:
To reduce transport costs for products, such as washing powder, which
although used and manufactured in the wet state are shipped in a dry form.
To maintain suitable handling properties. For example, sand will flow easily
when dry yet will be difficult to handle if wet.
To stop corrosion and the growth of microorganisms.
The removal of water from a product requires an input of energy to overcome
the latent heat of the solvent. Here, we concentrate on the evaporation of water
into air. We need to consider the amount of water that a sample of air can support,
the energies involved, and the rates of mass transfer. There are a number of terms
relating to water dissolved in air. They are defined in Appendix C and graphically
displayed on a psychrometric chart, shown in Fig. C.1.

3.1 Humidity and Evaporation

3.1.1 Background

Atmospheric air has water dissolved in it, with the amount dissolved being called
the humidity. High humidity may feel uncomfortable to those in it, and anyone
trying to exercise in a hot, humid climate knows how difficult it is to stay cool
under these conditions. The body regulates its temperature by sweating and using
the latent heat of the evaporating water to cool the body. In humid conditions the
96 3 Drying of Latex Films

evaporation rate slows, and hence the cooling effect of perspiration is reduced.
The latent heat of vaporisation for water is 2260 kJ/kg. This is a large number as is
illustrated in the following example. If 1 kg of latex, cast on a substrate, contains
about 400 g of water, 904 kJ of energy is required to evaporate the water. The
remaining 600 g of latex would need to cool by approximately 350C to provide
this amount of energy. Of course, this is not observed in practice because heat
transfer from the surroundings provides the necessary energy. But, it is illustrative
to note the significance of the large magnitude of the latent heat of vaporisation
for water. Evidence for evaporative cooling leading to a reduction in evaporation
rate is given by Erskelius et al. (2007).

3.2 Evaporation Rate from Pure Water

The evaporation rate, E, of a liquid may be expressed in units of mass per unit
area of surface per unit of time. In models of evaporation, it is often useful to
describe the evaporation rate as a velocity, E , which represents the distance that
the surface drops downward per unit of time. E can be easily found through an
experimental value of E divided by the density of the liquid, , to give E = E/. In
the case of water, the calculation is trivial when the value used for the density of
water is 1 g cm3.
As discussed in Appendix C, the evaporation rate of pure water is determined
by the difference in the chemical potential of the water vapour just above the
liquid water and in the bulk. Expression C.3 gives an overall mass transfer rate,
although the value of the mass transfer coefficient, km, can vary widely depending
on the specific situation. The simplest example of water evaporating into stagnant
air has a mass transfer coefficient on the order of 0.004 m/s (as found below from
an estimate). A convenient visualisation of the evaporation process is to consider a
boundary layer of thickness Lb above the pool of liquid. As sketched in Fig. 3.1,
the water partial pressure changes linearly over the boundary layer thickness from
the saturated vapour pressure at the liquid surface, to the bulk value at the edge.
The flux of water vapour is then written as the diffusion coefficient of the water
vapour through the air, Dvap, multiplied by the water vapour concentration
gradient, which is the partial pressure difference divided by the thickness of the
boundary layer. It is therefore simple to recognise that the mass transfer coeffi-
cient is the diffusion coefficient divided by the boundary layer thickness:
km=Dvap/Lb .
For stagnant air, the diffusion coefficient of water vapour can be estimated as
2 x 105 m2s1 (Cussler 1984). Inserting an evaporation rate of 0.3 cm per day into
expression C.3, and using a differential pressure of 1/76 of atmospheric pressure,
results in a mass transfer coefficient of 0.004 m/s. The boundary layer thickness
for stagnant air is therefore estimated to be 5 mm. It will be thinner when there is
flowing air.
Evaporation Rate from Pure Water 97

c
Pw=Pw Flux = D
y
Boundary D Mw
Lb layer = P
Lb RT
Mw
Pw=Pvap* = k m P
RT
Liquid water

Fig. 3.1 The boundary layer above a pool of liquid water sets the evaporation rate. The water
vapour immediately above the liquid is at the saturated vapour pressure, Pvap*. The pressure
varies linearly across the boundary layer. The flux of water from the liquid surface is then set by
Ficks law of diffusion. See Appendix C for a discussion.

Any other hindrance to mass transfer will lead to a reduction in the overall mass
transfer coefficient and hence the evaporation rate. The increased mass transfer
resistance is incorporated into the overall mass transfer coefficient through

1 1 1 1
= + + + .... (3.1)
km k1 k2 k3

where each step through which the water must pass has its own discrete mass
transfer coefficient, ki .
Possible methods for reducing the evaporation rate include spreading densely
packed monolayers of long-chain oils across the water surface (Barnes and Gentle
2006). As is intuitively expected, the structure of the surface-active agents is crucial
in determining the effectiveness of water transport reduction. Contaminants
that disrupt the structure have a large effect on the evaporation rate (McNamee et
al. 1998).
Air flow can significantly increase the evaporation rate. The concept behind
this effect is a reduction in the thickness of the boundary layer and consequently
an increase in the mass transfer coefficient. As a result, it is important to control
the air flow over a drying film, a fact well known to many practitioners. It has
been reported that merely opening a lab door can affect the air flow and the
evaporation rate (Croll 1986, Croll 1987).
Other methods to increase the evaporation rate are to increase the water tem-
perature or to reduce the humidity. Latex film drying is fast in hot, dry climates.
Because the temperature and humidity are typically out of the formulators control,
paint formulations can contain small water-soluble molecules, such as ethylene
glycol, to reduce the water vapour pressure and evaporation rate. It should be
noted that a high temperature alone does not cause fast drying, because a high
98 3 Drying of Latex Films

humidity can suppress evaporation rates. Conversely, a cold climate can contribute
to exceedingly slow drying. There is no obvious way to change a paint formula-
tion to raise the water vapour pressure to solve this problem.

3.3 Evaporation Rate from Latex Dispersions

Water loss from latex dispersions is easily followed gravimetrically. Vanderhoff et


al. (1973) observed a three-stage drying profile, with a constant rate of mass loss
followed by a decreasing rate period, asymptoting to a very slow rate of water
loss, which was consistent with diffusion of water through the polymer phase.
Vanderhoff et al. rationalise the results with a homogeneous wet film containing
particles that merely increase in volume fraction but remain uniformly distributed
across the film thickness. Once the particles consolidate at close packing, the
water recedes into the film and the evaporation rate decreases. An alternative
explanation is that a skin formed on the top of the film hindering evaporation. This
is shown as case (c) in Fig. 3.2.
In contrast, Croll (1986, 1987) observed a two-stage drying process, where a
constant rate of drying is followed by a falling rate period. Croll modelled the
drying process with a layer of packed particles above a still wet dispersion. Croll
argued that the resistance for water transport through packed particles is minimal
and that the particles do not have time to coalesce. Hence, the rate remains
constant until the amount of water in the film itself begins to diminish. It is
interesting to note that the initial rate of evaporation from the dilute dispersion,
studied by Croll, was 85% of the rate for pure water.
In terms of mass transfer resistance, a homogeneous (non-sedimented) dispersion
will have a boundary layer on its surface and the corresponding rate of water loss that
is close to pure water. If a layer of non-deformed particles accumulates at the top
surface, then an additional hindrance to evaporation is introduced and the rate of
water loss will diminish, although, as argued by Croll, this effect may be minimal. If
the top surface particles fuse together to form a continuous film, the rate of water loss
will be considerably reduced.
The results of Vanderhoff et al. and Croll demonstrate the importance of under-
standing the arrangement of particles vertically through the film during the drying
process. As will be shown, the distribution of particles is controlled by a range of
parameters including the particle size, film thickness and evaporation rate. The two
different behaviours, where the particles may either remain well dispersed or
accumulate at the top interface, are the two differing limits in the vertical distribution
problem. These scenarios are sketched in Fig. 3.2.
For a well-mixed dispersion, wherein the volume fraction is uniform through-
out the depth of the film thickness and the air surface is mainly water, the evapora-
tion rate may be taken as constant and equal to the evaporation rate of pure water.
The important question arises as to whether the film is uniform vertically (perpen-
dicular to the substrate). This is the subject of the next section.
Vertical Drying Profiles 99

Case A: Uniform Case B: Accumulation Case C: Accumulation


particle distribution of rigid particles at top of film-forming particles
results in constant interface results in at top interface results in
evaporation rate, as minimal increase in polymer layer and a
observed by Croll mass transfer resistance significant drop in
and a constant evaporation rate, as
evaporation rate, as observed by Vanderhoff
observed by Croll

Fig. 3.2 Side-view sketches of differences between experiments of Vanderhoff et al. (1973) and
Croll (1986, 1987). Croll observes a constant evaporation rate, close to that of pure water. This is
either due to a uniform particle distribution or particles that are too rigid to undergo wet
sintering. Vanderhoff et al. observe a slowdown in evaporation rate, consistent with a polymer
layer on top of the film. This occurs from non-uniform drying and particle deformation at the top
surface.

3.4 Vertical Drying Profiles

The spatial distribution of particles in latex films, which are subject to evapora-
tion, is important in determining the evaporation rate as well as the resulting film
microstructure. The observation is that inhomogeneities are observed both in the
vertical and in the horizontal (in the film plane) directions. We split these two
scenarios to examine each separately. Horizontal drying is considered in Section
3.5, and the vertical distribution of particles, sketched in Fig. 3.3, is considered
first. As the top surface reduces in height, particles that are near the surface are
captured, forming a close-packed region. Fig. 3.4 shows a series of freeze-
fracture SEM images from Ma et al. (2005). The accumulation of particles at the
top surface is seen above a dilute dispersion below. The packed particle layer
grows in size as drying progresses.
100 3 Drying of Latex Films

Fig. 3.3 Arrangement of particles in a drying film. The top surface recedes because of evapora-
tion, collecting particles. Diffusion of particles away from this concentrated region then results in
a distribution throughout the film. (Reproduced from Routh and Zimmerman (2004) with
permission)

(a) (b)

(c) (d)

Fig. 3.4 Freeze fracture SEM images of film cross-sections showing a front of close-packed
particles growing from the top surface of a drying film at successive lengths of times after
casting: a 1 minute, b 2 minutes and c 6 minutes. d A close-packed structure results after an hour
of drying. (Reproduced, with permission, from Ma et al. (2005))
Vertical Drying Profiles 101

3.4.1 Scaling Argument

The physics involves a balance between two competing timescales. One timescale
is associated with the time required for the water to evaporate in a film of initial
thickness H. The second timescale is how long it takes for a particle to diffuse
from the top of the film to the bottom. If diffusion is slow compared with the rate
of evaporation, then an accumulation of particles at the top surface is expected as
the water level falls. Conversely, if diffusion is rapid in relation to evaporation,
then a uniform vertical profile is predicted as the particles have sufficient time to
move towards the bottom of the film without accumulation at the top. The time for
evaporation is simply the film thickness, H divided by the evaporation rate, E (ex-
pressed as a velocity), hence

tevap ~ H .
E

For a film with initial thickness, H, the characteristic time for the diffusion of a
particle a distance H from the top to the bottom of the film is inversely related to
the particles diffusion coefficient:
2
tdiff ~ H .
D0

A comparison of the two times defines a dimensionless group, the Peclet number,
Pe as

t diff HE 6 RHE
Pe = ~ = , (3.2)
tevap D0 kT

where the final expression uses the Stokes-Einstein relation for the particle
diffusion coefficient in the dilute limit,

kT
D0 =
6R

with kT thermal energy, the solvent viscosity and R the particle radius. For Pe
>> 1, diffusion is relatively weak, and we expect a non-uniform particle distribu-
tion with a layer of close-packed particles accumulating at the top surface. The
particles do not move fast enough to stay ahead of the downward-moving surface.
If the particles are subject to deformation and coalescence, then a skin layer may
form and the evaporation rate will diminish considerably. (This is Case C in Fig.
3.2.) For Pe << 1 the particles are predicted to remain dispersed uniformly in the
film. The particles redistribute themselves fast enough to avoid being captured at
the wet film surface.
102 3 Drying of Latex Films

Equation (3.2) teaches us that uniform drying (usually what is desired) may be
achieved with slow evaporation rates and thin films. Small particles, in a solvent
with low viscosity, will have a high diffusivity and likewise encourage uniform
vertical drying. The addition of thickeners, such as the water-soluble polymers
often added to paint formulations, will have the effect of raising , slowing
particle diffusion, and ultimately leading to less uniform drying in the vertical
direction.

3.4.2 Governing Equations

Colloidal particles dispersed in a wet film obey a diffusion equation. Here, we


follow the analysis of Routh and Zimmerman (2004), although similar analyses
have been described by Brown et al. (Brown et al. 2002, Brown and Zukoski
2003) and Shimmin et al. (2006). In a scaled form, the local particle volume
fraction, , obeys

1 d
= K ( ) [Z ( )] , (3.3)
t Pe y d y

where K(), the sedimentation coefficient, accounts for hydrodynamic interactions


between particles resulting in a reduction in the particle diffusion coefficient. The
driving force for diffusion is the increase in particle chemical potential, captured
in the particle compressibility, Z(). Time is scaled with an evaporation time


t = tE
H

and distances are scaled with the original film thickness

y=y .
H

Numerical solutions to Eq. (3.3) describe how the volume fraction of solids
will vary with the vertical distance. The expected behaviour is shown in Fig. 3.5.
For large Peclet numbers (a value of 10 is shown), a sharp transition in the volume
fraction profile is seen with the original value in the wet dispersion persisting over
time in the lower regions of the film. For lower Pe (here a value of 0.2 is shown),
the particle concentration profiles are more uniform, with evaporation reducing
the film height and diffusion keeping the film profile mostly constant from top to
bottom. Yet, there is still a slightly higher concentration of particles near the top
surface.
Vertical Drying Profiles 103

(a)

(b)

Fig. 3.5 a Particle volume fraction profile in the vertical distance of a drying film with a Peclet
number of 10. The particles accumulate with a volume fraction of 64% at the top of the film. The
bottom of the film is unaffected by the evaporation because of the slow particle diffusion. There
is a sharp transition between the two regions. Successive times, normalised by the total drying
time, are shown. b Particle volume fraction profile in the vertical distance of a drying film with a
Peclet number of 0.2. Some particles accumulate at the top of the film, but the strong diffusion
results in a shallow volume fraction gradient. (Reproduced from Routh and Zimmermann (2004)
with permission)
104 3 Drying of Latex Films

Shimmin et al. (2006) include a sedimentation term in (3.3). This complicates


the numerical solution but leads to results similar to those presented in Fig. 3.5.
Narita et al. (2004, 2005) consider concentrated systems, where the diffusion of
particles is hindered by their neighbours. A revised Peclet number, in which the
diffusion coefficient is replaced by a transport coefficient, correlates well with
gravimetric measurements of water loss. The transport coefficient is derived from
Darcys law for flow of water through a bed of particles, and it also includes the
particle osmotic pressure and bed permeability.
In concentrated dispersions, the scaling of (3.3) becomes more complicated.
The time for evaporation is reduced to (m -)H/E, and the diffusion coefficient is
reduced to D0K(). Hence, a modified Peclet number may be written as Pe =
HE/(m -) D0K() = Pe/(m -)K(). Since (m -) and K() are both less than
unity, the modified Peclet number, Pe, will be considerably higher than the
original Pe for the case of a dilute dispersion. These corrections should introduce a
Peclet number that displays a transition in the simulated vertical profile around a
value of unity. The results should correlate with the results of Narita et al. (2005).

3.4.3 Experimental Studies

The vertical dependence of water concentration in drying films has been examined
by a number of techniques. The most powerful direct imaging technique is
cryogenic SEM. Its use is reported in depth by Ma et al. (2005) with some
remarkable images being displayed. As explained in Chapter 2 (Section 2.1.2.1),
the technique itself is non-trivial and prone to artefacts, but a direct guide to the
individual particle arrangement throughout the film depth can be obtained. Ma et
al. confirm the accumulation of particles at the top surface. Despite the appeal of
being able to visualise the particles, no numerical measure of the packing density
was directly obtained. It may be possible to use image analysis to determine the
local volume fraction although this is likely to be a manual and time-consuming
process.
Magnetic resonance (MR) profiling provides a quantitative measure of the
water content as a function of spatial position. The GARField magnet (described
in Chapter 2, Section 2.1.4.3) provides one-dimensional water profiles with a
spatial resolution of less than 10 m. MR profiling experiments have verified that
the water distribution is uniform in the vertical direction when Pe < 1, as predicted
from the arguments already outlined. A concentration gradient, with less water
near the film surface, is observed when Pe > 1 (Gorce et al. 2002). The results are
shown in Fig. 3.6. For a Peclet number of 0.2 the water profile is seen to be flat
and the film height decreases uniformly with time. For a Peclet number of 16, the
water profile is slanted with a lower water content at the top surface.
Vertical Drying Profiles 105

Fig. 3.6 GARField MR profiles showing the water content as a function of position in films at times
through the drying process. The magnetisation is proportional to water content. Series A has a
Peclet number of 0.2 (obtained with experimental parameters of H ~ 255 m, E ~ 0.2 108 m/s; D
~ 3.23 x 1012 m2/s). Notice how the water content is uniform throughout the film. Series B has a
Peclet number of 16 (obtained with H ~ 340 m, E ~ 15 108 m/s; D ~ 3.23 x 1012 m2/s). The
water content is higher towards the base of the film, indicating an accumulation of particles near the
top surface. (Reproduced with permission from Gorce et al. (2002))
106 3 Drying of Latex Films

In follow-on experiments, Pe was systematically varied over a wide range by


adjusting film thickness and drying conditions. In systems with Pe >10, the water
concentration gradient between the upper packed region and the lower dilute
region was found to be proportional to Pe to a power of 0.8 (Ekanayake et al.
2009). A scaling argument predicted a weaker dependence with the gradient being
proportional to Pe0.5 (Routh and Zimmerman 2004). The precise reasons for the
discrepancy are not clear and deserve further study.
Two-dimensional MR images provide further direct evidence for heterogene-
ous drying in thick (H = 5 mm) latex films (Rottstegge et al. 2003). Flowing air at
60C ensured a fast evaporation rate, leading to a very high Pe (value not re-
ported). The observation of a more dry layer at the film surface, growing over
time, is entirely in agreement with expectations. It was found that the addition of
excess non-ionic surfactant resulted in a more uniform water distribution. One
explanation for this effect of surfactant is that particle segregation is hindered by
the formation of a weak gel in the particle network.
The structure of final dried films can be determined by small angle neutron or
X-ray scattering (SANS or SAXS), where a long range order is detected with
peaks in the structure factor (Ballauff 2001). An example is given by Dingenouts
and Ballauff (1999) who used X-rays to show that the packed particles in a
partially dried film create an amorphous glass with a particle volume fraction of
around 64% and the remaining volume filled with water. SANS has also been used
to probe latex particle distributions as a function of time over a four-hour drying
period (Belaroui et al. 2003).
Liao et al. (2000) report Brownian dynamics simulations to explain segregation
of poly(methyl methacrylate) (PMMA) particles in a matrix of polystyrene (PS).
The PMMA is seen to clump or aggregate, motivating the interparticle potential
for electrostatically stabilised particles to be examined. If the particles are
aggregated, they diffuse slower and therefore accumulate at the top interface more
readily (i.e., Pe is higher). Indeed, the addition of excess salt to a charge-stabilised
latex leads to greater non-uniformity in drying (Knig et al. 2008). Salt causes a
screening of charge effects between particles that will reduce the co-operative
diffusion coefficient of the particles. This slowing down of the particle diffusion
could explain the particle accumulation at the top of the films with higher salt
contents. On the other hand, when the salt content is extremely high, particles will
sediment. The top surface then consists of water, and the evaporation rate is fast
(Erkselius et al. 2008).
A different numerical technique was employed by Reyes and Duda (2005) who
used a Monte Carlo simulation to examine the ordering of colloidal hard spheres
in thin films about eight particle radii thick. As might be expected, a fast evapora-
tion rate resulted in randomly packed films, and a slow evaporation rate showed
ordered colloidal domains.
Horizontal Packing and Drying Fronts 107

3.4.4 Consequence of Inhomogeneous Vertical Drying: Skin Formation

If particles accumulate at the top surface of the film, and they coalesce to form a con-
tinuous polymer film, then a barrier to further evaporation forms and drying becomes
extremely slow. This skin formation is a common observation when painting thicker
films, because of the greater Peclet number. The coalescence must be by a wet
sintering mechanism, because capillary stresses are not operational in a film with a
fluid below. (The mechanisms of skin formation are discussed further in Section 4.2.)
Therefore, skin formation is only observed for thicker films with soft polymers.
Skin formation causes the drying time to extend considerably and is therefore
an undesirable consequence of vertical inhomogeneities. Experiments have shown
that there is a correlation between particle coalescence and the formation of a skin
layer, which slows down water loss (Mallgol et al. 2006). The reason for the
slowdown is rather obvious. The transport of water by diffusion through the
polymer becomes the rate-limiting step in the water loss process. The prediction of
the diffusion coefficient of water through a polymer is a difficult task that is
complicated by swelling of the polymer by water. The sorptive capacity of
polymers may be predicted from knowledge of the chemical make-up. More
hydrophilic groups in a polymer lead to a higher sorptive capacity of water (van
Krevelen 1990). For instance, the sorption of a phenyl ring, a non-polar group, is
250 times lower than a hydroxyl group. In a poly(vinyl acetate) latex, as an
example, water concentrations of 34 wt.% are found (Rottstegge et al. 2003).
Diffusivity in a polymer with hydrogen-bonding groups increases with increasing
water content. The opposite trend a strong decrease is found in less hydro-
philic polymers, such as polymethacrylates (van Krevelen 1990). The flux of
water through the polymer is a function of the product of solubility and diffusivity,
and is low for a polymer that will be used in latex.
In applications of latex as a binder in a coating, there is often a need to smooth out
with a brush or rework a wet layer after its initial application. The time available to
rework a latex surface is called the open time by industrial scientists (Overbeek et
al. 2003). Latex films tend to have a shorter open time in comparison to polymer
solutions, because of the strong rise in the dispersion viscosity as the solids content
increases. Non-uniform drying will have the effect of concentrating the polymer
particles at the top surface. The open time will be even shorter in such a case.

3.5 Horizontal Packing and Drying Fronts

A common observation is that latex films do not dry uniformly, but instead have
fronts of differing appearance (partially transparent, clear etc.) passing laterally across
the film as it dries. Fig. 3.7 shows a picture of a dispersion containing 200 nm
108 3 Drying of Latex Films

Evaporation

x
A B C Fluid Packed
dispersion particles

Water P*
pressure

Pmax
x
Fig. 3.7 Top-view of a drying film displaying horizontal drying fronts. The particles consolidate
at the edge of the film and this establishes a horizontal flow. Particles are carried to the edge,
propagating the drying front laterally across the film. The region marked A is still fluid with a
particle volume fraction below close packing. The region marked B consists of close packed
particles, although still saturated with water. Region C is presumed to be dry. Notice the cracks
propagating into the film from the edge. On the right-hand side, the pressure of the fluid in the
film is sketched as a function of lateral distance, x. The sketch shows a cross-sectional view from
the edge of the film. In this case, the pressure generated by flow through the consolidated
particles, P*, is less than the maximum capillary pressure, Pmax, and hence the water remains
pinned at the edge of the film.

Fig. 3.8 Cryogenic SEM image of particles consolidating in a packing front at the edge of a
drying film (left-hand side). The right-hand side shows a dilute dispersion in water, which is now
in the form of ice. (Reproduced with permission from Ma et al. (2005))
Horizontal Packing and Drying Fronts 109

Fig. 3.9 A schematic diagram showing the two different types of horizontal drying front. The
particles consolidate into a close-packed solid near the films edge. The boundary with the dilute
central region defines the packing front. Further towards the edge, the space between particles
becomes devoid of water. The boundary with the wet particle bed is called the drying front.

polystyrene particles in water, which was cast on a substrate and allowed to dry.
The milky white region, marked with an A, is still fluid, with a volume fraction
below the point of close-packing. The region marked B is a solid in which the
particles have consolidated in a close-packed structure. Continued evaporation
from the solidified region leads to a flux of material from the fluid centre towards
the consolidated edge, carrying particles towards the edge and propagating the
front of close-packed particles across the film. This flow pattern is sketched on the
right-hand side of Fig. 3.7. This phenomenon can be visualised in experiments.
Fig. 3.8 shows a cryogenic SEM image, from the work of Ma et al. (2005), which
displays the accumulation of particles at the edge of a film.
Horizontal drying fronts have been described since the 1960s (Hwa 1964),
although more as a curiosity than anything else. There are usually at least two
different types of front in a latex film. The boundary between a dilute dispersion
and the packed particles is called a particle packing front or simply the particle
front. In a bed of packed particles, there is a boundary between wet and dry
regions, which is called the drying front. These two fronts are shown schemati-
cally in Fig. 3.9. Winnik and Feng (1996) measured the propagation rate for the
particle front for a range of blends between hard and soft particles and offered the
explanation of continued drying from the consolidated regions of the film as being
the driver of the packing front.
Parisse and Allain (1997) reported a model based on maintaining a constant
shape of a still wet drop of a colloidal dispersion. This model captured the
observed drying front behaviour and is similar to the model outlined below, in
section 3.5.1. The difference arises in the boundary condition at the drop edge,
which is assumed to retain a constant shape in the Parisse and Allain model and is
free to take up any shape in the model described below. The underlying physics of
evaporation from the consolidated region causing a lateral flow towards the edge
is however common between the two models. Deegan et al. (1997) report on drops
110 3 Drying of Latex Films

of dilute dispersions containing latex particles or particulate additives, such as


coffee. The horizontal drying front results in an accumulation of particles at the
edge of the drop and hence the common observation of coffee rings on kitchen
worktops! Subsequent work explored and exploited the role of the contact line at
the edge of the wet film as determining the patterning of the dry particles (Deegan
2000, Deegan et al. 2000).

3.5.1 Model for Horizontal Drying Fronts

As just noted, the experimental observations of fronts passing across a drying film
are numerous. The important questions are: why does the front form in the first
place? And how can the fronts position be predicted? Routh and Russel modelled
the flow in a dispersion of initially dilute dispersions (Routh and Russel 1998,
Routh et al. 2001) and subsequent experiments using MR imaging verified the
approach (Salamanca et al. 2002).
The starting point for the analysis is the lubrication approximation, which as-
sumes that the film is thin, such that vertical variations in pressure and particle
volume fraction are small and therefore negligible. The result is a flow based on
surface tension, with the film height also reducing because of evaporation. The
characteristic horizontal length scale in such a problem, L, is termed the capillary
length and represents the distance over which surface tension, , can cause a flow
in a film that is evaporating at a constant rate E . The capillary length is given by

1
4
L = H (3.4)
3E

where is the viscosity of the colloidal dispersion (not the solvent).


Surface tension seeks to reduce corners, so that material will flow away from
tight bends. Figure 3.10 shows how such a flow, away from an edge, will lead to a
reduction in the film height in the edge region. The evaporation rate is assumed to
remain constant across the film surface. In a given time interval, each part of the
film will lose the same amount of mass and have the same reduction in thickness.
The effect on particle volume fraction is far more significant in the thin regions.
(Think of a 1 m thick layer at a volume fraction of 25% losing 1 m in film height
because of water evaporation. The effect on the particle volume fraction will be
minimal. Compare that case to the effect on the volume fraction in a film of
thickness 2 m losing 1 m in thickness because of water evaporation.)
Because of the reduced height, the particles near the edge will begin to consoli-
date and to form a solid region prior to the central region. Continued evaporation
from the solidified region induces fluid to flow from the bulk to the edge. Particles
are carried to the edge, and the close-packed region propagates laterally across the
film. This mechanism is as sketched in Fig. 3.7. When the drying front forms, the
Horizontal Packing and Drying Fronts 111

lateral flow due to evaporation in the close-packed region dominates. The capil-
lary length scale is no longer relevant because the drying front propagates across
the entire film. The edge region, over which height variations are locked into the
film, is however the size of a few capillary lengths.

3.5.2 Lapping Time and Open Time

The time it takes for a latex film to start drying at the edge is referred to as the
lapping time, particularly by industrial scientists. When coating a large area,
such as a wall, a painter might stop halfway across it. If the lapping time is too
short, then there will be a seam at the border between the two halves when the
other half is coated (Overbeek et al. 2003). In some of the scientific literature, the
time at which water recedes from the edge of a latex film has been called the
open time. A model has been developed to predict this time.
As fluid flows through the consolidated region, there is a resulting pressure
drop along the packed bed. It is sketched in Fig. 3.7, showing the lowest value of
the pressure occurring at the film edge. The pressure at the film edge is termed a
capillary pressure. There is a maximum value of pressure that the particle network
can support. An estimate for this maximum value is approximated as

10
Pmax = (3.5)
R

although the value of 10 is for mathematical convenience, and a range from 2 to 12.9
is used by different groups depending on the assumed particle-packing geometry
(White 1982, Dunstan and White 1986, Brown 1956, Holmes et al. 2006).

Surface tension
induced flow

Initial state

L
Capillary length is the distance
that surface tension drives the
shape in the time that evaporation
takes to reduce the film height

Fig. 3.10 Surface tension causes fluid to flow from a tight corner to a more rounded shape. The
shape is changed over a capillary length.
112 3 Drying of Latex Films

The pressure drop along the particle bed is controlled by the bed permeability,
kp. A less permeable bed will require a greater pressure gradient to achieve the
same flow as in a more open and permeable bed. For Darcy flow we can write

dP
= ux (3.6)
dx kp

where ux is the horizontal fluid velocity in the film, with x representing the lateral
direction. From (3.6) the characteristic pressure drop along the particle bed, P*,
can be estimated by using the characteristic horizontal length scale, L, from (3.4)
and the characteristic horizontal velocity, LE / H . After substituting in L for the
distance x, and the characteristic horizontal velocity for ux, one finds

E L2
P* = (3.7)
Hk p

Comparing the maximum possible capillary pressure (3.5) with the pressure
achievable by flow through the particle network defines a dimensionless group,
called the reduced capillary pressure, Pcap. It is written as

Pmax 10Hk p
Pcap = = (3.8)
P* E RL2

If the pressure drop reaches this maximum value, then water will recede from
the edge of the film and the particles will become dry. This point will occur
when Pcap is less than around unity. A film with a lower Pcap will display a short
open time. Conversely, for the cases with Pcap >>1, the maximum capillary
pressure will never be reached by the horizontal flow, and the film will remain
wet at the edge of the film until the last stages of evaporation. One can think of
the water as being pinned at the film edge. In these cases, the film will display
a long lapping time.
Equation (3.8) teaches us what parameters are important in determining the rate
of horizontal drying. Drying from the edges may be avoided with slow evapora-
tion rates, and thick films, since these contribute to a high Pcap. The effect of
particle size is not immediately clear. The radius appears on the denominator in
(3.8) implying that a small particle size leads to a large Pcap. The complication
arises however in that the permeability, kp, is a strong function of particle size,
with a small particle radius reducing kp significantly. Hence, reducing the particle
size reduces Pcap and exacerbates the development of edge drying.
The recession of water from the edge of films was followed experimentally by
Salamanca et al. (2001) using MR imaging. Their experimental results are shown
in Fig. 3.11. The water in the film gives an NMR signal and appears white in the
images at its maximum concentration. The top image shows a film with water
throughout as found immediately after casting. After a period of time, the particles
Horizontal Packing and Drying Fronts 113

have collected at the edge. The grey region indicates a lower water concentration,
and the boundary between the white and grey region defines the position of the
packing front. The existence of a water signal indicates that the particles in the
packed region remain wet. At later times still, the water has receded from the edge
of the film and no signal is then detected from this dry film. The series of images
are consistent with the sketch of the two fronts in Fig. 3.9.

0 hr. 2.6 mm

3 hr.
Wet,
colloidal
dispersion Particle
packing front
6 hr.

Drying front

22 mm

Fig. 3.11 A series of NMR images obtained from a slow-drying latex layer after casting (zero
hours), after three hours of drying, and after six hours. The grey scale corresponds to water
content, with white representing higher water concentration, and black meaning no water.
Particle packing fronts (separating a wet particle bed and a dilute region) are seen after three
hours, and drying fronts are seen after six hours. (Reprinted from Salamanca et al. (2001);
copyright 2001 American Chemical Society)

The time before a drying front recedes from the edge of the film is easily de-
termined, and it corresponds to the lapping time. Salamanca et al. followed this
time as a function of the dimensionless capillary pressure Pcap, and their results are
reproduced in Fig. 3.12. The upward jump in open time around a Pcap of 10 agrees
broadly with theoretical predictions, which are shown for comparison.
A further consequence of the horizontal drying is that the pressure drop along
the packed region places the particle network into compression. The particles can
be deformed and film form under the action of this pressure. The result is a front
of optical clarity passing across the film a little behind the packed particle front,
where particles first come into close contact (Routh et al. 2001). Visually, one can
see a turbid film become partially transparent when the packing front passes, and
then become fully transparent when the optical clarity front passes.
114 3 Drying of Latex Films

Pcap
Fig. 3.12 Open time of a drying latex film plotted as a function of the dimensionless capillary
pressure Pcap. The points are experimental measurements, whilst the lines are theoretical fits
using the theory outlined in the text. A variety of particle sizes and film thicknesses were used to
achieve a range of Pcap values. (Reproduced with permission from Salamanca et al. 2001;
copyright 2001 American Chemical Society)

3.6 Colloidal Stability

As evaporation proceeds, the concentration of electrolyte in the remaining


aqueous phase will increase. A natural expected consequence is a loss of particle
stability and an aggregation between particles. Depending on the concentrations
involved, the particles may already be arranged in a close-packed form in the fluid
phase, and so the induced attraction will merely lead to a small consolidation of
the bed. Alternatively, if the loss in stability occurs early in the drying process,
bulk aggregates may form and will hinder uniform film formation. An example of
the effect of colloidal stability is given by Juhu and Lang (1993). Adding
surfactant to latex and achieving a full coverage of the particles, thereby ensuring
maximum colloidal stability, resulted in optimum packing of the particles and led
to a smooth film. Poorer colloidal stability resulted in the settling of particle flocs
and a rougher film.
There are two ways that a charge-stabilised dispersion can become unstable
during drying. Fig. 3.13 shows a theoretical consideration of how the Debye
length, 1, changes during the drying process. As outlined in Section 1.6 of
Chapter 1, the Debye length is the distance over which electrostatic repulsions
operate between particles. As the electrolyte concentration increases, the charges
become screened and the Debye length decreases. The maximum in the interparti-
Colloidal Stability 115

cle potential, which imparts stability, is reduced (see Fig. 1.14). If the maximum
potential is less than a few kT (say 5), then the thermal energy will enable the
aggregation of the particles. The energetic barrier between particles is no longer
high enough to oppose the attractive van der Waals forces.

wmax = 5kT

40 0.05 M
0.10 M
0.15 M
Particle Spacing (nm)

30 Time
Unstable

20

10 Stable

0
0.5 1.0 1.5
Debye screening length (nm)

Fig. 3.13 Evolution of Debye length and average interparticle spacing during drying for 90 nm
particles at an initial volume fraction of 0.4 and a critical coagulation concentration of 0.2M
NaCl. Aggregation is assumed to occur when the interparticle potential has a maximum repulsion
of 5 kT. Here the interparticle spacing refers to the distance between the particle surfaces.

Secondly, as the particles become more concentrated, the average interparticle


spacing decreases. One could imagine that if the particles are forced to be closer
than the distance corresponding to the maximum in the barrier, then aggregation
will occur.
The calculations in Fig. 3.13 relate to 180 nm latex particles with a critical coagu-
lation concentration (CCC) of 200 mM NaCl. It can be seen that for initial salt
concentrations below 50 mM, the dispersion remains stable through the drying
process. It is noticeable that right until the point of particle packing (a particle
spacing of 0), the barrier is sufficiently high to prevent aggregation. At higher initial
salt concentrations, which bring the dispersion closer to its CCC, it is predicted that
the barrier height could become so low that aggregation will start before particle
contact has been achieved. In reality, colloidal dispersions are designed to be highly
stable, which make it unlikely to be subject to flocculation during drying. However,
there have been reports of fast drying that are explained by particle flocculation and
sedimentation at high salt contents (Erkselius et al. 2008). It has also been proposed
that the presence of salt influences the stability of surfactant layers at particle
interfaces in a skin layer. The presence of some particular anions and cations
favoured particle coalescence in a skin layer (Knig et al. (2008).
116 3 Drying of Latex Films

3.7 Film Cracking

The flow of solvent through the packed-particle region of the film causes a
pressure drop. The capillary pressure may result in compression of the film as
discussed above. If the particles are too rigid to deform, the film can display
cracks, as shown in Fig. 3.7, where the cracks follow a short distance behind the
particle front through the film.
The idea that capillary pressure is the driving force for cracking is now widely
accepted (Dufresne et al. 2003, Dufresne et al. 2006). However, the actual
mechanism remains in dispute. There are a number of experimental observations
that make film cracking a fascinating area of study.

Crack spacing Position of crack front

Fig. 3.14 A crack front propagating across a drying film. Notice that the cracks all extend up to a
common crack front position. The spacing between the parallel cracks is roughly the same.

3.7.1 Do the Cracks Follow the Drying Front or Propagate Quickly Over
the Entire Film?

Two possibilities have both been reported. The simplest case consists of cracks
following the drying fronts across the film (Lee and Routh 2004). Fig. 3.14 shows
a screen grab from a video of 20 nm silica particles drying in a film. The cracks
propagate as a front across the film in a stick-slip type motion, where a plot of the
position of the crack front against time resembles a staircase. The same observa-
tion has been made by Dufresne et al. (2003, 2006). Others, however, report that
the crack front progression is much more uniform (White 2006) with a linear front
progression.
Further complicating the observations, some authors have reported a second
case in which the films remain intact until the entire film has consolidated, with
failure occurring thereafter (Tirumkudulu and Russel 2005). The most likely
117

explanation for this discrepancy is the capillary pressure generated by the horizon-
tal flow being too small to cause failure. The late-time failure might only be
observed in the cases with Pcap >>1. This conjecture, however, awaits experimen-
tal verification.

3.7.2 What Sets the Crack Spacing?

A common observation is that a film displays a characteristic distance between its


parallel cracks (Fig. 3.14). The derivation of a prediction of this crack spacing is
the subject of many papers. The most common analytical approach is to apply an
energy balance between the elastic energy released (or gained), when cracking
the film and the surface-free energy cost of creating new surfaces on either side
of the crack. The resultant prediction for crack spacing is determined by the size
of material needed to provide enough elastic energy to make cracking energeti-
cally favourable (Thouless 1990). The underlying assumption for a latex film is
that packed particles constitute an elastic solid. However, this assumption might
not be valid in a colloidal film in which a fluid is being transformed into a solid.
An alternative explanation was put forward by Lee and Routh (2004) who
proposed that the flow of solvent away from the crack can also relieve the
capillary pressure. The flow provides an additional relaxation mechanism that sets
the crack spacing.

References

Barnes G.T. and Gentle I.R. (2005) Interfacial Science, Oxford University Press.
Ballauff M .(2001) SAXS. and SANS studies of polymer colloids. Curr Opin Coll
Interf Sci 6:132-139
Belaroui F., Cabane B., Dorget M., Grohens Y., Marie P., Holl Y .(2003) Small-
angle neutron scattering study of particle coalescence and SDS desorption
during film formation from carboxylated acrylic latices. J Coll Interf Sci 262:
409-417
Brown G.L. (1956) Formation of films from polymer dispersions, Journal of
Polymer Science, 22: 423-434.
Brown L.A., Zukoski C.F. and White L.R. (2002) Consolidation during drying of
aggregated suspensions. AIChE J 48 (3) 492-502.
Brown L.A. and Zukoski C.F. (2003) Experimental tests of two phase fluid
models of drying consolidation. AIChE J,. 49(2) 362-372.
Ciampi E. and McDonald P.J. (2003) Skin formation and water distribution in
semicrystalline polymer layers cast from solution: A magnetic resonance imag-
ing study, Macromolecules, 36:8398-8405 2003.
118 3 Drying of Latex Films

Croll S.G. (1986) Drying of latex paint. Journal of Coatings Technology 58(734):
41-49.
Croll S.G. (1987) Heat and mass transfer in latex paints during drying. Journal of
Coatings Technology, 59(751): 81-92.
Cussler E.L. (1984) Diffusion: mass transfer in fluid systems, Cambridge Univer-
sity Press.
Deegan R.D., Bakajin O., Dupont T.F., Huber G., Nagel S.R. and Witten T.A.
(1997) Capillary flow as the cause of ring stains from dried liquid drops. Na-
ture, 389: 827-829.
Deegan R.D. (2000) Pattern formation in drying drops. Physical Review E,. 61(1):
475-485.
Deegan R.D., Bakajin O., Dupont T.F., Huber G., Nagel S.R. and Witten T.A.
(2000) Contact line deposits in an evaporating drop. Physical Review E 62(1):
756-765
Dingenouts N. and Ballauff M. (1999) First stage of film formation by latexes
investigated by small angle x-ray scattering, Langmuir, 15: 3283-3288.
Dufresne E.R., Corwin E.I., Greenblatt N.A., Ashmore J., Wang D.Y., Dinsmore
A.D., Cheng J.X., Xie X.S., Hutchinson J.W. and Weitz D.A. (2003) Flow and
fracture in drying nanoparticle suspensions, Physical Review Letters, 91(22):
4501-4504.
Dufresne E.R., Stark D.J., Greenblatt N.S., Cheng J.X., Hutchinson J.W., Ma-
hadevan L. and Weitz D.A. (2006) Dynamics of fracture in drying suspensions,
Langmuir 22(17): 7144-7147
Dunstan D. and White L.R. (1986) A capillary pressure method for measurement
of contact angles in powders and porous media. Journal of Colloid and Inter-
face Science, 111(1): 60-64.
Ekanayake, P., McDonald P.J., Keddie, J.L. (2009) An experimental test of the
scaling prediction for the spatial distribution of water during the drying of
colloidal films. European Physical Journal Special Topics 166: 21-27.
Erkselius S., Wads L. and Karlsson O.J. (2007) A sorption balance based method
to study the initial drying of dispersion droplets, Colloid and Polymer Science
285: 1707-1712.
Erkselius S., Wads L., Karlsson O.J. (2008) Drying rate variations of latex
dispersions due to salt induced skin formation. J Coll Interf Sci 317: 83-95.
Gorce, J.-P., McDonald, P.J. and Keddie, J.L. (2002) Vertical water distribution
during the drying of polymer films formed from emulsions. Eur Phys J E 8:
421-429.
Holmes D.M., Kumar R.V. and Clegg W.J. (2006) Cracking during lateral drying
of alumina suspensions, Journal of the American Ceramic Society 89 (6): 1908-
1913.
Hwa J.C.H. (1964) Mechanism of film formation from lattices. Phenomenon of
flocculation, Journal of Polymer Science: Part A 2(2): 785-796.
Juhu D. and Lang J. (1993) Effect of surfactant post added to latex dispersion on
film formation: A study by atomic force microscopy. Langmuir 9: 792-796.
References 119

Knig A.M., Weerakkody, T.G., Keddie, J.L. and Johannsmann, D. (2008)


Heterogeneous drying of colloidal polymer films: Dependence on added salt.
Langmuir 24: 7580-7589.
Lee W.P. and Routh A.F. (2004) Why do drying films crack? Langmuir 20(23):
9885-9888.
Liao Q., Chen L., Qu X. and Jin X. (2000) Brownian Dynamics simulation of film
formation of mixed polymer latex in the water evaporation stage, Journal of
Colloid and Interface Science 227: 84-94.
Ma Y., Davis H.T. and Scriven L.E. (2005) Microstructure development in drying
latex coatings. Progress in Organic Coatings 52: 46-62.
Mallgol J., Bennett G., McDonald P.J., Keddie J. and Dupont O. (2006) Skin
development during film formation of waterborne acrylic pressure sensitive
adhesives containing tackifying resins. The Journal of Adhesion, 82: 217-238.
McNamee C.E., Barnes G.T., Gentle I.R., Peng J.B., Steitz R. and Probert R.
(1998) The evaporation resistance of mixed monolayers of octadecanol and
cholesterol. Journal of Colloid and Interface Science, 207: 258-263.
Reyes Y. and Duda Y. (2005) Modelling of drying in films of colloidal particles.
Langmuir 21: 7057-7060.
Overbeek, A., Bckmann, F., Martin, E., Steenwinkel, P., Annable T. (2003) New
generation decorative paint technology. Progress in Organic Coatings 48: 125
139.
Narita T., Beauvais C., Hebraud P. and Lequeux F. (2004) Dynamics of concen-
trated colloidal suspensions during drying aging, rejuvenation and overaging,
European Physics Journal E 14: 287-292.
Narita T., Hebraud P. and Lequeux F. (2005) Effects of the rate of evaporation and
film thickness in nonuniform drying of film-forming concentrated colloidal
suspensions. European Physics Journal E 17: 69-76.
Parisse F. and Allain C. (1997) Drying of colloidal suspension droplets: Experi-
mental study and profile renormalisation. Langmuir, 13(14): 3598-3602.
Rottstegge J., Traub B., Wilhelm M., Landfester K., Heldmann C. and Spiess
H.W. (2003) Investigations on the film formation process of latex dispersions
by solid state NMR spectroscopy. Macromolecular Chemistry and Physics, 204
(5/6): 787-802.
Routh A.F. and Russel W.B. (1998) Horizontal drying fronts during solvent
evaporation from latex films. AIChE J 44(9): 2088-2098.
Routh A.F., Russel W.B., Tang J. and El-Aasser M.A. (2001) Process model for
latex film formation: Optical clarity fronts. Journal of Coatings Technology
73(916): 41-48.
Routh A.F. and Zimmerman W.B. (2004) Distribution of particles during solvent
evaporation from films, Chemical Engineering Science 59: 2961-2968.
Salamanca J.M., Ciampi E., Faux D.A., Glover P.M., McDonald P.J., Routh A.F.,
Peters A.C.I.A., Satguru R. and Keddie J.L. (2001) Lateral drying in thick films
of waterborne colloidal particles. Langmuir 17: 3202-3207.
Shimmin R.G., DiMauro A.J. and Braun P.V. (2006) Slow vertical deposition of
colloidal crystals: A Langmuir-Blodgett process?, Langmuir 22 6507-6513.
120 3 Drying of Latex Films

Thouless M.D. (1990) Crack spacing in brittle films on elastic substrates. Journal
of the American Ceramic Society, 73(7): 2144-2146.
Tirumkudulu M.S. and Russel W.B. (2005) Cracking in drying latex films.
Langmuir, 21(11) 4938-4948
Vanderhoff J.W., Bradford E.B. and Carrington W.K. (1973) The transport of
water through latex films, Journal of Polymer Science, part C: Polymer Sympo-
sium, 41:155-174
van Krevelen D.W. (1990) Properties of Polymers: Their Correlation with
Chemical Structure, Their Numerical Estimation and Prediction from Additive
Group Contribution, 3rd edition Elsevier: Amsterdam 569-573.
Wallin M., Glover P.M., Hellgren A.C., Keddie J.L. and McDonald P.J. (2000)
Depth profiles of polymer mobility during the film formation of a latex disper-
sion undergoing photoinitiated crosslinking. Macromolecules 33: 8443-8452
White L.R. (1982) Capillary rise in powders. Journal of Colloid and Interface
Science, 90(2).536-538
White L.R. (2006) Personal communication.
Winnik M.A. and Feng J. (1996) Latex blends: An approach to zero VOC
coatings. Journal of Coatings Technology, 68(852):39-50.
Chapter 4
4 Particle Deformation

4.1 Introduction

When soft particles come into close packing, they will usually start to deform.
The ideal result is a structure without voids although with the individual particles
still distinguishable, prior to interdiffusion. For such a deformation there must be a
driving force for compaction, and there will be a mechanical response from the
particles to balance it. The primary driving force is essentially a reduction in the
surface free energy and hence the surface area. Depending on the controlling
mechanism, there are three interfacial areas in a latex film that can be reduced:
polymer/air; polymer/water; and water/air. For polymeric materials, the rheological
response is complex and is crucially dependent on temperature. It is possible to have
an instantaneous elastic response or a time-dependent viscous creep.
After the drying stage, the packing of the particles will be into a form of close
packing, with the evaporation rate crucial in determining whether colloidal
crystallisation can occur (Russel 1990, Russel et al. 1990). If the particles pack
into a face-centred cubic structure, then after deformation each particle assumes a
rhombic dodecahedral geometrical shape when all void space is filled (as in
Fig. 1.10b). A TEM image of a close-packed array of dodecahedra is shown in
Fig. 4.1. The deformation of particles enables an opaque, powdery film to be
transformed into a clear, crack-free material. While subsequent interparticle
polymer diffusion is required to achieve mechanical strength (described in
Chapter 5), the deformation step is a necessary precursor. The success or other-
wise of this deformation transition is the deciding factor in whether the film is
useable as a protective coating. Hence, the deformation step is crucial in achieving
film formation, and industrial formulators will go to great lengths to ensure its
success.
This chapter is split up as follows. Firstly, the different possible driving forces, as
have been proposed over the previous 50 years, are outlined. Experimental evidence
for each mechanism is presented. The mechanical response of two particles is
discussed, ranging from the well-known limit of viscous particles sintering under the
122 4 Particle Deformation

action of surface tension, through to the response of elastic particles under the
influence of a force. Instead of considering two isolated particles, the stress-strain
response of a bed comprising many particles is derived for the case of viscoelastic
particles under the influence of surface tension and a compressive force. This
volume-averaged model then provides the basis for the process conditions under
which a deformation mechanism may operate. Finally, experimental observations
concerning different deformation mechanisms are outlined.

Fig. 4.1 Particles in an FCC array will deform into rhombic dodecahedra. This image was
obtained from transmission electron microscopy of a carbon replica of a fractured latex film.
(Reproduced with permission from Wang et al. (1992); copyright 1992 American Chemical
Society)

4.2 Driving Forces for Particle Deformation

There are a number of possible driving forces for particle deformation. A


common theme is the reduction in interfacial area as the particles deform,
eliminating the boundaries between particles and the interparticle voids. For a
100 m thick, the number of particles (with a radius of R = 100 nm) in the film
is 104/ (4/3 R3) = 2.4 x 1016. The surface area associated with these particles,
prior to their deformation, is 4R2 x 2.4 x 1016 = 3,000 m2, which is comparable
4.2 Driving Forces for Particle Deformation 123

to a football field or a cricket pitch. For a polymer surface tension of 0.01 N/m,1
the energy reduction associated with particle deformation to reduce their area is
therefore 30 J. This simple example considers the air and polymer interface, but
various other driving forces are discussed next. Sintering is a word used to
describe the reduction of interfacial area in order to reduce the surface free
energy of a particulate material. In the literature on latex, sintering is described
as being either wet or dry depending on whether the particles are in the
presence or absence of water. These and other mechanisms are sketched in Fig.
4.2 and will be discussed individually.

4.2.1 Wet Sintering

In the wet sintering mechanism, the surface tension between the polymer and
water phases is the dominant driving force. If the particles are pressed together in
water, the interfacial area between them is eliminated, and the interfacial free
energy is decreased. (Think of what happens when two oil droplets bump together
in a salad dressing.) The surface tension between a latex polymer and the aqueous
phase is estimated to be about pw = 0.015 N/m (Eckersley and Rudin 1990). The
wet sintering deformation mechanism was first proposed by Vanderhoff et al.
(1966). Experimental evidence to support wet sintering was presented by both
Sheetz (1965) and then by Dobler et al. (1992), who held latices under water for a
considerable length of time and observed a consolidation. Gradual increases in
optical transparency (van Tent and te Nijenhuis 2000) and in refractive index
(Keddie et al. 1995, 1996b) have been explained by a wet sintering mechanism.

4.2.2 Dry Sintering

This mechanism is analogous to wet sintering, except that the polymer-air surface
tension, pa, causes the deformation. Dry sintering is a common mechanism in
other types of particulate materials, such as ceramics made from oxide powders. It
was originally proposed by Dillon et al. (1951) and given strong experimental
backing by Sperry et al. (1994) who followed the cloudy-clear transition of a film
by using a minimum film formation temperature bar. Light is scattered from voids
within the film, and hence, as these voids become much smaller than the wave-
length of light, the film becomes clear (van Tent and te Nijenhuis 2000). The work
of Sperry et al. showed that the position of the transition was similar for two films:

1
All forms of condensed matter have a surface (or interfacial) tension, depending on their
surroundings (solid, liquid or vapour.) Surface usually refers to an interface with air. It is
sometimes more intuitive to think in terms of a surface energy (in units of energy per unit area),
rather than a tension (in units of force per distance). Note that 1 N/m = 1 J/m2.
124 4 Particle Deformation

one drying normally and another that had been fully dried at a far lower tempera-
ture before being placed on the temperature gradient of the bar. Particle deforma-
tion occurred in the dry particles at the same rate and temperatures as in particles
cast in the wet film. Their result proves that water is not relevant in the deforma-
tion around the cloudy-clear transition, at least for these acrylic films. Particles
were able to deform in air even when they were fully dry.

1. Wet Sintering (Vanderhoff et al., 1966)


polymer/water surface tension

2. Dry Sintering (Dillon et al,1951)


polymer/air surface tension

3. Capillary Deformation (Brown,1956)


pressure at air-water interface meniscus

4. Capillary Rings (Lin and Meier,1995)


Residual water as pendular rings around pendular rings
particle contacts

5. Sheetz deformation (Sheetz, 1965)


Solid polymer layer at surface

Fig. 4.2 Possible mechanisms for particle deformation.

4.2.3 Capillary Deformation

At a wet film surface, the water between particles will take up a meniscus, which
is a curved interface. The radius of curvature depends on the wetting of the
particles (indicated by the water contact angle on the polymer surface.) Such a
meniscus is shown in Fig. 4.3 with the curvature of the air-water interface creating
a negative pressure in the fluid, called the capillary pressure. (The pressure is
always lower on the concave side of the meniscus.) The capillary pressure is
4.2 Driving Forces for Particle Deformation 125

considered in some detail in Chapter 3 in the discussion of lateral drying (Section


3.5). A key point is that the pressure differential across the meniscus is inversely
proportional to its radius of curvature. A more tightly curved water surface will
create a greater capillary pressure. The pressure is a direct manifestation of the
water-air surface tension, wa, as a greater tension will require a greater pressure to
create a curved surface. The accepted value of wa is 0.073 N/m.

Fig. 4.3 Cryogenic TEM image of consolidated latex particles demonstrating the curvature of the
air-water interface. (The water is presently in the form of ice, but the meniscus from the liquid
state has been captured.) Reproduced with permission from Ma et al. (2005).

In a mechanism originally proposed by Brown (Brown 1956), the negative


fluid pressure is equivalent to a weight being placed on the top surface of a film,
compressing the particles. Particle compression pushes water toward the film
surface and thereby reduces the curvature of the meniscus. This reduces the
capillary pressure. For this reason, the rate of particle deformation is necessarily
set to be concurrent with evaporation. Brown balanced the elastic response of
particles (determined by the polymers shear modulus, G) with the capillary
pressure and came up with a criterion for film formation. For particles of radius R,
deformation is predicted to occur when G < 35 wa/R.
Numerous other authors have performed similar analyses, deriving slightly
different limits, although the prediction of deformation when G < A wa/R, where
A is some number in the range from 30 to 300, is recurrent (Mason 1973, Ecker-
sley and Rudin 1994). Experimentally, there are examples of latex systems in
which water evaporation has been found to occur simultaneously with particle
deformation, and it acts as the rate-limiting step, as expected for a capillary
deformation mechanism (Keddie et al. 1996b).
126 4 Particle Deformation

4.2.4 Capillary Rings

A complication arises with residual water in latex films. For dry sintering to be
operative, there must be a polymer-air interface. Any residual water, present as
pendular rings around particle-particle contacts, will rule out the possibility of
dry sintering. This argument was put forward strongly by Lin and Meier in a
series of papers (1995, 1996). Sperry et al. (1994) argue that residual water will
be adsorbed into the latex particles, making their surfaces dry and therefore
liable to dry sintering. Regardless of the local physical chemistry, the capillary
pressure exerted by pendular rings operates in exactly the same way as the
polymer-air surface tension, pulling the contact region outwards and hence
deforming the particles. The result is that deformation in a humid atmosphere,
with hydrophobic particles, is likely to be due to this moist sintering mechanism,
although deformation will proceed in exactly the same fashion as the particles
undergoing dry sintering.

4.2.5 Sheetz Deformation

Sheetz noted that film formation is often vertically inhomogeneous, with a skin
of polymer encasing a fluid dispersion below (Sheetz 1965). Sheetz argued that
diffusion of water through the continuous polymer produces a compressive
force on the film below. It seems more likely, however, that the polymer skin
significantly reduces the evaporation rate. Consequently, particles accumulated
at the top surface because of inhomogeneous drying (as discussed in Section
3.4) have considerable time to sinter by a wet sintering mechanism. The
vertical inhomogeneity has important implications on the final film morph-
ology, and hence Sheetz deformation is retained as a deformation mechanism in
its own right.
As will be discussed in Section 4.9, all of the deformation mechanisms may
be observed in a latex film. The easiest control parameters are evaporation rate,
which sets the time over which drying occurs, and temperature, which sets the
viscosity of the polymer (and its time response to shear forces). Irrespective of
the driving force for deformation, the sintering of particles is common. The next
section describes the different models for sintering of two isolated particles. The
limitation of this single pair viewpoint is then examined and methods to gener-
ate bulk film models are outlined.
4.3 Particle Deformations 127

4.3 Particle Deformations

4.3.1 Hertz Theory Elastic Spheres with an Applied Load

The simplest model for two sintering particles is due to Hertz (1881). In this
model, two elastic particles of radius R are pushed together by a force F. The
exact elastic solution gives the radius of contact a0 as

3 RF (1 v )
2

a0 = (4.1)
4 E

where is Poissons ratio2 and E is the materials elastic (Youngs) modulus. The
geometry is sketched in Fig. 4.4. A further result from the same analysis is that the
particle centres approach each other by a distance given by

9 F (1 )
2 2 2

=
3
(4.2)
2 RE 2

The Hertz theory has been verified for many systems where there is a consider-
able applied force. At small applied loads, however, discrepancies have been
observed. These are because of surface forces (or surface tension) pulling the
particles together and leading to an increase in the contact radius. The first
analysis of the effects of surface forces in the Hertz theory was carried out by
Johnson, Kendall and Roberts (JKR) (1971).

4.3.2 JKR Theory Elastic Spheres with an Applied Load and Surface
Tension

Surface energy will act to pull the contact points between particles outward and
hence increase the area of contact. (As with dry sintering, an increased contact
area will reduce the surface free energy.) Another way to visualise this mechanism
is that by increasing the radius of contact between particles, the area of contact
with the external fluid decreases, and hence the energy of the system decreases as
well. The increase in contact area is sketched in Fig. 4.4. Johnson, Kendall and
Roberts used this energy argument to derive the radius of the area of contact, a, as

2
The Poissons ratio, , relates strain in orthogonal directions. If a strain is applied in one
direction, then a strain of is observed in a direction at right angles to the direction of the
applied strain. For a non-compressible rubbery polymer, is 0.5.
128 4 Particle Deformation

3 (1 2 ) R
1
2

F + 3 R + 3 RF + 3 R
2

a =
3 (4.3)

4E 2 2

where is the surface tension between the particles and the external fluid. It is
interesting to note that if is set to zero, then (4.3) reverts to the Hertz solution,
(4.1). In addition, the surface tension provides an adhesion between the particles
that requires a (negative) force to ensure separation.

F R a F

Fig. 4.4 Two elastic spheres compressed by a force F have a radius of contact a0. Dotted line
shows how contact area will appear, if surface tension is operating.

An interesting aside is that the approach of Johnson, Kendall and Roberts is a


so-called macroscopic one, with a surface tension derived from a bulk material
property. An alternative, but necessarily analogous, approach is to use surface
forces with a van der Waals attraction operating between the particles. This
alternative approach is followed by Fogden and White (1990) in deriving both the
Hertz and JKR results. The discussion about a surface forces versus macroscopic
surface tension approach was taken further in a series of papers by Derjaguin and
co-workers (Derjaguin et al. 1975, Muller et al. 1980, Muller et al. 1983) and used
by Mazur et al. to explain why the deformation of smaller particles is seen to be
essentially instantaneous, and therefore elastic (Mazur et al. 1994, 1997; Jagota et
al. 1997). Larger particles in contact are predicted to have an initial elastic
deformation followed by a slower viscous flow.

4.3.3 Frenkel Theory Viscous Spheres with Surface Tension

The Hertz and JKR approaches assume that the particles are elastic, to derive the
material response. The opposite end of the material spectrum concerns viscous
particles. The deformation of viscous particles was considered by Frenkel (1945),
who estimated the angle of contact between two identical spheres, , under the
action of surface tension alone. The result is sketched in Fig. 4.5. The spheres are
assumed to remain spherical with the radius at any time, R, related to the original
4.3 Particle Deformations 129

radius, R0 by a mass balance. Equating the change in surface energy, dS/dt,


where is the surface tension and S is the total surface area, to the viscous
dissipation, which is estimated as V 2 , where V is the total volume, is the
polymer viscosity, and is the shear rate, estimated for small areas of contact as
d/dt ( 2/2), results in an expression for the angle of contact as

3 t
2 = . (4.4)
2 R0

The geometric analysis is then simple to give the approach of particle centres as
R0 2/2 and the area of contact between particles as R02 2. There have been
numerous variations to Frenkels theory (Kuczynski et al. 1970, Pokluda et al.
1997, Rosenzweig and Narkis 1981) although the basic rationale remains as
outlined above. Experimental evidence to support Frenkels theory was given by
Dillon et al. (1951) who measured the angle , for Saran latex particles, and
observed a linear growth rate for 2.

(a)

(b)

Fig. 4.5 a Two viscous spheres of radius R sintering because of surface tension. is the surface
tension between the polymer and the external phase. The degree of sintering can be characterised
by the angle . b An AFM image (0.5 m 0.5 m) of soft latex particles shows that the
boundaries between them is flat. An angle of contact is identified on the image for two particles.
(Image courtesy of Argyrios Georgiadis, University of Surrey)
130 4 Particle Deformation

4.3.4 Viscoelastic Particles

Many authors comment that polymer particles are necessarily viscoelastic


(Greenwood and Johnson 1981, Bellehumeur et al. 1998). Lee and Radok (1960)
derive the viscoelastic equivalent of the Hertz problem for two spheres com-
pressed by a force F. In a series of papers, Eckersley and Rudin assumed the
radius of the circle of contact between particles to be additive between the elastic
and viscous components (1990, 1993, 1994). A capillary deformation mechanism
was assumed to be balanced with an elastic material response, and a wet sintering
mechanism balanced with a viscous material response. This leads to a viscoelastic
prediction for the response of two particles, which was tested by measuring the
contact area for a range of acrylic particles.

4.4 The Problem with ParticleParticle Approach

While a model for sintering of two particles is relevant on the particle length scale,
the measurement that is of interest is not the area of contact between two particles
but the local volume fraction in the film. That is, to what extent have the particles
deformed to fill all available space? Instead of considering the force between two
particles, we are more interested in the local stresses (forces per unit area). This is
equivalent to saying that the stress-strain relation for a bed of sintering particles is
the quantity of most interest, and it happens to be the most easily measured as
well. The constitutive relation for the film is necessarily a volume average of
sintering particle pairs, and hence a model for the deformation of viscous or elastic
particles subject to surface tension and applied forces is needed. The final relation
needs to relate the local volume fraction of particles to the applied (capillary)
stress and surface tension.

4.4.1 Routh and Russel Film Deformation Model

There is considerable debate in the early literature on latex film formation. One
topic of debate is how to describe the mechanical response to a force (elastic,
viscous or viscoelastic) and whether particle pairs or arrays should be considered.
A second topic concerns the dominant mechanism for particle deformation (wet or
dry sintering, capillary, etc.). Within recent years, the issues have been resolved
with the development of a film deformation model (Routh and Russel 1999,
2001a, 2001b). The model considers viscoelastic particles in a bed. It also brings
together a description of the many possible mechanisms to show that each is
possible, depending on a few key parameters. A particle deformation model is
built to describe the deformation of a particle array. Parameter regimes are then
4.4 The Problem with ParticleParticle Approach 131

identified where each deformation mechanism applies. Here, we follow the


arguments in the original papers to derive the parameter map. A dimensional
analysis argument is also presented in Section 4.6 to explain how the parameter
map may be understood.

4.4.1.1 ParticleParticle Deformation

The Routh and Russel model for film deformation starts with a particle sintering
approach. A generalised viscoelastic model is used with a stress relaxation
modulus G(t) defining the polymer rheology. This function relates the stress
experienced by the polymer for given applied strain. The simplest way to deter-
mine the stress relaxation modulus is to apply a constant strain to a material and to
follow the stress as a function of time. With polymers, a large stress is seen at
early times, typically followed by a relaxation to a low value.
The constitutive equation is integrated over the volume of two sintering spheres
compressed by a force F and surface tension (applicable to interfaces with air or
water). To perform the integration, the particles are taken to remain spherical, as
in the Frenkel approach. The deformation is quantified in terms of a strain along
the particle centres R (analogous to 2/2 in the Frenkel approach). To lowest
order in strain, the sintering is defined by

t
2 F 2
+
R0 R0
2
d
( )
R = G (t t ') R2 dt '
dt '
(4.5)
0

For viscous particles, sintering under surface tension alone, (4.5) yields
R=t/2R0 which is of the same form as the Frenkel result (4.4). For elastic
particles under a compressive force, (4.5) yields R= (F/GR02)1/2 which differs
from the exact Hertz result (4.2) in the exponent being 1/2 rather than 2/3. (In this
comparison, R is related to in (4.2) as = R0 R.) This difference occurs because
of the simplifying assumption of the particles remaining spherical.

4.4.1.2 Integration to Film Deformation

Once a pair deformation model is obtained, it may be integrated over a suitable


volume to obtain the stress-strain relation for the film. Choosing a single particle
as the representative volume, the average stress may be expressed as the summa-
tion of all moments (Batchelor 1970). Hence, the stress may be related to the
forces compressing the particles. The strain along the lines of centres, R, is related
to the overall strain field . Hence, for a given strain in the film, we may deter-
mine the stress through the relaxation modulus. The simplest strain is one-
dimensional, and the resulting equation for the film compaction is
132 4 Particle Deformation

t
3 m 3m d 2
t + = G (t t ')
dt ' (4.6)
20 R0 56 dt '
0

This equation relates the stress at the top surface t and surface tension , to the
strain in the film and material response G(t).

4.4.1.3 Assumption of a Viscoelastic Fluid

Taking a particular mechanical response allows (4.6) to be solved and hence the
deformation to be predicted as a function of time. The mechanical response
chosen is a simple viscoelastic fluid with a single relaxation time

G' t
G (t ) = G' exp (4.7)

0

where G' is the high frequency modulus and 0 is the zero shear-rate viscosity.
Substituting (4.7) into (4.6) allows derivation of a non-dimensional differential
equation governing the compaction of a particle bed as

d t 7 d d
+ Gt + + G = G , (4.8)
dt 5 dt dt

where the dimensionless groups controlling the compaction, t , G , and are


defined in Table 4.1. The first two terms on the left-hand side relate to the stress
applied at the top surface of the film. The third term relates to the surface tension
causing the film to compact, and the term on the right-hand side describes the
mechanical response.

Table 4.1. Definition of dimensionless groups used to determine film compaction


Symbol Definition Physical Meaning

28 t R0 Stress at top surface (t)


t divided by capillary stress
3m wa (wa/R0)
Surface tension divided by

wa water-air surface tension

G' H Evaporation time ( H / E )


G divided by polymer relaxation
E
0 time ( 0 / G )
'

0 R0 E Time for film to compaction


(0R0/wa) divided by
wa H evaporation time ( H / E )
4.5 Deformation Maps 133

4.5 Deformation Maps

Using (4.8), it is possible to generate parameter maps that determine which


deformation mechanism applies in each situation. The final result is shown in Fig.
4.6, where four regimes are seen in the vertical direction on the map.

4.5.1 Wet Sintering

In the wet sintering mechanism, the particles remain in water. The deformation
must be quick enough so that no capillary pressure is established. Hence > t ,
and t = 0 . This condition defines a line in a graph of / versus G , with wet
sintering operating below the line.

4.5.2 Capillary Deformation

Here, the deformation is concurrent with evaporation, = t , and the capillary


pressure must be below the maximum value that can be achieved: t < max .
Choosing representative parameter values, max is around 20. Ignoring the effect
of wet sintering on particle deformation defines a second line on a graph of
versus G .

4.5.3 Dry Sintering

In dry sintering, the particles are assumed to not undergo significant deformation
until the water has evaporated. The condition for dry sintering is that the deforma-
tion achieved by the capillary stress as the water evaporates is minimal. Setting the
capillary stress at its maximum value, and ensuring small deformation ( < 0.054)
at the end of the evaporation, defines a line above which dry sintering applies.

4.5.4 Receding Water Front

In Fig. 4.6, there is a receding water front regime between the capillary deforma-
tion and dry sintering regimes. In this intermediate regime, significant deformation
of the particles occurs because of capillary deformation, but final compaction of
the film occurs by a dry sintering mechanism only after complete water evapora-
tion. Hence, this is an inhomogeneous regime, where a water front passes verti-
134 4 Particle Deformation

cally through the film. Capillary deformation occurs below the meniscus of the
water front, and dry sintering occurs in the particles above the front. In experi-
ments on latices at temperatures not far above their Tg, it has been found that air
voids between particles are present for a short time before decreasing in size and
disappearing entirely (Keddie et al. 1995). Although it was first proposed as an
additional stage in the film formation process (II*), the observed phenomenon is
an example of a receding water front mechanism.

Dry/moist sintering
R E
= 0 0
wa H

Receding water front

Capillary deformation

Wet Sintering (pw/wa=0.2)

G' H
G=
E 0

Fig. 4.6 Deformation regimes for different values of the key parameters. (Reproduced with
permission from Routh and Russel 1999; copyright 1999 American Chemical Society)

4.5.5 Use of the Deformation Maps

These three lines (separating four regimes) are shown in Fig. 4.6. The map demon-
strates how the dimensionless group is crucial in determining the deformation
mechanism. For any given experimental conditions, the value of may in principle
be calculated. An engineer may use the maps to investigate how changing a parame-
ter, such as the evaporation rate, film thickness or particle size, will influence the
particle deformation mechanism. Section 4.7 will consider the effect of temperature,
which has a pronounced effect because it can change the polymer viscosity by
several orders of magnitude. Other parameters, such as the surface tension, may
typically be adjusted only in real situations by a factor of two or three. The extent of
particle deformation will determine the final film morphology.
4.6 Dimensional Argument for Figure 4.6 135

4.6 Dimensional Argument for Figure 4.6

A way to view Fig. 4.6 is as a process where each possible driving force is
sampled sequentially. The condition for wet sintering is that the particles deform
faster than evaporation reduces the water level in the vertical direction, thereby
keeping the particles wet. For capillary deformation, the stress at the top surface
must be less than the maximum level that can be supported by the particle bed. For
dry sintering, we insist that the particles are essentially undeformed by the time
that the water has completely evaporated. The particle deformation rate must be
slow, relative to evaporation rate. These basic concepts may be used to derive
simple scaling predictions to aid in the understanding of the deformation maps.
The symbols and concepts follow on from Section 4.5.

4.6.1 Wet Sintering

In the case where particles are consolidated at close packing in the wet state, they
will be deformed because of the polymer-water surface tension, pw, in a wet
sintering mechanism. Equally, evaporation will continue at the same rate. For the
particles to remain under water, the film must decrease in height faster than
evaporation reduces the water height. The time for film compaction under a wet
sintering mechanism may be estimated as 0R0/pw and the time for evaporation is
H / E . Therefore the requirement for wet sintering is

0 R0 E
H pw < 1,

which follows from the evaporation time being greater than the sintering time.
Assuming a value of the group pw/wa to be 0.2 gives a requirement of hav-
ing a value around 1 at the transition to the wet sintering regime. This is shown
mathematically as ~O(1) or being on the order of 1, in the upper range of
the wet sintering regime.

4.6.2 Capillary Deformation

If capillary pressure is created at the top surface of a film, the height of the film
will decrease at a rate set by the evaporation rate. Therefore, the shear stress in the
film is given by the polymer viscosity multiplied by the shear rate set by this
vertical deformation, or

0 E .
H
136 4 Particle Deformation

This shear stress will balance the capillary pressure, given approximately by
wa/R0 at equilibrium. Deformation will occur when the capillary pressure is
greater than the shear stress. Thus, the condition for capillary deformation
becomes

0 E < A wa/R0
H

where A is a scaling parameter. Upon rearrangement, this becomes <A. The full
model, detailed above, provides the limit as <100.

4.6.3 Dry Sintering

For dry sintering, the argument is similar to the wet sintering line, i.e., merely
replacing polymer-water surface tension with the polymer-air value, pa. The
requirement is that when the water has completely evaporated, the deformation in
the film must be small. The deformation, as water leaves the film, is by a capillary
mechanism with the capillary pressure being at its maximum value. Equation (4.8)
with set to zero and t set to a constant provides

d
t = .
dt

Solving this yields

2 t t
2 =

and hence setting the film strain as less than = 0.05 (an arbitrarily chosen small
value) at the time of complete evaporation ( t = 0.36 assuming an initial solids
volume fraction of 64%) provides a condition for dry sintering of >104.

4.6.4 Sheetz Deformation

The model outlined above applies to a bed of particles that is uniform vertically.
This initial assumption ignores the possibility of skin formation and hence the
Sheetz deformation mechanism. To be able to predict this mechanism, it is necessary
to consider the evaporation step. A vertical inhomogeneity in water concentration
must be present for skin formation. As discussed in Chapter 3 (Section 3.4.1), this is
the predicted case for a large Peclet number, Pe. Once particles have packed
together at the films top surface, with a fluid dispersion below, the only deforma-
4.7 Effect of Temperature 137

tion mechanism that can cause a skin to form is wet sintering. The fluid below the
compacting particles is at atmospheric pressure, and the upper packed layer is too
thin to generate a large capillary pressure. This is because the generated capillary
pressure is equal to the pressure drop across the top polymer layer due to solvent
evaporation. For a thin layer, this pressure drop is minimal. An important conclusion
from the preceding argument is that a skin can only form when <1 (wet sintering)
and for Pe > 1 (non-uniform drying). Because Pe and are controlling the defor-
mation process, it is important that the parameter map includes both variables. These
ideas lead to the two-dimensional parameter map shown in Fig. 4.7. The value of
G is taken as large, so that only the value of controls the deformation in the low
Peclet number case. Fig. 4.7 may be used in a fashion similar to 4.6. For a given set
of experimental conditions, the magnitude of the Peclet number and may be
estimated. This gives a position in the figure and an estimate for the deformation
mechanism that will be observed. If the conditions of film formation are operating in
the bottom right-hand corner of Fig. 4.7, then it is likely that a skin will form in the
film. For small values of Pe, the film is taken as vertically homogeneous and the
deformation map reverts to the case of Fig. 4.6.

Dry/moist
E R0 0
= sintering
wa H 104
Receding Water Front
2
10
101 Capillary
Partial skin E
100
Skinning slows
Wet Sintering Sheetz deformation
1 6R0 HE
Pe =
kT

Fig. 4.7 Deformation map considering the effects of vertical inhomogeneity as predicted by the
Peclet number. (Reproduced with permission from Routh and Russel (2001); copyright 2001
American Chemical Society)

4.7 Effect of Temperature

The parameter controls the deformation mechanism that applies in the film.
For <1, wet sintering is dominant, whereas for 1< <102, capillary deformation
applies, and the rate of film formation is set by the evaporation rate. For >104
the film forms by dry sintering alone, and for 102< <104 the film forms by a
138 4 Particle Deformation

mixture of capillary deformation, as a water front recedes through the film; dry
sintering occurs once the evaporation is complete. The definition of

0 R0 E
=
wa H

includes a number of parameters that depend on temperature. The evaporation rate


will approximately double with a 10C temperature increment, and surface tension
will vary by a factor of around 50% over extremes of temperature. However, the
overwhelmingly dominant effect is the change in the low shear viscosity, 0. At
the glass transition temperature, the low shear viscosity is around 1013 Pa s, and at
lower temperatures it is not measurable on laboratory time scales. For tempera-
tures well above the glass transition temperature, the polymer viscosity is orders
of magnitude lower, and the deformation mechanism will be by wet sintering. As
the temperature is lowered, the viscosity increases sharply before diverging at the
glass transition temperature (as in a Vogel-Fulcher equation). The range of
deformation mechanisms are observed in sequence, with dry sintering being seen
around 2C below the glass transition temperature.
A way to sample a range of temperatures is to use an MFFT bar. The photo-
graph in Fig. 4.8 shows a typical experiment from Sperry et al. (1994) where a
film is cast on the bar, which has been preheated to create a gradient to span
across the polymers glass transition temperature. The cloudy-clear transition is
clearly visible and is marked on the photo. The crack-point transition temperature
is invariant with time but is impossible to see in this picture, although its position
is marked.
For the cloudy-clear transition, a film is predicted to become clear when the
solids content approaches 1 (or the void size is far smaller than the wavelength of
light). Using a known value of the polymer glass transition temperature, film
thickness and particle size, the temperature of the cloudy-clear transition tempera-
ture may be predicted (Routh et al. 2001).
Although temperature is the easiest control parameter to ensure polymer malle-
ability, it is not always possible to control at the point of application. Hence, it is
common to add coalescent that plasticises the polymer, reducing its glass transi-
tion temperature, and ensuring easier particle deformation. These volatile organic
compounds (VOCs) subsequently evaporate over a period of days, resulting in a
hard, rigid polymer film. Even a glassy polymer, such as poly(styrene) with a Tg
of 100C, can be film-formed at room temperature with enough added plasticiser.
The Tg of a plasticised film may be estimated from the concentration of the
plasticiser using standard models (Toussaint et al. 1997). Environmental legisla-
tion requires the reduction of the amount of VOCs present in coatings. Under-
standing and controlling the particle deformation step is crucial in achieving
optimum formulations.
The models presented here do not consider explicitly the effects of particle
packing on the particle deformation stage. Chapter 7 considers how size polydis-
4.8 Effect of Particle Size 139

persity affects the density of particle packing. Furthermore, the models do not
consider heterogeneities in the particles, such as a core-shell particle or particles
with an inorganic inclusion. Particle deformation in these more complicated
systems presents a topic for future modelling and development of theory.

Cloudy-clear point Crack point

Cold Hot

Film previously dried at


low temperatures

Film dried from wet


with bar at temperature

Temperature Tg 4 oC T g + 4 oC Tg + 14 oC
Predicted Dry/moist Receding Capillary Wet
deformation sintering deformation sintering
water front
mechanism

Fig. 4.8 Photograph of an MFFT bar showing the approximate temperature ranges and the
deformation mechanisms predicted in the ranges. (Reproduced with permission from Sperry et
al. 1994; copyright 1994 American Chemical Society.

4.8 Effect of Particle Size

While temperature has the most dramatic effect on determining a film formation
mechanism, the effect of particle size is important and will often be used as a
design parameter in industrial applications. The controlling group,

0 R0 E
= ,
wa H

has the particle size appearing in the numerator. Reducing the particle size has the
same effect as decreasing the low shear viscosity (achieved by increasing the
temperature). This is equivalent to saying that film formation is easier with
smaller particles, leading to a downward shift on the deformation map. Latex that
undergoes dry sintering will be subject to capillary deformation when the particle
size is reduced sufficiently.
Particle size effects are well known and may be ascribed to the magnitude of
the driving pressure scaling as the inverse of particle size. For example, the
magnitude of the capillary pressure is /R0 and so is increased for smaller parti-
cles. In sintering mechanisms, the initial ratio of surface area to volume varies as
1/R0, such that there is a greater surface energy (per unit mass of material)
associated with smaller particles. Size effects have been examined experimentally
140 4 Particle Deformation

by Jensen and Morgan (1991), where the minimum film formation temperature
was followed using particles of identical chemistry but different diameters. As
expected, smaller particles (in this case, of diameter 63 nm) formed films at
temperatures 10C below larger particles (diameter of 458 nm). However, the
particle size also affects the optical transparency, and so MFFT data must be
treated with care (van Tent and te Nijenhuis 2000), as discussed in Section 2.1.1.1.

4.9 Experimental Evidence for Deformation Mechanisms

There are a number of possible techniques to determine experimentally which


deformation mechanism is operational in a film. These are discussed in detail in
Chapter 2 and are briefly reviewed here in relation to how they may be used to
identify the operative mechanism.

4.9.1 Inferring Deformation Mechanisms from Water Distributions

Magnetic resonance profiling can detect the local water concentration with a
spatial resolution of less than 10 m. The various deformation mechanisms are
expected to have the different profiles shown in Fig. 4.9, and many of these have
been observed experimentally. A fluid film will display a decreasing height, due to
evaporation, until the particles come into contact. At this point, the profiles will
depend upon the deformation mechanism. Fig. 4.9A sketches the expected profile
for dry sintering. After the particles have packed together, the film remains near
this height (decreasing gradually with particle deformation), but the water level
decreases below the film surface (defined by the top of the packed bed of parti-
cles). The water content in the wet regions of the film remains constant and is set
by the available space between particles. Fig. 4.9B shows the idealised water
content for wet sintering and capillary deformation. The water content in the film
remains spatially uniform and decreases over time simultaneously with the film
height. The difference between capillary deformation and wet sintering is in their
time-dependence. For capillary deformation, the film height will decrease at
exactly the rate of evaporation; as water leaves by evaporation, the particles
deform to fill the space that is vacated by the water. The water will remain through
the depth of the film until the completion of drying. On the other hand, for wet
sintering, the deformation rate must be faster than evaporation, but this inequality
does not affect the water profile. Fig. 4.9C shows the inhomogeneous case in
which a skin is forming. The water content in the lowest part of the film is
constant with depth, and the skin simply grows in thickness with increasing time.
4.9 Experimental Evidence for Deformation Mechanisms 141

A
B
Water content Water content

time

time

Hfilm
Vertical distance Vertical distance

C
Water content

time

Vertical distance

Fig. 4.9 Idealised water distribution profiles for A dry sintering (where Hfilm represents the
thickness of the packed particle bed); B wet sintering and capillary deformation; and C skin
formation.

The experimental data in Fig. 4.10 shows a number of the idealised features.
Profiles in A show the water height decreasing uniformly and receding below the
polymer film surface. This pattern is indicative of dry sintering. Profiles in B show
the water content increasing in the lower regions of the film and the water
becoming pinned at the top surface of the film. The film then decreases its water
content while the water remains at the top of the film. This pattern is indicative of
capillary deformation. Profiles in C show a non-uniform drying profile. But, in
this case, there is evidence for a surface layer with very low water content (< 2%),
and the rate of water loss is severely slowed. This series of profiles is consistent
with a mechanism of skin formation.
142 4 Particle Deformation

A1
A2

Substrate

0.7 B
7'
0.6
Air
Relative intensity

0.5 2'
0.4 5'

0.3
0.2
0.1
0.0
-25 0 25 50 75 100 125
Height (m)
0.16
0.14 C 31 minutes
54 minutes
0.12
Water fraction

94 minutes
0.1 139 minutes
0.08 194 minutes
264 minutes
0.06
365 minutes
0.04
0.02
0
0 25 50 75 100 125 150 175 200 225 250 275
(m)
Height (m)
Vertical distance

Fig. 4.10 Typical GARField profiles obtained from drying latex films. The NMR intensity
(vertical axis) is proportional to the water concentration. A1 One horizontal axis shows the
height in the film and the other shows time ranging from 0 to 90 min. A2 The time-dependence
of the water level thickness for this data set, showing a constant rate of decrease. The water
height decreases linearly and drops below the film thickness, which is estimated from the solids
content to be about 200 m. This result is indicative of dry sintering, which is expected for this
latex, which has a glass transition temperature that is very near to the temperature of the room.
Data courtesy of Andr Utgenannt. B This series of profiles shows an initial drop in water height
until the point of particle packing (at about 7 min.), followed by a very gradual decrease. During
the later stages of drying, there is evidence that some water stays at the film surface, albeit with
more water near the substrate. This trend is indicative of capillary deformation. The numbers
(with primes) refer to the time, in minutes, after casting of the film. The polymers Tg (45 C is
far below room temperature, so wet sintering is expected. A NMR signal is obtained from the
polymer in this case. Data reproduced with permission from Mallgol et al (2002), Copyright
American Chemical Society C These profiles show skin formation. There is a step in the water
content, with minimal water in the upper skin layer and more than 10% below it. Note how
drying takes almost six hours. (Data reproduced with permission from Gorce et al. (2002))
4.9 Experimental Evidence for Deformation Mechanisms 143

4.9.2 Determination of Deformation Mechanisms Using an MFFT Bar and


Optical Techniques

Film formation on an MFFT bar can provide indirect evidence for dry sintering.
Sperry et al. (1994) dried a number of films for this purpose, and a representa-
tive sample is shown in Fig. 4.8. The cloudy-to-clear transition is easily observ-
able and is found to move to lower temperatures with time. The authors initially
dried a film at low temperatures, thus ensuring no particle deformation, and
obtained a cloudy, cracked film. The bar was then placed on the MFFT bar and
brought up to the range of test temperatures; a second film was applied in its wet
state. The observation is that the cloudy-clear transition occurs at the same
temperature for the wet film and the one previously dried. This implies that
water is not present in the film at the cloudy-clear transition, meaning that dry
sintering must be the relevant deformation. Sperry et al. went further to explain
the dynamic data with a model for bubbles decreasing in size in a viscous liquid,
driven by surface tension.
The dry sintering mechanism requires the creation of interparticle voids after
the water has evaporated. Optical techniques are able to identify the onset of air
void formation and to observe how the void size decreases as dry sintering
proceeds. The apparent refractive index of a film drops when air voids appear
(Keddie et al. 1995). (The refractive index of air is 1.0) A drop in the refractive
index followed by a rise (referred to as stage II*) may be explained by a
receding water front followed by a dry sintering mechanism. Similarly, a film
with air voids, undergoing dry sintering, has a lower optical transparency and
may show interparticle interference. As particles deform more over time, the
transparency increases and the interference disappears (van Tent and te Nijen-
huis 2000).
Optical transmission measurements are supportive of a mechanism of wet
sintering in latex films that are formed at a temperature far above their MFFT (van
Tent and te Nijenhuis 2000). The changes in transmission can be correlated with
the extent of particle deformation in the presence of water. In such a case, the
refractive index (measured by ellipsometry) rises steadily during drying, as the
polymer fraction increases as the water fraction decreases (Keddie et al. 1995,
1996b). The evaporation rate has been found to be the rate-limiting step in
determining the optical changes, as would be expected from a capillary deforma-
tion process (Keddie et al. 1996b).

4.9.3 Microscopy of Particle Deformation

Direct imaging of the film surface provides information on the extent of particle
deformation. Transmission electron microscopy of film cross-sections and surface
replicas (as in Fig. 4.1) provides undisputable evidence that latex particles create
144 4 Particle Deformation

polyhedral shapes when they deform to fill all available space. There are several
examples of film cross-sections that show the expected hexagonal honey comb
structure (Rharbi et al. 1996). Most previous electron microscopy studies are
performed on dry films after particle deformation, and hence they do not provide
evidence for the mechanism of the deformation.
The top surface of a film is easily imaged using atomic force microscopy
(AFM) to provide quantitative information on particle structure. The technique
has been applied to dry latex films by many groups (Park et al. 1998, Lee et al.
1999, Goh et al. 1993, Joanicot et al. 1997, Gerharz et al. 1997, Gilicinski and
Hegedus 1997, Butt and Kuropka 1995, Wang et al. 1992). The extent to which
the particles are no longer spherical and are becoming polyhedral can be found.
Of course, the top surface may not always be representative of the bulk. Micro-
toming and cross-sectioning methods can be used to obtain an exposed cross-
section of a film to provide a guide of the vertical homogeneity and structure in
the bulk.
In dry films, AFM can be used to measure the vertical distance between the top
of the particles and the valley region at their point of contact. Fig. 4.11 shows how
a peak-to-valley distance is defined and obtained from an AFM image. In a latex
film, at a temperature above the polymers glass transition temperature, the peak-
to-valley height decreases over time. The rate of flattening increases as the
annealing temperature increases and the polymer viscosity (which resists polymer
flow) drops (Goudy et al. 1995). As in the case of dry sintering, the driving force
is the reduction of surface area; a flat surface has a lower area than a corrugated
one. Models of sintering have been adapted to describe the time and temperature
dependence of the peak-to-valley height (Goh et al. 1993) and surface roughness
(Perez and Lang 1999) with good agreement with experimental results. Studies of
particle deformation in the presence of varying humidity levels provide data to
support arguments for particle deformation under capillary forces in pendular
rings (Lin and Meier 1995, 1996).
Whereas, AFM is usually applied only to dry or damp surfaces, an environ-
mental SEM can provide images of films during the deformation process with
varying levels of water content (Keddie et al. 1995). Fig. 4.12 shows an example
of typical results. The loss of distinct particles is easily followed. The fact that
particle deformation can be observed prior to the loss of water rules out a dry
sintering mechanism and provides evidence for either wet sintering or capillary
deformation. Scanning transmission electron microscopy may be applied to dilute
dispersions within an environmental chamber. The technique is able to observe the
particle deformation in an aqueous dispersion that is characteristic of wet sintering
(Bogner et al. 2005).
4.9 Experimental Evidence for Deformation Mechanisms 145

160
Vertical height (nm)

140

120

100

80

60

40

20

0
0 200 400 600 800 1000 1200

Lateral distance (nm)

Fig. 4.11 Measurements of topographical cross-sectional profiles of particles in a dry latex film
obtained with atomic force microscopy. The profile of a hard acrylic latex (with a glass
transition temperature of 40 C with particle size of 420 nm after drying at room temperature is
presented (solid line). The peak-to-valley height is identified by the double-headed arrow on the
profile and on the sketch above it. For comparison, the profile obtained from the same latex after
film formation with the addition of a plasticizer is shown as the dashed line. The plasticised
polymers glass transition temperature is below room temperature, and the peak-to-valley height
has decreased. (Data courtesy of A. Georgiadis, University of Surrey)

a b
c c

Fig. 4.12 Environmental SEM image of a film forming latex a at close packing before significant
deformation and b after particle deformation has occurred. Scale bar is 2 m in both images.
(Reproduced from Keddie et al. 1995 with permission; copyright 1995 American Chemical
Society)
146 4 Particle Deformation

4.9.4 Scattering Techniques

Scattering techniques have the advantage of being applicable to wet samples. As


such, they can be used to probe particle deformation by wet sintering and capillary
mechanisms. In films made from emulsions, the flattening of boundaries between
droplets essentially a wet sintering mechanism has been examined with small
angle neutron scattering (SANS, presented in Section 2.1.3.1). Estimates of the
particle curvature in the Plateau borders were obtained (Crowley et al. 1992).
Hydrophilic membranes in particles have been found in SANS experiments to
prevent direct contact initially between packed and deformed particles (Chevalier
et al. 1992). Labelled surfactants can be used in SANS experiments to determine
the interrelation between surfactant layers and particle deformation. There is
evidence that particle deformation acts as a trigger for surfactant desorption
(Belaroui et al. 2003).

4.9.5 Detection of Skin Formation

The formation of a solid layer at the top of a film is a common occurrence in the
drying process of a waterborne paint. To quantify the evolution of skin layers over
time, magnetic resonance profiling has been employed to measure the water
distribution, as discussed in Section 4.9.1. IR microscopy can also be used to
probe the growth of coalesced surface layers, provided layers are thick enough to
overcome the restrictions of the techniques resolution (Guigner et al. 2001). In
some experiments, the amount of particle deformation and connectivity in a skin
layer may be inferred from water concentration profiles in combination with
measurements of particle motion (Knig et al. 2008). An indirect indication of the
extent of particle coalescence in a skin layer is through the amount by which the
water loss rate is decreased. Additionally, physical probes of a film surface may
provide an indication of the integrity and mechanical properties of a surface skin
layer and distinguish it from a still fluid surface. A thin film analyser, described in
Chapter 2 (Section 2.1.1.2), can provide such measurements. Quite often, a skin
layer can even be lifted off by hand or with the use of a needle!

References

Batchelor G.K. (1970) The stress system in a suspension of force-free particles.


Journal of Fluid Mechanics 41: 545-570.
Bellehumeur C.T., Kontopoulou M. and Vlachopoulos J. (1998) The role of
viscoelasticity in polymer sintering. Rheological Acta 37:270-278.
References 147

Belaroui F., Cabane B., Dorget M., Grohens Y., Marie P., Holl Y. (2003) Small-
angle neutron scattering study of particle coalescence and SDS. desorption
during film formation from carboxylated acrylic latices. J Coll Interf Sci 262:
409-417
Bogner A., Thollet G., Basset D., Jouneau P.-H., Gauthier C. (2005) Wet STEM:
A new development in environmental SEM for imaging nano-objects included
in a liquid phase. Ultramicroscopy 104: 290-301
Brown G.L. (1956) Formation of films from polymer dispersions, Journal of
Polymer Science, 22: 423-434.
Butt H.J. and Kuropka R. (1995) Surface structure of latex films, varnishes and
paint films studied with an atomic force microscope. Journal of Coatings Tech-
nology 67(848): 101-107.
Chevalier Y., Pichot C., Graillat C., Joanicot M., Wong K., Maquet J., Lindner P.
and Cabane B. (1992) Film formation with latex particles. Colloid and Polymer
Science 270: 806-821.
Crowley T.L., Sanderson A.R., Morrison J.D., Barry M.D., Morton-Jones A.J.,
Rennie A.R. (1992) Formation of bilayers and Plateau borders during the dry-
ing of film-forming latexes as investigated by small-angle neutron scattering.
Langmuir 8: 2110-2123
Derjaguin B.V., Muller V.M. and Toporov Y.T. (1975) Effect of contact deforma-
tions on the adhesion of particles. Journal of Colloid and Interface Science,
53(2): 314-326.
Dillon R.E., Matheson L.A.and Bradford E.B. (1951) Sintering of synthetic latex
particles. Journal of Colloid Science 6:108-117.
Dobler F., Pith T., Holl Y., Lambla M. (1992) Synthesis of model lattices for the
study of coalescence mechanism. Journal of Applied Polymer Science 44:1075-
1086.
Eckersley S.T. and Rudin A. (1990) Mechanism of Film Formation from Polymer
latexes. Journal of Coatings Technology, 62(780):89-100.
Eckersley S.T. and Rudin A. (1993) The effect of plasticization and pH. on film
formation of acrylic latexes, Journal of Applied Polymer Science, 48:1369-
1381.
Eckersley S.T. and Rudin A. (1994) The film formation of acrylic latexes: A
comprehensive model of film coalescence, Journal of Applied Polymer Sci-
ence, 53:1139-1147.
Fogden A. and White L.R. (1990) Contact elasticity in the presence of capillary
condensation 1. the nonadhesive Hertz problem, Journal of Colloid and Inter-
face Science 138(2):414-430
Frenkel J. (1945) Viscous flow of crystalline bodies under the action of surface
tension, Journal of Physics, 9(5):385-391.
Gerharz B., Kuropka R., Petri H. and Butt J.J. (1997) Investigation of latex
particle morphology and surface structure of the corresponding coatings by
atomic force microscopy, Progress in Organic Coatings 32: 75-80.
148 4 Particle Deformation

Gilicinski A.G. and Hegedus C.R. (1997) New applications in studies of water-
borne coatings by atomic force microscopy, Progress in Organic Coatings 32:
81-88.
Goh M.C., Juhu D., Leung O.M., Wang Y.C., Winnik M.A. (1993) Annealing
effects on the surface structure of latex films studies by atomic force micros-
copy. Langmuir 9: 1319-1322
Goudy A., Gee M.L., Biggs S., Underwood S. (1995) Atomic-force microscopy
study of polystyrene latex film morphology Effects of aging and annealing.
Langmuir 11: 4454-4459
Gorce J.-P., Bovey D., McDonald P.J., Palasz P., Taylor D. and Keddie J.L.
(2002) Vertical water distribution during drying of polymer films cast from
aqueous emulsions, European Physics Journal E., (8): 421-429.
Greenwood J.A. and Johnson K.L. (1981) The mechanics of adhesion of viscoe-
lastic solids, Philosophical Magazine A 43(3):697-711.
Guigner D., Fischer C., Holl Y. (2001) Film formation from concentrated reactive
silicone emulsions. 1. Drying mechanism. Langmuir 17: 3598-3606
Hertz H. (1881) Reine Angew Math 92:156; English translation in Hertz, H.,
1896, Miscellaneous Papers by Heinrich Hertz, MacMillan & Co, London, UK,
146183.
Jagota A., Argento C.and Mazur S. (1998) Growth of adhesive contacts for
Maxwell viscoelastic spheres. Journal of Applied Physics 83(1): 250-259.
Jensen D.P. and Morgan L.W. (1991) Particle size as it relates to the minimum
film formation temperature of latices, Journal of Applied Polymer Science 42:
2845-2849.
Joanicot M., Granier V.and Wong K. (1997) Structure of polymer within the
coating: an atomic force microscopy and small angle neutrons scattering study,
Progress in Organic Coatings 32: 109-118.
Johnson K.L., Kendall K.and Roberts A.D. (1971) Surface energy and the contact
of elastic solids, Proceeding of the Royal Society of London A 324:301-313.
Keddie J.L., Meredith P., Jones R.A.L. and Donald A.M. (1995) Kinetics of film
formation in acrylic lattices studied with multiple angle of incidence ellipsome-
try and environmental SEM, Macromolecules 28: 2673-2682.
Keddie J.L., Meredith P., Jones R.A.L. and. Donald A.M. (1996) Film formation
of acrylic latices with varying concentrations of non-film forming latex parti-
cles, Langmuir 12: 3793-3801.
Keddie J.L., Meredith P., Jones R.A.L. and Donald A.M. (1996b) Rate-limiting
steps in film formation of acrylic latices as elucidated with ellipsometry and
environmental scanning electron microscopy. ACS Symp Ser: Film Formation
in Waterborne Coatings 648: 332-348
Knig, A.M., Weerakkody, T.G., Keddie, J.L., and Johannsmann, D. (2008)
Heterogeneous Drying of Colloidal Polymer Films: Dependence on Added
Salt. Langmuir 24: 75807589.
Kuczynski G.C., Neuville B. and Toner H.P. (1970) Study of sintering of
poly(methyl methacrylate), Journal of Applied Polymer Science 14:2069-2077.
References 149

Lee E.H. and Radok J.R.M. (1960) The contact problem for viscoelastic bodies,
Trans. ASME Journal of Applied Mechanics, 27: 438-444.
Lee D.Y., Choi H.Y., Park Y.J., Khew M.C., Ho C.C., Kim J.H. (1999) Kinetics
of film formation of poly(n-butyl methacrylate) latex in the presence of
poly(styrene/-methylstyrene/acrylic acid) by atomic force microscopy, Lang-
muir 15(23):8252-8258.
Lin F. and Meier D.J. (1995) A study of latex film formation by atomic force
microscopy. 1. a comparison of wet and dry conditions, Langmuir 11:2726-
2733.
Lin F. and Meier D.J. (1996) A study of latex film formation by atomic force
microscopy. 2. film formation vs rheological properties: Theory and experi-
ment, Langmuir 12:2774-2780.
Ma Y., Davis H.T. and Scriven L.E. (2005) Microstructure development in drying
latex coatings. Progress in Organic Coatings, 52 (1): 46-62.
Mallgol J., Gorce J.P., Dupont O., Jeynes C., McDonald P.J. and Keddie J.L.
(2002) Origins and effects of a surfactant excess near the surface of waterborne
acrylic pressure-sensitive adhesives. Langmuir 18: 4478-4487.
Mason G. (1973) Formation of films from lattices a theoretical treatment, British
Polymer Journal, 5: 101-108.
Mazur S., Beckerbauer R. and Buckholz J. (1997) Particle size limits for sintering
polymer colloids without viscous flow. Langmuir 13:4287-4294.
Mazur S. and Plazek D.J. (1994) Viscoelastic effects in the coalescence of
polymer particles. Progress in Organic Coatings 24:225-236.
Mizutani T., Arai K., Miyamoto M. and Kimura Y. (2006) Preparation of spheri-
cal nanocomposites consisting of silica core and polyacrylate shell by emulsion
polymerization. Journal of Applied Polymer Science 99: 659-669.
Muller V.M., Yushchenko V.S. and Derjaguin B.V. (1983) General theoretical
consideration of the influence of surface forces on contact deformation and the
reciprocal adhesion of elastic spherical particles. Journal of Colloid and Inter-
face Science 92(1): 92-101.
Muller V.M., Yushchenko V.S. and Derjaguin B.V. (1980) On the influence of
molecular forces on the deformation of an elastic sphere to a rigid plane. Jour-
nal of Colloid and Interface Science 77(1): 91-101.
Park Y.J., Khew M.C., Ho C.C. and Kim J.H. (1998) Kinetics of latex formation
of PBMA latex in the presence of alkali soluble resin using atomic force mi-
croscopy. Colloid and Polymer Science 276: 709-714.
Perez, E. and Lang, J. (1999) Flattening of latex film surface: theory and experi-
ments by atomic force microscopy. Macromolecules 32: 1626-1636.
Pokluda O., Bellehumeur C.T. and Vlachopoulos J. (1997) Modification of
Frenkels model for sintering. AIChE Journal 43(12): 3253-3256.
Rharbi Y., Boue F., Joanicot M., Cabane B. (1996) Deformation of cellular
polymeric films. Macromolecules 29: 4346 - 4359.
Rosenzweig N. and Narkis M. (1981) Dimensional variations of two spherical
polymeric particles during sintering. Polymer engineering and science
21(10):582-585.
150 4 Particle Deformation

Routh A.F. and Russel W.B. (1999) A process model for latex film formation:
Limiting regimes for individual driving forces. Langmuir 15(22):7762-7773.
Routh A.F. and Russel W.B. (2001a) A process model for latex film formation:
Limiting regimes for individual driving forces. Langmuir 17:7446-7447.
Routh A.F. and Russel W.B. (2001b) A process model for latex film formation:
Experimental evidence, Industrial & Engineering Chemistry Research 40 (20)
4302-4308.
Routh, A.F., Russel, W.B., Tang, J.S., El-Aasser, M.S. (2001) Process model for
latex film formation: Optical clarity fronts. Journal of Coatings Technology 73:
41-48.
Russel W.B. (1990) On the dynamics of the disorder-order transition. Phase
Transitions 21:127-137.
Russel W.B., Saville D.A. and Schowalter W.R. (1990) Colloidal Dispersions,
Cambridge University Press.
Sheetz D.P. (1965) Formation of films by drying of latex. Journal of Applied
Polymer Science, 9: 3759-3773.
Sperry P.R., Snyder B.S., ODowd M.L., Lesko P.M. (1994) Role of water in
particle deformation and compaction in latex film formation. Langmuir 10:
2619-2628.
Toussaint A., DeWilde M., Molenaar F., Mulvihill J. (1997) Calculation of Tg and
MFFT depression due to added coalescing agents. Prog Organ. Coatings 30:
179-184.
Tzitzinou A., Keddie J.L., Geurts J.M., Peters A.C.I.A. and Satguru R. (2000)
Film formation of latex blends with bimodal particle size distributions: Consid-
eration of particle deformability and continuity of the dispersed phase. Macro-
molecules 33(7): 2695-2708.
Vanderhoff J.W., Tarkowski H.L., Jenkins M.C., Bradford E.B. (1966) Theoreti-
cal Consideration of the interfacial forces involved in the coalescence of latex
particles, Journal of Macromolecular Chemistry, 1(2): 361-397.
Wang Y., Juhu D., Winnik M.A., Leung O.M. and Goh M.C. (1992) Atomic
force microscopy study of latex film formation, Langmuir 8: 760-762.
Wang Y.C., Kats A., Juhu D., Winnik M.A., Shivers R.R., Dinsdale C.J. (1992)
Freeze-fracture studies of latex films formed in the absence and presence of
surfactant, Langmuir 8: 1435-1442.
Chapter 5
5 Molecular Diffusion Across Particle Boundaries

In the previous chapters, we have reviewed how particles pack together during the
drying process and why they become deformed from their initial spherical shape.
Yet, the film formation process does not stop there: an important third stage
follows.
In 1958, only a few years after the theories of particle deformation were first
proposed (as outlined in Chapter 4), Voyutski published his reaction to this
previous work. His short paper contained some impressive insights but did not
have quantitative experiments to support them. His ideas foreshadowed many of
the future experimental developments that will be presented in this chapter.
Voyutski realised that particle deformation was not the final required step in
latex film formation. He wrote that to obtain a stable film it is necessary that the
segments of chain polymers diffuse from one globule to another, thus forming a
stable link. The importance of diffusion was thus highlighted. It is worth noting
that by 1958 the first measurement of a polymer self-diffusion coefficient had
been reported previously for a bulk material (Bueche, Cashin and Debye 1952).
But, there were no measurements in colloidal systems. In this chapter, we shall
consider the details of the crossing of polymer molecules from one particle to
another and shall see that, somewhat contrary to Voyutskis view, it takes more
than a few segments to achieve a stable link.
Calling the process autohesion, Voyutski went on to comment on its time
dependence, saying that the final strength is not achieved until some length of
time which is necessary for the diffusion of polymer molecules. In fact, the study
of the dynamics of diffusion would occupy many of the most eminent polymer
physicists in the later decades of the twentieth century.
A fundamental observation is that the mechanical strength of latex films is
found to increase over time. In some of the early literature, the final stage of film
formation was therefore called further gradual coalescence in recognition of the
fact that the process was slow (Vanderhoff 1966). The word coalescence refers
here to the loss of particle boundaries to create a homogenous film, as was
observed with electron microscopy in early studies. It is now fully accepted that
152 5 Molecular Diffusion Across Particle Boundaries

the strengthening of films is the result of interdiffusion at the molecular level


across the boundaries between particles as illustrated in Fig. 5.1. We use the word
interdiffusion rather than merely diffusion to indicate that molecules are
travelling between particles rather than only within them.
Readers who have been introduced to modern soft matter physics will take the
concepts of interdiffusion and interface broadening for granted. It is readily
apparent why adhesion will develop between the surfaces of polymers above their
glass transition temperature, Tg, when placed into contact. At the molecular level,
the narrow gap between the initial surfaces vanishes as molecules reach across.
The welding of plastic laminates and the healing of cracks through heating are just
two examples of familiar processes that rely on such a diffusion process.
However, for many years after Voyutskis first comments, the concept of
autohesion as a third stage of latex film formation was not universally accepted
(Voyutski and Ustinova 1977). One reason for scepticism amongst scientists was
because of reports of latex films in which particle identity was retained even after
aging for up to two years at room temperature or after heating up to 150C (Distler
and Kanig 1978). As we shall see later in this chapter, there are good explanations
for these confusing results.
An appreciation of the details at the molecular level rather than just the mac-
roscopic has emerged from the theory of reptation. In this way, the subject has
been shaped by some of the worlds greatest soft matter physicists. Our under-
standing of interdiffusion in latex films has been helped enormously by the many
SANS and FRET experiments that have tested the theories and confirmed the
mechanism of polymer diffusion in latex films. The principles of these techniques
were described in Chapter 2.

Fig. 5.1 Molecular level view (not to scale) of the interdiffusion of molecules across particle
boundaries in a latex film made up of flattened particles.
5.1 Essential Polymer Physics 153

5.1 Essential Polymer Physics

5.1.1 Interface Width at Polymer-Polymer Interfaces

Latex film formation has long been linked to the theory of crack healing and
polymer interface welding, as the basic physics is the same. In a crack healing
experiment, two fracture surfaces on either side of a crack are placed together and
heated above the polymer Tg. Welding experiments, in contrast, use two smooth
polymer surfaces. After heating for a sufficient time at a sufficient temperature,
the strength of the interface between the two sides is equivalent to the strength of
the bulk material.
At the end of the particle deformation stage, the boundaries between particles
are flattened, but at the molecular level there may not be an overlap. Consider the
planar interface between the deformed particles shown in Fig. 5.1. In blends of
identical particles, the polymers will be miscible, and the interface will broaden as
a result of diffusion. This is a process of mass transport enabled by the random
thermal motion of the centre-of-mass of the molecules. Diffusion requires motion
of the polymer backbone and not merely the side groups. If a label such as an
isotope or a fluorescent marker is put on the molecules in a particle, a concentra-
tion profile can be visualised to develop across the interface (Fig 5.2). Drawing a
tangent across the symmetric concentration profile and finding the inflection
points, as demonstrated in Fig. 5.2f, will define an interfacial width, wI. Over time,
wI will broaden as a result of diffusion. Note that this process requires molecular-
level contact, also known as wetting, between the two halves of the interface.
Initially, there may be an air gap or other phase at the boundaries.
In a bulk polymer over long time scales, the self-diffusion coefficient D is used
to predict the average distance that the centre-of-mass of a molecule travels, i.e.,
its root-mean square displacement <r2>1/2, over a time of travel, t in a matrix of
itself:

<r2>1/2 = (6Dt)1/2 (5.1)

The important feature is that the distance varies with the square-root of time,
such that when doubling the time of travel, the molecule only travels 2 = 1.41
times further. Such diffusion is referred to as Fickian, and we shall see that it
applies to the later stages of diffusion at interfaces. This time-dependence results
from the fact that diffusing molecules do not travel in a straight trajectory. Instead,
they follow a random path.
Occasionally, a latex film is made from blends of particles that have no or partial
miscibility. Whereas the kinetics mainly determines wI for miscible systems,
thermodynamics dictates wI for immiscible systems. The interface profile at equilib-
rium is described mathematically by a hyperbolic tangent function, and wI depends
only on the interaction parameter, , for the two polymer phases, the statistical
154 5 Molecular Diffusion Across Particle Boundaries

segment length, b, and the degree of polymerization, N, for the two phases (Broseta
et al. 1990). In the limit of infinite N, the equilibrium wI is given as

b
wI ~ . (5.2)

Here (and elsewhere) the symbol ~ represents a proportionality and not an


equality. Although immiscible polymers make useful systems with which to study
the relationship between interfacial width and adhesion energy (Schnell et al.
1998), most latex films are made from miscible polymers. Hence, time-dependent
interface broadening is of greater interest to us here.

(a) (b) (c)

(f) (d) (e)

pol

z
wI
wi
z
Fig. 5.2 Formation of a broadened interface in latex films. a Particle deformation creates a
flattened boundary between particles. b Any initial gap between particles is closed c as wetting is
achieved. d Diffusion of molecules over a distance of broadens the interface, and e the
interface continues to broaden over time. f An interfacial width, wI, can be defined from the
profile of the concentration, p, of one of the polymers across the interface.

5.1.2 Polymer Reptation

When polymer molecules are sufficiently long, they become intertwined and
entangled, so that there can be stress transfer between them. There is an analogy to
how strings of pearls or long segments of rope may become knotted. The many
5.1 Essential Polymer Physics 155

entanglement points within a polymer define the constraints that restrict the
motion of the chain-like molecules. In the process, known as reptation, polymer
chains move along their contours in a snake-like fashion, but they cannot move
laterally from this direction because of the effect of the surrounding network of
entanglements (Fig. 5.3). Indeed, the word reptation is derived from the Latin
word reptare, which refers to how a snake crawls along a surface. Only molecules
that are long enough to entangle, as measured by the entanglement molecular
weight, Me, are subject to reptation. The theory of polymer diffusion by reptation
was developed in seminal work (de Gennes 1971, 1989, Doi and Edwards 1986),
and only the salient points will be outlined here.

Fig. 5.3 A visualisation of a network of entanglements between long chain molecules (in gray).
The motion of the chain represented with the black line is confined by these entanglements so
that it can only move along in a snake-like fashion along its tube, indicated in white. (Reprinted
with permission from www.nobel.org; copyright The Royal Swedish Academy of Sciences)

One could imagine many ways for high molecular weight polymers to move
across an interface. The reptation model tells us that when a polymer chain crosses
an interface with another polymer, one of the chain ends will travel first, followed
naturally by its middle and then its tail end. Some elegant experiments, in which
the ends of the polymer chains were labelled to distinguish them from the middle
section, have provided direct evidence for this type of motion (Russell et al. 1993).
The topological constraints that are created by the network of entanglements
that surround each polymer chain in a glassy polymer are commonly referred to as
a tube. The chain must initially snake along the path that is defined by the tube
until it emerges from it. Following the work of Doi and Edwards (1986), we
divide the reptation process into four time regimes separated by three characteris-
tic times, as presented in Table 5.1.
We will consider each of the characteristic times, starting from the shortest. To
gain a feel of the orders of magnitude, Table 5.1 gives values for the times for
polystyrene (PS) with a molecular weight of M = 2.33 x 105 g mole1 and at a
temperature of 120C. It is important to remember that these times are dependent
on both the temperature and the molecular weight.
156 5 Molecular Diffusion Across Particle Boundaries

Table 5.1. Scaling predictions from reptation theory


Value for polystyrene with M
Time dependence of the r.m.s.
Time regime = 233,000 g mole1 and T =
segmental displacement
120 C
t < e <r2>1/2 ~ t1/4 e 0.08 min.
e < t < R <r2>1/2 ~ t1/8 R 11 min.
R < t < d <r2>1/2 ~ t1/4 d 1080 min
t > d <r2>1/2 ~ t1/2

The Rouse entanglement time, e, is the time over which the segmental diffu-
sion distance is comparable to the diameter of the confining tube. For our refer-
ence PS, the tube diameter is 5.7 nm (Karim et al. 1990). At times less than e, the
segmental motion is not affected by the presence of the entanglements. e tends to
be very short for most systems. It is only five seconds in our reference PS! Hence,
the first time regime is difficult to access experimentally, and it is not important
for interfacial strength development.
The Rouse relaxation time, R, refers to the time over which the motion of a
segment becomes correlated over an entire chain. In the time regime between e
and R, the segmental displacement distance is only weakly dependent on time,
being proportional to t to a power of 1/8.
A third characteristic time, the reptation time, d, is defined as the time re-
quired for the centre-of-mass of a molecule to move by one radius of gyration. In
travelling this distance, the chain disengages from its original confining tube.
At times beyond d, the displacement varies with t1/2, and the process is said to
be Fickian. Karim et al. (1990) demonstrated conclusively that the Fickian self-
diffusion coefficient obtained in the bulk polymer, DB, holds when t > d. At times
much less than d, they found D in PS to be almost five times greater than DB, and
it decreased over time until d was reached.
In the time regime between R and d, the segmental displacement distance is
proportional to t to a power of 1/4. Specifically, Doi and Edwards (1986) predict
that the root mean square displacement of the polymer segments is

<r2>1/2 = (6Nb2DBt)1/4. (5.3)

where all symbols have been previously defined. In high-molecular-weight


polymers near their Tg, such that molecular motion is relatively slow, d can be on
the order of 103 minutes, as it is for our reference PS. Hence, the time regime
between R and d is significant on experimental time scales and must be consid-
ered during the film formation process.
Doi and Edwards (1986) present this expression for d:

d = Nb2/(32DB). (5.4)
5.1 Essential Polymer Physics 157

For the reference PS, DB = 5.45 x 1018 cm2/sec and b = 0.67 nm (Karim et al.
1990). As shall be discussed in greater detail later in this chapter, DB varies
approximately linearly with N 2, so that d is seen to be strongly dependent on N:
d ~ N 3.
For a polymer diffusing across a planar interface into the same polymer, an
interpenetration distance, , across an interface (Aradian et al. 2002) can be
shown to scale as

~ Rg (t/ d)1/4. (5.5)

Through the scaling relation for Rg ~ bN1/2 and from the definition of d as
given in (5.4), this expression for can be shown to be equivalent, within a
constant, to the Doi-Edwards expression (5.3) for <r2>1/2 for t between R and d.
In Section 5.2, we shall see how (and the resulting wI) are correlated with the
fracture energy and strength of polymer-polymer interfaces.
When two surfaces are first placed in close contact, there are no chains bridg-
ing the interface. As reptation proceeds, however, more chains will cross the
boundary and thereby connect the two halves. Aradian et al. (2002) have argued
that the number of chains bridging the interface per unit area is proportional to the
value of . We shall see the importance of these connector chains in determining
the strength of the interface.
It is difficult to measure wI of polymer interfaces at times shorter than d be-
cause its magnitude is on the order of 10 nm or less. Neutron reflectivity data
obtained from planar film interfaces over a wide range of times have had too much
scatter to adequately test the predicted time dependence of wI (Karim et al. 1990).
For interfaces in immiscible systems (Sferrazza et al. 1997) and in systems
approaching miscibility (Carelli et al. 2005), an additional complication is that the
interface is broadened by thermally induced capillary waves (Binder et al. 1999).
In latex films, there have been several reports of measurements of the diffusion
distance as a function of time using SANS. One would expect to find for
diffusion times less than d but greater than R that ~ t1/4, whereas at times
greater than d to find ~ t1/2. In many experimental systems, however, d is only
a few minutes, and it is not possible to obtain many data points in the required
range (Yoo et al. 1990). A solution to the problem is to combine data from a wide
range of molecular weights and diffusion temperatures.
Following this approach, the most conclusive evidence for reptation in latex
was obtained by Kim et al. (1994). In their study of interdiffusion in poly(styrene)
latex films, they found a transition from a t1/4 dependence of the diffusion distance
to a t1/2 dependence at a time corresponding to the reputation time. This result is
just as is predicted from the reptation theory as outlined above.
158 5 Molecular Diffusion Across Particle Boundaries

5.2 Development of Mechanical Strength and Toughness

In classic experiments, Zosel and Ley (1993) showed that a brittle-to-tough


transition occurred when uncrosslinked latex films were annealed for short times.
They studied PBMA with a Tg of 29C, which could be film-formed at room
temperature (23C) to make brittle films with no measurable toughness. But after
just five minutes of heating at 90C, the films had sufficient strength to allow a
measurement of fracture energy (i.e., energy required to break the film). When
heated for longer times, there was a further gradual increase in toughness.
These experiments teach us two important lessons. First, when two surfaces are
brought into ultra-close contact (that is, distances on the order of a few nm or
less), there will be attraction from van der Waals forces, specifically dispersive
forces. This alone is enough to give a measurable adhesion energy, such as when
cling-film is pressed into contact on glass. In latex films, the particle deformation
stage creates flat particle boundaries and brings the particles into close contact
with a gap on the order of atomic length scales. When two surfaces make such
close contact, they are said to be wetting, which is illustrated from a molecular
viewpoint in Fig. 5.4. After Zosel and Leys latex films were heated for just five
minutes, there was sufficient particle deformation to enable wetting to give a
measurable fracture energy. The adhesion energy of wetting surfaces is relatively
low, however.
Secondly, these experiments reveal the effects of diffusion of molecules across
the particle boundaries to create entanglements that increase the mechanical
strength of the interfaces. As a diffusive process, the strengthening is gradual.
In the remainder of this section, we shall discuss the details of how diffusing
molecules bridge the gap across the interfaces and thereby strengthen them. But
first, it is useful to distinguish between properties that are determined by the bulk
of a latex film and those that are influenced by the particle-particle interfaces
within a film (Aradian et al. 2002). For instance, the shear modulus of an elas-
tomer is directly proportional to the density of chemical crosslinks in the bulk
material. Similarly, the extent of swelling in the elastomer, when exposed to a
good solvent, is influenced by the crosslink density. A films hardness will be
determined by the strength of bonds in the bulk material and by its glass transition
temperature in relation to the test temperature.
On the other hand, in a fresh film in which there has been limited diffusion
across the particle interfaces, the stress at which the film fails under tension (i.e.,
the tensile strength) depends on the strength of the weakest link that of the
interfaces. Likewise, the barrier properties the resistance to the penetration of
water or other solvents is a strong function of the interfacial structure. If the
interface is not healed, there will lower resistance to liquid flow, as the gap will
provide a path for transport.
At the interface, there can be fracture from chains pulling out from one of the
two sides of the interface via a pull-out mechanism. Alternatively, both ends may
stay on either side of the interface, but the molecule can break in its middle in the
5.2 Development of Mechanical Strength and Toughness 159

mechanism called chain scission. To obtain chain scission, the two ends of the
molecules must be entangled so that they can support stress along the chain. The
differences between interpenetrating and entangled chains are illustrated in Fig. 5.4.
There are numerous examples of experiments that investigated the strengthening
of miscible polymer interfaces and correlated this process, whenever possible, with
the extent of interdiffusion. Experiments have been performed on planar interfaces
of bulk materials in healing experiments. The general approach is either to crack a
specimen and then to place the cracked surfaces in contact or to place two polished
surfaces in contact. After heating at various times and temperatures to allow
interdiffusion, the mechanical properties of the healed interface are evaluated using
a fracture mechanics test. The key parameters from fracture mechanics studies and
some common test geometries are presented in Appendix A4.
The result of a classical fracture mechanics test (Williams 1977) is usually
presented as the fracture toughness, KIC, which is a material property with SI units
of Nm3/2. The critical energy release rate, Gc with SI units of Jm2, is a related
parameter that represents the critical energy required to create a fractured interface
of unit area. A crack will not propagate at an energy release rate below Gc. This
property is sometimes called the fracture energy or adhesion energy (Schnell et
al. 1999). Finally, the fracture strength, f, of an interface can be measured.
Details of these parameters and their experimental determination are provided in
Appendix A4.

(a) (b) (c)

Interfacial wetting: weak Chain extension Chain entanglement


adhesion from van der across the interface: across the interface:
Waals attraction likely failure by chain possible failure by chain
pull-out scission

Fig. 5.4 Molecular-level view of polymer/polymer interfaces that a have contact at the molecular
level (wetting), b have molecules penetrating across the interface, and c have molecules crossing
the interfaces over a sufficient distance to become entangled. The weakest interface is a,
followed by b in which interfacial failure is likely to be by chain pull-out, and c in which chain
scission is possible.
160 5 Molecular Diffusion Across Particle Boundaries

In comparing the various types of measurements, it is necessary to know the


relationship between them: KIC ~ f ~ GC1/2. In a latex film undergoing brittle
fracture, the fracture strength in a tensile test, T, is defined as the stress at the
point of failure. It should follow the same time dependence as f obtained in
fracture mechanics experiments. It is important to remember that a latex polymer
that is film-forming at room temperature (without the aid of coalescent), will not
be glassy. Fracture in cracks along the particle interfaces will be blunted by plastic
deformation. In this case, fracture stresses may be greater. Great care must be
taken when applying concepts from the fracture mechanics of planar interfaces to
latex films.
Zosel and Ley (1993) commented that the fracture energy per unit volume to
break a sample, WB (obtained from the integration under a stress-strain curve) is a
better parameter to characterise mechanical strength. It measures the toughness,
which is the total energy to deform a material, both elastically and plastically. It
provides more information than the tensile stress at break, T.
The time-dependence of these properties has been explored by several research
groups, and Table 5.2 shows rather good agreement in their findings. In experi-
ments on bulk materials, nearly all experiments find that KIC ~ t1/4 (and hence GC ~
t 1/2). In latex films, the results are somewhat less consistent, but undoubtedly, T
and WB increase over time. In these experiments, there is naturally a large degree
of scatter in the results. Often only four or five data points are obtained within the
reptation regime, and hence it is exceedingly difficult to properly test the power
dependence on time.
Undoubtedly, there is a correlation between diffusion time and latex film strength
(and toughness), and undoubtedly there is interface broadening over time. But, there
has been considerable debate and uncertainty around how to explain these correla-
tions. There are two main schools of thought, which we will consider here.
The first approach is to consider that KIC and Gc depend in some way on the
density of chains crossing the interface, . The second approach relates KIC and Gc
to the depth of interpenetration, , and assumes there is a certain minimum
distance required for bulk strength to be achieved. The two approaches are not
entirely at odds, however, since the distance of diffusion will increase simultane-
ously with (Aradian et al. 2002). The relation must be treated cautiously,
however, as it has been hypothesised that can reach its maximum value early on
in the diffusion process (Schell et al. 1998).
It is generally accepted that for a chain crossing a polymer-polymer interface to
contribute significantly to the strength, it must entangle with both halves. Thus, its
size must be at least twice the entanglement molecular weight, Me. Short chains
that interpenetrate but do not entangle will lead to interface failure by chain pull-
out. Entangled chains, on the other hand, are more prone to fail by chain scission.
An outcome of polymer-polymer interdiffusion is unwanted adhesion also
known as blocking between polymer surfaces placed in contact. Inorganic
powders, such as carbonates or talcum, are sometimes put on latex glove or
condom surfaces so that they do not fuse together. The fine particles of powder
prevent wetting and blocking.
5.2 Development of Mechanical Strength and Toughness 161

Table 5.2 Experimental findings on the time dependence of interfacial strength and adhesion
energy
Fracture Mechanics Experiments

Polymer system Result Reference

Poly(methyl methacrylate);
Poly(styrene-co-acrylo-nitrile) KIC ~ t1/4 Jud et al. (1981)

Poly(styrene-co KIC ~ t1/4 Nguyen et al. (1982)


acrylonitrile)

Hydroxy-terminated f ~ t1/4 Wool and OConnor (1982)


polybutadiene

Poly(styrene)/poly KIC ~ t1/4 Kausch et al. (1987)


(phenylene oxide)

Poly(styrene) weak time dependence Schnell et al. (1998)


of GC then sudden jump
to bulk value

Latex Films

Polymer system Result* Reference

Poly(styrene)
M ~ 1.8 x 106 g/mole T ~ t0.69 Yoo et al. 1991
Poly(styrene)
M ~ 2.6 x 105 g/mole T ~ t0.27 Yoo et al. 1990
(but t > d)

Direct mini-emulsified Poss. transition from Kim et al. 1993


Poly(styrene) f ~ t0.32 to f ~ t0.51
M ~ 2 x 105 g/mole but only four data points

Poly(styrene) T ~ t0.62 Kim et al. 1994


M ~ 1.5 x 105 g/mole

Poly(methyl methacrylate) T ~ t0.88 Yoo et al. 1995


M ~ 2.5 x 106 g/mol

Poly(butyl methacrylate) WB ~ t0.5 Zosel and Ley 1993


But diffusion data
indicate t > d.
*In interpreting these results, note that KIC ~ f and that KIC ~ GC1/2
162 5 Molecular Diffusion Across Particle Boundaries

5.2.1 Dependence on the Density of Chains Crossing the Interface

Behind the model of Jud et al. (1981) is an assumption that Gc is directly propor-
tional to . Arguing that ~ t 1/2, they then predicted that Gc ~ t 1/2, in agreement
with experiments. This derivation uses centre-of-mass Fickian dynamics, which
do not apply when t < d, and thus can bring the results into question. But Prager
and Tirrell (1981) pointed out the relevance of chain end position whether fixed
along the interface or randomly distributed within Rg of the interface in deter-
mining the rate of chain crossing. They predicted that ~ t1/2 for the latter case.
Fracture surfaces are likely to be covered with dangling chain ends albeit from
chains that are shorter than the average. In latex systems, chain ends often have
functional groups or initiators that may be thermodynamically favoured at the
particle surface. Hence, it is useful to consider the predictions of Prager and Tirrell
(1981) for chains isolated at the surface. In this case, they predicted that ~ t1/4,
which is at odds with the arguments of Jud et al. (1981). It may then be argued that
Gc1/2 ~ KIC ~ ~ t1/4, in agreement with most experiments. de Gennes (1989b) agreed
that GC ~ and that the work to pull-out or fracture each connecting chain should be
considered. He went on to consider the relevant fracture mechanics of a crack
travelling at an interface. In the case where chain ends are isolated at an interface,
then Gc ~ t1/2 is predicted, in agreement with experiments for planar interfaces. Using
the minor chain reptation model, Zhang and Wool (1989) predicted a cross-over in
the time dependence of for times increasing above d. At shorter times, they
predicted ~ t3/4 but when t > d, they found ~ t1/2.

5.2.2 Dependence on Interdiffusion Distance,

At the heart of the minor chain reptation model by Kim and Wool (1983) is the
assumption that interfacial strength, f, is proportional to the penetration depth of
monomer segments, which is predicted by reptation theory to vary as t1/4 when t <
d. They therefore predict that f ~ t1/4 (and the fracture toughness is likewise
predicted to vary in the same way).
These questions over the time-dependence of interfacial strength development
may be seen as rather academic. The more important point is that, no matter what,
latex films will get stronger over time. Another relevant question to consider is
how the interfacial strength depends on wI and .
For immiscible glassy polymer interfaces, which are not usually found in latex
films, the required interface width to achieve bulk adhesion energy depended on
the polymer molecular weight (Schnell et al. 1998). For lower M, the maximum
adhesion energy was achieved when wI was comparable to the Rg. But for larger
M, wI did not need to be as high as Rg but instead corresponded to the average
distance between entanglement points.
5.2 Development of Mechanical Strength and Toughness 163

A similar argument was made for miscible interfaces, which were found to
depend on the value of M in relation to the entanglement molecular weight, Me
(Schnell et al. 1998). It was proposed that for M < 12 Me, full strength will be
achieved when wI = Rg. But for M > 12Me, wI only needs to reach a value corre-
sponding to Rg for a polymer with M = 12Me. This seems to provide a good rule
of thumb for users of latex films. In the case of high M polystyrene, the critical wI
was found experimentally to be in the range between 10 and 12 nm, which is only
slightly less than the predicted value of 13.1 nm (Schnell et al. 1998).
There have been several elegant experiments in which diffusion distances, , in
latex films have been measured simultaneously with the tensile strength, T. In one
example, in polystyrene with M 1.5 x 105 g/mole (or about 8Me) the maximum
strength was achieved when reached about 9 or 10 nm, which is just slightly less
than the Rg of 10.6 nm (Kim et al. 1994). Their previous experiments (Kim et al.
1993) obtained a similar result. But this result is close to the prediction of Zhang and
Wool (1989) that maximum strength will be achieved when = 0.81 Rg.
A similar correlation has been found for fracture energy (per unit volume), WB,
and . Recall that WB represents the total energy required to break a film and is
thus more than an indication of the elastic modulus. Rather, it indicates the
toughness. While increases with interdiffusion time, so too does WB. During the
interdiffusion of a poly(butyl methacrylate) latex film, WB increased from 25
mJ/mm2 to a plateau value of about 30 mJ/mm2 (Zosel and Ley 1993). With
further interdiffusion, the WB did not increase further because the polymer chains
at the particle interfaces became entangled. Further diffusion offers no benefit to
the mechanical properties. The results in Fig. 5.5 show how the plateau in WB is
reached when the polymer chains have diffused a distance of = 40 nm, which is
approximately the radius of gyration for this high molecular weight polymer.

WB
(mJ/mm3) (nm)

Square-root of time (s1/2)


Fig. 5.5 Simultaneous increase in the fracture energy per unit volume, WB, and the diffusion
distance, (as determined from SANS experiments (Hahn et al. 1986)) for poly(butyl methacry-
late films). The onset of the plateau in WB corresponds to a value of that is comparable to the
polymers radius of gyration. The films were annealed at a temperature of 90C, which is
approximately 60C above the polymers Tg. (Reprinted with permission from Zosel and Ley
(1993), copyright 1993 American Chemical Society)
164 5 Molecular Diffusion Across Particle Boundaries

We close this section by noting that a film does not have to be entirely dry for
particles to make molecular contact in order to allow interdiffusion. FRET experi-
ments have shown chain interpenetration at average solids contents as low as 90%
(Feng et al. 1998). In films that dry from their edges (as discussed in Section 3.5),
particles can be in any one of three states (wet and dispersed, wet and packed, or dry
and packed) depending on their lateral position. The presence of water at the
boundaries between particles was found to prevent interdiffusion. The extent of
interdiffusion has been found to depend on the amount of time elapsed since the
passing of the drying front. Hence, during the film formation process, there is more
molecular mixing near the edge of a film compared to its centre (Haley et al. 2008).
Other experiments have led to the conclusion that there can be greater molecu-
lar mobility near particle surfaces to make a soft shell. Interdiffusion over
distances up to 20 nm has been measured even at temperatures near the nominal Tg
of the bulk latex polymer. Diffusion of short chains and plasticisation by water
were ruled out for their system (Turshatov et al. 2008), but water has been found
to enhance diffusion by the plasticisation of other systems (Haley et al. 2008).

5.3 Factors that Influence Diffusivity

We have established that diffusion is important in determining the mechanical


strength of a latex film. Users of latex films require ways to control the diffusion
process to achieve the desired properties on the relevant time scale. In this section,
several key factors that influence D will be considered. A discussion of the effects
of surfactant is deferred to Chapter 6.

5.3.1 Molecular Weight and Chain Branching

The theory of the dynamics of linear polymer chains is essentially complete, whereas
the understanding of branched chains is well advanced but has some remaining
questions (McLeish 2002). The scaling prediction from reptation theory is that the
polymer self-diffusion coefficient of linear polymers above their entanglement
molecular weight, Me, varies inversely with the square of their molecular weight: D ~
M 2 (Doi and Edwards 1986). Rather understandably, larger molecules are predicted
to diffuse far more slowly than smaller ones! There are some shortcomings, however,
in this model, and various corrections to the theory predict a stronger M dependence.
Lodge (1999) compiled experimental data for seven different polymers, and he found
that they collectively supported a scaling relation of D ~ M 2.3. For lower molecular
weights when chains are too short to entangle, i.e., when M < Me, in the Rouse
regime, then D is greater in value and varies less strongly, D ~ M 1. Regardless of the
precise value in the scaling relationship, it is clear that molecular weight is an
important parameter influencing the diffusion stage in latex film formation.
5.3 Factors that Influence Diffusivity 165

However, the situation is far more complicated in latex. First, the molecular
weight distribution in latex particles is often quite broad. It can be difficult to
control molecular weight precisely during emulsion polymerisation, and so
tuning diffusion is not a trivial exercise. Secondly, the lower molecular weight
fraction is often near the particle surface. In this case, the surface molecules,
which are the ones that will first cross the particle boundaries, will also be the
fastest diffusers, in comparison to the molecules behind them. Wang and Winnik
have used this idea to explain why the rate of interdiffusion is often faster in the
earlier part of the stage and slows down over time (1993). A third complication is
that the polymer chains in the latex often are not linear but highly branched. The
diffusivity of a branched molecule is exponentially slower in comparison to a
linear one of the same molecular weight (de Gennes 1975). The diffusion coeffi-
cient of a star polymer (in which several chain arms emanate from a central
point) depends exponentially on the molecular weight of the arms, rather than on
the total molecular weight (McLeish 2002).
Experimentally, the expected M 2.3 dependence of diffusion in high-M
poly(butyl methacrylate) latex films has been supported (Wang and Winnik 1993).
D was found to be 3.05 x 1014 cm2 s1 for M = 7.6 x 104 g mole1, whereas it was
significantly lower (3.3 x 1016 cm2 s1) for a higher M of 5.9 x 105 g mole1. A
strong molecular weight dependence of chain diffusion was also found when the
diffusion is through a gel network, created by light crosslinking of latex particles
(Oh et al. 2003). It is worth noting here that de Gennes original idea for the
reptation model (1971) was developed for chains diffusing in a gel.

5.3.2 Temperature Dependence

Diffusion is a thermally activated process and hence increases with increasing


temperature. In polymers, the motion of the centre-of-mass of a molecule does not
occur until temperatures above the glass transition temperature, Tg, when the
polymer is in the molten state. Below Tg, only segmental motion is possible at
least in the bulk material as there are lingering questions about sub-Tg polymer
mobility at surfaces. The diffusion coefficient is strongly dependent on tempera-
ture. Using poly(butyl methacrylate) as an example, at temperatures well above
the Tg of ca. 35C, an increase in temperature by 20C results in an increase in the
diffusion coefficient by about one order of magnitude. Values range from less than
1016 cm2s1 to about 1013 cm2s1 over the temperature range from 70C to 120C
(Juhu and Lang 1994). To study diffusion on realistic experimental time scales
(hours not weeks), most experiments use high temperatures. But, for applications
in coatings or adhesives, the latex film is usually stored at room temperature.
Diffusion in latex has been confirmed to occur at room temperature, even when
the Tg of latex was also at that temperature. In such a case, the diffusion coeffi-
cient had the low value of 5 x 1018 cm2s1 (Feng et al. 1998).
166 5 Molecular Diffusion Across Particle Boundaries

There are two primary ways to describe mathematically the temperature-


dependence of D, and these will be outlined here. As we shall see, the latter
equation is a more generalised expression for the first.
Arrhenius temperature dependence. A so-called Arrhenius equation is
commonly used to describe the thermal activation of chemical reactions in which
the rate varies exponentially with temperature. The strength of the dependence is
characterised by an activation energy, Ea. The temperature dependence of the
diffusion coefficient is accordingly given as
D = Doexp(-Ea/RT) (5.6)
1 1
where Do is a pre-factor and R is the ideal gas constant (8.314 J mole K in SI
units). This equation assumes that Ea is independent of temperature, which is an
acceptable assumption only over a narrow temperature range that is far above the
polymer Tg. As T decreases towards Tg, Ea increases as D diverges towards
infinity. For polymers, Ea represents an apparent activation energy.
If the natural logarithm of D is plotted against the reciprocal of T, the data
points will fall on a straight line, when Eq. (5.6) applies; the slope of this line
provides a measure of Ea. Fig. 5.6 gives an example of how measurements of the
fraction of molecular mixing between particles (obtained in FRET experiments)
are used to calculate D as a function of temperature. The temperature dependence
of D expected from Eq. (5.6) is then found.
Knowledge of Ea is useful for making predictions of D at a certain temperature
when it is known at a different temperature. To provide an indication of the
temperature-dependence of D that is typically encountered in latex films, Table
5.3 lists some Ea values. The table includes the temperature range over which the
values were obtained, expressed in relation to the polymers Tg, so as to indicate
the range of applicability.

Fig. 5.6 a Volume fraction of mixing between latex particles as a function of increasing time annealed
at three temperatures (45, 55 and 65C), as indicated. b From these experiments, the apparent diffusion
coefficients, Dapp, are calculated at each temperature and shown as open squares. Dapp increases
strongly with increasing temperature in a way that can be described by an Arrhenius equation. (Data
from a second latex series with a higher molecular weight is shown as filled circles for comparison.)
(Reprinted with permission from Wu et al. (2004), copyright 2004 American Chemical Society)
5.3 Factors that Influence Diffusivity 167

Table 5.3 Apparent Activation Energy for Diffusion in Various Latex Compositions
Apparent Activa-
Temperature
Polymer tion Energy, Ea Reference
Range, T Tg (C)
(kJ/mole)
Poly(styrene) 218 8 25 55 Kim et al. 1994
Poly(n-butyl methacry- 164 Wang and Winnik
40 70
late) 1993
Poly(methyl methacry-
181 38 68 Liu et al. 2001
late-co-butyl acrylate)
Poly(vinyl acetate) 151 26 112 Wu et al. 2004b
Poly(vinyl acetate-co-
143 0.8 23 63 Oh et al. 2003b
butyl acrylate)
Poly (vinyl acetate-co-
155 8 31 56 Wu et al. 2004b
dibutyl maleate) High M
Poly (vinyl acetate-co-
dibutyl maleate) M = 2.5 189 8 27 47 Wu et al. 2004b
x 105 g/mole

Williams-Landel-Ferry (WLF) Treatment. The WLF equation was developed


as a means of comparing dynamic moduli data obtained at differing temperatures
and time scales using the time-temperature superposition principle. For instance,
at high temperatures, well above Tg, viscosity is exceedingly low and must be
measured over short time scales (at high frequencies). On the other hand, at
temperatures near Tg, it is impossible to measure the viscosity over short time
scales, because of the exceedingly high viscosity. The WLF treatment allows a
wide variety of such measurements to be shifted in time so that they fall on a
smooth master curve.
The WLF equation has therefore been adapted to predict D at an arbitrary tem-
perature T, provided that the diffusion coefficient is known to be DR, at a reference
temperature, TR. The two temperatures and diffusion coefficients are related by a
shift factor, aT:

C1 (T TR ) DT
log aT = = log R (5.7)
C2 +T TR DRT

C1 and C2 are factors that depend on the particular polymer and the TR. As an
example, C1 = 14.5 and C2 = 255 K for the often-studied poly(n-butyl methacry-
late) when TR = 373 K (Kobayashi et al. 2002). The shift factors for D have been
shown to coincide with the shift factors (and C1 and C2 values) obtained for the
dynamic moduli of the same latex polymer at the same reference temperature (Wu
et al. 2004b). Specifically, a master curve of dynamic moduli as a function of
frequency may be constructed from measurements obtained over a range of
temperatures by shifting the frequencies by aT according to the temperature of the
168 5 Molecular Diffusion Across Particle Boundaries

measurement. Viscoelastic properties and diffusion coefficients can thus be inter-


related through the WLF treatment.
The rearrangement of (5.7) reveals an Arrhenius-type (i.e., exponential) de-
pendence of D on temperature, which is of similar form to (5.6) (Wang and
Winnik 1993). The exponential behaviour of the time-temperature shift factor aT
may be manipulated (Ferry 1980) to yield an expression for the temperature-
dependence of the activation energy, Ea, explicitly as:

2.302 RC1C2T 2
Ea = (5.8)
( C2 + T TR )
2

Hence, the WLF and Arrenhius methods, outlined here to describe the tempera-
ture dependence of D, are linked.

5.3.3 Influence of Hard Particles

In various applications of latex films, desired properties are achieved through the
addition of hard (i.e., non-deformable) particles. Such particles can be inorganic,
e.g., calcium carbonate or titanium dioxide pigments, or they can be polymeric,
such as glassy acrylics. Table 5.4 gives a few examples of the properties and
characteristics that can be achieved with hard particles. There will be further
discussion of the effects of hard particle addition in Chapter 7.

Table 5.4 Properties that can be achieved through the addition of hard particles to latex binders
Type of particle Property or Characteristic
TM
Hollow glassy polymers (Ropaque ) Increased opacity because of air voids
Silica particles Anti-static; stain resistance
Carbon black, Metals Electrical conductivity; fully opaque
Increased elastic modulus (stiffness), high
Any durability and abrasion resistance; greater
hardness
Various inorganic (talc, clays, alumina) Reduced tackiness
Any at high fractions Increased film porosity
Any with a refractive index different than the
Opacity because of light scattering
latex

There are numerous examples that provide evidence that the mobility and me-
chanical properties of a polymer near an interface with a hard wall, such as a
filler particle, differ from what is found in the bulk of the polymer. As reviewed
by Feng et al. (1998), measurements of NMR relaxation, dynamic mechanical
moduli, elasticity, and the glass transition temperature, all point to reduced
5.3 Factors that Influence Diffusivity 169

mobility at the interface between a polymer melt and a hard wall or particle. There
has also been increasing evidence in recent years that the addition of hard particles
can have an effect on polymer diffusion in latex films. Two categories of explana-
tion can be offered, and each will be considered here.
1. Interfacial Blocking and Tortuosity Effects. As we have seen, diffusion of
molecules between particles requires interfacial wetting. A hard filler particle can
sit at the interface between two or more latex particles when it is much smaller
than them. Filler particles can thereby reduce the total area of contact between
latex particles in a film. Indeed, FRET experiments have indicated that the contact
area between poly(butyl methacrylate) latex particles (diameter of 90 nm) de-
creased as very small silica particles were blended in (Kobayashi et al. 2002).
There was only a significant effect, however, when the silica was much smaller
than the latex (16 or 27 nm) as shown in Fig. 5.7. When the filler particles are
significantly smaller than the latex particles, the number of filler particles will be
higher than the number of latex binder particles, even when the volume fraction of
filler is lower. It follows that nanoparticles will present a physical barrier to
interdiffusion at the interfaces between latex particles.

Fig. 5.7 Initial efficiency of energy transfer, FET, which indicates the amount of close contact
between poly(butyl methacrylate) particles, immediately after they first pack together. The effect
of silica nanoparticle concentration is shown. Four particle sizes (16 nm (open square), 27 nm
(open circle), 54 nm (filled square) and 73 nm (filled circles)) are used. The number ratio of
silica-to-latex particles are indicated for the smaller particles. (Reprinted with permission from
Kobayashi et al. (2002), copyright 2002 American Chemical Society)

When filler particles are larger than the latex particles, on the other hand, the
number of fillers is relatively low, and these particles are too large to fill the
interstitial space between the latex particles. It is then much more likely that full
contact will be achieved between the latex particles without any blocking by the
filler. The latex particles can pack together and fill in the space around the larger
filler (Kobayashi et al. 2001). Figures 5.8 a and b distinguish between the cases
when the filler particles are smaller than and greater than the latex particle size,
respectively.
170 5 Molecular Diffusion Across Particle Boundaries

As we have seen, polymer diffusion occurs at the molecular level. Film


strengthening only requires diffusion over distances comparable to the molecular
radius-of-gyration. It is important to remember these issues when considering the
effect of hard particles on interdiffusion in latex films. We can argue that when the
hard particles are smaller than the polymers end-to-end diffusion distance, a
molecule will need to travel around the particles (as shown in Fig. 5.8c). The path
is said to be tortuous. As a result, the pathway for diffusion over a certain end-to-
end distance is longer, and the diffusion time will be longer. In an experiment that
measures the centre-of-mass diffusion coefficient, the diffusion rate will appear
slower. In many practical systems, the size of the filler particles is substantially
larger than the diffusion distances, and so tortuosity effects can be ruled out (Feng
et al. 1998). One can imagine a scenario in which the spacing between the hard
particles is less than the Rg of the polymer (Fig. 5.8d). For the chain to move
through the gap between particles, it would need to extend from its random coil
conformation and decrease its entropy. Consequently, diffusion would not be
favoured by entropic considerations in such a case.

(a) Inorganic particles (b)

(d)
(c)

Tortuous path Diffusion distance is short


relative to filler particle size

Fig. 5.8 Illustrations of the effects of filler particles on polymer interdiffusion. a When the filler
particles are smaller than the latex particles, they are trapped between particles and prevent
polymer-polymer contact. b When the filler particles are larger than the latex particles, contact
between the latex particles is not affected in the same way. c When a polymer chain has to
diffuse around a hard filler particle, the path is described as tortuous. But when the diffusion
distance is short relative to the size and spacing of the filler particles, the chains path is
unaffected. d Chains must stretch to diffuse between filler particles that are less than a radius of
gyration apart.
5.3 Factors that Influence Diffusivity 171

2. Reduced Mobility Effects. Filler particles rarely have an inert surface. Instead,
their surface often has polar moieties (such as carboxylic acid or hydroxyl groups)
or an electrostatic charge. It is common, therefore, for polymers to be attracted to
filler particles as a result of hydrogen bonding, polar interactions or charge. Monte
Carlo simulations of soft polymers at the interface with hard particles (Papakon-
stantopoulos et al. 2005) have found that an attraction between a nanoparticle and
a polymer melt creates a glassy layer at the interface. In a filled latex, such
interfacial regions would prevent polymer interdiffusion.
Experiments using FRET to measure mixing and diffusion in filled latex films
provide evidence for the effects of restricted mobility at interfaces with hard
particles. An important piece of evidence comes from the examination of the
effect of hard particle size, Rhp on polymer mobility. The surface area-to-volume
ratio of a spherical particle is proportional to 1/Rhp, telling us that interfaces
become more significant with smaller particles. The mobility of latex polymers
becomes reduced in the presence of hard acrylic particles (Feng et al. 1998) and
silica particles (Kobayashi et al. 2002) as 1/Rhp increases. In both systems, it is
believed that there is a positive attraction between the polymer and filler particle.
In the case of silica, 16 nm particles at 40 wt.% loading reduced the apparent
diffusion coefficient by a factor of 10, compared to pure latex. This result is
consistent with an increase in the apparent glass transition temperature of the
polymer by up to 15C (Fig. 5.9). At present, there is no fully satisfactory theo-
retical explanation for the effects of interfaces on polymer mobility, but this is an
active area of research.

Fig. 5.9 Calculated increases in the polymer glass transition temperature as the volume fraction
of silica particles (16 nm and 27 nm) is increased. (Reprinted with permission from Kobayashi et
al. (2002), copyright 2002 American Chemical Society)
172 5 Molecular Diffusion Across Particle Boundaries

A second important piece of evidence, in support of the notion of restricted


mobility at filler interfaces, as a means to explain the reduced diffusivities is
provided by the temperature dependence of diffusion in composite films. If the
slowing-down of diffusion is the result of tortuosity, the effect should still be
apparent at higher temperatures when the polymer has greater mobility. On the
other hand, if the diffusive slowdown is the result of restricted mobility (or even
glass formation) along interfaces, the effect should be lost at elevated temperatures
where thermal energy can overcome such restrictions. Experiments on poly(butyl
methacrylate) latex show that silica nanoparticles have little effect on the extent of
diffusion at 120C in comparison to a much greater effect at 60C. Therefore, the
data support the concept of reduced mobility at interfaces (Kobayashi et al. 2002).

5.3.4 Latex Particle Size

It has been proposed that diffusivity will be enhanced when the latex particle
radius, R, is comparable to the radius of gyration of the polymer, Rg. One reason is
that there will be a higher probability that one or both chain ends are near the
particle surface, allowing reptation to proceed sooner. A second reason is that
molecules confined near a surface are perturbed from their random coil conforma-
tion. It is thermodynamically favourable for a chain to relax by extending across
an interface to achieve a random coil and thereby increase its entropy. Yoo et al.
(1991) reported measurements of diffusion for a high molecular weight
poly(styrene) latex having an Rg = 17 nm that was confined in particles with R =
27 nm. They found that D was higher than expected and attributed the result to the
confining effect of the particle size.
In latex used for most applications, however, one usually finds that R >> Rg,
and these effects are not relevant. Experiments have revealed that D for 100 nm
latex particles is equal to that for 300 nm particles of the same polymer. However,
it is worth noting that mixing between particles is achieved faster with a smaller R,
because a greater fraction of molecules is near the particle surface (Boczar et al.
1993). Another way to think about this is to realise that smaller particles will have
a greater interfacial area per unit volume and, therefore, will mix with their
neighbours to a greater extent when travelling a certain distance. In larger parti-
cles, the same amount of interfacial overlap will correspond to a lower fraction of
mixing.

5.3.5 Particle Structure and Hydrophilic Membranes

Despite the contrary images in many cartoon drawings of latex film formation,
colloidal particles are often not homogeneous. Sometimes, the particles are
deliberately made to have two phases, such as the core-shell particle structures
5.3 Factors that Influence Diffusivity 173

that will be discussed in Chapter 7. In many cases, though, the situation is more
subtle. The latex particle surface can be enriched in hydrophilic components that
essentially create a shell layer, sometimes called a membrane. The most often
studied hydrophilic group is carboxylic acid, such as is found in copolymers
containing acrylic acid and methacrylic acid. At neutral values of pH, these acid
groups are negatively charged, introducing a repulsion to impart greater colloidal
stability.
When such particles are packed together in an array, the bilayers create a foam
structure. Small-angle neutron scattering experiments have revealed the periodic
structure of latex films containing hydrophilic membranes of acrylic acid (Joanicot
et al. 1993). Instead of bilayers of surfactant sandwiching a thin layer of water, as
in a conventional foam structure, the bilayers are made of hydrophilic polymer
(Fig. 5.10). In latex foam structures, sandwiched water layers that are no more
than 2 nm thick retain stability, whereas in oil-in-water emulsions, rupture occurs
when the water layer is less than about 10 nm (Joanicot et al. 1993). During film
formation, the hydrophilic membranes are gradually fragmented to allow the
core polymers to come into contact. When these cellular latex films are
deformed, the strength of the membranes determines whether the particles slip
past each other or the foam structure is elongated in the stress direction (Rharbi et
al. 1996). Carboxylic acid groups can be introduced into the latex particles by two
main methods. A hydrophobically modified acrylic acid copolymer can be used as
a steric stabiliser in the emulsion polymerisation of a hydrophobic monomer.
Alternatively, an acidic comonomer can be added in the final stage of polymerisa-
tion so that it is localised in the particle shell.
The precise effect that acrylic acid membranes have on interdiffusion depends
on the particular circumstances of the latex. An acrylic acid-rich phase, when in a
fully dry latex film, has an identifiable glass transition temperature over 100C
(Zosel et al. 1990). At room temperature, therefore, there is a glassy barrier
between particles. Hence, one important factor is the composition of the shell
layer in determining whether it is glassy or rubbery at the temperature of film
formation. Furthermore, the composition determines the miscibility of the shell
with the core polymer. Miscibility allows interdiffusion when it is otherwise
retarded or prevented altogether. A thicker shell layer is intuitively expected to
create a better barrier to interdiffusion.
Neutralisation of carboxylic acid-rich particle surfaces through the addition of
inorganic bases substantially retards diffusion in dry films. Divalent bases retard
diffusion more than monovalent ones (Kim and Winnik 1994). It is likely that the
ionomer phase, formed by the carboxylic acid groups and cations, has a greater Tg
in comparison to the shell polymer and therefore provides resistance to diffusion
across it (Kim and Winnik 1995). But in the presence of water, diffusion is about
two orders of magnitude faster even when there is neutralisation (Feng and
Winnik 1997).
174 5 Molecular Diffusion Across Particle Boundaries

COOH rich shell


Fig. 5.10 Example of the structure that can develop when particles have a copolymer shell that
is rich in acrylic acid groups.

5.4 Faster Diffusion with Coalescing Aids

The conflicting requirements of some latex film applications are sometimes called
the film formation paradox. For applications in household varnishes and paints,
film formation at room temperature is required, which necessitates the use of
polymers with a Tg below the ambient temperature. In these same applications,
there is also a requirement for hard, scratch-resistant surfaces. To achieve this
property, the addition of hard particles, as discussed already in Section 5.3, can be
employed. But hard particles still require a soft polymer binder to hold the
particles together. Films cannot be created solely from latex particles that have a
Tg above the ambient temperature. The paradox is how to make hard films through
the coalescence of hard particles at ambient temperatures.
A common resolution to the paradox is to reduce the Tg through the addition of
diluent molecules, called plasticisers or coalescing aids. Usually, a volatile
molecule is chosen so that it can enable film formation but then evaporate to leave
behind a film with a higher Tg. A common example of a coalescing aid that has
been used widely in the literature is 2,2,4-trimethyl-1,3-pentanediol monoisobu-
tyrate (TPM, sold under the trade name of TexanolTM).
The effect of a coalescing aid is more profound than merely lowering the
polymer latex Tg. As we have seen, diffusivity increases strongly as temperature
increases above Tg, and it is enhanced diffusivity that enables continuous and
strong films to be created. FRET experiments have revealed the profound effect of
coalescing aids on diffusivity. D for poly(butyl methacrylate), as measured with
FRET, increases by roughly four orders of magnitude when 12 weight percent
TPM is added. Just three weight percent TPM is enough to increase D by one
order of magnitude (Juhu et al. 1993).
5.5 Simultaneous Crosslinking and Diffusion: Competing Effects 175

In selecting a coalescing aid, it is important to consider several factors, as out-


lined below.
Evaporation Rate. The rate at which the coalescing aid leaves the film will
determine the time taken for the film to achieve its maximum hardness. Small
amounts of remnant molecules can soften the coating. The effects of evaporation
are apparent in FRET experiments showing that D for poly(butyl methacrylate)
decreases over time as the coalescing aid is lost (Winnik et al. 1992).
Glass Transition Temperature. The extent to which a polymer is softened
depends not only on the amount of added coalescing aid, but also on its type. To
achieve a greater effect of plasticisation, a coalescing aid with a lower Tg should
be used. (The Tg can be estimated to be about 2/3 of the melting temperature.) As
an example, 10 weight percent of TPM will reduce the Tg of poly(butyl methacry-
late) to only 10C, whereas the same concentration of diacetone alcohol will
reduce the Tg to 26C (Juhu and Lang 1994).
Solubility. In the wet latex, coalescing aids are likely to be partitioned between
the polymer and aqueous phases. If the solubility is higher in the aqueous phase,
then the plasticising effect will be delayed until the coalescent partitions in the
polymer. The partitioning determines the amount of solvent in the polymer and
aqueous phases and hence the extent of plasticisation. TPM is poorly soluble in
water and hence will mainly reside in the polymer phase. It can thereby encourage
diffusivity when particles first make contact with each other, even in the presence
of water.

5.5 Simultaneous Crosslinking and Diffusion: Competing Effects

Crosslinked latex films are stiffer than their uncrosslinked counterparts, and the
elastic modulus increases in proportion to the density of the crosslinks. Conse-
quently, crosslinking offers a means of increasing the durability and wear resis-
tance of coatings and to adjust the elasticity and tack of adhesives. Crosslinked
networks will swell when exposed to good solvents but they will not dissolve, and
crosslinking thereby imparts good chemical resistance. Yet, without proper control
of the crosslinking process, a weaker not stronger film can result.
Two main approaches can be used to introduce crosslinking into latex films.
Figure 5.11 compares these two types of approach. In the first type, crosslinking
molecules are free in the aqueous phase and then partition within and react with
the polymer during drying. In this type of system, the latex polymer is copolymer-
ised with appropriate reactive groups. The crosslinker molecules create bonds
between adjacent polymer chains. Fig. 5.11a illustrates that if crosslinking occurs
too quickly, crosslinked gel particles will be created. To prevent premature
crosslinking, giving what is called a short pot life, the external crosslinker can be
blended in immediately before film formation. In this case, it is considered a two-
176 5 Molecular Diffusion Across Particle Boundaries

pack system. Examples include various polyurethane systems and epoxy-amine


systems (Guerts et al. 1996). Taylor and Winnik (2004) have published a beauti-
fully written review of crosslinking chemistry, which gives further details.
Crosslinking reactions can also be activated by heat or light in functionalised
latex. Examples of functional groups for thermally activated crosslinking in
coatings include aziridine, epoxy, isocyanate, acetoacetoxy, and carbodiimide
groups. Fig. 5.11a illustrates that when interdiffusion precedes crosslinking, a
uniform crosslinked network can be created.
In the two pack in one pot approach, blends of complementary reactive parti-
cles are used. Crosslinking occurs when the two types of polymer come into
physical contact. A few recent examples are (1) blends of epoxy- and amino-
functional latex (Geurts et al. 1996), (2) blends of carbodiimide-containing and
carboxylic acid-containing latex (Pham and Winnik 2000), and (3) blends of
acetoacetoxy and amine-containing latex (Guerts et al. 1996). In two-pack in one-
pot systems, there is no reaction unless chains diffuse from one particle into a
neighbouring particle of the opposite type. Some miscibility is therefore required
(see Fig. 5.11b). A somewhat surprising experimental result has been that the
crosslinking reactions can promote miscibility in systems that are otherwise
immiscible (Pham et al. 2000b and Mohammed et al. 1997).

(a) X X
X X X

X
X
X
X X
X Gel particles from fast crosslinking
X X X
X
X
X X X
X

X = crosslinker molecules

Crosslinking after interdiffusion


(b) A
A
B A
B

A
B
A
B Particle deformation
A Interdiffused region of
reactive A + B polymers

Fig. 5.11. a Two-pack systems in which chemical crosslinkers are blended with the latex
dispersion prior to film formation. If crosslinking is faster than interdiffusion, weak interfaces
will result. b Two-pack in one-pot systems in which complementary particles (A and B) are
blended. When reactive A groups come into contact with B groups, a crosslinking reaction
occurs.
5.5 Simultaneous Crosslinking and Diffusion: Competing Effects 177

It is helpful to preface further discussion by distinguishing here between the


diffusion of free chains into a gel (i.e., crosslinked network) and interdiffusion
between two gel networks. The diffusion of a chain through a gel proceeds by
reptation (de Gennes 1971). In experiments where gel particles were blended with
ordinary, uncrosslinked particles, the diffusivity was similar in magnitude to the
diffusion in the ordinary particles (Wu et al. 2004a). That is, the free chains diffused
around the obstacles of chemical crosslinks in much the same way as they moved
through the obstacles created by entanglements, as shown in Fig. 5.12a.

(a)

Free chains into gel particles: interfaces dissolve

(b)

Contact between gel particles: limited interpenetration

Fig. 5.12 a Free chains are able to diffuse into crosslinked particles, leading to entanglements
and the dissolution of the interface. b When particles are crosslinked prior to film formation,
dangling chain ends can cross the particle-particle interface. The film will not achieve full
strength because of limited (or no) interfacial entanglements.
178 5 Molecular Diffusion Across Particle Boundaries

The case of diffusion between two crosslinked particles is different. The centre-
of-mass of polymer chains near a particle surface cannot diffuse to a significant
distance when those chains are attached to an extensive crosslinked network. Yet,
even in crosslinked particles, there will be some dangling ends that can extend
into neighbouring particles (Fig. 5.12b). In most cases, it would be expected that
chain ends will be too short to become entangled, so that the interface will never
achieve the strength of the bulk material.
In experiments on latex particles that are crosslinked prior to drying, films with
measurable strength can be obtained (Tamai et al. 1999, Zosel and Ley 1993).
Polymer chains dangling from the network, with sizes greater than the entangle-
ment molecular weight, are able to form entanglements across the interface. As the
density of the crosslinking increases, the fraction of polymer that is mixed
between particles decreases. Solvent resistance in such films is poor, however
(Tamai et al. 1999), and the films become fragile when swollen, which can be
taken to mean that there is disentanglement induced by swelling. In comparison to
similar crosslinking systems in which there is some interdiffusion, the strength and
toughness of the pre-crosslinked latex is much lower (Tronc et al. 2002b). Once
the crosslinking density is raised to a certain threshold, all film cohesion is lost,
even after annealing at high temperatures (Zosel and Ley 1993).
When crosslinking reactions occur simultaneously with interdiffusion, the two
processes compete. For instance, when crosslinking occurs relatively rapidly, the
diffusion has been found to cease entirely when a gel content of only 42% was
achieved (Liu et al. 2001). But, when the crosslinking rate was slowed down (by
removing the catalyst), then nearly full mixing through diffusion could be
achieved prior to crosslinking. This effect is explained by the fact that branched
chains diffuse more slowly than their linear counterparts, leading to a slowing
down of diffusion as crosslinking proceeds. When a chain is tethered to a larger
network, its long-range diffusion is halted entirely. Scaling arguments predict that
complete mixing between particles can be achieved when the reaction rate is slow
enough, but with faster reactions, very little mixing is achieved (Aradian et al.
2000). Arguing that the amount of mixing between particles is proportional to the
adhesion energy between particles, Gc, and hence the entire film toughness,
Aradian et al. derived scaling relations for the dependence of Gc on molecular
weight, M, in the case where there are a small number of crosslinks per chain
(such that M ~ Mc, the molecular weight between crosslink points):
For fast crosslinking reactions: Gc ~ M 7/4

For slow crosslinking reactions: Gc ~ M1/2.


These are completely opposite dependencies on M. In the slow reaction regime,
it is straightforward to see that longer chains will be better able to form entangle-
ments on either side of the interface and therefore lead to greater adhesion energy.
On the other hand, with fast reactions, the slower diffusivity of longer chains is a
hindrance. Shorter molecules are more successful in crossing the interface to
provide some reinforcement before being locked in place by crosslinks.
References 179

A control parameter, , has been introduced to allow comparison of the rates of


crosslinking and diffusion, in order to see which is dominant (Aradian et al. 2002).
Experimental work has shown that a polymer-polymer interface achieves the
strength of the bulk material when there is molecular diffusion of distances on the
order of the radius of gyration of the molecule. Reptation theory tells us that the
time required to diffuse this distance is on the order of the reptation time, d,
which varies as N 3. A characteristic time for the crosslinking reaction, xl, was
derived and found to vary as McN 1. The control parameter is merely the ratio of
these two times:

= d /xl ~ N 3/McN-1 ~ N4/Mc (5.9)

For << 1, diffusion occurs much faster than crosslinking, and the system is in
the slow reaction regime. Maximum adhesion energy can be achieved as
interdiffusion proceeds without hindrance. If N is increased such that = 1, the
system moves into the fast reaction regime. Within this latter regime, values of
closer to 1 favour interfacial crosslinking (Aradian et al. 2002). At higher values
of , there is very little interface crosslinking, leading to weak interfaces with a
pronounced inverse relation on chain length: Gc ~ N 5/2. An important point to
remember is that this model does not consider a molecular weight distribution. In
many real systems, there are smaller molecules that interdiffuse rapidly and
contribute to interfacial development and strengthening at shorter times (Tronc et
al. 2002b).
As noted in the introduction to this chapter, the experiments of Distler and
Kanig (1978) created confusion because the particles under investigation retained
their identity. In the light of recent understanding gained from the recent research
outlined here, these results are readily explained. Their poly(n-butyl acrylate) latex
was crosslinked near the particle surfaces because of the addition of 2% meth-
oxymethyl acrylamide, which created a robust cellular structure. It took some time
for scientists to appreciate fully the effect that crosslinking has on the structural
development of latex films.

References

Aradian A., Raphal E., de Gennes P.-G. (2000) Strengthening of a polymer


interface: Interdiffusion and cross-linking. Macromolecules 33: 9444-9451.
Aradian A., Raphal E., de Gennes P.-G. (2002) A scaling theory of the competi-
tion between interdiffusion and cross-linking at polymer interfaces. Macro-
molecules 35: 4036-4043.
Binder K., Muller M., Schmid F., Werner A. (1999) Interfacial profiles between
coexisting phases in thin films: Cahn-Hilliard treatment versus capillary waves.
J Stat Phys 95: 1045-1068
180 5 Molecular Diffusion Across Particle Boundaries

Boczar E.M., Dionne B.C., Fu Z.W., Kirk A.B., Lesko P.M., Koller A.D. (1993)
Spectroscopic studies of polymer interdiffusion during film formation. Macro-
molecules 26: 5772-5781
Broseta D., Fredrickson G.H., Helfand E., Leibler L. (1990) Molecular weight and
polydispersity effects at polymer-polymer interfaces. Macromolecules 23: 132-
139
Bueche F., Cashin W.M., Debye P. (1952) The measurement of self-diffusion in
solid polymers. J Chem Phys 20: 1956-1958
Carelli C., Jones R.A.L., Young R.N., Cubitt R., Dalgliesh R., Schmid F., Sfer-
razza M. (2005) Approaching criticality in polymer-polymer systems. Phys Rev
E. 72: 031807 (5 pages)
de Gennes P.G. (1971) Reptation of a polymer chain in the presence of fixed
obstacles. J Chem Phys 55: 572-579
de Gennes P.-G. (1975) Reptation of stars. J Phys (Paris) 36: 1199-1203.
de Gennes P.-G. (1989) Formation of a diffuse interface between two polymers.
CR Acad Sci Paris, II. 308: 13-17
de Gennes P.-G. (1989b) Weak adhesive junctions. J Phys France 2551-2562
Distler D., Kanig G. (1978) Fine structure of polymers from aqueous dispersion.
Colloid Polym Sci 256: 1052-1060 (in German)
Doi M., Edwards S.F. (1986) The Theory of Polymer Dynamics. Clarendon Press:
Oxford.
Feng J.R., Winnik M.A. (1997) Effect of water on polymer diffusion in latex
films. Macromolecules 30: 4324-4331.
Feng J.R., Pham H., Stoeva V., Winnik M.A. (1998) Polymer diffusion in latex
films at ambient temperature. J Polym Sci Pt B Polym Phys 36: 1129-1139.
Feng J.R., Odrobina E., Winnik M.A. (1998b) Effect of hard polymer filler
particles on polymer diffusion in a low-Tg latex film. Macromolecules 31:
5290-5299.
Ferry J.D. (1980), Viscoelastic properties of polymers, third edition, John Wiley
and Sons
Geurts J.M., van Es J.J.G.S., German A.L. (1996) Latexes with intrinsic crosslink
activity. Prog Org Coatings 29: 107-115
Hahn K., Ley G., Schuller H., Oberthr R. (1986) On particle coalescence in latex
films. Colloid Polym Sci 264: 1092-1096.
Haley J.C., Liu Y., Winnik M.A., Lau W. (2008) The onset of polymer diffusion
in a drying acrylate latex: how water initially retards coalescence but ultimately
enhances diffusion. J Coat Technol Res 5(2): 157-168.
Joanicot M., Wong K., Richard J., Maquet J., Cabane B. (1993) Ripening of
cellular latex films. Macromolecules 26: 3168-3175
Jud K., Kausch H.H., Williams J.G. (1981) Fracture mechanics studies of crack
healing and welding of polymers. J Mater Sci 16: 204-210
Juhu D., Wang Y., Winnik M.A., Haley F. (1993) Influence of a coalescing aid
on polymer diffusion in poly(butyl methacrylate) latex films. Makromol Chem
Rapid Commun 14: 345-349
References 181

Juhu D., Lang J. (1994) Latex film formation in the presence of organic solvents.
Macromolecules 27:695-701
Karim A., Mansour A., Felcher G.P., Russell T.P. (1990) Short-time relaxation at
polymeric interfaces. Phys Rev B 42: 6846-6849
Kausch H.H., Petrovska D., Landel R.F., Monnerie L. (1987) Intermolecular
interaction in polymer alloys as studied by crack healing. Polym Eng Sci 27(2):
149-154
Kim H.B., Winnik M.A. (1994) Effect of surface acid group neutralisation on
interdiffusion rates in latex films. Macromolecules 27: 1007-1012
Kim H.B., Winnik M.A. (1995) Factors affecting interdiffusion rates in films
prepared from latex particles with a surface rich in acid groups and their salts.
Macromolecules 28: 2033-2041
Kim H.K., Wool R.P. (1983) A theory of healing at a polymer-polymer interface.
Macromolecules 16: 1115-1120
Kim K.D., Sperling L.H., Klein A., Wignall G.D. (1993) Characterisation of film
formation from direct mini-emulsified polystyrene latex films via SANS Mac-
romolecules 26: 4624-4631
Kim K.D., Sperling L.H., Klein A., Hammouda B. (1994) Reptation time,
temperature, and cosurfactant effects on the molecular interdiffusion rate dur-
ing polystyrene latex film formation. Macromolecules 27: 6841-6850
Kobayashi M., Rharbi Y., Winnik M.A. (2001) Effect of inorganic pigments on
polymer interdiffusion in a low-Tg latex film. Macromolecules 34: 1855-1863.
Kobayashi M., Rharbi Y., Brauge L., Cao L., Winnik M.A. (2002) Effect of silica
as fillers on polymer interdiffusion in poly(butyl methacrylate) latex films.
Macromolecules 35: 7387-7399
Liu R., Winnik M.A., Di Stefano F., Vanketessan J. (2001) Interdiffusion vs cross-
linking rates in isobutoxyacrylamide-containing latex coatings. Macromole-
cules 24: 7306-7314
Lodge T.P. (1999) Reconciliation of the molecular weight dependence of diffusion
and viscosity in entangled polymers. Phys Rev Lett 83: 3218-3221
McLeish T.C.B. (2002) Tube theory of entangled polymer dynamics. Adv. in
Phys. 51: 1379-1527
Mohammed S., Daniels E.S., Sperling L.H., Klein A., El-Aasser M.S. (1997)
Isocyanate-functionalised latexes: Film formation and tensile properties. J Appl
Polym Sci 66: 1869-1884
Nguyen T.Q., Kausch H.H., Jud K., Dettenmaier M. (1982) Crack-healing in
crosslinked styrene-co-acrylonitrile. Polymer 23: 1305-1310
Oh J.K., Yang J., Tomba J.P., Rademacher J., Farwaha R., Winnik M.A. (2003)
Molar mass effect on the rate of polymer diffusion in poly(vinyl acetate-co-
butyl acrylate) latex films. Macromolecules 36: 8836-8845
Oh J.K., Tomba P., Ye X., Eley R., Rademacher J., Farwaha R., Winnik M.A.
(2003b) Film formation and polymer diffusion in poly(vinyl acetate-co-butyl
acrylate) latex films. Temperature dependence. Macromolecules 36: 5804-
5814.
182 5 Molecular Diffusion Across Particle Boundaries

Papakonstantopoulos G.J., Yoshimoto K., Doxastakis M., Nealey P.F., de Pablo


J.J. (2005) Local mechanical properties of polymeric nanocomposites. Phys
Rev E. 72: 031801 (6 pages)
Pham H.H., Winnik M.A. (2000) Synthesis, characterisation, and stability of
carbodiimide groups in carbodiimide-functionalised latex dispersions and
films. J Polym Sci: Part A: Pol Chem 38: 855-869
Pham H.H., Farinha J.P.S., Winnik M.A. (2000b) Crosslinking, miscibility, and
interface structure in blends of poly(2-ethylhexyl methacrylate) copolymers.
An energy transfer study. Macromolecules 33: 5850-5862
Prager S., Tirrell M. (1981) The healing process at polymer-polymer interfaces. J
Phys Chem 75: 5194-5198.
Rharbi Y., Bou F., Joanicot M., Cabane B. (1996) Deformation of cellular
polymeric films. Macromolecules 29: 4346-4359
Russell T.P., Deline V.R., Dozier W.D., Felcher G.P., Agrawal G., Wool R.P.,
Mays J.W. (1993) A direct observation of reptation at polymer interfaces. Na-
ture 365: 235-237
Schnell R., Stamm M., Creton C. (1998) Direct correlation between interfacial
width and adhesion in glassy polymers. Macromolecules 31: 2284-2292
Sferrazza M., Xiao C., Jones R.A.L., Bucknall D.G., Webster J., Penfold J. (1997)
Evidence for capillary waves at immiscible polymer/polymer interfaces. Phys
Rev Lett 78: 3693-3696
Stamm M., Httenbach S., Reiter G., Sprinter T. (1991) Initial stages of polymer
interdiffusion studied by neutron reflectometry. Europhys Lett 14: 451-456.
Tamai T., Pineng P., Winnik M.A. (1999) Effect of cross-linking on polymer
diffusion in poly(butyl methacrylate-co-butyl acrylate) latex films. Macromole-
cules 32: 6102-6110
Taylor J.W., Winnik M.A. (2004) Functional latex and thermoset latex films. JCT
Research 1(3) 163-190
Tronc F., Liu R., Winnik M.A., Eckersley S.T., Rose G.D., Weishuhn J.M.,
Meunier D.M. (2002a) Epoxy-functionalised, low-glass-transition-temperature
latex. I. Interdiffusion versus crosslinking in the presence of a diamine. J Polym
Sci: Pt A: Pol Chem 40: 2609-2625
Tronc F., Chen W., Winnik M.A., Eckersley S.T., Rose G.D., Weishuhn J.M.,
Meunier D.M. (2002b) Epoxy-functionalised, low-glass-transition-temperature
latex. II. Interdiffusion versus crosslinking in the presence of a diamine. J Po-
lym Sci: Pt A: Pol Chem 40: 4098-4116
Turshatov A., Adams J., Johannsmann D. (2008) Interparticle contact in drying
polymer dispersions probed by time resolved fluorescence. Macromolecules
41: 5365-5372
Voyutski S.S. (1958) Amendments to the papers by Bradford, Brown and co-
workers: Concerning mechanism of film formation from high polymer disper-
sions. J Polym Sci 32: 528-530
Voyutski S.S., Ustinova Z.M. (1977) Role of autohesion during film formation
from latex. J Adhesion 9: 39-50.
References 183

Wang Y., Winnik M.A. (1993) Polymer diffusion across interfaces in latex films.
J Phys Chem 97:2507-2515
Williams, J.G. (1977) Fracture mechanics of polymers. Polym. Engin. Sci. 17:
144-149
Winnik M.A., Wang Y., Haley F. (1992) Latex formation at the molecular level:
The effect of coalescing aids on polymer diffusion. J Coatings Techn 64(811):
51-61
Wool R.P., OConnor K.M. (1982) Time dependence of crack healing. J Polym
Sci: Pol Lett Ed 20: 7-16
Wu J., Tomba J.P., Winnik M.A., Farwaha R., Rademacher J. (2004a) Effect of
gel content on polymer diffusion in poly(vinyl acetate-co-dibutyl maleate) latex
films. Macromolecules 37: 4247-4253.
Wu J., Tomba J.P., Winnik M.A., Farwaha R., Rademacher J. (2004b) Tempera-
ture dependence of polymer diffusion in poly(vinyl acetate-co-dibutyl maleate)
latex films. Macromolecules 37: 2299-2306.
Yoo J.N., Sperling L.H., Glinka C.J., Klein A. (1990) Characterisation of film
formation from polystyrene latex particles via SANS 1. Moderate molecular
weight. Macromolecules 23: 3962-3967
Yoo J.N., Sperling L.H., Glinka C.J., Klein A. (1991) Characterisation of film
formation from polystyrene latex particles via SANS 2. High molecular weight.
Macromolecules 24: 2868-2876
Yoo S., Harelle L., Daniels E.S., El-Aasser M.S., Klein (1995) Interfacial aspects
of strength development in poly(methyl methacrylate)-based latex systems. J
Appl Polym Sci 58: 367-374
Zhang H.Z., Wool R.P. (1989) Concentration profile for a polymer-polymer
interface. 1. Identical chemical composition and molecular weight. Macro-
molecules 22: 3018-3021
Zosel A., Heckmann W., Ley G., Mchtle W. (1990) Influence of chemical
heterogeneity on structure and properties of emulsion copolymers. Makromol
Chem Macromol Sym 35/36: 423-446
Zozel A., Ley G. (1993) Influence of crosslinking on structure, mechanical
properties and strength of latex films. Macromolecules 26: 2222-2227
Chapter 6
6 Surfactant Distribution in Latex Films

6.1 Introduction

Colloidal polymer particles are made by techniques of emulsion polymerisation,


as described in Chapter 1. To emulsify the monomer phase in conventional
emulsion polymerisation, and to create micelles that serve as seed particles, one or
more surfactants are used. Surfactants are amphiphilic molecules that have a
hydrophilic head group and a hydrophobic tail, causing them to adsorb at inter-
faces and to self-assemble. They impart colloidal stability (through steric or
charge effects) and encourage monodispersity of particle size. Although it is
possible to perform surfactant-free emulsion polymerisation (Goodwin 1971),
which avoids the complications of surfactant migration and segregation, it is
difficult to obtain high volume fraction latices with this method. Hence, industrial
latices typically contain a large amount of surfactant (on the order of 2 wt. % of
the polymer). Surfactants are usually classified as either non-ionic, cationic or
anionic, depending on the charge on the hydrophilic head of the molecule.
An often desired scenario is for uniform distribution of surfactant through a
film, although this is rarely observed in practice. This chapter examines how and
why surfactants are distributed non-uniformly in latex films. Previous studies into
surfactant stratification are summarised, as well as the experimental techniques
commonly used to observe the distribution. A model is presented for predicting
the distribution of surfactant, by following the convective and diffusive mass
transport. This allows the distributions to be controlled through the adsorption
isotherm of surfactant on the particle surfaces.
There is a large and extensive literature relating to surfactant distributions in
latex films. The earliest work dates back to 1936 (Wagner and Fischer 1936) and
an overview of twentieth century research is provided by Hellgren et al. (1999).
As will be discussed in this chapter, surfactants are implicated in diminishing film
properties and performance, and are suspected of influencing the film formation
process itself. This chapter aims to correlate and explain the vast array of experi-
mental observations.
186 6 Surfactant Distribution in Latex Films

An example of non-uniform surfactant distribution in an adhesive film is shown


in Fig. 6.1. Despite being film formed at 60C (over 100C above the glass
transition temperature of 45C), the latex particles are not coalesced but remain
as discrete entities. Upon rinsing the film with water, however, the particle
boundaries become blurred, as shown in Fig. 6.1b, as polymer interdiffusion starts
to take place. The conclusion (supported by chemical analysis) is that a water-
soluble component consisting of surfactants, polymers and oligomers in the latex
serum phase is partitioning between the particles and preventing them from
coalescing. This is an instance where the surfactant distribution affects the final
morphology at a film surface.

a b

Fig. 6.1 Phase contrast AFM images of an acrylic latex (based on a EHA-BMA-MMA copoly-
mer) film-formed at 60C for three minutes. The copolymer Tg is about 45C. a An image of
the film as-dried, showing the particles hindered in their deformation and coalescence. b An
image of the same film is shown after rinsing with water. The particle boundaries are less distinct
as a result of particle coalescence. (Images courtesy of J. Mallgol)

6.1.1 Where Can Surfactant Go in a Dried Film?

A useful starting point is to consider the possible fate of a surfactant in a dried


film. Fig. 6.2 shows schematically five possibilities for the final location of
surfactant. For instance, when various types of surfactant (anionic, cationic and
ionic) are added to a latex dispersion, the surfactant might be enriched at the air
interface, at the substrate interface, or at both (Zhao et al. 1987, Belaroui et al.
2003). These two positions are shown as (i) and (ii), respectively, on the diagram.
Strongly absorbed surfactants will remain on the particle surfaces during film
formation and thereby be trapped along particle boundaries (iii). Trapped surfac-
tant is suspected of creating hydrophilic pathways in a film. Indeed, in nominally
dry latex films, water has been found to be bound to the ionic headgroups of
6.1 Introduction 187

surfactants at particle boundaries (Rottstegge et al. 2003). Such boundary layers


have been implicated in the faster rates of drying observed in the presence of
excess surfactant (Winnik and Feng 1996). In some latex systems, the surfactant is
pushed away from the particle boundaries to create aggregates or clumps, shown
as (iv). Micrometre-size clumps of ionic surfactant have been observed in dry
films with confocal Raman imaging (Belaroui et al. 2003) (see Fig. 6.3) and with
energy filtering TEM (Du Chesne et al. 1997). Finally, there are a few instances,
in which the surfactant is soluble in the polymer and is distributed throughout a
film (v). One example is the anionic surfactant, sodium dodecyl sulphate, in
poly(vinyl acetate) (Vijayendran et al. 1981). Plasticisation of acrylic latex films
has been attributed to the presence of non-ionic surfactants (Kientz and Holl 1993,
Eckersley and Rudin 1993), indicating at least some solubility of the surfactant in
the polymer phase.
It should be noted that the surfactant distribution in a latex film is not static.
After a film has been dried, the surfactant can continue to relocate and hence a
considerable aging effect can be seen (Hellgren et al. 1999). The process by which
surfactant is pushed to an interface is called exudation. The surfactant distribu-
tion is not only a function of the polymer and surfactant chemistry. For a particu-
lar latex, the distribution can vary depending on the film formation conditions. For
instance, films dried well above the MFFT have been found to have surfactant
clumps throughout the film, while films dried at and close to the MFFT had an
accumulation of surfactant at the film-air interface. The difference in surfactant
distributions, in this particular instance, was ascribed to different surfactant
mobility in the polymer phase as a function of temperature (Du Chesne et al.
1997). There is also evidence that polymer mobility has an influence on the
surfactant desorption and segregation process (Belaroui et al. 2003). It is only with
the development of an improved theoretical framework that some of this rich
variety of results can be explained.

Fig. 6.2 Possible locations of surfactant in a dry latex film. Segregated at interfaces with (i) air or
(ii) the substrate as either continuous or discontinuous layers; (iii) adsorbed along particle-
particle boundaries. (iv) Surfactant can aggregate into pockets. Finally, some surfactants are
partially miscible with the polymer phase and are (v) dissolved within the film matrix.
188 6 Surfactant Distribution in Latex Films

1 m

Fig. 6.3 Confocal microscopy imaging showing the distribution of sodium dodecyl surfactant
laterally within a latex film. The grey scale is proportional to surfactant concentration, with white
being the highest. Surfactant is seen to be concentrated in a region that is several m across.
(Reprinted from Belaroui et al. (2003), with permission from Elsevier)

6.1.2 Effect of Non-Uniform Surfactant Distributions

There are many detrimental consequences from non-uniform surfactant distribu-


tions. A number of these are explained below.

6.1.2.1 Gloss and Appearance

An experimental system comprising an acrylic polymer stabilised by sodium


dodecyl sulphate displayed an accumulation of surfactant at the polymer-air
interface (Amalvy and Soria 1996). One consequence of this accumulation
(position (i) in Fig. 6.2), is that the gloss of the coating may be diminished
(Asua and Schoonbrood 1998). The gloss of a coating is determined by the
flatness of the top surface. High gloss requires specular reflection in which the
angle of incident light equals the angle of the reflected light. If the surface is
6.1 Introduction 189

rough over distances comparable to the wavelength of visible light, then light
will be scattered in many directions rather than being reflected specularly
(Jarnstrom et al. 2008).
Hulden et al. (1994) show that the type of surfactant affects a coatings gloss.
The use of non-ionic surfactants leads to a higher gloss than when ionic surfac-
tants are used. Surfactants do not necessarily decrease the gloss. For instance,
Heldmann et al. (1999) show an increase in gloss for a vinyl acetate coating with
an increasing amount of non-ionic poly(ethylene oxide) surfactant. The explana-
tion is complicated because the increase in surfactant concentration leads to a
decrease in particle size and also a narrower particle size distribution. This effect
facilitates denser particle packing and the likelihood of a smoother film surface. It
was also proposed that a surfactant layer at the coating surface could fill the
valleys between particles and decrease roughness. Reflection from the surfactant-
polymer interface will not occur if the refractive indices are matched.
Keslarek et al. (2002) showed that films from extensively dialysed latex, where
surfactant is removed, have a smoother surface when compared to the correspond-
ing films made from the surfactant-containing latex. The effect is likely to result
from the influence of surfactant on particle packing. Particle packing has been
found to be densest when particles are stabilised by surfactant (Juhu and Lang
1993). Topographical feature sizes approaching the wavelength of light will lead
to significant off-specular reflection and a loss of gloss. A further consequence of
accumulation at the film-air interface is surface tension gradients driving lateral
transport. This results in a surface rippling (Gundabala 2008), which will be
detrimental to the appearance of the film. This phenomenon is discussed in
Section 6.4.

6.1.2.2 Aesthetic Qualities and Dirt Pick-Up

An example of the importance in understanding the accumulation of surfactant at


the film-air interface is found in the art world. Acrylic latex paints have been used
in modern art since the 1960s. Celebrated artists, including Pablo Picasso, Cy
Twombly and Jackson Pollock, are known to have used household paints in their
artwork, and the worlds art museums hold an enormous collection of paintings
having latex binders (Learner 2005). Art conservators have found that surfactant
exudation to a latex coating surface has detrimental effects on the gloss and
aesthetic qualities of the paintings. Furthermore, tacky surfactant layers have been
implicated in increasing the pick-up of dirt and various contaminants and in the
growth of mould (Jablonski et al. 2003). The omnipresence of surfactant has
raised questions about whether and how to clean the painting surfaces (Ormsby et
al. 2006).
190 6 Surfactant Distribution in Latex Films

6.1.2.3 Adhesion and Viscoelasticity

Excess surfactant at either film interface can have dramatic effects on the film
performance. Pressure-sensitive adhesives (PSAs) offer a good example. One
measure of the adhesion of a PSA to a substrate is its peel strength. This is the
stress required to pull the film from the substrate and is necessarily a function of
time, temperature and peeling rate. Yang et al. (2003) looked at the water resis-
tance and peel strength of acrylic PSAs with different surfactants and related their
results to different amounts of surfactant segregation. Zosel and Schuler (1999)
showed that different surfactants in poly (ethyl hexyl methacrylate) films result in
different peel strengths. A series of papers from the group of Holl (Charmeau et al.
1996, 1999, 1999b, Gerin et al. 1999) looked at the adhesion of latex films to a
substrate and showed that a layer of surfactant at the film-substrate interface
(location (ii) in Fig. 6.2) dramatically reduces the peel strength.
The adhesion energy required to de-bond a PSA film is dependent largely on
the bulk viscoelastic properties of the polymer and not just the interface bonding.
Therefore, the observed effects of non-ionic surfactants, such as plasticisation
(Eckersley and Rudin 1993) and accelerated interdiffusion rates (Farinha et al.
1996), will undoubtedly influence the adhesion energy and other mechanical
properties. Phase-separated surfactants within latex films have also been impli-
cated in the reduction of tensile strength (Bindschaedler et al. 1987).

6.1.2.4 Barrier Properties and Water Whitening

A pathway of surfactant will lead to enhanced transport of water. Accumulation of


surfactant at the particle boundaries could produce such a route and this is
probably the worst case scenario for surfactant arrangement (location (iii) in Fig.
6.2). The pathway will result in hydrophilic routes for moisture, making the film a
poor barrier. In many cases, the coating is applied to protect the substrate, and this
channelling constitutes a severe failure of the coating. Examples of surfactant
causing enhanced moisture transport are given by Steward et al. (1995) and
Roulstone et al. (1992a and 1992b).
Any lumps of surfactant (location (iv) in Fig. 6.2) will create hydrophilic pock-
ets within a film that will contribute to water absorption. Butler et al. (2004, 2005)
examined the effect of surfactant type on the uptake of water into poly (butyl
acrylate-co-methyl methacrylate) films. Surfactant migration, segregation,
interaction with the polymer, and polarity were all shown to be relevant factors. It
was proposed that surfactant molecules can invert to create pockets in which the
hydrophilic heads of surfactant point inwards. When the surfactant boundary
layers break up, the barrier resistance increases.
It has been shown that more deformable polymers (at temperatures above their
glass transition temperature) are more prone to create pockets of water in compari-
son to more rigid polymers (Agarwal and Farris 1999). Any pockets of water will
display a lower refractive index in comparison to the bulk polymer and lead to the
6.1 Introduction 191

scattering of light. The film will appear hazy or cloudy. Water whitening, as the
effect is sometimes called, is particularly problematic in pressure-sensitive
adhesive applications, such as labels and decals (Donkus 2007). Water pockets
ranging in size from 25 to 200 nm have been found to create opacity in adhesives.

6.1.2.5 Film Formation Process

Vandezande and Rudin (1996) found that the surfactant type has a significant
effect upon film drying. Using an ethoxylated nonyl phenol surfactant that
plasticised their latex particles, Vanderxande and Rudin show that the MFFT
temperature increases as the surfactant chain length is increased. This is because
of reduced permeation of long chain surfactants into the latex particles. In
addition, because surfactants influence the colloidal stability of latex particles,
they will affect particle packing. The packing of particle flocs creates a more open
and rougher film surface (Juhu and Lang 1993). Subsequently, plasticisation of
the polymer can increase the amount of particle deformation (Eckersley and Rudin
1993). Plasticisation effects can also explain an observed decrease in the minimum
film formation temperature (MFFT) of hydrophilic latex films in the presence of
ionic surfactant, whereas anti-plasticisation might explain an increase in the
MFFT in a non-polar hydrophobic latex. It has also been argued that these effects
on MFFT are related to changes in the polymers interfacial energy and hence the
driving force for particle deformation (Powell et al. 1968).

6.1.3 Mechanisms of Surfactant Transport

As discussed, surfactant exudation can occur in dry films. Yet, the distribution of
surfactant immediately after a film is formed is determined by the surfactant
transport in the wet state. Fig. 6.4 shows the possible location for surfactant in a
wet film. The molecules can be
freely diffusing in solution (positions i and ii),
adsorbed at interfaces (iii and iv),
adsorbed or attached to particles (v).
Accordingly, there are multiple transport mechanisms for surfactant.
At concentrations below the critical micelle concentration (CMC), free surfac-
tant molecules exist mainly as unimers, but at values above it, they create mi-
celles. Unimers and micelles are both subject to convective and diffusive flows.
The surfactant on the latex particles will obey an adsorption isotherm linking the
amount adsorbed to the concentration in solution. Likewise, the amount adsorbed
at the interfaces with the substrate and air will be determined by an adsorption
isotherm. Interfaces will be saturated at concentrations above the CMC. Further-
more, the particles will convect and diffuse, thus carrying surfactant with them.
192 6 Surfactant Distribution in Latex Films

Factors that influence the distribution of particles in a drying latex film are
presented in Chapter 3.
A common argument is that surfactant exudation is driven by a reduction in the
interfacial energy (Urban 1997). The magnitude of this mechanism, however, will
lead only to a low interface accumulation (Kientz and Holl 1993). Surfactants
form a monolayer (or sometimes a bilayer) at an interface. For a surface area of 1
m2, we can estimate the number of adsorbed surfactant molecules as 1018; it is
common to assume an adsorbed area of 1 nm2 per molecule (Cussler 1984). This
translates to 1.6 mole per m2, and for sodium dodecyl sulphate it corresponds to
0.5 mg of surfactant adsorbed at the interface. For a film of final dried thickness
100 m, the surfactant at the interface corresponds to approximately 5 x 103 wt%.
This is far less than the amount of surfactant typically added to the system. It can
be concluded that the large amount of segregation commonly observed in fresh
films is due to other driving forces, such as the convection and diffusion outlined
above, and not to a lowering of interface energy. Interface accumulation of
surfactant over time in aged films is the result of a phase separation.

Fig. 6.4 Possible locations of surfactant in a wet latex film. Freely diffusing in the aqueous phase
as (i) unimers or (ii) micelles; Adsorbed at the interface between the aqueous phase and (iii) air,
(iv) the substrate, or (v) a latex particle. There is a dynamic equilibrium such that the positions of
surfactant molecules are constantly changing.

6.2 Adsorption Isotherms

Although the free surfactant is subject to diffusion in the aqueous phase, the adsorbed
surfactant is carried along with the diffusing particles. It is necessary therefore to
consider what determines the amount of surfactant to be transported with the
particles. At equilibrium, the amount of adsorbed surfactant will be set by the
concentration of surfactant in the aqueous phase. There is a thermodynamic balance
6.2 Adsorption Isotherms 193

between the amount of adsorbed surfactant and the amount in solution, described by
the adsorption isotherm, with an equal exchange between the two states.
There are a number of models for adsorption isotherms, and the simplest is the
Langmuir isotherm. The rate of surfactant adsorption is assumed here to be
proportional to the solution concentration of surfactant, Cs, and the proportion of
free sites on the particles. Hence, it is given by k Cs ()/, with being the
adsorbed surface concentration (expressed in mol m2), the maximum amount
that can be adsorbed on the particles, and k is a constant of proportionality. The
rate of surfactant desorption is assumed to be proportional to the number of
occupied sites on the particles: k /. Equating the two rates provides the
adsorption isotherm as
Cs
= (6.1)
A + Cs

where the critical concentration

Ak'
k

relates to the bulk concentration (in mol m3) at which the particles have a surface
concentration of half their saturated value. A typical Langmuir isotherm is
sketched in Fig. 6.5. For a given latex and surfactant system, the values of A and
can be readily determined from experimental data. Typically, a known mass of
particles is mixed in a solution of known surfactant concentration and volume.
After equilibration is achieved, the particles are centrifuged out. The concentration
of surfactant remaining in solution is then used to calculate the amount adsorbed
on the particle surfaces (Gundabala 2004).


(mol m-2)

A
Cs (M)

Fig. 6.5 Typical plot of a Langmuir adsorption isotherm. The maximum loading of surfactant is
given the symbol and the bulk concentration at which the particles have half their maximum
loading is called the critical adsorption concentration, A.
194 6 Surfactant Distribution in Latex Films

6.3 Modelling of Surfactant Distribution during the Drying Stage

A model aids in the understanding of what determines the surfactant distribution


during film formation. Consider a latex film that is initially spread in the liquid
state and allowed to dry. Fig. 6.6 sketches the geometry. The top surface of the
film reduces in height because of evaporation. The free surfactant is at a concen-
tration of Cs in the serum. The movement and distribution of the surfactant is
determined by the transport of the surfactant via convection and diffusion.
Convection refers to flow associated with a moving fluid, and diffusion refers to
thermal motion of molecules in the absence of net flow. There may also be
thermal gradients and the corresponding vortices, leading to surfactant motion
(Tiwari 2007).

Fig. 6.6 Sketch of a wet film applied to a substrate. The surfactant is either free in solution or
attached to particles. The amount of surfactant at the film-air and film-substrate interfaces is
small and so is ignored for the purpose of the mass balance. An adsorption isotherm links
the amount of surfactant adsorbed on the particle surfaces, , to the amount of surfactant in
solution, Cs.

The surfactant is adsorbed on the particles at a concentration of . The particles


are likewise free to convect and diffuse, and they will carry adsorbed surfactant.
Hence, non-uniformity in the particle distribution during the drying stage will
contribute to a non-uniform surfactant distribution. Finally, as water evaporates,
the surfactant concentration in the serum, Cs, will naturally rise. The system is
dynamic. This higher concentration will drive adsorption onto the latex particles,
as described by the adsorption isotherm, such as in (6.1).
To write a conservation equation for surfactant, it is necessary to describe the
motion of both the particles and the surfactant. Gundabala et al. (2004) wrote a
one-dimensional surfactant balance equation that accounted for the diffusion of
both of these components. For particles of radius R at a particle volume fraction of
, their equation is
6.3 Modelling of Surfactant Distribution during the Drying Stage 195

(
Cs (1 ) + 3
R ) = D (1 ) C + 3D ( ) ( )
s p

t y y y y
s
R
(6.2)

In this case, y represents the direction normal to the substrate. The term on the
left-hand side accounts for the accumulation of surfactant over time t, either in the
serum with a concentration Cs or attached to particles with an adsorbed amount .
The first term on the right-hand side corresponds to the diffusion of free surfactant
in the serum phase. It is explicitly assumed that there is a single diffusion coeffi-
cient for surfactant, Ds, and so different transport properties of unimers in com-
parison to micelles are ignored. This is a reasonable approximation, since the
lower diffusion coefficient of a micelle is countered by the increased number of
monomers in the micelle. Hence, on a molar basis of monomers, the error is small
(Cussler 1984). The final term on the right-hand side relates to the diffusion of the
latex particles. The diffusion coefficient of colloidal particles, Dp, is strongly non-
linear, especially near a phase transition, and hence the volume fraction depend-
ence of the diffusion must be considered (Routh and Zimmerman 2004).
The classical way to attempt the solution of (6.2) is to scale the variables and to
extract the relevant dimensionless groups. In exactly the same strategy as em-
ployed in Section 3.4 for the one-dimensional drying of colloidal dispersions, the
relevant time scale is taken as the time for evaporation, H / E where H is the film
thickness and E is the evaporation rate in m/s. Two dimensionless groups appear,
which relate the relative magnitudes of the diffusion rates (surfactant (s) or
particles (p)) to the evaporative-induced convection rate. These relations are
captured in the Peclet numbers for the surfactant and the particles:

HE
Pes =
Ds
(6.3)
HE
Pe p =
Dp

The surfactant Peclet number, Pes compares the rate of surfactant diffusion to the
evaporation rate. For large Pes, the diffusion rate is slow in comparison to
evaporation, and a gradient in surfactant concentration vertically across the film is
expected. For Pes <<1, the diffusion is strong compared to evaporation, and
uniform vertical profiles of surfactant are expected. The particle Peclet number,
Pep, is exactly as defined in Chapter 3 Eq. (3.2), and determines the vertical
profile of particles during drying.
The other key system parameter is the adsorption isotherm. Gundabala et al.
(2004) assumed a Langmuir isotherm as given by (6.1). In addition, they assumed
that the system is in equilibrium at all times. The validity of this assumption depends
on the surfactant mobility compared to the time scale of the problem, namely the
196 6 Surfactant Distribution in Latex Films

evaporation time. There must be sufficient time for the surfactant to retain equilib-
rium. Hence, it is more likely to be valid when films are dried slowly.
Fig. 6.7 to 6.9 display solutions obtained from solving Eq. (6.2). Solutions were
obtained using a commercial finite element package (Comsol), although any
numerical approach will work. Results are presented at the point when water
evaporation has been completed. (Any surfactant movement in the dry state is not
considered in the model.) The results are plotted as the percentage of surfactant
excess or depletion against the scaled distance from the substrate. The surfactant
excess and depletion is determined as the surfactant concentration at any point
compared to the value that would have occurred if the concentration profile had
been uniform with depth. For example, if the average surfactant concentration is 1
mol/m3 and the concentration at some point is 2 mol/m3, the surfactant excess is
100%. The scaled distance relates to vertical distances in the film divided by the
film thickness. Thus, a value of unity refers to the film-air interface, and a value of
zero indicates the substrate interface.
Fig. 6.7 displays the final surfactant profile from solution of Eq. (6.2) for a
range of surfactant Peclet numbers and a fixed adsorption isotherm. As is intui-
tively pleasing, low Peclet numbers, which relate to large values of the surfactant
diffusion coefficient, result in uniform surfactant profiles, while a surfactant Peclet
number of 10 results in a 300% excess surfactant at the top surface of the film. It
is clear that the diffusion of surfactant acts to smooth out inhomogeneities. In this
particular example, the particle Peclet number was taken as infinite, although this
assumption was relaxed in subsequent work (Gundabala et al. 2006).

30 300
% Surfactant Excess/Depletion (PeS=0.1;PeS=1)

Pes=0.1
% Surfactant Excess/Depletion (PeS=10)

25 Pes=1
Pes=10
20 200

15

10 100

0 0

-5

-10 -100
0.0 0.2 0.4 0.6 0.8 1.0

Scaled Distance from Substrate

Fig. 6.7 Simulations of the final surfactant profiles for three different surfactant Peclet numbers,
Pes. In each case, the adsorption isotherm is the same for each simulation and the particle Peclet
number, Pep, is taken as infinite. Depletion and excess relates to the amount of surfactant
compared to the amount that would be present if the concentration were uniform across the film.
(Reproduced from Gundabala et al. 2004, copyright 2004 American Chemical Society)
6.3 Modelling of Surfactant Distribution during the Drying Stage 197

Figures 6.8 and 6.9 demonstrate the effect of the adsorption isotherm on the
final profile. If more surfactant is bound to the particles, then as the particles are
carried to the top of the film, surfactant is likewise transported. Consequently, the
amount of surfactant adsorbed on the particles, and the dependence on the
changing surfactant concentration in the serum phase Cs, has a pronounced effect
on the final distribution at the end of the drying stage. This effect is shown in Fig.
6.8 with the parameter inf quantifying the maximum amount of surfactant that can
be bound to the particles. inf is a non-dimensional version of the maximum in the
adsorption isotherm and is defined as inf = 3 /RCs0. The term Cs0 relates to the
original concentration of surfactant in solution. In Fig. 6.9, the critical surfactant
concentration at which adsorption occurs is quantified by the parameter A . This
parameter is a non-dimensional version of the critical surfactant concentration, A
and is defined as A = A / C s 0 . For a higher value of A , the particles initially carry
less surfactant, and hence transport less of it to the top surface. This results in a
more uniform distribution of surfactant in the final profile.

Fig. 6.8 Effect of maximum adsorbed amount on the surfactant distribution. More surfactant
adsorbed to the particles (large k) results in more surfactant at the top surface at the end of
drying. The particle Peclet number is taken Pep = and the surfactant Peclet number as Pes = 1.
Depletion and excess refers to the amount of surfactant compared to the amount that would be
present if the concentration were uniform across the film. (Reproduced from Gundabala et al.
2004, copyright 2004, American Chemical Society)

A striking feature of the results of the calculated profiles in Fig. 6.8 and Fig.
6.9 is that an excess of surfactant at the film-air interface is expected under
essentially any conditions. These results are in line with experimental observa-
tions, which almost always see surfactant accumulation at the film surface. The
198 6 Surfactant Distribution in Latex Films

implication is that end-users of latex films should never expect a clean film
surface. On the other hand, it is certainly possible to obtain a depletion of surfac-
tant near the interface with the substrate. In particular, when the surfactant absorbs
on the particles as a dense layer ( is large), the surfactant is not expected to
accumulate at the substrate-film interface.

Fig. 6.9 Effect on the critical surfactant concentration, parameter A in Eq. (6.1), on the final
surfactant distribution. The particle Peclet number is taken Pep = and the surfactant Peclet
number as Pes = 1. Depletion and excess refers to the amount of surfactant compared to the
amount that would be present if the concentration were uniform across the film. (Reproduced
from Gundabala et al. 2004, copyright 2004 American Chemical Society)

The solution of Eq. (6.2) allows the prediction of surfactant distributions only
at the end of the water evaporation stage. Once the particles have come into
contact and begin to deform, further development of the surfactant profile is
possible. Conceptually at least, the transport of surfactant through the polymer
phase will be orders of magnitudes slower than through the aqueous phase,
although the time constraint in the evaporation stage will no longer apply. The
effects of non-uniformities during drying, therefore, are expected to be profound
and long-lasting. Tzitzinou et al. (1999) demonstrate that surfactant segregation at
the film surface in their system developed during the drying stage (as considered
in the model here), rather than in a later exudation stage in a dry film. The effects
of the transport during the drying stage persisted at later times.
A further complication arises during the deformation stage in that surfactant
can be desorbed from particle surfaces and hence become forced into the latex
serum. For instance, an anionic surfactant has been found to desorb into the
aqueous phase of latex over a period of hours during slow coalescence (Kientz et
6.4 Effect of Surfactants Vertical Distribution on Film Topography 199

al. 1994). Desorption will negate the adsorption isotherm argument expounded in
the model presented here and might cause localised regions of surfactant within
the film. To tackle the problem analytically, an initial condition for the distribution
of surfactant will be required. One possibility is to use the solution proposed by
Gundabala et al. (2004), which relates to the concentration profile at the point at
which the particles reach close packing. To date, no modelling has been done to
account for this forced exudation.
A discussion in the literature relates to the effect of surfactant desorption in the
development of concentration profiles. Kientz and Holl (1993) argue that surfac-
tant desorption is necessary so that the particles can deform and interdiffuse as
they form a continuous film. This desorption is predicted to create localised
regions of high surfactant concentration and lead to segregation of surfactant. In
experiments using environmental STEM, small regions of surfactant have been
observed to develop in thin latex films at the point where particle coalescence
occurs (Arnold et al. 2009). As particle coalescence takes place, surfactant appears
to be pushed away into clusters within the film.
Based on experimental studies using SANS, Belaroui et al. (2003b) proposed
that particle deformation increased particle-particle contact and therefore induced
surfactant desorption. Surfactant exudation to the surface was greater in a latex
film having a lower glass transition temperature in comparison to harder, less
deformable latex particles. This idea was given further support in recent experi-
ments in which the effect of aging temperature on surfactant exudation was
examined. At lower temperatures, at which the polymers were below their glass
transition temperature, there was less surfactant exudation (Kessel et al. 2008).
Upon heating films above this temperature, the rate of exudation increased.

6.4 Effect of Surfactants Vertical Distribution on Film Topography

If surfactant migrates to the wet film-air interface, it will lower the surface
tension. Non-uniformity in the lateral concentration of surfactant results in lateral
pressure gradients, which will drive flow from regions of low to high surface
tension. There is a large body of literature relating to such surface-tension driven
motions, and they are collectively called Marangoni flows. The flow most
directly relevant to latex films corresponds to fronts propagating away from areas
of high surfactant concentration. An excellent review of Marangoni flows is given
by Oron et al. (1997).
It is interesting to note that any vertical surfactant distribution will result in
lateral non-uniformities in film height. As outlined in Section 6.3 there is a
tendency for surfactant to accumulate at the film-air interface when either Pes or
Pep >> 1. The first report of such lateral surface non-uniformity was by Juhu et
al. (1995). The effect is demonstrated in Fig. 6.10 where an AFM image of a dried
film is shown. The height image shows many crater-like (or valley-like) struc-
tures, and these correspond to the regions of the film that are less energy dissipa-
200 6 Surfactant Distribution in Latex Films

tive, as indicated by the phase contrast image. The distance between these
undulations and their number across the surface are a strong function of the film
thickness, as shown in Fig. 6.11, where films with a range of height are imaged.
The peak-to-peak distance of the undulations increases with increasing film
thickness. The cause of the film height undulations is exudation of surfactant to
the film-air interface during drying. Marangoni instabilities then result in lateral
flow away from regions of higher surfactant concentration and the subsequent
thinning of the film in this region. This problem has recently been modelled, and
predictions for the spacing of the instabilities as a function of the adsorption
isotherm and film parameters have been made in agreement with the experimental
observations (Gundabala et al. 2008).

a b

Fig. 6.10 AFM a height and b phase-contrast images of a dried latex film showing surfactant
islands. The surfactant is detected as lighter regions in the phase-contrast image and as lower
regions in the height image. Image sizes are 20 m x 20 m. (Images reprinted from Gundabala
et al. 2008; copyright Wiley-VCH Verlag GmbH & Co. KGaA; reproduced with permission)

Fig. 6.11 Phase contrast AFM images for a range of films with increasing wet thickness, from
left to right: 30 m, 60 m, 90 m, and 120 m. All image sizes are 100 m x 100 m. Notice
how both the size and number of the film surface corrugations scale with the original film
thickness. (Image reprinted from Gundabala et al. 2008; copyright Wiley-VCH Verlag GmbH &
Co. KGaA; reproduced with permission)
6.5 Experimental Evidence for Surfactant Locations 201

6.5 Experimental Evidence for Surfactant Locations

Chapter 2 outlines experimental techniques used to study latex film formation.


This section lists a number of techniques that are specifically useful in examining
surfactant concentrations at different locations within the film. It is inevitable that
some overlap with Chapter 2 will occur, but we try here to concentrate on the
information obtained and its interpretation, rather than the technique per se.

6.5.1 Interfaces with Air and Substrates

Some techniques are designed to find the type and the amount of surfactant at a
film-air (or vacuum) interface. Experimentally, a number have been employed for
this purpose. X-ray photoelectron spectroscopy (XPS) and secondary ion mass
spectrometry (SIMS) provide chemical information on the first few nm of the
surface under study (Zhao et al. 1988). A surfactant excess can be determined by
comparison to the bulk composition. Rinsed surfaces and dialysed latex (in which
the surfactant has been removed) are good comparators. An estimate of the layer
thickness can be made either by varying the angle of the incident beam or by
milling the surface. Fourier transform infrared (FTIR) spectroscopy allows a
greater depth to be analysed, but provides a single value for the concentration near
the top surface (Kientz et al. 1993, Kientz and Holl 1993, Zhao and Urban 2000,
Amalvy and Soria 1996). A series of papers from the group of Urban uses FTIR
techniques to examine surfactant exudations at the surface of a range of latex films
(Thorstenson 1993, Dreher et al. 2003, Dreher et al. 2005, Lestage et al. 2005).
Information about the lateral distribution of surfactant across a film surface can be
ascertained using atomic force microscopy (AFM), where the use of phase-
contrast imaging yields a strong image contrast between surfactant and polymer
phases (Mallgol et al. 2002), as demonstrated in Fig. 6.1. The combination
of AFM with spatially resolved SIMS allows unambiguous chemical mapping
of the top surface and simultaneous comparison with AFM phase images
(Gundabala 2008).
In addition to examining the air interface, a common practice is to peel the film
from its substrate and then to probe the interface formerly in contact with the
substrate. Despite the simplicity of this approach, there is a problem in the
possibility of surfactant loss to the substrate, and hence the possibility of an
inaccurate measurement of concentration.
202 6 Surfactant Distribution in Latex Films

6.5.2 Surfactant in the Bulk of the Film

Freeze-fracture TEM allows the arrangement of particles in the film to be fol-


lowed through the drying process (Wang et al. 1992), and surfactants have been
found to alter the arrangement of particles within drying films. Energy-filtering
TEM provides unambiguous evidence for surfactant clumps within the bulk of
films (Du Chesne et al. 1997). Surfactants can also be detected using NMR
spectroscopy (Rottstegge et al. 2003), and the use of deuterium-labelled surfactant
allows analysis using small-angle neutron scattering (Belaroui et al. 2003).

6.5.3 Depth Profiling and Mapping

It is desirable to employ techniques that provide concentrations as a function of


depth into the film and laterally across it. Fourier transform Raman spectroscopy
(Zhao and Urban 2000b; 2000c) provides information on symmetric bands in the
vibration spectra. Hence, quantitative information on surfactant concentration is
obtained. Arnold and Holl (Arnold 2009) use confocal Raman microscopy to
detect surfactant concentrations through a films thickness. By taking measure-
ments at various locations, they demonstrate both vertical and lateral inhomoge-
neities in surfactant concentrations. Representative results are shown in Fig. 6.12.
At all lateral positions, the vertical profiles show an accumulation of surfactant
towards the top (air) interface and then reasonably uniform concentration profiles
from about 5 m below the surface. The concentration of surfactant in the centre
of the film is higher than at the edge. This is counter-intuitive since it can be
expected that the horizontal drying will carry surfactant towards the edge. One
likely mechanism opposing such an outward flow could be Marangoni flows that
result from surfactant concentration gradients. (Section 6.4 gives more informa-
tion about Marangoni flows). Confocal Raman spectroscopy also provides
compositional data into the film depth. There is evidence that, in some instances,
surfactant gathers in large pockets within the bulk of latex films. Confocal Raman
microscopy has provided vivid evidence for such an effect (Belaroui et al. 2003),
as shown in Fig. 6.3.
Rutherford backscattering spectrometry (RBS) has been employed by numer-
ous researchers to investigate surfactant concentration as a function of depth into
the top few hundred nanometres of a film (Tztitzinou et al. 1999, Aramendia et al.
2003, Mallgol et al. 2002, Lee et al. 2006). The energy of helium ions after the
elastic scattering from a surface provides the elemental composition, and the
energy loss of the ions passing through a polymer matrix allows an elemental
depth profile to be acquired. Fig. 6.13 shows a typical RBS spectrum for a
sulphur-free latex film with post-added lithium dodecyl sulphate surfactant. A
signal from sulphur can only be attributed to the surfactant. As can be seen, the
result is a surfactant structure comprising about 50 mol. % surfactant in the top
6.5 Experimental Evidence for Surfactant Locations 203

100 nm of the film. Beneath this layer, there is a much lower and uniform concen-
tration of surfactant in the bulk of the film, at least up to the depth limit of the
RBS depth profiling, which is a few hundred nm. The thickness of the surfactant-
enriched layer is far more than is expected from a surfactant monolayer to reduce
the interface energy. Hence, this is evidence for surfactant transport during drying.

Fig. 6.12 Sodium dodecyl sulphate (SDS) concentration profiles obtained from confocal Raman
microscopy of a latex film comprising a butyl acrylate-methyl methacrylate copolymer. Depth
profiles are taken at two different locations in the film. The dashed line indicates 6 wt.% SDS,
which is the concentration that would be present if the concentration profile was uniform
throughout the film. A profile taken near the centre (point A) has a higher surfactant concentra-
tion than the profile taken near the edge (point B). Images courtesy of Arnold and Holl.
204 6 Surfactant Distribution in Latex Films

latex

surfactant

Fig. 6.13 (top) Typical RBS spectrum for a film doped containing 0.72 wt. % Lithium surfactant.
The data are shown as points, and the line represents the predictions of a best-fit model of the
surfactant profile. (bottom) The depth profiles are obtained by analysis of the spectra. Notice the
large surfactant excess in the first hundred nm of the film then the much lower constant
concentration into the bulk of the film. Reproduced with permission from Lee et al. 2006.
Copyright 2006 American Chemical Society.

The RBS technique is not readily able to probe lateral inhomogeneities. How-
ever, chemical maps of latex film surfaces, obtained with secondary ion mass
spectrometry (SIMS), provide evidence for island-like clusters of surfactant on
film surfaces (Gundabala et al. 2008).

6.6 Reactive Surfactants

Instead of being bound to the particle surface through electrostatic or van der
Waals interactions, a reactive surfactant is covalently bound to the polymer phase.
Chemical bonds prevent the possibility of desorption and exudation of surfactant
at least without the grafted polymer also moving.
6.6 Reactive Surfactants 205

6.6.1 Reactive Surfactant Chemistry

There are three main ways to chemically bind a surfactant to a polymeric surface.
In each approach, a chemical bond is created between an aliphatic chain and the
polymer chains at the particle surface.
An initiator, such as a temperature-cleavable cyanopentanoic acid group, can
provide a reactive group at the surfactants tip. Polymerisation will then be
initiated by the surfactant, and the polymer chains will grow with the surfactant
at one chain end. These types of reactive surfactants, wherein bonds between
surfactant and polymer are made with the help of an initiator group, are called
inisurfs.
Use of a chain transfer agent, such as a thiol group, will enable bond formation
with a growing polymer chain. These types of surfactant have been termed
transurfs.
The third possibility, and the one that has received the most attention, relies on
the use of a monomer group within the surfactant. Examples include maleates
and crotonates. This type of molecule is known as a surfmer. The use of
surfmers in polymerisations is reviewed by Asua and Schoonbrood (1998), and
there is an extensive literature concerning the synthesis of surfmers (Abele et
al. 1999, Montoya-Goni et al. 1999, Unzue et al. 1997, Sindt et al. 2000,
Thenoz et al. 1999, Guyot et al. 1999), which emerged from a European Com-
mission project on the topic. Sketches of chemical structures that are examples
of inisurfs, transurfs and surfmers are shown in Fig. 6.14.

6.6.2 Effect of Surfmers on Film Properties

As might be expected, there is a reduction in the accumulation of surfactant at the


film surface when surfmers are used in a latex, in comparison to a conventional
surfactant, which shows an excess concentration at the surface (Aramendia et al.
2005). This result supports the viewpoint of Kientz and Holl (1993), who state that
exudation of surfactant from particle surfaces during the particle deformation step
cause the formation of surfactant clumps, and these lead to migration of surfactant
to the surface. The chemical linkage of surfmers to polymer chains eliminates the
possibility of exudation, and no aging effect is expected. Similar results, indicating
that surfmers do not migrate to the film surfaces, are also found by AFM analysis
(Lam et al. 1997).
There is growing literature relating to the effects of surfmers on film properties
(Gauthier et al. 2002, Amalvy et al. 2002, Keslarek et al. 2004). Gauthier et al.
(2002) demonstrate that films made using particles stabilised with surfmers have
similar mechanical properties and decreased water uptake in direct comparison
with films made using conventional surfactants. Amalvy et al. (2002) demonstrate
206 6 Surfactant Distribution in Latex Films

an increased resistance to water uptake and also a dramatic decrease in the


exudation of surfactant to the film-air interface.
A possible downside to the use of surfmers is that the surfmers themselves
could provide a continuous path through the film for moisture penetration (see
location iii in Fig. 6.2 for a sketch). The determining factor will be the surface
density of surfactant chains on the particle surfaces and whether the density is
sufficient to form such a pathway. It should, however, be stressed that there are
numerous articles reporting reduced moisture uptake in latex films made with
surfmers (Asua and Schoonbrood 1998). Another consequence of using surfmers
is that the lack of desorption may hinder the particle deformation and interdiffu-
sion. It therefore seems likely that a sufficient amount of surfmer to enable
colloidal stabilisation, but not too much to enable hydrophilic pathways nor to
hinder film formation, will lead to superior film properties.

Fig. 6.14 Structures for examples of A inisurfs, B transurfs and C surfmers. Structure A
represents the diester of 4,4-azobis(4-cyanopentanoic acid), which is an initiator, attached to a
poly(ethylene oxide) nonylphenol tail. This inisurf is used in emulsion polymerisation studies by
Kusters et al. (1992). Structure B represents the transurf used by Vidal et al. (1995) in the
emulsion polymerisation of styrene. It contains a transfer agent at one end. Structure C shows
potassium 11-crotonoyloxyundecan-1-ylsulfate. This surfmer is synthesised by Lam et al. (1997)
and used in emulsion polymerisation. It contains a monomer with a reactive group that partici-
pates in polymerisation.
6.7 Summary 207

6.7 Summary

Numerous experimental studies have demonstrated that surfactant is a necessary


evil in latex films. The surfactant is added to ensure colloidal stability during
particle synthesis, but its presence in the final dried film can have detrimental
effects on the film properties. The peel strength, gloss, and film surface smooth-
ness are all affected. The use of surfmers can alleviate the detrimental effects by
stopping surfactant exudation from particle surfaces. A number of experimental
techniques have been developed to examine the surfactant distribution, with
confocal Raman spectroscopy able to provide depth profiles of Raman-active
surfactant molecules up to depths of about 100 m. From a theoretical perspec-
tive, the modelling of surfactant distributions is still at a preliminary stage. One of
the most pressing open problems is to examine the effect of desorption of surfac-
tant from particles as they deform. Although this chapter has concerned surfac-
tants, other water-soluble species, including oligomers and salts can be distributed
non-uniformly in a film and contribute to the detriment of properties.

References

Abele S., Zicmanis A., Graillat C., Monnet C. and Guyot A. (1999) Cationic and
zwitterionic polymerisable surfactants: quaternary ammonium dialkyl maleates.
1. Synthesis and characterisation, Langmuir 15: 1033-1044.
Agarwal N. and Farris R.J. (1999) Water absorption by acrylic based latex blend
films and its effect on their properties. Journal of Applied Polymer Science
72(11): 1407-1419.
Amalvy J.L. and Soria D.B. (1996) Vibrational spectroscopic study of distribution
of sodium dodecyl sulphate in latex films. Progress in Organic Coatings 28(4):
279-283.
Amalvy J.L., Unzue M.J., Schoonbrood A.S. and Asua J.M. (2002), Reactive
surfactants in heterophase polymerisation: Colloidal properties, film-water
adsorption, and surfactant exudation, Journal of Polymer Science part A:
Polymer Chemistry 40(17): 2994-3000.
Aramendia E., Mallgol J., Jeynes C., Brandiaran M.J., Keddie J.L. and Asua J.M.
(2003) Distribution of surfactants near acrylic latex film surfaces: A compari-
son of conventional and reactive surfactants (surfmers). Langmuir 19( 8): 3212-
3221.
Aramendia E., Barandiaran M.J., Grade J., Blease T. and Asua J.M. (2005)
Improving water sensitivity in acrylic films using surfmers. Langmuir 21(4):
1428-1435.
Arnold C. (2009), PhD Thesis, CNRS, Institut Charles Sadron, Strasbourg.
Arnold C. et al. (2009) Mapping the route from wet to dry. European Coatings
Journal 11: 28-32
208 6 Surfactant Distribution in Latex Films

Asua J.M. and Schoonbrood H.A.S. (1998) Reactive surfactants in heterophase


polymerisation. Acta Polymer 49: 671-686.
Belaroui F., Hirn M.P., Grohens Y., Marie P. and Holl Y. (2003) Distribution of
water-soluble and surface-active low molecular weight species in acrylic latex
films, Journal of Colloid and Interface Science 261:336-348.
Belaroui F., Cabane B., Dorget A., Grohens Y., Marie P. and Holl Y. (2003b)
Small-angle neutron scattering study of particle coalescence and SDS desorp-
tion during film formation from carboxylated acrylic lattices, Journal of Colloid
and Interface Science 262(2): 409-417.
Bindschaedler C., Gurny R., Dlker E. (1987) Influence of emulsifiers on film
formation from cellulose acetate latexes. Experimental study of phase-
separation phenomena due to sodium dodecyl sulphate. J Appl Polym Sci 34:
2631-2647
Butler L.N., Fellows C.M. and Gilbert R.G. (2005) Effect of surfactants used for
binder synthesis on the properties of latex paints. Progress in Organic Coatings
53(2): 112-118.
Butler L.N., Fellows C.M. and Gilbert R.G. (2004), Effect of systems on the water
sensitivity of latex films. Journal of Applied Polymer Science 92(3): 1813-
1823.
Charmeau J.Y., Kientz E. and Holl Y. (1996) Adhesion of latex films; Influence of
surfactants. Progress in Organic Coatings 27(1-4): 87-93.
Charmeau J.Y., Satre A., Vovelle L. and Holl Y. (1999) Adhesion of latex films.
Part II. Loci of failure. Journal of Adhesion Science and Technology 13(5):
593-614.
Charmeau J.Y., Gerin P.A., Vovelle L., Schirrer R.and Holl Y. (1999) Adhesion
of latex films. Part III. Surfactant effects at various peel rates, Journal of Adhe-
sion Science and Technology 13(2): 203-215.
Cussler E.L. (1984) Diffusion: Mass transfer in fluid systems, Cambridge Univer-
sity Press.
Donkus L.J. (2007) Solvent-like performance from emulsion PSAs: Advances in
water-whitening resistance. Adhesives Age 40(10) 32-35.
Du Chesne A., Gerharz B. and Lieser G. (1997) The segregation of surfactant
upon film formation of latex dispersions: an investigation by energy filtering
transmission electron microscopy. Polymer International 43 187-196.
Dreher W.R., Zhang P., Urban M.W., Porzio R.S. and Zhao C.L. (2003), Sty-
rene/2-ethylhexyl acrylate/methacrylic acid (Sty/E.H.A/M.A.A) coalescence
and response-driven mobility of sodium dodecyl sulphate (S.D.S) in colloidal
films. 22. A spectroscopic study, Macromolecules 36(4): 1228-1234.
Dreher W.R., Jarrett W.L. and Urban M.W. (2005), Stable nonspherical fluorine-
containing colloidal dispersions: Synthesis and film formation, Macromole-
cules 38(6): 2205-2212.
Eckersley, S.T., Rudin A. (1993) The effect of plasticisation and pH. on film
formation of acrylic latexes. J Appl Polym Sci 48: 1369-1381
Farinha J.P.S., Martinho J.M.G., Kawaguchi S., Yekta A., Winnik M.A. (1996)
Latex film formation probed by nonradiative energy transfer: Effect of grafted
References 209

and free poly(ethylene oxide) on a poly(n-butyl methacrylate) latex. J Phys


Chem 100: 12552-12558
Gauthier C., Sindt O., Vigier G., Guyot A., Schoonbrood H.A.S., Unzue M. and
Asua J.M. (2002) Reactive surfactants in heterophase polymerisation. XVII.
Influence of the surfactant on the mechanical properties and hydration of the
films. Journal of Applied Polymer Science 84(9): 1686-1700.
Gerin P.A., Grohens Y., Schirrer R. and Holl Y. (1999) Adhesion of latex films.
Part IV. Dominating interface effect of surfactant, Journal of Adhesion Science
and Technology 13(2): 217-236.
Goodwin J.W., Hearn J., Ho C.C. and Ottewill R.H. (1973) The preparation and
characterisation of polymer lattices formed in the absence of surface active
agents. British Polymer Journal 5: 347-362.
Gundabala V.R., Zimmerman W.B., and Routh A.F. (2004) A model for surfac-
tant distribution in latex coatings. Langmuir 20 (20): 8721-8727.
Gundabala V.R. and Routh A.F. (2006) Predicting surfactant distributions in dried
latex films, ACS Symposium Series 941: 53-65.
Gundabala V.R., Lei C., Ouzineb K., Dupont O., Keddie J.L., and Routh A.F.
(2008) Lateral surface non-uniformities in drying latex films. AIChE J 54(12):
3092-3105.
Guyot A., Tauer K., Asua J.M., Van Es S., Gauthier C., Hellgren A.C., Sherring-
ton D.C., Montoya-Goni A., Sjoberg M., Sindt O., Vidal F., Unzue M.,
Schoonbrood H., Shipper E. and Lacroix-Desmazes P. (1999) Reactive surfac-
tants in heterophase polymerisation. Acta Polymer 50: 57-66.
Heldmann C., Cabrera R.I., Momper B., Kuropka R. and Zimmerschied K. (1999)
Influence of non-ionic emulsifiers on the properties of vinyl acetate/VeoVa10
and vinyl acetate/ethylene emulsions and paints Progress in Organic Coatings
35: 69-77.
Hellgren A.C., Weissenborn P. and Holmberg K. (1999) Surfactants in water-
borne paints. Progress in Organic Coatings 35(1-4): 79-87.
Hulden M., Sjoblom E. and Saarnak A. (1994), Adsorption in latex paints in
relation to some of the properties of the mill base and the final coating, Journal
of Coatings Technology, 66(836): 99-105.
Jablonski E., Hayes J., Learner T. and Golden M. (2003) Conservation concerns
for acrylic emulsion paints. Reviews in Conservation 4: 312
Jarnstron J., Ihalainen P., Backfolk K. and Peltonen J. (2008), Roughness of
pigment coatings and its influence on gloss, Applied Surface Science 254:
5741-5749.
Juhu D. and Lang J. (1993) Effect of surfactant post added to latex dispersion on
film formation: A study by atomic force microscopy. Langmuir 9: 792-796
Juhu D., Wang Y.C., Lang L., Leung O.M., Goh M.C. and Winnik M.A. (1995)
Surfactant exudation in the presence of a coalescing aid in latex films studied
by atomic force microscopy. Journal of Polymer Science Part B: Polymer Phys-
ics 33(7) 1123-1133.
Keslarek A.J., Costa C.A.R. and Galembeck F. (2002) Latex film morphology and
electrical potential pattern dependence on serum components: A scanning
210 6 Surfactant Distribution in Latex Films

probe microscopy study. Journal of Colloid and Interface Science 255(1):


107-114.
Keslarek A.J., Leite C.A.P. and Galembeck F. (2004) Surfactant and counter ion
distribution in styrene-butyl acrylate-acrylic acid dry latex submonolayers.
Journal of the Brazilian Chemical Society 15(1): 66-74.
Kessel N., Illsley D.R., and Keddie J.L. (2008), The Influence of Interdiffusion
and Crosslinking on the Film Formation of an Acrylic Latex, JCT Research 5
285-297.
Kientz E. and Holl Y. (1993) Distribution of surfactants in latex films. Colloids
and Surfaces A 78: 255-270.
Kientz E., Danicher L., Lambla M. and Holl Y. (1993) Distribution of surfactants
in poly(2-ethyl hexyl methacrylate) latexes. Polymer International 31 (3):
297-304.
Kientz E., Dobler F., Holl Y. (1994) Desorption of the surfactant from the particle
surface during latex film formation. Polymer International 34: 125-134.
Kusters J.M.H., Napper D.H., Gilbert R.G. and German A.L. (1992) Kinetics of
particle growth in emulsion polymerisation systems with surface-active initia-
tors. Macromolecules 25: 7043-7050.
Lam S., Hellgren A.C., Sjoberg M., Holmberg K., Schoonbrood H.A.S., Unzue
M.J., Asua J.M., Tauer K., Sherrington D.C. and Goni A.M. (1997) Surfactants
in heterophase polymerisation: A study of film formation using atomic force
microscopy. Journal of Applied Polymer Science 66(1): 187-198.
Learner T.J.S. (2005) Analysis of modern paints. Getty Conservation Institute,
Los Angeles
Lee W.P., Gundabala V.R., Akpa B.S., Johns M.L., Jeynes C., Routh A.F. (2006)
Distribution of surfactants in latex films: A Rutherford backscattering study.
Langmuir 22 (12): 5314-5320.
Lestage D.J., Yu M. and Urban M.W. (2005) Stimuli-responsive surfac-
tant/phospholipids stabilised colloidal dispersions and their film formation.
Biomacromolecules 6(3): 1561-1572 2005.
Mallgol J., Gorce J.P., Dupont O., Jeynes C., McDonald P. and Keddie J.L.
(2002) Origins and effects of a surfactant excess near the surface of waterborne
acrylic pressure-sensitive adhesives. Langmuir 18 (11): 4478-4487.
Montoya-Goni A., Sherrington D.C., Schoonbrood H.A.S. and Asua J.M. (1998)
Reactive surfactants in heterophase polymerization. XXIV. Emulsion polym-
erisation of styrene with maleate and succinate containing cationic surfactants.
Polymer 40: 1359-1366.
Ormsby B., Learner T., Schilling M., Druzik J., Khanjian, Carson D., Foster G.,
Sloan M. (2006) The effects of surface cleaning on acrylic emulsion paintings:
A preliminary investigation. Tate Papers, Issue 6 Autumn. Available on-line at:
http://www.tate.org.uk/research/tateresearch/tatepapers/06autumn/ormsby.htm#
is6note6
Oron A., Davis S.H., Bankoff S.G. (1997) Long scale evolution of thin liquid
films. Reviews of Modern Physics 69(3): 931-980.
References 211

Powell E., Clay M.J., Saunston B.J. (1968), Effect of surfactant and copolymer
polarity on film-formation temperatures of some vinyldene chloride copoly-
mers, J Appl Polym Sci 12: 1765-1773
Rottstegge J., Traub B., Wilhelm M., Landfester K., Heldmann C. and Spiess
H.W. (2003) Investigations on the film formation process of latex dispersions
by solid state NMR spectroscopy. Molecular Chemistry and Physics 204(5-6):
787-802.
Roulstone B.J., Wilkinson M.C. and Hearn J. (1992a) Studies on polymer latex
films. 2. Effect of surfactant on the water-vapour permeability of polymer latex
films. Polymer International 27(1): 43-50.
Roulstone B.J., Wilkinson M.C. and Hearn J. (1992b) Studies on polymer latex
films. 3. Permeability of latex films to aqueous organic solutes. Polymer Inter-
national 27(1): 51-55
Routh A.F. and Zimmerman W.B. (2004) Distribution of particles during solvent
evaporation from films. Chemical Engineering Science 59 (14): 2961-2968.
Sindt, O., Gauthier C., Hamaide T. and Guyot A. (2000) Reactive surfactants in
heterophase polymerisation. XVI. Emulsion copolymerization of styrene-butyl
acrylate-acrylic acid in the presence of simple maleate reactive surfactants.
Journal of Applied Polymer Science 77: 2768-2776.
Steward P.A., Hearn J. and Wilkinson M.C. (1995) Studies on permeation through
latex films. 1. Films containing no or low levels of additives. Polymer Interna-
tional 38(1): 1-12.
Thenoz F., Soula O., and Guyot A. (1999) Reactive surfactants in heterophase
polymerisation. XXV. Core shell lattices stabilised with a mixture of maleic
anionic and styrenic non-ionic surfactants, Journal of Polymer Science; Part A:
Polymer Chemistry 37: 2251-2262.
Thorstenson T.A., Evanson K.W. and Urban M.W. (1993) Film air and film
substrate interfaces of latex films monitored by Fourier-Transform Infrared
spectroscopy. Advances in Chemistry Series 236: 305-331.
Tiwari N., Mester Z. and Davis J.M. (2007) Stability and transient dynamics of
thin liquid films flowing over locally heated surfaces. Physical Review E.
76(5): 056306 (14 pages)
Tzitzinou A., Jenneson P.M., Clough A.S., Keddie J.L., Lu J.R., Zhdan P.,
Treacher K.E. and Satguru R. (1999) Surfactant concentration and morphology
at the surfaces of acrylic latex films Progress in Organic Coatings 35: 89-99.
Unzue M.J., Schoonbrood H.A.S., Asua J.M., Montoya-Goni A., Sherrington
D.C., Stahler K., Goegel K.H., Tauer K., Sjoberg M. and Holmberg K. (1997),
Reactive surfactants in heterophase polymerisation. VI. Synthesis and screen-
ing of polymerisable surfactants (surfmers) with varying reactivity in high
solids styrene-butyl acrylate-acrylic acid emulsion polymerisation, Journal of
Applied Polymer Science 66: 1803-1820.
Urban M.W. (1997) Surfactants in latex films; interactions with latex components
and quantitative analysis using ATR and step-scan PAS FT-IR spectroscopy
Progress in Organic Coatings 32 (1-4): 215-229.
212 6 Surfactant Distribution in Latex Films

Vandezande G.A. and Rudin A. (1996) Film formation of vinyl acrylic latexes;
Effects of surfactant type, water and latex particle size. Journal of Coatings
Technology 68(860): 63-73.
Vidal F., Guillot J. and Guyot A. (1995) Surfactants with transfer agent properties
(transurfs) in styrene emulsion polymerisation. Colloid and Polymer Science
273:999-1007.
Vijayendran B.R., Bone T., Gajria C. (1981) Surfactant interactions in poly(vinyl
acetate) and poly(vinyl acetate butyl acetate) latexes. J Appl Polym Sci 26:
1351-1359.
Wagner H. and Fischer G. (1936) Film formation from emulsions, Kolloid Z.,
77(1):12-20.
Wang Y.C., Kats A., Juhu D., Winnik M.A., Shivers R.R. and Dinsdale C.J.
(1992) Freeze fracture studies of latex films formed in the absence and pres-
ence of surfactant. Langmuir 8(5): 1435-1442.
Winnik M.A. and Feng J. (1996) Latex blends: An approach to zero VOC
coatings. Journal of Coatings Technology 68(852):39-50
Yang Y.K., Li H. and Wang F. (2003) Studies on the water resistance of acrylic
emulsion pressure-sensitive adhesives (P.S.As). Journal of Adhesion Science
and Technology 17(13): 1741-1750.
Zhao C.L., Holl Y., Pith T. and Lambla M., (1987), FTIR-ATR Spectroscopic
determination of the distribution of surfactants in latex Coll Polym Sci, 265:
823-829
Zhao C.L., Dobler F., Pith T., Holl Y. and Lambla M. (1989) Surface composition
of coalesced acrylic latex films studied by XPS and SIMS Journal of Colloid
and Interface Science 128(2): 437-449.
Zhao Y. and Urban M.W. (2000) Mobility of SDOSS powered by ionic interac-
tions in Sty/n-BA/MAA core-shell latex films. 21. A spectroscopic study.
Langmuir 16: 9439-9447.
Zhao Y. and Urban M.W. (2000b) Phase separation and surfactant stratification in
styrene/n-butyl acrylate copolymer and latex blend films. 17. A spectroscopic
study. Macromolecules 33: 2184-2191.
Zhao Y.Q. and Urban M.W. (2000c) Polystyrene/poly(n-butyl acrylate) latex
blend coalescence, particle size effect, and surfactant stratification: A spectro-
scopic study. Macromolecules 33(20): 7573-7581.
Zosel A. and Schuler B. (1999) The influence of surfactants on the peel strength
of water-based pressure sensitive adhesives, Journal of Adhesion 70(1-2):
179-195.
Chapter 7

7 Nanocomposite Latex Films and Control of Their


Properties1

7.1 Introduction

We have seen in previous chapters how identical, homogeneous particles may be


used as the building blocks of a homogeneous film. Within the past decade, there
has been enhanced interest in using colloidal particles in water to create nanocom-
posite films, which are defined as materials made of two or more phases blended
at nanometre length scales. A key advantage of the colloidal approach is that it
offers control of structure at the nanoscale (within particles) and at the meso- and
even macroscale through the creation of ordered assemblies of particles via film
formation. The fabrication of nanocomposites offers exciting new challenges and
opportunities for the science and technology of latex film formation. The film
formation process offers a way of tailoring the properties of nanocomposite films
through the control of the assembly of particles, just as is the case for conventional
latex films.
Polymer nanocomposite films made from waterborne colloids can be put into
one of two broad classifications:
1. Blends of latex particles and a second type of particle (inorganic or polymeric)
2. Hybrid or nanocomposite particles consisting of a polymer component and a
second component (either another polymer or an inorganic phase).
Each of these two types of nanocomposites will be considered here. We shall
use the word hybrid to refer to materials in which two or more phases are
covalently bonded together at the molecular level (Guyot et al. 2007). Nanocom-
posites shall refer to materials in which the phases are merely mixed at nanometre
length scales.

1 This chapter uses material from Wang and Keddie (2009) with permission from Elsevier.
214 7 Nanocomposite Latex Films and Control of Their Properties

7.1.1 Properties of Nanocomposites

The quest for novel and enhanced properties is a key driver in the development of
colloidal nanocomposite films. Structure can be built into colloidal particle
building blocks through the combination of two or more polymer phases or a
polymer and an inorganic phase. Nanocomposites are made from dissimilar
polymers or they contain nanoscale fillers, e.g., clay platelets (Giannelis 1996) and
carbon nanotubes (a form of carbon) (Ajayan and Tour 2007, Coleman et al.
2006). They have combinations of properties that are not found in the individual
components. The mechanical properties, electrical and thermal conductivity, and
flammability resistance all differ in nanocomposites in comparison to conven-
tional composites.
In nanocomposites prepared by non-colloidal routes, outstanding properties
have been reported, as outlined below. There are fewer examples of the studies of
the properties of colloidal nanocomposites made from latex, but these will be
considered later in this chapter.
Flammability Reduction and Flame Retardancy. The addition of inorganic
nanosized fillers to polymers reduces their flammability. The addition of clay
particles is particularly effective in reducing the flammability of polypropylene
(Gilman et al. 2000) and in poly(styrene) (Zhu et al. 2001) in which the degrada-
tion onset temperature was increased by 50C. The peak heat release rate in
poly(styrene) was decreased by up to 58% by the addition of clay, whereas the
addition of silica nanoparticles decreased this rate by about 50% in poly(methyl
methacrylate) (Kashimaji 2003).
Mechanical Strength and Stiffness. Strength refers to the maximum stress that
can be imposed on a material before it fails. Stiffness is measured by the elastic
modulus, which is determined by the stress required to strain a material in the
linear region. If a polymer matrix is reinforced with a second phase having a
higher elastic modulus, then the composite would be expected to be stiffer than the
pure polymer. Very stiff nanofillers, such as carbon nanotubes, can increase the
stiffness of polymers by huge amounts. For instance, single-wall carbon nanotubes
(SWNTs) dispersed at high fractions in poly(vinyl alcohol) lead to a remarkably
high elastic modulus of 85 GPa and a failure stress of 3.75 GPa (Dalton et al.
2003), whereas the polymer alone has a modulus of 15 GPa and a failure stress of
only about 100 MPa. When natural silk fibres are reinforced with single-wall
carbon nanotubes in an electrospinning process, the elastic modulus increases by
460% (Ayutsede et al. 2006). However, as is often the case, there is a trade-off
between achieving a high modulus and a high strain to failure. Stiffer materials
tend to fracture at lower strains.
When a filler with a high modulus is added to a matrix of a lower modulus,
analytical equations can be applied to predict the elastic modulus of the compos-
ite. The Halpin-Tsai equations, which are commonly employed, take into account
the aspect ratio (e.g., length over diameter) and orientation of the filler (Halpin
7.1 Introduction 215

and Kardos 1976). However, the modulus of nanocomposites often cannot be


adequately described by a two-phase model. Instead, a third phase, which repre-
sents the material at the interface between the filler and the matrix, must be
considered. This so-called interphase has a mechanical response that is different
than that of the other two phases. (Ji et al. 2002)
Toughness. Quite often, stiff materials fracture after being strained by a small
amount. They are referred to as brittle. In many applications, it is desirable to
retain stretchability along with stiffness. The ability of a material to withstand
large strains without breaking is called toughness. One particularly good
example of toughening was demonstrated by the addition of clay to
poly(vinylidene fluoride). The strain at breaking increased from 20% for the pure
polymer to 140% in the nanocomposite. This increase in the elongation translated
into an order of magnitude increase in the toughness (Shah et al. 2004). A
toughness of 870 J per gram (corresponding to the energy required to deform the
material) and a strain at failure of 430% have been reported in poly(vinyl alcohol)
fibres reinforced with single-wall carbon nanotubes (Miaudet et al. 2005). By
comparison, natures toughest fibre, spider silk, has a toughness of 160 J/g, and it
fails at a 30% strain (Dalton et al. 2003). In a unique nanoscale arrangement of
clay and polymer, recoverable strains as high as 3000% have been reported along
with the ability to tailor the elastic modulus and strength (Haraguchi et al. 2006).
Despite the promise of outstanding mechanical performance, nanocomposite
films and bulk materials often display mechanical strengths that are far below the
idealised theoretical predictions. This failure has been attributed to poor nanofiller
dispersion in the polymer matrix and to weak interface adhesion with the matrix,
as well as the random and uncontrolled morphology of the nanocomposites
(Moniruzzaman and Winey 2006, Vaia and Maguire 2007).
Viscoelasticity. Nonlinear elasticity refers to a deformation with large strains. In a
nanocomposite material with a continuous network of one phase, large strains are
likely to disrupt or break that network. Linear elasticity refers to the limit with
small strains (typically less than 1%), in which the deformation will not funda-
mentally alter the material. Dynamic mechanical analysis, in which a material is
strained in an oscillatory way, provides information on the viscoelasticity, which
is the combination of the elastic and viscous properties. Viscoelasticity is highly
sensitive to the arrangement of phases in a nanocomposite. For instance, the
viscoelasticity is affected by how well clay particles are dispersed in the continu-
ous matrix (Xu et al. 2005). When the clay platelets are at a sufficiently high
concentration to make contact with each other, there is a transition from a liquid-
like to a solid-like response.
When nanoparticles are dispersed in an entangled polymer melt, the viscosity
decreases. In particular, the zero-shear rate viscosity of molten polystyrene is
reduced upon the addition of polystyrene nanoparticles, which were made by
crosslinking polymer chains (Mackay et al. 2003). This result is at odds with the
expectations from the Einstein relation (Eq. (1.4)) that describes how the viscosity
216 7 Nanocomposite Latex Films and Control of Their Properties

of a fluid increases approximately monotonically with the volume fraction of


particles dispersed in it. The reduced viscosity effect holds only when the distance
between particles is smaller than the polymer chain size, i.e., only when the
entangled polymer chains are confined (Tuteja et al. 2005). On the other hand,
when the nanoparticles are spaced further apart, a viscosity increase is observed.
Furthermore, in entangled polymers in which the chains physisorbed onto dis-
persed nanoparticles, a viscosity increase has been found (Zhang and Archer
2003). Yet, reduced viscosity and associated property enhancements is found
with inorganic nanoparticles provided they are blended in via a rapid precipitation
process in a bad solvent (Tuteja et al. 2007).
The diffusion coefficient of nanoparticles in a viscous fluid is given by the
Stokes-Einstein relation (Eq. (1.3)), which says that it varies inversely with
particle size. However, in experiments in which inorganic nanoparticles are
smaller than the mesh size of a polymer melt (less than 10 nm in diameter), the
diffusivity has been reported to be 200 times faster than predicted (Tuteja et al.
2007b). The observation cannot be explained solely by the observed reduced
viscosity, but it might be related to the effect of the nanoparticles on the polymer
entanglement network.
Thermal and Electrical Conductivity. Electrically conducting nanofillers, such
as carbon nanotubes, impart conductivity to insulating polymers, even when added
at relatively low concentrations. The electrical conductivity can increase as much
as 1012 times when carbon nanotubes are dispersed in a polymer above the
percolation threshold (McCullen et al. 2007).
For coatings applications, other attractive properties of nanocomposites are
scratch-resistance and the ability to prevent the permeation of liquids and vapours,
known as barrier resistance. Architectural coatings also require block resistance,
which is the tendency to have low adhesion between coatings surfaces placed into
contact (Schuler et al. 2000). Think of the coating surfaces in contact with each
other inside a window frame.

7.1.2 Applications of Colloidal Nanocomposites

Colloidal nanocomposites with structures tailored on a nanoscale have applica-


tions in printed flexible electronics (Rogers 2001), photonic bandgaps (Cheng et
al. 2006, Stein et al. 2008), sensors (Lee and Asher 2000), data storage (Cumpston
et al. 1999), and displays (Arsenault et al. 2007). Of course, one of the oldest
applications of waterborne polymer colloids is for architectural coatings. Colloidal
nanocomposites of polyacrylates and silica have been developed by BASF for
commercial applications as a transparent, flame-retardant, scratch-resistant coating
(Xue and Wiese 2006, Tiarks et al. 2007)
There is also interest in using an assembly of colloidal particles to template
inorganic phases in a structure ordered on a nanoscale. Natural opal consists of
7.2 Types of Hybrid Particles 217

spherical silica particles arranged in an ordered, three-dimensional array. The void


space between particles contains air. If polymer core- inorganic shell particles are
arranged into an array, their shells impart a regular modulation of optical and
mechanical properties. If the polymer cores are removed by chemical etching or
calcination, a porous inorganic structure is left behind. This inverse opal struc-
ture, as it is called, can serve as a template for other nanomaterials or as three-
dimensional scaffolds for cell or tissue engineering (Liu et al. 2005, Kotov et al.
2004, Zhang et al. 2005, Lee et al. 2006).
Our emphasis in this chapter will be on (i) the types of hybrid particles that can be
made (but not how to make them) in Section 7.2; (ii) ways to assemble the particles
into structured films in Section 7.3; (iii) methods and advantages of blending particles
in Section 7.4; and (iv) the effects of particle structure on final film properties in
Section 7.5, summarised in three lessons. We will see how outstanding properties can
result when phases are blended on very short length scales.

7.2 Types of Hybrid Particles

We shall start by considering the variety of structures and material combinations


that can be found with hybrid particles. There are already many excellent descrip-
tions of synthesis methods in the literature, and the details of these methods will
not concern us here.

7.2.1 Polymer-Polymer Hybrid Particles

To make polymer-polymer hybrid particles, there are three commonly used


methods, as described below. (Guyot et al. 2007)
1. Miniemulsion polymerisation in which a polycondensate (such as an alkyd or
polyurethane) is dissolved in a monomer mixture, which is then mini-
emulsified and polymerised. In miniemulsion polymerisation, each droplet of
monomer serves as a nanosized reactor. There is no transport of the monomer
through the aqueous phase as in conventional emulsion polymerisation.
Miniemulsion polymerisation has been reviewed and is described in detail
therein (Asua 2002). Polymer-polymer hybrids have been made using pre-
formed alkyds, polyesters and polyurethanes.
2. Miniemulsion of all monomers followed by a poly-condensation or poly-
addition reaction to form a polycondensate and then a radical polymerisation of
the other monomer. There are two steps in one pot (i.e., in the monomer drop-
lets).
3. In a third method, the polycondensate can be emulsified and then used as a seed
particle in emulsion polymerisation. Particles of alkyd resins, epoxy resins,
218 7 Nanocomposite Latex Films and Control of Their Properties

silicones and polyurethanes have all been used as seeds for subsequent emul-
sion polymerisation of other monomers.
Miniemulsion polymerisation can also be used in the synthesis of nanocompo-
site particles containing polymer and inorganic phases. Through a large interna-
tional effort in recent years, there is now an improved understanding of how to
control the nanostructure of particles prepared by techniques of miniemulsion
polymerisation (Landfester 2009).
In hybrid particles, the two polymer phases are expected to be co-continuous,
as shown in Fig. 7.1b. Polymer-polymer nanocomposite particles can also have a
core-shell structure (Fig. 7.1a) in which one of the polymer phases encapsulates
the other in a thin layer. The polymer phase in the central core of the particle is
isolated from the water. Thermodynamics as well as polymer dynamics play a
large role in determining the final particle structure (Stubbs and Sundberg 2007,
2008). At thermodynamic equilibrium, the interfacial energy between all phases
will be at a minimum. This energy is proportional to the area between the phases.
In some cases, the shell coverage is incomplete, so that a half-moon particle is
created (Fig. 7.1c). Lobed particles with two separate phases are also found
sometimes. Another possibility is for one of the polymer phases to exist as
occlusions within a continuous phase of the particles. Such occluded structures
(Fig. 7.1d) are sometimes found in miniemulsion polymerisations using pre-
formed polymers as one of the phases.

(a) (b)

Core-shell Polymer/resin
miniemulsion

(c) (d)

Half-moon Occluded
Fig. 7.1 An illustration of some of the types of polymer-polymer nanocomposite particles, shown
in cross-section. a Core-shell particle; b A hybrid particle synthesised by miniemulsion
polymerisation using a pre-formed resin; c A half-moon structure in which one phase only
partially encapsulates the core. d An occluded structure in which small phases are distributed
within the particle.
7.2 Types of Hybrid Particles 219

Core-shell particles can be synthesised in a two-stage seeded emulsion polym-


erization process. For instance, a polystyrene seed was made by standard batch
emulsion polymerisation, and then a poly(n-butyl acrylate) shell was grown on the
seed in the second stage (Chevalier et al. 1999). Particles containing two immis-
cible polymers, such as poly(methyl methacrylate) and a poly(n-butyl acrylate)
copolymer, can be made via a two-stage semi-batch emulsion polymerisation.
Varying the mass ratio of the two polymer phases (Schuler et al. 2000) and the
fraction of core and shell phases (Suresh et al. 2007) enables a rich variety of
particle morphologies. For the same polymer pairs, reversing the order of the
stages of the polymerization allows the core polymer to become the shell
polymer and vice-versa (Colombini et al. 2005), leading to a difference in the
viscoelastic properties of the films.

7.2.2 Inorganic and Polymer Nanocomposite Particles

There are many methods to make nanocomposite or hybrid particles of polymer


and inorganic phases. Some of these methods have been outlined elsewhere
(Wang and Keddie 2009, Castelvetro and De Vita 2004). Fig. 7.2 illustrates the
various types of particle structures that can be created. We will consider each of
these types of particles separately.

Raspberry Core-shell Core-Shell


Currant bun

Core-Shell Encapsulated by Encapsulated by Encapsulation of


nanotubes clay inorganic by polymer

Fig. 7.2 Illustrations of some of the types of polymer-inorganic nanocomposite particles. The
drawings show the cross-sections of the particles. The inorganic phase is shown in black in all
drawings.
220 7 Nanocomposite Latex Films and Control of Their Properties

1. Core-shell particles have either an inorganic or a polymer phase in the core, and
the opposite phase in a continuous shell. There are a few examples of a single,
inorganic nanoparticle being encapsulated with a polymer shell. One approach to
making this type of particle is by the use of controlled or living atom transfer
radical polymerization (ATRP) to grow the polymer chains from the surface of
inorganic particles dispersed in water. The polymerisation starts from an initiating
group on the inorganic surface, such as an activated carbon-halogen bond. (Pyun
and Matyjaszewski 2001)
Typically, initiators are adsorbed on the particles and the monomers polymerise
from the surface (Von Werne and Patten 1999). Hydrophilic polymers are more
commonly used, as they provide stability in water. Chains of polystyrene and
poly(methyl methacrylate) have been successfully grown from the surfaces of
silica nanoparticles using ATRP (Von Werne and Patten 2001).

(a) (b)

(c) (d)

Fig. 7.3 Examples of polymer-inorganic nanocomposite particles. a TEM image of core-shell


particles in which the shell is made from 20 nm silica particles and the core is P2VP-PMMA
copolymer. (Reprinted from Dupin et al. 2007. Copyright 2007 American Chemical Society); b
SEM image of core-shell particles in which the core is made from silica, and the shell consists of
PS nodules (Reprinted from Reculusa et al. 2002. Copyright 2002 American Chemical Society);
c Cryogenic TEM images of particles in which laponite encapsulates a poly(styrene-co-butyl
acrylate) core (reprinted from Negrete-Herrera et al. 2007; copyright Wiley-VCH Verlag GmbH
& Co. KGaA; reproduced with permission); d Hollow silica shells obtained from the calcination
of a polystyrene particle core. (Reprinted from Schmid et al. 2007, copyright 2007 American
Chemical Society)
7.2 Types of Hybrid Particles 221

In the opposite type of core-shell particles, there is a polymer core which is


fully surrounded by a shell of an inorganic phase, usually made up of densely
packed nanoparticles. Good examples are colloidal silica nanoparticles decorating
the surface of a poly(2-vinylpyridine) copolymer (Dupin et al. 2007), as in Fig.
7.3a, or a polystyrene core (Schmid et al. 2006). Interestingly, the exact inverse
structure of silica particles with nodules of polystyrene on the surface (Reculusa et
al. 2002) has also been created (Fig. 7.3b).
When a second phase envelopes a particle, it is said to be encapsulated. For
instance, colloidal silica has been encapsulated within epoxy particles by phase
inversion emulsification (Yang et al. 2002). In the opposite type, the inorganic
phase creates an outer shell, sometimes called armour because the hard inorganic
phase creates a protective exterior to the softer interior, just like a steel uniform
protects human flesh.
A popular inorganic material for the armour is exfoliated clay. Clay consists of
crystals of layered alumino-silicates, similar to sheets of paper in a book (or folio).
With the right chemical treatment, the sheets separate from each other (i.e., exfoliate)
to create platelets that are only a few nanometres thick (e.g., 1 nm for laponite, which
is a synthetic material) and up to hundreds of nanometres in diameter, depending on
the type of clay. If the polymer enters the gallery space between the clay sheets, the
clay is said to be intercalated. The creation of clay-polymer nanocomposite particles
is building on a long tradition of using blends of clay and latex in porous coatings for
paper, suitable for printing applications (Watanabe and Lepoutre 1982).
With techniques of emulsion polymerisation, the dispersion of clays is better in
comparison to physical blending of clay and latex particles, leading to enhanced
mechanical, thermal and permeability properties (Diaconu et al. 2008). Emulsion
polymerisation in the presence of clay in the water phase was one of the first
reported methods to create polymer-clay nanoparticles (Noh and Lee 1999). But
better control of particle structure and the resulting properties is obtained with
chemical functionalisation of the clay to increase its incorporation into the
particle. Initiators and macromonomers can be used to modify clay particles to
enhance their compatibility with the polymer phase (Diaconu et al. 2009).
Particles have been produced in which modified clay platelets entirely encapsulate
the polymer core, as demonstrated in Fig. 7.3c (Negrete-Herrera et al. 2007).
Carbon nanotubes have also been used to encapsulate latex particles, but owing
to the relatively long length of the nanotubes, the particles were larger than found
in standard latex (Paunov and Panhuis 2005).
In some cases, the inorganic shell is sufficiently mechanically robust that it can
be used to create hollow nanocapsules through the removal of the polymer cores.
Fig. 7.3d shows silica shells, which have interesting optical properties, that have
been made through the calcination of the polystyrene core of core-shell particles
(Schmid et al. 2007).
2. Currant-bun particles have inorganic nanoparticles distributed inside the latex
particles but with some additional inorganic particles emerging from the particle
surface (see Fig. 7.2). There is an analogy here to a bun that is filled with currants
222 7 Nanocomposite Latex Films and Control of Their Properties

(or raisins) with a few of the currants visible from the outside. An example is silica
nanoparticles contained in poly(4 vinyl pyridine) particles (Tiarks et al. 2001, Percy
et al. 2000).
3. Raspberry particles have a surface that is decorated with smaller inorganic
nanoparticles, which are closely packed together and distributed inside the
polymeric core. The name is obvious when comparing the appearance at the
nanoscale to the fruit surface at the macroscale. Sometimes, core-shell particles
with a shell of inorganic nanoparticles are called raspberry particles.
The tendency of solid particles to adsorb strongly at liquid-liquid interfaces was
first described by Ramsden (1903) and Pickering (1907) near the turn of the
twentieth century. This phenomenon has led to a method of creating colloidal
nanocomposites, which has attracted interest recently, called Pickering emulsion
polymerisation. Inorganic nanoparticles are used as the emulsifier for monomers
in water in an emulsion polymerisation process (Cauvin et al. 2005, Bon and
Colver 2007). These nanoparticles are then assembled into a shell that encapsu-
lates the particle core. Clay has been used to encapsulate polystyrene (Cauvin et
al. 2005) and polyacrylamide (Voorn et al. 2006) latex particles in Pickering
emulsion polymerisations. Table 7.1 summarises various examples of latex
particles encapsulated by inorganic phases, made by various polymerisation
methods. The list is not exhaustive but is meant to provide examples.

Table 7.1 Polymer Particles Encapsulated with an Inorganic Phase


Polymer Inorganic phase Colloid size (m) Synthesis References
Method
PS Clay 0.3 Pickering Cauvin et al.
(Laponite) polymerisation 2005, Bon
and Colver
2007
PAAm Clay (Montmoril- 0.71 Pickering Voorn et al.
lonite) polymerisation 2006
PS TiO2 1, 2050 Pickering Liu et al.
polymerisation 2006, Chen
et al. 2007
PMMA Clay 0.1 Pickering Zhang et al.
(Montmorillonite) polymerisation 2008
P(St-BuA) Clay 0.1 In-situ polymeri- Negrete-
(Laponite) sation Herrera et al.
2007
PMMA Fe3O4 400 In-situ polymeri- Hasell et al.
sation 2007
PS Fe3O4 13 In-situ polymeri- Yang et al.
sation 2008
P2VP Silica 0.180.22 Emulsion Dupin et al.
polymerisation 2007
PS Silica 0.23.3 Dispersion Schmid et al.
polymerisation 2006, 2007
7.2 Types of Hybrid Particles 223

7.2.3 Self-Assembly of Nanocomposite Particles by Precipitation or


Flocculation of Pre-Formed Nanoparticles

In previous examples, the inorganic phase is incorporated into a particle in a


polymerisation process. Other types of nanocomposites can be created by coagula-
tion or flocculation of pre-formed inorganic particles onto larger polymer particles.
In one approach for creating nanocomposite particles, a polymer particle can be
used as a nanoscale substrate surface onto which inorganic particles are attached
through a hetero-flocculation process. In one of the first examples, a careful
control of particle charge was used to induce the adsorption of small colloidal
particles onto larger, oppositely charged colloidal particles (Harley et al. 1992).
When the hetero-flocculation occurs, a core-shell particle is created, in which a
polymer core is surrounded by an inorganic particulate shell. The thickness of the
shell can be effectively controlled through the concentration and size of the
inorganic nanoparticles.
Various inorganic nanoparticles, carbon nanotubes, and clay platelets have all
been precipitated or flocculated onto larger colloidal polymer spheres. Fig. 7.4
schematically illustrates the assembly of various inorganic objects, e.g., nanoparti-
cles, nanoclays and nanotubes, onto polymer colloid surfaces.

Precipitation
trigger

Functionalised
nanoparticles

+
Nanoclay
platelets
Functionalised
polymer
colloids
Nanotubes

Colloid/inorganic
suspension

Fig. 7.4 Methods for creating particles encapsulated with an inorganic phase: nanoparticles, clay
platelets, or nanotubes. The inorganic phase precipitates or aggregates onto the polymer particles
to create a shell. (Reprinted with permission from Wang and Keddie (2009))
224 7 Nanocomposite Latex Films and Control of Their Properties

Table 7.2 Examples of Hybrid Particles Obtained via Hetero-flocculation or Precipitation


Polymer Inorganic phase Polymer colloid Method Reference
size (m)
PS TiO2, SiO2, 0.210.64 opposite charge Caruso et
Clay(Laponite) al. 2001
PS Fe3O4 0.64 opposite charge Caruso et
al. 1999
PS Ag 0.230.605 specific reaction Mayer et
al. 2000,
Caruso et
al. 2001
PS Au 0.4, 1 specific reaction Tian et al.
2007, Khan
et al. 2001
PS ZnS 0.275 specific reaction Pich et al.
2005
PS CdS 0.13 opposite charge Sherman
and Ford
2005
PS CdTe 0.468 solvent treatment, Rad-
opposite charge tchenko et
al. 2001
PS Single-wall carbon 0.4 stacking Dionigi et
nanotubes interaction al. 2007
Sulfonated PS Multi-wall carbon 510 pH and charge Paunov and
nanotubes adjustment Panhuis
2005
Sulfonated PS CdTe(S) 0.64 opposite charge Susha et al.
2000,
Rogach et
al. 2000
P(St-MAA-AA) CaCO3 0.38 specific reaction Lu et al.
2005
P(St-MAA-AA) Cu2O 0.08 specific reaction Lu et al.
2005
PNIPAM Au 0.45 opposite charge Karg et al.
2007
PMMA Clay 0.2 ion exchange Rossi et al.
(Montmorillonite 2002
and Laponite)

The hetero-flocculation process can be driven by solvent treatment (Rad-


tchenko et al. 2001), pH adjustment (Paunov and Panhuis 2005), ion exchange
(Rossi et al. 2002), electrostatic attraction (Caruso et al. 1999), hydrogen bonding,
or specific chemical reactions (Mayer et al. 2000). When the charge or polarity of
colloid surfaces is used to attract nanocrystals, the process is usually referred to as
controlled precipitation. In the case of precipitation on polymer colloids, the
surface needs to be decorated with a layer containing -NH2, -COOH, -SO4H,
-SO3H, -OH, -SH, -CONH2, -CH2NH2, -CH2Cl or similar groups (Xia et al. 2000).
7.3 Colloidal Particle Deposition and Assembly Methods 225

Numerous types of inorganic phases, with nanoscale or microscale dimensions,


have been attached onto polymer colloids to create colloidal nanocomposite
particles, as summarised in Table 7.2.

7.3 Colloidal Particle Deposition and Assembly Methods

When a colloidal dispersion of particles in water is deposited onto a substrate and


water evaporation is allowed to proceed, either a continuous film or an array of
separate isolated particles will form under appropriate drying conditions. Highly
dilute dispersions are used to create isolated particles. Polymer hybrid or nano-
composite particles can be assembled to form nanocomposites with a structure that
is ordered at the nanoscale (Kumacheva et al. 1999). The hybrid particles are most
often packed on a face-centred cubic lattice (Wang et al. 2008b; Deng et al. 2008).
But, body-centred cubic, body-centred orthorhombic, space-filling tetragonal, or
body-centred tetragonal lattices are also possible (Dziomkina and Vansco 2005).
Polymeric core-shell particles can be film-formed to create nanocomposite
films. If the temperature of the deposition process is less than the polymer Tg, the
polymer particles will tend to remain spherical. At higher deposition and process-
ing temperatures, the particles will deform to fill available space, thus creating a
honey-comb structure with the shell phase comprising the walls (Fig. 7.5a). If the
temperature is above the Tg of the particle shells, film formation is facilitated
through the deformation of the shell. The cores remain isolated in the film
structure, demonstrating a memory of the particle structure. A honeycomb-like
film structure, such as is presented in the idealised process shown in Fig. 7.5a, has
been realised in practice (Fig. 7.6).
Even when there is particle deformation, the boundaries between shells can be a
weak point in the film, and in some cases, the introduction of crosslinking between
the shells has been found to increase the fracture strength of core-shell films (Schel-
lenberg et al. 2000). Film formation from homogeneous particles is not possible at
temperatures below the polymers Tg. Rather remarkably, however, film formation is
possible with a glassy shell polymer (below its Tg), provided that there is a deform-
able, low-Tg core to compensate (Domingues dos Santos and Leibler 2003).
In the case of encapsulated particles, an ordered structure is created with the
inorganic particles (e.g., spherical nanoparticles, nanoclay platelets, or nanotubes)
situated at the boundaries between the polymer cores (as in Fig. 7.5b). The ordered
nanocomposite structure can be adjusted through the particle radius, R (varying
from tens of nanometres to a few microns), the polydispersity of the size, the
chemical architectures in the polymer core, the composition of the shell, and the
shell thickness, ls. When using colloidal hybrid particles to produce 3-D polymer
nanocomposites, R and ls can be tailored during the emulsion polymerisation
process or the hetero-flocculation process to adjust the core particle density in the
final nanocomposite (Kumacheva et al. 1999).
226 7 Nanocomposite Latex Films and Control of Their Properties

(a)

(b)

Fig. 7.5 Film formation from core-shell particles. a Soft core-soft shell particles deform to fill
available space and thus create a honeycomb structure. The walls are created by the shell material.
b Raspberry particles with inorganic particles in the shell create a film in which the inorganic phase
is continuous throughout the film. If the particles are fully encapsulated by an inorganic phase, then
interdiffusion of the polymer between particles could be blocked. A weak film will result.

(a) (b)

Shell
Core
Fig. 7.6 a TEM image of two-phase particles, after being stained with RuO4. The particle core is
poly(ethyl hexyl methacrylate) and the shell is crosslinked poly(n-butyl acrylate). b After film
formation, the particles are deformed into polyhedra. In a TEM image of a film cross-section,
hexagonal cores are seen in a honeycomb-like network of the shell material. (Reprinted from
Schellenberg et al. 2000; copyright Wiley-VCH Verlag GmbH & Co. KGaA; reproduced with
permission)
7.3 Colloidal Particle Deposition and Assembly Methods 227

In instances where polymer particles are fully encapsulated by a shell of hard,


inorganic particles, satisfactory film formation is not expected. Interdiffusion of
polymer chains, which is required to knit the interfaces between particles, will be
blocked by the inorganic particles. In some cases, however, the inorganic phase
could undergo chemical bonding to other similar particles and thereby strengthen
the interfaces. Film formation and the development of mechanical strength can be
achieved when the particle surfaces are not fully covered with the inorganic phase,
so that polymer-polymer contacts can be made. Upon film formation, the inor-
ganic particles are dispersed in a seamless polymer matrix. Fig. 7.7 shows an
example in which the phases are finely enough distributed to result in optical
clarity (Amalvy et al. 2001).
Film formation of clay-encapsulated particles produces a structure in which the
clay platelets are isolated at the boundaries between the polymer particles. In a
cross-sectional image, such as in Fig. 7.8, the structure has similarities to a honey-
comb. There are imperfections in the walls of the structure and distortions from a
hexagonal shape, which reflect non-uniformity in the clay shells of the original
particles. To achieve good mechanical integrity, such imperfections are most likely
favourable, as they enable polymer interdiffusion across particle boundaries.

7.3.1 Deposition Methods

In previous chapters, film formation by casting concentrated dispersions of latex


particles onto horizontal surfaces has been considered in depth. Substrates can
alternatively be dipped into colloidal dispersion and withdrawn at a constant

(a)
Transmittance (%)

(b)

Wavelength (nm)
Fig. 7.7 a TEM image of a cross-section of a nanocomposite film made from copolymer-silica
nanocomposite particles. The black phase is attributed to the silica. The copolymer contains n-
butyl acrylate, styrene, and 4-vinyl pyridine. The particles contain 30% silica nanoparticles.
b The resulting optical transmittance spectra (UV to near IR) of similar nanocomposite films,
containing 42% (dashed line) and 56% (solid line) silica nanoparticles, confirm their clarity.
(Reprinted from Amalvy et al. (2001), copyright 2001 American Chemical Society)
228 7 Nanocomposite Latex Films and Control of Their Properties

Fig. 7.8 TEM image showing the honeycomb-like structure that has resulted from the film
formation of particles with a P(St-co-BA) core encapsulated by laponite clay (shown in Fig.
7.3c). (Reprinted from Negrete-Herrera et al. 2007; copyright Wiley-VCH Verlag GmbH & Co.
KGaA; reproduced with permission)

velocity in a vertical deposition process (Wang et al. 2008b; Kitaev and Ozin
2003, Wang and Zhao 2007). Vertical deposition is highly effective provided
that the evaporation velocity of the solvent exceeds the sedimentation velocity
of the particles (Shimmin et al. 2006). The horizontal deposition of colloidal
particles across a large area usually creates a film with defects (e.g., vacan-
cies) and with lateral variations in the type of packing (e.g., face-centred
versus body-centred). Both these types of packing are apparent in the thin
layer in Fig. 7.9. Of course, ordered packing is dependent on having particles
of the same size. The figure shows how particles of the wrong size can create
defects in the packing structure. In comparison to horizontal deposition (i.e.,
conventional film formation), vertical deposition has often been found to
produce superior quality colloidal films with fewer defects and a single
packing type.
A third deposition method is to spread the dispersion on a substrate that ro-
tates about a central axis in a spin-casting process (Wang and Mhwald 2004,
Mihi et al. 2006, Jiang and McFarland 2005). In spin-casting, outward flow
thins the film, and centrifugal shear forces smooth the film. There is growing
interest in the use of surface patterns to assist and guide the particle deposition,
and this constitutes a way to control particle deposition by the other three
methods. Unlike horizontal and vertical deposition methods, which usually
7.3 Colloidal Particle Deposition and Assembly Methods 229

result in close-packed colloidal particles, spin-casting can fabricate either close-


packed (Wang and Mhwald 2004, Mihi et al. 2006) or isolated (Venkatesh et
al. 2007, Jiang and McFarland 2005, Jiang 2006, Sun et al. 2007) particle arrays
depending on the process conditions. The methods of vertical deposition and
spin-casting are illustrated in Fig. 7.10.

mica BCC

FCC

Particles of wrong size


Fig. 7.9 Atomic force microscopy image of latex particles in colloidal crystals. Boundaries
between apparent FCC and BCC crystals are observed. Particles that are slightly larger than
the others contribute to packing defects and disruption of the colloidal crystal. Image size is
4 m x 4 m.

7.3.2 Vertical Deposition

A schematic diagram of the vertical deposition process is presented in Fig. 7.10a.


During vertical deposition, convective mass flow and capillarity are the two main
driving forces for the colloids to assemble at the substrate surface (Dziomkina and
Vancso 2005). The layer thickness can range from several nanometres (corre-
sponding to a particle monolayer) to several micrometers. Because of the tendency
for sedimentation of larger colloidal particles, there is a limitation to the maximum
particle size that can be deposited on a substrate by the vertical method. Using
water as the continuous phase, the maximum colloid size deposited on a substrate
is as large as 2 m (Dimitrov and Nagayama 1996).
230 7 Nanocomposite Latex Films and Control of Their Properties

(b)
Outward flow

(a) Substrate

(c)

Fig. 7.10 Alternative methods of film formation. a Vertical deposition in which a substrate is
lifted out of a colloidal dispersion. b Spin-casting in which the colloidal particles flow outward
as the substrate is spun about its central axis. c An AFM image of particles deposited onto a mica
substrate by spin-casting. The particles are spaced apart and arranged in a hexagonal array.
Image size is 2 m x 2 m. (Reprinted with permission from Wang and Keddie (2009))

The solvents evaporation rate and surface tension affect the deposition rate and
the packing order (Kitaev and Ozin 2003, Cong and Cao 2003). For example,
when a surfactant is added to decrease the surface tension, a simple cubic array of
spaced particles is more favourable, rather than a densely packed, hexagonal array,
which is favoured with a higher surface tension (Cong and Cao 2003). Particles
are drawn together as a result of the negative capillary pressures generated by a
concave meniscus in the neck region between particles. This pressure is directly
proportional to the surface tension, so that decreasing the surfactant concentration
below the critical micelle concentration decreases the extent to which particles are
drawn together into a dense array.

7.3.3 Surface Pattern-Assisted Deposition

Patterns on the surface of the substrate can be used to control the assembly of
colloidal particles. This method has gained considerable attention for reasons of
fundamental understanding and for advanced device fabrication (Lee et al. 2004,
Schaak et al. 2004, Masuda et al. 2004, Ren et al. 2007, Chen et al. 2000). Patterns
on the substrate can be either chemical or topographical. Particle assembly is
achieved by any of the three methods of particle deposition (horizontal, vertical or
spin-casting), but the surface pattern directs the particles during the process.
Spatial confinement leads to the colloidal assembly.
7.3 Colloidal Particle Deposition and Assembly Methods 231

Chemically patterned surfaces usually have a lateral variation or modulation of


electrostatic forces or charges, wettabilities, hydrophilicities, and so on. A scheme
of colloidal assembly with chemical patterning is provided in Fig. 7.11a. Colloidal
particles with a certain surface chemistry will selectively locate on the patterned
regions. For instance, electropositive particles will attach to negative regions, or
hydrophobic particles will preferentially adsorb onto hydrophobic regions of the
substrate. In topographical patterning (Xia et al. 2003), various arrays of holes,
grooves or micro-channels on substrates are used to guide particle assembly, as
illustrated schematically in Fig. 7.11b. Both methods of surface patterning require
a sequence of complex manufacturing processes to fabricate the substrate.
Fustin et al. (2004) investigated the parameters influencing the templated
growth of colloidal particles on chemically patterned surfaces by vertical deposi-
tion. In this case, silanol groups were used to create hydrophilic surface regions,
and fluoroalkysilane monolayers were used to create hydrophobic surface patterns.
The particles assembled onto the hydrophilic substrate by capillary forces.
Unlike deposition on a non-patterned surface, on which the colloid concentra-
tion is uniform along the air-solvent-substrate contact line, there is a concentration
variation along the contact line when colloids are deposited over a hydrophilic
stripe. In addition to the usual colloid flux, from the bulk of the dispersion to the
drying zone, there is a lateral particle flux from the hydrophobic region to the
hydrophilic region to counter the discontinuous colloidal concentration. In vertical
deposition, the liquid in contact with the hydrophilic stripe rises above the
dispersion to a certain height, which is defined as the meniscus length, L. The
liquid in contact with the hydrophobic stripe sinks below the solution. At the two
sides where the hydrophilic and the hydrophobic regions meet, there is an abrupt
change of the wettability and consequently a strong deformation of the meniscus
shape. When the colloids are deposited on a patterned substrate, the pattern
dimensions (line widths) have a strong influence on the deposited layer thickness
and colloidal crystal quality.

Fig. 7.11 a Colloidal particle assembly on a chemically patterned surface using the vertical
deposition process. Redrawn after Fustin et al. (2004). b Colloidal particle assembly on surface
channels using the horizontal deposition process. (Reprinted with permission from Wang and
Keddie (2009))
232 7 Nanocomposite Latex Films and Control of Their Properties

In addition to substrate patterning, electrostatic (Aizenberg et al. 2000), exter-


nal electric field (Zhang and Liu 2004, Hayward et al. 2000), floating self-
assembly (Liu et al. 2005), centrifugation (Deng et al. 2008), filtration (Hoa et al.
2006), and other methods (Prevo and Velev 2004, Pan et al. 2006, Malaquin et al.
2007) have all been investigated as a means for colloidal particle assembly.

7.3.4 Long-Range Order from Self-Assembled Core-Shell Particles

When nanocomposite particles are assembled into a regular array, it is referred to as


a colloidal crystal. There is ordering of the nanosized phases over longer length
scales, as determined by the symmetry of the lattice structure. The result of particle
ordering is a regular modulation of phases that can yield attractive properties.
Depending on the contrast and relative sizes of the two phases, the optical, thermal
and mechanical properties will modulate throughout the bulk of the material.
Particles have created highly ordered structures across areas as large as 2 cm x 2 cm
in films made by vertical deposition and slow evaporation (Wohlleben et al. 2007).
The iridescence observed in natural opal results from the modulation of refrac-
tive index that stems from the silica particles that make up its structure. Hard
homogeneous latex particles can achieve a modulation in the refractive index
(from the difference in index between a polymer and air in the interstitial voids).
However, synthetic opal made from glassy particles would have poor mechanical
integrity. Hard particles do not form a cohesive film.
Core-shell particles, in which the shell and core polymers have distinctly dif-
ferent refractive indices, can be used to create synthetic opal structures. A particle
shell can create a continuous phase, in which the particle cores are distributed in
an ordered way. In an early attempt, film formation at elevated temperatures and
under pressure (so-called melt compression), was used to achieve the ordering of
particles into a regular array (Ruhl et al. 2003). To prevent the rearrangement of
the two polymer phases or phase separation, the particle cores were crosslinked
and the shell polymer was grafted to the core. The structure was locked into
position. The colour of light that is scattered from an artificial opal depends on the
angle of incidence of the light, as described by Braggs law. The reflected colour
can be adjusted through the particle size, in a way analogous to how X-ray
diffraction is affected by the spacing between atoms. Any colour of the rainbow
or indeed in the near infrared is possible. To achieve light scattering, a refractive
index difference between the core and shell is required. Polystyrene is a good
candidate polymer for the core, owing to its higher index in comparison to acrylic
polymers.
Artificial opal films can be created at room temperature using a vertical deposi-
tion method, when the particles have a thin rubbery shell (less than 20 nm thick)
around a glassy core (135 nm). The shell deforms to fill the void space and
imparts mechanical strength (Wang et al. 2006b). Deformable opal films have the
interesting characteristic that their colour changes as the material is strained. The
7.4 Colloidal Nanocomposites from Particle Blends 233

distance between the particle cores increases and hence the lattice spacing.
Recoverable strain requires some deformability of the particle cores. Otherwise,
surface roughening and film cracking results (Wohlleben et al. 2007).

7.4 Colloidal Nanocomposites from Particle Blends

A straightforward way to create nanocomposites is through the blending of latex


particles with other nanoparticles suspended in a common media (usually water).
This method can essentially be considered a nanoscale version of the processing of
pigmented latex films, which contain micrometer-sized particles of inorganic
oxides, such as titania, silica or calcite.

7.4.1 Advantages of Particle Blends

As an alternative to the colloidal approach, of special interest here, nanocompo-


sites can be created by mixing inorganic nanoparticles with a polymer dissolved in
a good solvent. In this solution route, the polymer and inorganic phase must be
soluble in the same solvent. Hydrophobic polymers therefore require that the
inorganic phase is suspended in an organic solvent for the polymer. Inorganic
particles can also be blended with a polymer melt to create a nanocomposite.
However, polymer melts have relatively high viscosities, so that blending with
inorganic phases is energy intensive. Lower melt viscosities are only obtained at
temperatures well above the polymer Tg, and inorganic particle agglomeration can
still be problematic. It is a challenge to disperse nanoparticles finely without
clustering in a polymer melt.
In comparison, the colloidal route of blending inorganic nanoparticles in water
with latex particles avoids the problems of solution and melt-processing (Gros-
siord et al. 2005). The latex can be made from a hydrophobic polymer; there are
no restrictions on its solubility in a solvent. The processing temperature only
needs to be slightly higher than the glass transition temperature of the polymer. It
is easier to disperse the inorganic particles without clustering.

7.4.2 Dispersion of Nanoparticles

A key challenge is to finely disperse the inorganic nanoparticles in water, and this
topic will be considered next. To overcome the fundamental problem of aggrega-
tion and poor dispersability, inorganic nanoparticles can be decorated with a layer
of polymers. These molecules impart steric stability to prevent the nano-
inorganics from approaching separation distances in the attractive region
234 7 Nanocomposite Latex Films and Control of Their Properties

(Shvartzman-Cohen et al. 2004). Techniques of steric stabilisation have been


employed to stabilise various types of colloidal particles, including silica (Li et al.
2005), clay (Shah et al. 2005), gold (Ohno et al. 2002), Fe3O4 (Lattuada and
Hatton 2007) and high aspect-ratio carbon nanotubes. This is through either
covalent bonding (Pyun and Matyjaszewski 2001) of the stabiliser onto the
inorganic surface or through physisorption onto the surfaces. To fabricate nano-
composites from waterborne polymer colloids, nanofillers are usually modified
with water-soluble molecules, such as polymers and surfactants, to disperse them
in water. Table 7.3 summarises the ordered colloidal nanocomposites deposited
from various colloidal blends with latex, which have been reported recently.

Table 7.3 Examples of Blends of Polymer Colloids with Inorganic Nanoparticles


Polymer Inorganic Dispersant for Film formation References
nanoparticle the inorganic method
particles
PVAc SWNT GA Horizontal Grunlan et al.
deposition 2004, 2006
P(St-BuA) MWNT SDS Horizontal Dufresne et al.
deposition 2002
PS, PMMA SWNT SDS, GA Horizontal Regev et al.
deposition 2004, Grossiord
et al. 2005
P(BuA-co-AA) SWNT PVA Horizontal Wang et al. 2006
deposition
P(BuA-co-AA) MWNT PVP, PVA, SDS, Horizontal Wang et al. 2008
TX deposition
PVAc Carbon black - Horizontal Grunlan et al.
deposition 2001
Sulfonated latex Au - Vertical Tessier et al.
deposition 2000, 2001
PS SiO2 - Vertical Wang et al.
deposition 2008b, Kitaev
and Ozin 2003
PVP = poly(vinyl pyrrolidone); PVA = poly(vinyl alcohol); SDS = sodium dodecyl sulfate;
GA = gum arabic; TX = Triton X-100 (surfactant)

Stabilisation and dispersion of carbon nanotubes is of particular interest in


high-strength nanocomposites, and review articles have been written on the use of
surfactants (Vaisman et al. 2006) and chemical functionalisation (Lin et al. 2007).
Note that because of the short-ranged and steep interaction potentials between
nanotubes, separating nanotubes by just a few nm is highly effective in stabilising
them in a dispersion (Shvartzman-Cohen et al. 2004b, Yerushalmi-Rozen and
Szleifer 2006).
Aggregates of inorganic nanoparticles in particle blends lead to a loss of optical
clarity because of the introduction of heterogeneity in the refractive index at
7.4 Colloidal Nanocomposites from Particle Blends 235

length scales approaching the wavelength of light (Vandervorst et al. 2006).


Aggregates of hard particles without a polymer binder between the particles can
create weak points in nanocomposites. Hence, effective methods of dispersing
nanoparticles to prevent agglomerates or bundles are essential. Polymer nanocom-
posites by colloidal routes have the potential to resolve the problem with melt-
processing of achieving uniform mixing at nanometre length scale. Nevertheless,
blends of latex and inorganic nanoparticles are prone to phase separation, with the
inorganic particles forming clusters (Tiarks et al. 2007). Hence, there is continuing
interest in, and a preference for, nanocomposite or hybrid particles.

7.4.3 Long-Range Order in Particle Blends

An ordered, nanostructured nanocomposite can be created from blends of


nanoparticles in a common media via horizontal or vertical deposition. Taking
nanotubes as an example, a nanocomposite is created by blending finely dispersed
nanotubes and colloidal polymer particles uniformly in water. After deposition on
the substrate, the water evaporates and the polymer particles assume a close-
packed arrangement with the nanotubes occupying the interstitial void space
(Grunlan et al. 2004 and 2006). Finally, the polymer particles will coalesce to
form a coherent film, locking the SWNTs within a so-called segregated network
(Grunlan et al. 2006). This method was first demonstrated by Grunlan et al. and
has been increasingly used by others.
Blends of hard (high Tg) and soft (low Tg) are used to tailor the mechani-
cal properties of latex films. Film formation can be achieved under ambient
temperatures, owing to the soft latex, but the elastic modulus and hardness will
be raised, owing to the hard particles. There is also an effect on film cracking.
Whereas hard particles alone will produce a cracked film, cracking is avoided
when the particles are incorporated in a film-forming matrix. In a recent study,
cracking was avoided below a critical volume fraction of hard particles of about
0.45 (Pauchard et al. 2009).
In blends of spherical particles, the size ratio of the particles largely determines
the continuity of the two phases and the type of packing. As a general rule, a high
size ratio of large-to-small particles will allow the smaller particles to pack
completely around the larger particles, even when the volume fraction of the
smaller particles is very low. In this case, the number ratio of small-to-large
particles is high. There can be many more small particles than large, even though
their volume fraction is low.
Fig. 7.12 makes these concepts clearer. The grey particles make up the majority
of the particles by volume, and the black particles make up the same volume
fraction in both illustrated examples. However, when the black particles are
consistently smaller than the grey particles, they can pack around them and
comprise a continuous phase. For a given size ratio, there is a critical volume
fraction, Vc, at which there are enough of the smaller particles to fit around the
236 7 Nanocomposite Latex Films and Control of Their Properties

larger particles. Fig. 7.13 compares structures that are below and above the Vc. For
a size ratio of 6:1, it has been found that the small particles must make up about
20% by volume in order for them to create a continuous phase in the film (Tzitzi-
nou et al. 1999). There is more than one way to define Vc. It can be calculated
from the number of small particles to create a monolayer on all of the large
particles. In the film, there will then be a bilayer of the small particles between the
larger ones. Alternatively, it can be calculated from the number of small particles
to create a continuous path around the large particles (Kusy 1977).
Note that for large and small particles to contribute an equal volume in the blend,
the number of the small particles must be greater. As an example, let us imagine a
blend in which the large:small size ratio is 10:1. In this case, only one out of every
thousand particles needs to be large in order to achieve 50% by volume of the large
particles. If the number fraction of large particles is only 0.00011, the large particles
will constitute 10% by volume in the blend. A number fraction of large particle of
0.01865 will surprisingly achieve 95 volume percent of large particles.
A nice example of the effect of particle size ratio in bimodal latex blends is
presented in Fig. 7.14. The minimum film formation temperature (MFFT) was
measured as a function of the size ratio of soft to hard particles. (In this context,
soft particles have a glass transition temperature, Tg, below room temperature, and
hard particles have a Tg above it.). In films in which the hard and soft particles are
blended with equal masses, there is a steady increase in the MFFT as the soft
particle size becomes larger (Colombini et al. 2004). The reason is that the smaller
hard particles pack more efficiently around the larger soft particles in comparison
to equal-size particles. With greater continuity of the hard particle phase, the film
acts harder and has a higher MFFT. The small hard particles have voids between
them that scatter light, unless the soft polymer phase has sufficiently low viscosity
to flow into the voids.

Low ratio of large-to-small sizes:


Isolation of small particles High ratio of large-to-small sizes:
Continuity of small particles

Fig. 7.12 A comparison of particle packing in particle blends of two phases (grey and black).
When the black particles are only slightly smaller than the grey particles, they remain isolated
when the particles pack together. When the black particles are much smaller than the grey
particles, they can create a continuous layer.
7.4 Colloidal Nanocomposites from Particle Blends 237

(a)

(b)

Fig. 7.13 Illustration of the concept of the critical volume, Vc, for continuity of the small
particles. a When the volume fraction of the small particles is < Vc, the particles are isolated. b
When the volume fraction is > Vc, the particles create a continuous path.

Fig. 7.14 Minimum film formation temperatures (MFFT) for films made from hard and soft
particles blended in a 1:1 ratio by mass. The soft particles had a Tg of ca. 2 C, and the hard
particles had an average Tg of 116 C. The surface-average particle size of the two particle types
was varied, while the mass ratio was fixed. MFFT increases with an increasing ratio of the
surface-average soft particle radius, Rsoft, to the surface-average hard particle radius, Rhard.
(Reproduced from Colombini et al. 2004)
238 7 Nanocomposite Latex Films and Control of Their Properties

Colloidal crystals can be created from blends of large and small particles, pro-
vided that the size distribution of each population is narrow. There is a direct
analogy with ionic crystals that are made up of small cations and large anions. The
smaller particles fit into the void space between the large particles in a densely
packed array. Each particular crystal type requires a certain size ratio of
large:small particles. Table 7.4 lists the required ratios for three different crystal
types. Unlike the previous examples, in which the number of small particles was
much greater than the number of large particles, these colloidal crystal structures
dictate that the number ratio of large:small is 1:1.

Table 7.4 Large:Small Size Ratios Required for Colloidal Crystals Made from Bimodal Blends
of Latex Particles
Crystal Salt Analogue Space Occupied Number of Large:Small
Structure by Small Neighbours Size Ratio
Particle Around Each
Small Particle
Simple CsCl Cubic 8 1.37:1
Face-centred NaCl Octahedral 6 2.41:1
cubic
Diamond ZnS Tetrahedral 4 4.45:1

7.5 Three Lessons about the Properties of Waterborne


Nanocomposite Films

The recent spurt of growth of research on waterborne nanocomposites allows us to


learn some lessons for future developments. Our interest here is on how the
particle structures and the control of the particle arrangement in film formation
provide a means of tailoring film properties.

7.5.1 Lesson One

A percolating phase, although a minority in the composition, has a strong


influence on the properties of the film.
Percolation is defined as the situation in which a minority phase creates a con-
tinuous path across the majority phase. It is possible to travel from one side of
the material to the other without leaving the percolating phase. The lowest
concentration of the minority phase at which percolation is achieved is called the
percolation threshold.
7.5 Three Lessons about the Properties of Waterborne Nanocomposite Films 239

7.5.1.1 Percolation of Spherical Particles

Computer simulations have found that when equal-sized particles of two types
(say A and B) are randomly packed together, the particles in the minority will
create a percolating cluster when a fraction of 0.326 of the sites are occupied by
the minority particles (Frith and Buscall 1991). This cluster is defined by condi-
tions where there are continuous contacts between the same types of particles
across the entire layer.
At the point when a rigid percolating network is created, there is a strong change
in the mechanical properties. The network acts like a skeleton that raises the elastic
modulus of the composite. Chevalier et al. (1999) studied hard particles blended in a
matrix created by soft particles (film-forming at room temperature). They showed
that percolation of the particles had some effect on the modulus, but the effect was
more pronounced after the nanocomposite was heated above the Tg of the hard
phase. Interdiffusion between particles fused the particles into a rigid skeleton. By
comparison, when contact between the hard particles was prevented by encapsulat-
ing the particles with a soft shell, the hard particles had a much smaller effect.
A percolation model was able to predict the observed increase in the elastic
modulus by about three orders of magnitude (Fig. 7.15), when the volume fraction
of hard particles rose above the percolation threshold (Chevalier et al. 1999). The
percolation threshold is estimated to be at 0.3 in this system. The hard particles were
smaller than the soft particles comprising the continuous matrix, so the percolation
threshold is lower than the predictions for the packing of equal-sized particles.

Fig. 7.15 Dynamic shear modulus (real part) of latex films made from blends of hard
poly(styrene) (PS) particles and soft poly(n-butyl acrylate) particles, for varying volume
fractions of PS. The experimental points are shown with the filled circles. The solid line is the
prediction of a phenomenological mechanical model for the percolation assuming a percolation
threshold of 0.30. (Reprinted with permission from Chevalier et al. 1999, copyright 1999
American Chemical Society)
240 7 Nanocomposite Latex Films and Control of Their Properties

7.5.1.2 Percolation of Rod-Like Particles

Fig. 7.16 shows the structure of a film in which the concentration of rod-like
nanofillers is below and above the percolation threshold. As a general rule, low
percolation thresholds of rod-like fillers require particles that are randomly
oriented and are finely dispersed (not agglomerated or bundled). The colloidal
route can satisfy both these requirements. Relatively large filler particles, i.e.,
micrometer-sized carbon black particles, have a percolation threshold of 2 vol.%
in blends with a latex (Grunlan et al. 2001). On the other hand, the percolation
threshold for the onset of electrical conductivity in a nanocomposite of colloidal
poly(vinyl acetate) latex and single-wall carbon nanotubes (SWNTs) has been
found to be as low as 0.04 wt% (Regev et al. 2004).
The percolation threshold of randomly oriented, needle-like particles is in-
versely related to their aspect ratio, defined as the ratio of their length over their
width. Therefore, because nanotubes have a high aspect ratio, they can achieve a
low percolation threshold (Foygel et al. 2005). A small number of longer nano-
tubes and a weak attraction between the nanotubes both reduce the percolation
threshold significantly. However, the flexibility and straightness of the nano-
tubes have only a minor effect (Kyrylyuk and van der Schoot 2008). The percola-
tion of electrical conductivity in nanocomposites of polymers and carbon
nanotubes has been reviewed (Bauhofer and Kovacs 2009).

(a)

(b)

Fig. 7.16 Illustrations of rod-like, inorganic particles dispersed in a polymer film. a At concen-
trations below the percolation threshold, some particles are in contact but there is not a path
across the film. b Above the percolation threshold, there a paths in all directions moving across
the film.

The 3-D percolating network of a conducting inorganic phase can impart high
electrical conductivity, and so measurements of conductivity are a good probe of
the structure. A three-dimensional honeycomb network of SWNTs created around
PS colloidal particles has an electrical conductivity on the order of 104 Sm1, but it
is insulating below the threshold (Dionigi et al. 2007). A network of a conducting
polymer in a colloidal nanocomposite achieved a percolation threshold of 0.011,
7.5 Three Lessons about the Properties of Waterborne Nanocomposite Films 241

which is a factor of 10 lower than in a conventional blend, accompanied by an


electrical conductivity that is several orders of magnitude greater (Mezzenga et al.
2003). Elsewhere, a soft colloidal nanocomposite of poly(butyl acrylate) and
SWNTs achieved a percolation threshold of only 0.03 wt.% (Wang et al. 2008).
These colloidal nanocomposites can serve as electrically conductive adhesives and
coatings and find applications in electronic devices and displays.

7.5.1.3 Properties in Percolating Systems

Percolation can also be achieved by the film formation of core-shell particles. If


the shell is thin, then its volume fraction will be in the minority. When the
particles are packed together to create a film, this minority shell phase will be
continuous through the material. Even though it is a minority phase, it can be
dominant in determining the properties. By comparison, in blends of particles, the
minority phase is often isolated.
Colloidal methods of nanocomposite processing use building blocks with a
high surface area, which result in nanocomposites with an extremely high internal
interface area. For instance, the interfacial area between particles with a radius of
125 nm (assuming particle deformation but not coalescence) is greater than 10 m2
for every gram of polymer. A shell material will cover this high interface area and
therefore be efficiently dispersed.
The presence of a percolating shell network can have a pronounced effect on
mechanical properties. An excellent example is provided by films made from
particles with a soft, rubbery core and a stiff, glassy shell (Domingues dos Santos
and Leibler 2003). The nanocomposite films have a continuous glassy phase. The
elastic modulus of the film is much higher than that of the core polymer. When the
elastic limit is exceeded, the films show a yield point followed by plastic deforma-
tion. The mechanical response of the core-shell films differs remarkably from
blends of particles made from the same polymers as in the core and the shells. The
volume fraction of the glassy phase in the particle blends was kept the same as in
the core-shell material, so that a fair comparison could be made. In the particle
blend films, the glassy particles are isolated in the soft matrix. In this case, the
films are tacky and flow under a low applied stress. Fig. 7.17 shows schematically
the two types of structures and the resulting mechanical properties.
We can also consider an example of inorganic-polymer nanocomposite parti-
cles. Laponite clay platelets have been used to encapsulate particles with a
styrene-acrylate copolymer core. When these particles are assembled onto a
nanocomposite, a percolating network of clay platelets is created (Negrete-Herrara
et al. 2007) in which the clay platelets are arranged in a three-dimensional honey-
comb-like structure. The laponite decorates the boundaries between the deformed
particles. This honeycomb network is quite efficient in mechanical reinforcement,
giving a 50-fold increase in Youngs modulus in comparison with the polymer
alone. There is also a strong increase in the maximum stress before failure and in
the toughness. See Fig. 7.18 for a comparison of the stress-strain relationship of
242 7 Nanocomposite Latex Films and Control of Their Properties

the nanocomposite and a copolymer film without clay. With the addition of clay,
the elastic modulus (indicated by the slope at small strain) increases. More
importantly, the nanocomposite can be strained by a similar amount as the
copolymer without failure. Thus, the addition of clay increases the area under the
stress-strain curve and hence the toughness.
To see a pronounced effect on the mechanical properties of a core-shell film, a
combination of glassy and rubbery polymers is not necessarily required. There is a
recent example of a soft-soft nanocomposite. This material is made from
particles that have a shell that is capable of crosslinking. When film-formed,
percolating elastic, crosslinked network is created. This nanomaterial was specifi-
cally designed for adhesives to achieve a softening of the material under moderate
strains and a stiffening of the material at high strains (Deplace et al. 2009).

Glassy
Yield point Soft core
shell
Core-shell latex
Glassy
Soft matrix particles

Particle blend

Strain
Fig. 7.17 A comparison of the deformation of two types of nanocomposite films: core-shell
structure versus particular blend structure. The phase shown in black is a glassy polymer, and the
grey phase is a rubbery polymer. Both films contain 20% of the hard phase. The core-shell
structure is stiffer and is tougher. After its elastic limit is reached, the film yields and deforms
plastically. The blend film does not show yielding but instead flows under low stress. (Adapted
from Domingues dos Santos and Leibler (2003); copyright 2003; reprinted with permission of
John Wiley & Sons, Inc)

7.5.1.4 Properties of Hybrid and Blend Systems

Core-shell particles are not the only way to create continuous phases in a film.
Hybrid latexes prepared by miniemulsion polymerisation are expected to have the
two polymer phases mixed at the molecular level. By comparison, blends of
particles tend to have a continuous phase created by the majority phase. Domains
of the other phase are expected to be the size of the particles or larger.
7.5 Three Lessons about the Properties of Waterborne Nanocomposite Films 243

Fig. 7.18 Tensile stress-strain curves for loading and unloading of latex films: a poly(styrene-co-
butyl acrylate) and b nanocomposite of the same copolymer and Laponite. Film nanostructure is
shown in Fig. 7.8. (Reproduced from Negrete-Herrera et al. 2007; copyright Wiley-VCH Verlag
GmbH & Co. KGaA; reproduced with permission)

These expectations were met in hybrid latex made from polyurethane and an
acrylic copolymer via miniemulsion polymerisation. The two phases were co-
continuous in the film. By comparison, films made from blends of the two
particles had larger, more distinct phases (Wang et al. 2005). These differences in
structure had an influence on the mechanical properties. The hybrid film had a
higher yield stress (the point where the material is plastically deformed), and a
larger strain at break in comparison to the film made from blends of particles. This
example shows that there can be real advantages from mixing phases at the
molecular level to create a co-continuous structure.
In blends of hard (glassy) and soft (rubbery) particles, the continuity of the two
phases is determined by their particle size ratio and volume fractions, as already
pointed out. For instance, in a blend that consists of a 1:1 ratio by volume of the hard
and soft phases, the elastic modulus at intermediate temperatures is a function of the
particle size (Colombini et al. 2004). In an example, illustrated in Fig. 7.19, when
the size of the soft particles is 0.36 times the size of the hard particles, there is good
continuity of the soft phase, and the modulus is relatively low, but higher than for
the pure soft phase. When the soft particles were larger than the hard particles, they
were less continuous, and the hard particles made a greater contribution to increas-
ing the modulus. The modulus can be raised substantially through the addition of a
hard phase in a particle blend film, even when the hard phase is in the minority by
volume. The smaller the size of the hard particles in comparison to the larger ones,
the more effective they are in raising the modulus.
244 7 Nanocomposite Latex Films and Control of Their Properties

Fig. 7.19 Storage modulus as a function of temperature for films made from blends of hard (Tg
120 C and soft (Tg 20 C particles. Results from two films are compared each with equal
amounts by mass of the hard and soft polymer. The data shown with the open squares are from a
film in which the radius of the soft particles, Rsoft, is lower than for the hard particles, Rhard. The
data shown with the filled circles are from a film in which Rsoft > Rhard, as indicated. In the latter
case, the hard particles have more connectivity, and the modulus is higher in the temperature
region between the two Tg values. (Data from Colombini et al. 2004)

7.5.2 Lesson Two

In nanocomposite particles for coatings, the surface of the particles has a domi-
nant influence on the surface properties of the coating. This is an important lesson
when designing particles for particular coatings applications.
Inorganic phases are often introduced into coatings to increase scratch resis-
tance. The friction of a coating is set by the material at the surface. Thus, if an
inorganic phase is encapsulated inside a particle, it will be buried slightly beneath
the outermost surface of the coating. The inorganic material will then make a
lower contribution to the films hardness. For instance, nanocomposite films have
been made from acrylic polymers and silica (SiO2) nanoparticles. High scratch
resistance was achieved, because the hard silica particle decorated the latex
particle and hence film surfaces. The hard particles were therefore at the film
surface and not buried. (Tiarks et al. 2007). Blends of inorganic nanoparticles and
latex usually have a patchier surface with the inorganic appearing in clusters. In
this case, some of the polymer is left exposed and is sensitive to scratching.
7.5 Three Lessons about the Properties of Waterborne Nanocomposite Films 245

Another surface-sensitive characteristic is blocking resistance. This is defined


as the tendency to avoid adhesion between polymer surfaces when they are placed
in contact. Blocking (adhesion) results from polymer diffusion across the inter-
face, in a process similar to the crack healing discussed in Chapter 5. Silica (or
other inorganic particles) at the surface of particles is believed to prevent interdif-
fusion of polymers across contacting surfaces (Tiarks et al. 2007). The silica
nanoparticles presumably create a physical barrier to diffusion across the film-film
interface. Hence, blocking resistance is increased.
Nanocomposite particles are now being produced industrially for use as a
binder in architectural paints. Inorganic nanoparticles that are homogeneously
incorporated into polymer latex particles are being sold under the trade name of
Col.9 (Weitenkopf 2009). After film formation, a three-dimensional network of
nanoparticles is created. It is reported that the inorganic particles make the binder
less tacky especially at higher temperatures in comparison to standard thermo-
plastic binders. Hence, there is less pick-up of dirt. However, the polymer phase
prevents the binder from being too brittle, so that cracking and loss of pigment
particles (called chalking) is prevented. Formulated paints using the nanocompo-
site particles as binders are now being produced and sold.
Polymer-polymer nanocomposite latex coatings have similarly been used to
achieve blocking resistance. Latex particles that contain a hard polymer and a soft
polymer phase have been used. The hard polymer phase creates a rougher surface,
does not wet other surfaces, and inhibits interdiffusion. The soft polymer phase
enables particle coalescence and film formation. When there is too much of the
hard polymer present, film formation is not possible (Schuler et al. 2000).
The same sort of argument, regarding the importance of particle surfaces, ap-
plies to the wettability of surfaces. Poly(dimethyl siloxane), or PDMS, has been
incorporated into acrylic particles to increase the hydrophobicity. The contact
angle of water, which is used as a measure of a surfaces hydrophilicity, is
determined by the composition of the material near the surface. Hence, the
contribution of PDMS is lost if the phase is localised inside of the particle cores.
The greatest hydrophobicity has been found when the PDMS is copolymerised
with the acrylic polymer such that PDMS is found at the surface of the colloidal
particles (Rodriguez et al. 2009).

7.5.3 Lesson Three

The interface between a nanofiller and the matrix can have a profound impact on
the macroscale properties.
In the case of nanocomposites manufactured from blends of particles, Wang et al.
(2008) have pointed out that the dispersant for waterborne inorganic nanoparticles,
such as a surfactant or a water-soluble polymer, becomes localised at the interface
with the matrix. Fig. 7.20 shows the effect schematically. One does not expect to
find a clean interface between an inorganic nanoparticle and the polymer matrix.
246 7 Nanocomposite Latex Films and Control of Their Properties

MWNT
(a)

Water

Dispersant Latex particles

(b)

Film
MWNT

Interface without PVP

(c)

Fig. 7.20 a Carbon nanotubes are dispersed in water using hydrophilic polymers. b After film
formation, the latex polymer creates a matrix in which the carbon nanotubes (CNT) are
dispersed. A clean interface between the CNT and the matrix is not possible. c Polymers are
adsorbed on the CNTs. In the nanocomposite film, the polymers are isolated at the interface
between the CNT and the matrix. (Reprinted with permission from Wang et al. 2008c; copyright
2008 American Chemical Society)

Wang et al. found that the value of the adhesive criterion, tan /E', of the
poly(butyl acrylate) P(BuA)/MWNT colloidal nanocomposites can be tuned
through control of the interface structure. A small molecule (surfactant) or short
chain polymer yielded a lower tan/E' value compared to high molecular polymer
chains at the nanotube-matrix interface, and consequently lower adhesion energy
resulted. The adhesion energies of the nanocomposite adhesives can be adjusted
by the extent of chain entanglement and molecular friction at the interface
between the filler and the matrix.
Fundamentally, the performance and fracture mechanism of polymer nanocom-
posites containing high aspect-ratio fillers (e.g., nanofibres (Callister 2007), carbon
nanotubes (Coleman et al. 2006b), and nanoplatelets (Bonderer et al. 2008)) are
controlled by the interface strength between the matrix and the filler and by the
fillers aspect ratio. A nanofiller will break during deformation of the nanocomposite
only when the stress transferred is larger than its fracture strength. If the transferred
stress does not exceed the fillers fracture strength, the filler will pull out from the
matrix. Fig. 7.21 illustrates these two fracture mechanisms. The critical length, Lc,
7.5 Three Lessons about the Properties of Waterborne Nanocomposite Films 247

defined as the minimum length at which the stress transferred to a nanofiller equals
its fracture strength, is calculated for fibres and platelets as

fD
Lc = (7.1)
2

where f is the filler fracture strength, D the filler diameter, and the interfacial
strength. The critical length of carbon nanotubes (Coleman et al. 2006b) is
calculated as

f D Di 2
Lc = 1 (7.2)
2 D 2

with Di being the inner diameter of the nanotubes. The nanofiller will pull out
from the matrix when its length LNF is less than Lc, and it will fracture when LNF >
Lc. It might be expected that the fracture of most colloidal nanocomposites will
occur by nanofiller pull-out, because of the difficulty in achieving a high .

(a)

(b)

(c)

Fig. 7.21 The failure mechanisms in nanocomposites that contain nanosized fillers. a The interface
between the filler particles and the polymer matrix has an important impact on the macroscopic
properties. b If the interface is strong, the filler particles will fracture along with the matrix. c With
a weaker interface, the filler particles will pull out of the matrix but will not fracture.

For a nanofiller to increase the strength of a nanocomposite, stress must clearly


be transferred from the matrix to the filler. Recent work has used covalent bonding
of the matrix to the filler to ensure high interfacial strength (Lin et al. 2003). There
are various examples showing that an increase in leads to increased macroscale
strength and toughness. For instance, spherical nanoparticles in PP or PS systems
248 7 Nanocomposite Latex Films and Control of Their Properties

(Zhou et al. 2007) are highly effective in toughening the nanocomposite when the
testing temperature is above the glass transition temperature (Tg) of the polymer
matrix. At the higher temperatures, the polymer can wet the nanofillers and provide
good interface adhesion. A second example of interface effects is when a polymer is
grafted onto MWNTs to improve miscibility with the matrix. The efficient load
transfer from the matrix to the nanotubes increases the mechanical performance (Shi
et al. 2007). Yet another example is when a high density of polymer chains grafted
on SWNT surfaces leads to more effective interfacial stress transfer and results in a
pronounced mechanical reinforcement (Xie et al. 2007).

0.30
1
a Pure PBA 12341234
2 -3
0.25 c =2.5 x10
3 -3
=3.6 x10
Stress/MPa

0.20 4
=6.5 x10
-3

0.15 m

0.10

0.05
1 2 3 4
0.00
0 5 10 15 20 25 30
Strain

0.30 1
ba 2
N=9
N=90
= 0.170
= 0.157
0.25 3 N=3200 = 0.026
4 N=14144 = 0.010
0.20
Stress/MPa

0.15

0.10

0.05
1 2 3 4
0.00
0 5 10 15 20 25 30
Strain
Fig. 7.22 a Stress-strain relationship in nanocomposite films made from blends of poly(n-butyl
acrylate) latex and carbon nanotubes. The estimated density of the adsorbed poly(vinyl pyrroli-
done) (PVP) dispersant, (units of chains/nm2) is shown by each curve for a fixed degree of
polymerisation, N, for the PVP. Results from a film without nanotubes (just pure poly(butyl
acrylate)) is shown for comparison. b Similar data in which the degree of polymerisation of the
PVP is varied. Values are shown by each curve. (Reprinted with permission from Wang et al.
2008c, copyright 2008 American Chemical Society)
References 249

It is well established that polymer chain length (N) and density (, number of
chains per unit area) at interfaces influence fracture (Creton et al. 1992, Norton et
al. 1995, Dai et al. 1996), adhesion (Creton et al. 1994), and friction (Bureau and
Lger 2004, Casoli et al. 2001). Experiments by Wang et al. (2008c) systemati-
cally investigated the interface control of in a soft polymer-nanotube nanocom-
posite through variation of the chain length and density of the polymers at the
interface between the nanotubes and the continuous matrix. The macroscale
strength and toughness of the nanocomposites were profoundly affected by and
N, as demonstrated in the stress-strain curves in Fig. 7.22.
Analysis of these results showed that increases with , but then levels off
above a critical value. The value of (per chain) increases with increasing chain
length. The results are explained through an analogy between the force required in
the nanotube pull-out process and the friction force of polymer brushes on solid
surfaces sliding along a rubber. It is predicted that the monomeric friction coeffi-
cient of different polymers could be used to tune the interface strength and
consequently to adjust the macroscopic properties of colloidal nanocomposites.

References

Aizenberg J., Braun P.V., Wiltzius P. (2000) Patterned colloidal deposition


controlled by electrostatic and capillary forces. Phys Rev Lett 84: 2997-3000.
Ajayan P.M., Tour J.M. (2007) Materials science Nanotube composites. Nature
447: 1066-1068.
Amalvy J.I., Percy M.J., Armes S.P., Wiese H. (2001) Synthesis and characterisa-
tion of novel film-forming vinyl polymer/silica colloidal nanocomposites.
Langmuir 17 : 4770-4778.
Arsenault A.C., Puzzo D.P., Manners I., Ozin G.A. (2007) Photonic-crystal full-
colour displays. Nature Photonics 1: 468-472.
Asua J.M. (2002) Mini-emulsion polymerisation. Prog Polym Sci 27: 1283-1346.
Ayutsede J., Gandhi M., Sukigara S., Ye H., Hsu C.-M., Gogotsi Y., Ko F. (2006)
Carbon nanotube reinforced Bombyx mori silk nanofibres by the electrospin-
ning process. Biomacromolecules 7: 208 -214.
Bauhofer W. and Kovacs J.Z. (2009) A review and analysis of electrical percola-
tion in carbon nanotube polymer composites. Composites Science and Tech-
nology 69 (Spec. Issue S.I): 1486-1498
Bon S.A.F., Colver P.J. (2007) Pickering miniemulsion polymerisation using
Laponite clay as a stabiliser. Langmuir 23: 8316-8322.
Bonderer L.J., Studart A.R., Gauckler L.J. (2008) Bioinspired design and assem-
bly of platelet reinforced polymer films. Science 319: 1069-1073.
Bureau L., Lger L. (2004) Sliding friction at a rubber-brush interface. Langmuir
20: 4253-4529.
Callister W.D. (2007) Materials Science and Engineering, An Introduction. 7th ed.
John Wiley & Sons, Inc. 585-595.
250 7 Nanocomposite Latex Films and Control of Their Properties

Casoli A., Brendl M., Schultz J., Auroy P., Reiter G. (2001) Friction induced by
grafted polymeric chains. Langmuir 17: 388-398.
Castelvetro V., De Vita C. (2004) Nanostructured hybrid materials from aqueous
polymer dispersions. Adv Coll Interf Sci 108-109: 167-185.
Cauvin S., Colver P.J., Bon S.A.F. (2005) Pickering stabilised miniemulsion
polymerisation: Preparation of clay armoured latexes. Macromolecules 38:
7887-7889.
Caruso F., Susha A.S., Giersig M., Mhwald H. (1999) Magnetic core-shell
particles: Preparation of magnetite multilayers on polymer latex microspheres.
Adv Mater 11: 950-+.
Chen K.M., Jiang X., Kimerling L.C., Hammond P.T. (2000) Selective self-
organisation of colloids on patterned polyelectrolyte templates. Langmuir 16:
7825-7834.
Chen T., Colver P.J., Bon S.A.F. (2007) Organic-inorganic hybrid hollow spheres
prepared from TiO2-stabilised Pickering emulsion polymerisation. Adv Mater
19: 2286-+.
Cheng W., Wang J.J., Jonas U., Fytas G., Stefanou N. (2006) Observation and
tuning of hypersonic bandgaps in colloidal crystals. Nat Mater 5: 830-836.
Chevalier Y., Hidalgo M., Cavaill J.-Y., Cabane B. (1999) Structure of water-
borne organic composite coatings. Macromolecules 32: 7887-7896
Coleman J.N., Khan U., Gunko Y.K. (2006) Mechanical reinforcement of
polymers using carbon nanotubes. Adv Mater 18: 689-706.
Coleman J.N., Khan U., Blau W.J., Gunko Y.K. (2006b) Small but strong: A
review of the mechanical properties of carbon nanotube-polymer composites.
Carbon 44: 1624-1652.
Colombini D., Hassander H., Karlsson O.J., Maurer F.H.J. (2004) Influence of the
particle size and particle size ratio on the morphology and viscoelastic proper-
ties of bimodal hard/soft latex blends. Macromolecules 37: 6865-6873.
Colombini D., Ljungberg N., Hassander H., Karlsson O.J. (2005) The effect of
polymerisation route on the amount of interphase in structured latex particles
and their corresponding films. Polymer 46: 1295-1308.
Cong H.L., Cao W.X. (2003) Colloidal crystallisation induced by capillary force.
Langmuir 19: 8177-8181.
Creton C., Kramer E.J., Hui C.Y., Brown H.R. (1992) Failure mechanisms of
polymer interfaces reinforced with block copolymers. Macromolecules 25:
3075-3088.
Creton C., Brown H.R., Shull K.R. (1994) Molecular-weight effects in chain
pullout. Macromolecules 27: 3174-3183.
Cumpston B.H., Ananthavel S.P., Barlow S., Dyer D.L., Ehrlich J.E., Erskine
L.L., Heikal A.A., Kuebler S.M., Lee I.-Y.S., McCord-Maughon D., Qin J.Q.,
Rckel H., Rumi M., Wu X.-L., Marder S.R., Perry J.W. (1999) Two-photon
polymerisation initiators for three-dimensional optical data storage and micro-
fabrication. Nature 398: 51-54
References 251

Dai C.A., Kramer E.J., Washiyama J., Hui C.Y. (1996) Fracture toughness of
polymer interface reinforced with diblock copolymer: Effect of homopolymer
molecular weight. Macromolecules 29: 7536-7543.
Dalton A.B., Collins S., Munoz E., Razal J.M., Ebron V.H., Ferraris J.P., Coleman
J.N., Kim B.G., Baughman R.H. (2003) Super-tough carbon-nanotube fibres.
Nature 423: 703.
Deng Y., Liu C., Liu J., Zhang F., Yu T., Zhang F., Gu D., Zhao D. (2008) A
novel approach to the construction of 3-D. ordered macrostructures with poly-
hedral particles. J Mater Chem 18: 408-415.
Deplace F., Rabjohns M.A., Yamaguchi T., Foster A.B., Carelli C., Lei C.-H.,
Ouzineb K., Keddie J.L., Lovell P.A., Creton C. (2009) Deformation and adhe-
sion of a soft-soft nanocomposite designed with structured polymer colloid
particles. Soft Matter 5: 1440-1447.
Diaconu G., Paulis M., Leiza J.R. (2008) Towards the synthesis of high solids
content waterborne poly(methyl methacrylate-co-butyl acrty-
late)/montmorrillonite nanocomposites. Polymer 49: 2444-2454.
Diaconu G., Micusik M., Bonnefond A., Paulis M., Leiza J.R. (2009) Macromole-
cules 42: 3316-3325.
Dionigi C., Stoliar P., Ruani G., Quiroga S.D., Facchini M., Biscarini F. (2007)
Carbon nanotube networks patterned from aqueous solutions of latex bead
carriers. J Mater Chem 17: 3681-3686.
Domingues dos Santos F., Leibler L. (2003) Large deformation of films from soft-
core/hard-shell hydrophobic latices. J Polym Sci: Pt B: Polym Phys 41: 224-
234.
Dupin D., Schmid A., Balmer J.A., Armes S.P. (2007) Efficient synthesis of
poly(2-vinylpyridine)-silica colloidal nanocomposite particles using a cationic
azo initiator. Langmuir 23: 11812-11818.
Dziomkina N.V., Vancso G.J. (2005) Colloidal crystal assembly on topologically
patterned templates. Soft Matt 1: 265-279.
Frith W.J. and Buscall R. (1991) Percolation and critical exponents on randomly
close-packed mixtures of hard spheres. J Chem Phys 95: 5983-5989.
Foygel M., Morris R.D., Anez D., French S., Sobolev V.L. (2005) Theoretical and
computational studies of carbon nanotube composites and suspensions: Electri-
cal and thermal conductivity. Phys. Rev. B 71: 104201.
Fustin C.A., Glasser G., Spiess H.W., Jonas U. (2004) Parameters influencing the
templated growth of colloidal crystals on chemically patterned surfaces. Lang-
muir 20: 9114-9123.
Giannelis E.P. (1996) Polymer layered silicate nanocomposites. Adv Mater 8:
29-&.
Gilman J.W., Jackson C.L., Morgan A.B., Harris R., Manias E., Giannelis E.P.,
Wuthenow M., Hilton D., Phillips S.H. (2000) Flammability properties of
polymer-layered-silicate nanocomposites. Polypropylene and polystyrene nano-
composites. Chem Mater 12: 1866-1873.
Grossiord N.G., Loos J., Koning C.E. (2005) Strategies for dispersing carbon
nanotubes in highly viscous polymers. J Mater Chem 15: 2349-2352.
252 7 Nanocomposite Latex Films and Control of Their Properties

Grunlan J.C., Gerberich W.W., Francis L.F. (2001) Lowering the percolation
threshold of conductive composites using particulate polymer microstructure. J
Appl Polym Sci 80: 692-705.
Guyot A., Landfester K., Schork F.J., Wang C. (2007) Hybrid polymer latexes.
Prog Polym Sci 32: 1439-1461.
Halpin J.C., Kardso J.L. (1976) Halpin-Tsai equations review. Polymer Engi-
neering and Science 16: 344-352
Haraguchi K., Ebato M., Takehisa T. (2006) Polymer-clay nanocomposites
exhibiting abnormal necking phenomena accompanied by extremely large
reversible elongations and excellent transparency. Adv Mater 18: 2250-2254.
Harley S., Thompson D.W., Vincent B. (1992) The adsorption of small particles
onto larger particles of the opposite charge Direct electron-microscope stud-
ies. Colloids and Surfaces 62: 163-176.
Hasell T., Yang J.X., Wang W.X., Li J., Brown P.D., Poliakoff M., Lester E.,
Howdle S.M. (2007) Preparation of hybrid polymer nanocomposite microparti-
cles by a nanoparticle stabilised dispersion polymerisation. J Mater Chem 17:
4382-4386.
Hayward R.C., Saville D.A., Aksay I.A. (2000) Electrophoretic assembly of
colloidal crystals with optically tunable micropatterns. Nature 404: 56-59.
Hoa M.L.K., Lu M.H., Zhang Y. (2006) Preparation of porous materials with
ordered hole structure. Adv Colloid Interface Sci 121: 9-23.
Ji X.L., Jing J.K., Jiang W., Jiang B.Z. (2002) Tensile modulus of polymer
nanocomposites. Polymer Engineering and Science 42: 983-993.
Jiang P. (2006) Large-scale fabrication of periodic nanostructured materials by
using hexagonal non-close-packed colloidal crystals as templates. Langmuir
22: 3955-3958.
Jiang P., McFarland M.J. (2004) Large-scale fabrication of wafer-size colloidal
crystals, macroporous polymers and nanocomposites by spin-coating. J Am
Chem Soc 126: 13778-13786.
Jiang P., McFarland M.J. (2005) Wafer-scale periodic nanohole arrays templated
from two-dimensional nonclose-packed colloidal crystals. J Am Chem Soc
127: 3710-3711.
Karg M., Pastoriza-Santos I., Prez-Juste J., Hellweg T., Liz-Marzn L.M. (2007)
Nanorod coated PNIPAM microgels: Thermoresponsive optical properties.
Small 3: 1222-1229.
Kashiwagi T., Morgan A.B., Antonucci J.M., Van Landingham M.R., Harris R.H.,
Awad W.H., Shield J.R. (2003) Thermal and flammability properties of silica-
poly(methylmethacrylate) nanocomposite. J Appl Polym Sci 89: 2072-2078.
Khan M.A., Perruchot C., Armes S.P., Randall D.P. (2001) Synthesis of gold-
decorated latexes via conducting polymer redox templates. J Mater Chem 11:
2363-2372
Kitaev V., Ozin G.A. (2003) Self-assembled surface patterns of binary colloidal
crystals. Adv Mater 15: 75-+.
References 253

Kotov N.A., Liu Y.F., Wang S.P., Cumming C., Eghtedari M., Vargas G.,
Motamedi M., Nichlos J., Cortiella J. (2004) Inverted colloidal crystals as
three-dimensional cell scaffolds. Langmuir 20: 7887-7892.
Kyrylyuk A.V., van der Schoot P. (2008) Continuum percolation of carbon
nanotubes in polymeric and colloidal media. Proc. Nat. Acad. Sci - USA. 105:
8221-8226.
Kumacheva E., Kalinina O., Lilge L. (1999) Three-dimensional arrays in polymer
nanocomposites. Adv Mater 11: 231-+.
Kusy R.P. (1977) Influence of particle size ratio on continuity of aggregates. J
Appl Phys 48: 5301-5305.
Landfester K. (2009) Miniemulsion polymerisation and the structure of polymer
and hybrid nanoparticles. Angewandte Chemie International Edition 48:4488-
4507
Lattuada M., Hatton T.A. (2007) Functionalisation of monodisperse magnetic
nanoparticles. Langmuir 23: 2158-2168
Lee K., Asher S.A. (2000) Photonic crystal chemical sensors: pH. and ionic
strength. J Am Chem Soc 122: 9534-9537.
Lee W., Chan A., Bevan M.A., Lewis J.A., Braun P.V. (2004) Nanoparticle-
mediated epitaxial assembly of colloidal crystals on patterned substrates.
Langmuir 20: 5262-5270
Lee J., Shanbhag S., Kotov N.A. (2006) Inverted colloidal crystals as three-
dimensional microenvironments for cellular co-cultures. J Mater Chem 16:
3558-3564
Li, D.J., Sheng X., Zhao B. (2005) Environmentally responsive hairy nanoparti-
cles: Mixed homopolymer brushes on silica nanoparticles synthesised by living
radical polymerization techniques. J Am Chem Soc 127: 6248-6256.
Lin Y., Zhou B., Fernando K.A.S., Liu P., Allard L.F., Sun Y.-P. (2003) Poly-
merica carbon nanocomposites from carbon nanotubes functionalised with
matrix polymer. Macromolecules 36: 7199-7204.
Lin Y., Meziani M.J., Sun P. (2007) Functionalised carbon nanotubes for poly-
meric nanocomposites. J Mater Chem 17: 1143-1148.
Liu Y.F., Wang S.P., Lee J.W., Kotov N.A. (2005) A floating self-assembly route
to colloidal crystal templates for 3D. cell scaffolds. Chem Mater 17: 4918-
4924.
Liu Y.Y., Chen X.Q., Wang R.H., Xin J.H. (2006) Polymer microspheres stabi-
lised by titania nanoparticles. Mater Lett 60: 3731-3734.
Lu C.H., Qi L.M., Cong H.L., Wang X.Y., Yang J.H., Yang L.L., Zhang D.Y., Ma
J.M., Cao W.X. (2005) Synthesis of calcite single crystals with porous surface
by templating of polymer latex particles. Chem Mater 17: 5218-5224.
Mackay M.E., Dao T.T., Tuteja A., Ho D.L., Van Horn B., Kim H.-C., Hawker
C.J. (2003) Nanoscale effects leading to non-Einstein-like decrease in viscos-
ity. Nature Materials 2: 762-766.
Malaquin L., Kraus T., Schmid H., Delamarche E., Wolf H. (2007) Controlled
particle placement through convective and capillary assembly. Langmuir 23:
11513-11521.
254 7 Nanocomposite Latex Films and Control of Their Properties

Masuda Y., Itoh T., Itoh M., Koumoto K. (2004) Self-assembly patterning of
colloidal crystals constructed from opal structure or NaCl structure. Langmuir
20: 5588-5592.
Mayer A.B.R., Grebner W., Wannemacher R. (2000) Preparation of silver-latex
composites. J Phys Chem B 104: 7278-7285.
McCullen S.D., Stevens D.R., Roberts W.A., Ojha S.S., Clarke L.I., Gorga R.E.
(2007) Morphological, electrical and mechanical characterisation of elec-
trospun nanofibre mats containing multiwalled carbon nanotubes. Macromole-
cules 40: 997-1003.
Mezzenga R., Ruokolainen J., Fredrickson G.H., Kramer E.J., Moses D., Heeger
A.J., Ikkala O. (2003) Templating organic semiconductors via self-assembly of
polymer colloids. Science 299: 1872-1874.
Miaudet P., Badaire S., Maugey M., Derr A., Pichot V., Launois P. (2005) Hot-
drawing of single and multiwall carbon nanotube fibres for high toughness and
alignment. Nano Lett 5: 2212-2215.
Mihi A., Ocana M., Miguez H. (2006) Oriented colloidal-crystal thin films by
spin-coating microspheres dispersed in volatile media. Adv Mater 18: 2244-+.
Moniruzzaman M., Winey K.I. (2006) Polymer nanocomposites containing carbon
nanotubes. Macromolecules 39: 5194-5205.
Negrete-Herrara N., Putaux J.-L., David L., De Haas F., Bourgeat-Lami E. (2007)
Polymer/laponite composite latexes: Particle morphology, film microstructure,
and properties. Macromol Rapid Commun 28: 1567-1573.
Noh M.H., Lee D.C. (1999) Comparison of characteristics of SAN-MMT nano-
composites prepared by emulsion and solution polymerisation. J Appl Polym
Sci 74: 2811-2810
Norton L.J., Smigolova V., Pralle M.U., Hubenko A., Dai K.H., Kramer E.J.
(1995) Effect of end-anchored chains on the adhesion at a thermoset-
thermoplastic interface. Macromolecules 28: 1999-2008.
Ohno K., Koh K., Tsujii Y., Fukuda T. (2002) Synthesis of gold nanoparticles
coated with well-defined, high-density polymer brushes by surface-initiated
living radical polymerisation. Macromolecules 35: 8989-8993.
Pan F., Zhang J.Y., Cai C., Wang T.M. (2006) Rapid fabrication of large-area
colloidal crystal monolayers by a vertical surface method. Langmuir 22: 7101-
7104
Paunov V.N., Panhuis M.I.H. (2005) Fabrication of carbon nanotube-based
microcapsules by a colloid templating technique. Nanotechnology 16: 1522-
1525.
Pauchard L., Abou B., Sekimoto K. (2009) Influence of mechanical properties of
nanoparticles on macrocrack formation. Langmuir 25: 66726677
Percy M.J., Barthet C., Lobb J.C., Khan M.A., Lascelles S.F., Vamvakaki M.,
Armes S.P. (2000) Synthesis and characterisation of vinyl polymer-silica nano-
composites. Langmuir 16: 6913-6920.
Pich A., Hain J., Prots Y., Adler H.J. (2005) Composite polymeric particles with
ZnS. shells. Polymer 46: 7931-7944.
Pickering S.U. (1907) Emulsions. J Chem Soc 91: 2001-2021.
References 255

Prevo B.G., Velev O.D. (2004) Controlled, rapid deposition of structured coatings
from micro- and nanoparticles suspensions. Langmuir 20: 2099-2107
Pyun J., Matyjaszewski K. (2001) Synthesis of nanocomposite organic/inorganic
hybrid materials using controlled/living radical polymerisation. Chem Mater
13: 3436-3438.
Radtchenko I.L., Sukhorukov G.B., Gaponik N., Kornowski A., Rogach A.L.,
Mohwald H. (2001) Core-shell structures formed by the solvent controlled
precipitation of luminescent CdTe nanocrystals on latex spheres. Adv Mater
13: 1684-1687.
Ramsden W. (1903) Separation of solids in the surface-layers of solutions and
suspensions (Observations on the surface-membranes, bubbles, emulsions,
and mechanical coagulation). Preliminary account. Proc R Soc London 72:
156-164
Reculusa S., Poncet-Legrand C., Ravaine S., Mingotaud C., Duguet E., Bourgeat-
Lami E. (2002) Syntheses of raspberry-like silica/polystyrene materials. Chem
Mater 14: 2354-2359.
Regev O., El Kati P.N.B., Loos J., Koning C.E. (2004) Preparation of conductive
nanotube-polymer composites using latex technology. Adv Mater 16: 248-+.
Ren Z.Y., Li X., Zhang J.H., Li W., Zhang X.M., Yang B. (2007) Tunable two-
dimensional non-close-packed microwell arrays using colloidal crystals as
templates. Langmuir 23: 8272-8276.
Rodrguez R., de las Heras Alarcn C., Ekanayake P., McDonald P.J., Keddie J.L.,
Barandiaran M.J., Asua J.M. (2009) Correlation of Silicone Incorporation into
Hybrid Acrylic Coatings with the Resulting Hydrophobic and Thermal Proper-
ties. Macromolecules 41, 8537-8546
Rogach A., Susha A., Caruso F., Sukhorukov G., Kornowski A., Kershaw S.,
Mhwald H., Eychmller A., Weller H. (2000) Nano- and microengineering:
Three-dimensional colloidal photonic crystals prepared from submicometre-
size polystyrene latex spheres pre-coated with luminescent polyelectro-
lyte/nanocrystal shells. Adv Mater 12: 333-+.
Rogers J.A. (2001) Electronics Toward paper-like displays. Science 291: 1502-
1503.
Rossi G.B., Beaucage G., Dang T.D., Vaia R.A. (2002) Bottom-up synthesis of
polymer nanocomposites and molecular composites: Ionic exchange with
PMMA latex. Nano Lett 2: 319-323.
Ruhl T., Spahn P., Hellmann G.P. (2003) Artificial opals prepared by melt
compression. Polymer 44: 7625-34
Schaak R.E., Cable R.E., Leonard B.M., Norris B.C. (2004) Colloidal crystal
microarrays and two-dimensional superstructures: A versatile approach for
patterned surface assembly. Langmuir 20: 7293-7297.
Schellenberg C., Tauer K., Antonietti (2000) Nanostructured polymer films based
on core-shell latexes: Preparation and characterisation. Macromol Symp 151:
465-71.
256 7 Nanocomposite Latex Films and Control of Their Properties

Schmid A., Fujii S., Armes S.P. (2006) Polystyrene-silica nanocomposite particles
via alcohol dispersion polymerisation using a cationic azo initiator. Langmuir
22: 4923-4927.
Schmid A., Fujii S., Armes S.P., Leite C.A.P., Galembeck F., Minami H., Saito
N., Okubo M. (2007) Polystyrene-silica colloidal nanocomposite particles
prepared by alcoholic dispersion polymerisation. Chem Mater 19: 2435-2445.
Schuler B., Baumstark R., Kirsch S., Pfau A., Sandor M., Zosel A. (2000)
Structure and properties of multiphase particles and their impact on the per-
formance of architectural coatings. Prog Org Coatings 40: 139-150.
Shah D., Maiti P., Gunn E., Schmidt D.F., Jiang D.D., Batt C.A., Giannelis E.P.
(2004) Dramatic enhancements in toughness of polyvinylidene fluoride nano-
composites via nanoclay-directed crystal structure and morphology. Adv Mater
16: 1173-1177.
Shah D., Fytas G., Vlassopoulos D., Di J., Sogah D., Giannelis E.P. (2005)
Structure and dynamics of polymer-grafted clay suspensions. Langmuir 21:
19-25.
Sherman R.L., Ford W.T. (2005) Semiconductor nanoparticle/polystyrene latex
composite materials. Langmuir 21: 5218-5222.
Shi J.H., Yang B.X., Pramoda K.P., Goh S.H. (2007) Enhancement of the me-
chanical performance of poly(vinyl chloride) using poly(n-butyl methacrylate)-
grafted multi-walled carbon nanotubes. Nanotechnology 18: 375704.
Shimmin R.G., DiMauro A.J., Braun P.V. (2006) Slow vertical deposition of
colloidal crystals: A Langmuir-Blodgett process? Langmuir 22: 6507-6513.
Shvartzman-Cohen R., Levi-Kalisman Y., Nativ-Roth E., Yerushalmi-Rozen R.
(2004) Generic approach for dispersing single-walled carbon nanotubes: The
strength of weak interaction. Langmuir 20: 6085-6088
Shvartzman-Cohen R., Nativ-Roth E., Baskaran E., Levi-Kalisman Y., Szleifer I.,
Yerushalmi-Rozen R. (2004b) Selective dispersion of single-walled carbon
nanotubes in the presence of polymers: the role of molecular and colloidal
length scales. J Am Chem Soc 126:14850-14857.
Stein A., Li F., Denny N.R. (2008) Morphological control in colloidal crystal
templating of inverse opals, hierarchical structures, and shaped particles. Chem
Mater 20: 649-666.
Stubbs J.M., Sundberg D.C. (2007) The dynamics of morphology development in
multiphase latex particles. Prog Org. Coatings 61: 156-165.
Stubbs J.M., Sundberg D.C. (2008) Core-shell and other multiphase latex particles
confirming their morphologies and relating these to synthesis variables. JCT
Research 5: 169-180.
Sun C.H., Linn N.C., Jiang P. (2007) Templated fabrication of periodic metallic
nanopyramid arrays. Chem Mater 19: 4551-4556
Suresh K.I., Pakula T., Bartsch E. (2007) Synthesis, morphology and rheological
behaviour of fluoropolymer-polyacrylate nanocomposites. Macromol React
Eng 1: 253-263
Susha A.S., Caruso F., Rogach A.L., Sukhorukov G.B., Kornowski A., Mhwald
H., Giersig M., Eychmller A., Weller H. (2000) Formation of luminescent
References 257

spherical core-shell particles by the consecutive adsorption of polyelectrolyte


and CdTe(S) nanocrystals on latex colloids. Colloid Surface A 163: 39-44.
Tian C.G., Mao B.D., Wang E.B., Kang Z.H., Song Y.L., Wang C.L., Li S.H.
(2007) Simple strategy for preparation of core colloids modified with metal
nanoparticles. J Phys Chem C 111: 3651-3657.
Tiarks F., Landfester K., Antonietti M. (2001) Silica nanoparticles as surfactants
and fillers for latexes made by miniemulsion polymerisation. Langmuir 17:
5775-5780.
Tiarks F., Leuninger J., Wagner O., Jahns E., Wiese H. (2007) Nanocomposite
dispersions for water-based coatings. Surf Coatings Intern 5: 221-229.
Tuteja A., Mackay M.E., Hawker C.J., Van Horn B. (2005) Effect of ideal,
organic nanoparticles on the flow properties of linear polymers: Non-Einstein-
like behaviour. Macromolecules 38: 8000-8011
Tuteja A., Duxbury P.M., Mackay M.E. (2007) Multifunctional nanocomposites
with reduced viscosity. Macromolecules 40: 9427-9434
Tuteja A., Mackay M.E., Narayanan S., Asokan S., Wong M.S. (2007b) Break-
down of the continuum Stokes-Einstein relation for nanoparticle diffusion.
Nano Letters 7: 1276-1281
Vandervorst, P., Lei C., Lin Y., Dupont O., Dalton A.B., Sun Y.-P., and Keddie
J.L. (2006) Prog Organic Coat 57: 91-97
Vaia R.A., Maguire J.F. (2007) Polymer nanocomposites with prescribed mor-
phology: Going beyond nanoparticle-filled polymers. Chem Mater 19: 2736-
2751.
Vaisman L., Wagner H.D., and Marom G. (2006) The role of surfactants in
dispersion of carbon nanotubes. Adv Colloid Interface Sci 128-130: 37-46.
Venkatesh S., Jiang P., Jiang B. (2007) Generalised fabrication of two-
dimensional non-close-packed colloidal crystals. Langmuir 23: 8231-8235.
Von Werne T., Patten T.E. (1999) Preparation of structurally well-defined
polymer-nanoparticle hybrids with controlled/living radical polymerisations. J
Am Chem Soc 121: 7409-7410.
Von Werne T., Patten T.E. (2001) Atom transfer radical polymerisation from
nanoparticles: A tool for the preparation of well-defined hybrid nanostructures
and for understanding the chemistry of controlled/living radical polymerisa-
tions from surfaces. J Am Chem Soc 123: 7497-7505.
Voorn D.J., Ming W., van Herk A.M. (2006) Polymer-clay nanocomposite latex
particles by inverse Pickering emulsion polymerisation stabilised with hydro-
phobic montmorillonite platelets. Macromolecules 39: 2137-2143.
Wang D.Y., Mhwald H. (2004) Rapid fabrication of binary colloidal crystals by
stepwise spin-coating. Adv Mater 16: 244-+.
Wang C., Chu F., Graillat C., Guyot A., Gauthier C., Chapel J.P. (2005) Hybrid
polymer latexes: acrylics-polyurethane from miniemulsion polymerisation:
properties of hybrid latexes versus blends. Polymer 46: 1113-1124.
Wang T., Lei C.H., Dalton A.B., Creton C., Lin Y., Shiral Fernando K.A., Sun Y.-
P., Manea M., Asua J.M., Keddie J.L. (2006) Waterborne, nanocomposite
258 7 Nanocomposite Latex Films and Control of Their Properties

pressure-sensitive adhesives with high tack energy, optical transparency and


electrical conductivity. Adv Mater 18: 2730-2734.
Wang J., Wen Y., Ge H., Sun Z., Zheng Y., Song Y., Jiang L. (2006b) Simple
fabrication of full colour colloidal crystal films with tough mechanical strength.
Macromol Chem Phys 207: 596-604.
Wang L.K., Zhao X.S. (2007) Fabrication of crack-free colloidal crystals using a
modified vertical deposition method. J Phys Chem C 111: 8538-8542.
Wang T., Lei C.-H., Liu D., Manea M., Asua J.M., Creton C., Dalton A.B.,
Keddie J.L. (2008) A molecular mechanism for toughening and strengthening
waterborne nanocomposites. Adv Mater 20: 90-94.
Wang J., Ahl S., Li Q., Kreiter M., Neumann T., Burkert K., Knoll W., Jonas U.
(2008b) Structural and optical characterisation of 3D. binary colloidal crystal
and inverse opal films prepared by direct co-deposition. J Mater Chem 18: 981-
988.
Wang T., Dalton A.B., Keddie J.L. (2008c) Importance of molecular friction in a
soft polymer-nanotube nanocomposite. Macromolecules 41: 7656-7661.
Wang T., Keddie J.L. (2009) Design and fabrication of colloidal polymer nano-
composites. Adv Coll Interf Sci. 147-148: 319-332.
Watanabe J. and Lepoutre P. (1982) A mechanism for the consolidation of the
structure of clay-latex coatings. J Appl Polym Sci 27: 4207-4219
Weitenkopf D. (2009) www.col9.de (accessed July 1, 2009)
Wohlleben W., Bartels F.W., Altmann S., Leyrer R.J. (2007) Mechano-optical
octave-tuneable elastic colloidal crystals made from core-shell polymer beads
with self-assembly techniques. Langmuir 23: 2961-2969.
Xia Y.N., Yin Y.D., Lu Y., McLellan J. (2003) Template-assisted self-assembly of
spherical colloids into complex and controllable structures. Adv Funct Mater
13: 907-918.
Xie L., Qiu F., Lu H.B., Yang Y.L. (2007) Single-walled carbon nanotubes
functionalized with high bonding density of polymer layers and enhanced me-
chanical properties of composites. Macromolecules 40: 3296-3305.
Xu L., Reeder S., Thopasridharan M., Ren J., Shipp D.A., Krishnamoorti R.
(2005) Structure and melt rheology of polystyrene-based layered silicate nano-
composites. Nanotechnology 16: S514-S521
Xue Z. and Wiese H. (2006) U.S. patent U.S7094830B2 2006.
Yang Z.Z., Qiu D., Li J. (2002) Waterborne dispersions of a polymer-encapsulated
inorganic particle nanocomposite by phase-inversion emulsification. Macromol
Rapid Commun 23: 479-483.
Yang J., Hasell T., Wang W.X., Li J., Brown P.D., Poliakoff M., Lester E.,
Howdle S.M. (2008) Preparation of hybrid polymer nanocomposite microparti-
cles by a nanoparticle stabilised dispersion. J Mater Chem 18: 998-1001.
Yerushalmi-Rozen R., Szleifer I. (2006) Utilising polymers for shaping the
interface behaviour of carbon nanotubes. Soft Matt. 2: 24-28.
Zhang J., Chen K.Q., Zhao H.Y. (2008) PMMA colloid particles armoured by clay
layers with PDMAEMA brushes. J Polym Sci Part A Polym Chem 46: 2632-
2639.
References 259

Zhang K.Q., Liu X.Y. (2004) In situ observation of colloidal crystal monolayer
nucleation driven by an alternating electric field. Nature 429: 739-743.
Zhang Q., Archer L.A. (2002) Poly(ethylene oxide)/silica nanocomposites :
Structure and rheology. Langmuir 18: 10435-10442
Zhang Y.J., Wang S.P., Eghtedari M., Motamedi M., Kotov N.A. (2005) Inverted-
colloidal-crystal hydrogel matrices as three-dimensional cell scaffolds. Adv
Funct Mater 15: 725-731.
Zhou T.H., Ruan W.H., Rong M.Z., Zhang M.Q., Mai Y.L. (2007) Keys to
toughening of non-layered nanoparticles/polymer composites. Adv Mater 19:
2667-+.
Zhu J., Morgan A.B., Lamelas F.J., Wilkie C.A. (2001) Fire properties of polysty-
rene-clay nanocomposites. Chem Mater 13: 3774-3780.
Chapter 8

8 Future Directions and Challenges

This book has aimed to provide a comprehensive introduction to the main topics
relevant to film formation. It is interesting to note some of the topics that were
considered as future challenges and new directions in 1997 in a review of latex
film formation (Keddie 1997):
Film formation from latex blends;
Film formation from core-shell and composite particles;
Using reactive surfactants.
Progress on these topics during the intervening years is easily charted. Devel-
opments in the creation of nanocomposite films through latex blends and two-
phase particles were summarised in Chapter 7. Achievements in the use of
reactive surfactants were highlighted in Section 6 of Chapter 6.
Despite the vast amount of research over the past decades, unanswered ques-
tions about the film formation process remain. Interest in the topic therefore
remains strong. In numerous developments, polymer colloids are being used in
new applications. Here, we highlight a few exciting future directions in the field of
latex film formation.

8.1 Film Formation from Anisotropic Particles

So far, we have considered film formation only from spherical particles. Histori-
cally, techniques of emulsion and dispersion polymerisation have created particles
that are perfect spheres. The surfaces of the particles have a uniform composition
with homogeneous chemical functionality and reactivity. Latex particles are
usually isotropic. In recent years, techniques of synthesis have been developed to
create colloidal particles with controlled anisotropy. Particles that have a degree of
anisotropy lend themselves to self-assembly into complex structures (Glotzer and
Solomon 2007). One type of anisotropy can be broadly said to be geometric, such
262 8 Future Directions and Challenges

as having facets, a high aspect ratio, or branching. There have been reports, for
instance, of ellipsoidal latex particles with an aspect ratio (defined as the ratio of
the long axis to the short axis) as high as 12.8 (Mohraz and Solomon 2005). Some
ellipsoidal particles can be seen in Fig. 8.1. When the particles are sedimented,
some alignment of the long axes parallel to each other can be found.

Fig. 8.1 Ellipsoidal latex particles made from poly(methyl methacrylate) having an aspect ratio of
5.2. (Reprinted from Mohraz and Solomon (2005), copyright 2005 American Chemical Society)

Another type of anisotropy can be broadly classed as chemical in nature; ex-


amples include patchiness in surface coverage or chemical patterning. Spherical
particles in which opposite faces have differing polarity or charge are called
Janus particles (Roh et al. 2005). The name Janus is derived from the Roman
god who had two faces. In the same way, a Janus particle can have two faces, such
as one hydrophilic and one hydrophobic. This type of particle will spontaneously
assemble at an oil-water interface (Casagrande et al. 1989). Such activity can be
harnessed to create supra-particle assemblies. Janus particles can also have
opposite electric charge on their two faces. When the particle diameter is greater
than the charge screening length, particle clusters (containing up to 13 particles)
are formed in aqueous suspensions (Hong et al. 2006).
These new anisotropic particles offer fresh challenges in the field of film for-
mation. Numerous forces, as presented in previous chapters, act upon the particles
during the film formation process. Beyond these, attractions and repulsions
between opposite sides of the particles will influence the orientation of particles
during the packing stage. New types of packing, beyond FCC or BCC, will be
8.2 Assembly of Particles over Large Length Scales 263

found with oblong or asymmetric particles. Some earlier work has used particles
with two hemispherical phases to create heterogeneous films (Hagen et al. 1996).
The structures shown in Fig. 8.2 indicate that the orientation and ordering of the
particles was not controlled. In other experiments, individual latex particles that
are partly hydrophilic and partly hydrophobic have been shown to become
oriented at interfaces (Pfau et al. 2002). The more hydrophilic face was found to
absorb on surfaces that were quite hydrophilic. On less hydrophilic surfaces, the
hydrophobic face absorbed. In thicker films, the particles adsorbed with their
hydrophobic face along the interface, regardless of the substrates hydrophilicity.
There is plenty of scope for modelling and further experimentation, especially
with the aim of controlling particle orientation in thick films.

(a) (b)

Fig. 8.2 a Phase separation between polystyrene and poly(butadiene) leads to hemispherical
phases within particles, as shown in this TEM image in which one phase is stained. b This
particle structure is preserved in the films. Hemispherical structures can be seen in the TEM
image. A better control over the particle alignment and orientation will lead to films with unique
structure over long length scales. (Reprinted from Hagen et al. 1996; copyright Wiley-VCH
Verlag GmbH & Co. KGaA; reproduced with permission)

8.2 Assembly of Particles over Large Length Scales

With the growth of applications of polymer colloids moving well beyond tradi-
tional coatings and adhesives, and going toward new applications, such as tissue
scaffolds and photonic crystals, polymer colloid specialists have much to offer.
One of the biggest challenges is to achieve the ordering of particles on a lattice
264 8 Future Directions and Challenges

over large areas. Quite often, there are packing defects (missing particles or extra
particles) that disrupt the order. With a tight control over the particle size distribu-
tion, and with the right formulation and drying conditions, however, colloidal
crystals covering areas as large as 2.6 cm x 2.6 cm have already been achieved
(Wohlleben et al. 2007).
Good progress has been made in the creation of photonic crystals using latex
technology. Photonic crystals are used in all-optical integrated circuits that rely on
the transport of photons rather than electrons, as in a conventional circuit. In a
photonic crystal, a range of light frequencies are forbidden to exist within the
interior of the crystal, thus creating an optical band gap. Photonic crystals have a
periodic modulation of refractive index (or dielectric constant at the frequencies of
light) to manipulate the light (Joannopoulos et al. 1997). The requirements for
photonic crystals are rather strict. The size of the unit cell for a photonic crystal
must be comparable to the wavelength of light. For near-infrared light at 1.5 m,
the unit cell must be about 0.5 m, which puts the size scale in the colloidal
domain. To increase the band gap, the magnitude of the refractive index modula-
tions must be increased by using materials with greater index differences.
An inverse opal structure, which is found in some types of photonic crystals,
can be created from close-packed, hard colloidal spheres. The empty space around
the particles is filled in with a high-refractive index material; later the particles are
removed (dissolved or etched away), to leave spherical voids. Polymer particles
are effective templates for the creation of inverse opals. The challenges in creating
newer photonic crystal structures are even greater. In a double-inverse-opal
structure, a weakly scattering spherical particle is introduced inside each void
within a periodic, inverse opal structure (see Fig. 8.3). This design enables optical
switching from a reflector state to a photonic crystal state (Ruhl et al. 2006).

Silica particles

Titania honeycomb
structure

Fig. 8.3 SEM image of a double inverse opal structure created using latex particles packed into
an array. Titania was deposited in the void space between latex particles with a silica core (radius
of 60 nm). When the polymer was removed, the silica particles are caged inside of the titania
honey-comb structure. (Reprinted from Ruhl et al. 2006; copyright Wiley-VCH Verlag GmbH &
Co. KGaA; reproduced with permission)
8.3 Technique Development 265

Periodically structured composites that have unit cells smaller than a particular
wavelength of light (e.g., visible or infrared) are known as metamaterials. It has
been proposed that in future, anisotropic colloidal particles (including latex
particles) could be self-assembled to create metamaterials (Stebe et al. 2009). One
of the most exciting applications of metamaterials is to create cloaking devices to
render objects invisible at certain wavelengths of light.

8.3 Technique Development

The study of latex film formation has benefited enormously from new experimen-
tal techniques. Latex films are often used as a model system during technique
development. Latex systems were studied early on in the emergence of small
angle neutron scattering, fluorescence resonance energy transfer (FRET), and
atomic force microscopy, among others. This trend is likely to continue, consider-
ing the wide applicability of polymer colloids.
Drying processes will continue to be of interest. A particular technique devel-
oped recently, inverse-micro-Raman-spectroscopy, offers many attractive features.
It can determine water content non-invasively in drying latex films in the vertical
as well as the horizontal directions. A spatial resolution of 2 3 m and a time
resolution of about 1 s are better than other techniques and make the technique
particularly powerful (Ludwig et al. 2007). Along a similar vein, recent innova-
tions in the experiment apparatus for FRET allow measurements of interdiffusion
as a function of horizontal position in wet latex films (Haley et al. 2007).
The use of two or more complementary techniques simultaneously provides
rich information and is welcomed. In a recent example, diffusing wave spectros-
copy (DWS) has been employed to measure particle diffusivity during the drying
stage while magnetic resonance profiling determined the distribution of water
(Knig et al. 2008). DWS has also been successfully combined with a membrane-
bending apparatus to achieve simultaneous measurements of particle dynamics
and stress development during drying (Arnold et al. 2009). FRET measurements
have been coupled with optical imaging to correlate the position of the drying and
particle packing fronts with the extent of interdiffusion. The amount of molecular
mixing between particles can then be known in relation to how much water is
present locally and how long it has been since the drying and packing fronts have
passed (Haley et al. 2008). There is further scope to use FRET in combination
with other non-invasive techniques.

8.4 Nanocomposite Structure and Property Correlations

Chapter 7 explained how recent research has drawn upon basic latex technology in
the development of a class of nanocomposites. The use of hybrid particles provides
control at the nanoscale, whereas the arrangement of particles into colloidal crystal
266 8 Future Directions and Challenges

arrays offers controlled periodicity and symmetry at the mesoscale. The properties
of colloidal nanocomposites are modulated over these length scales.
Progress has been made in understanding the properties of latex films contain-
ing carbon nanotubes. For instance, electrically conducting networks of carbon
nanotubes in a soft latex film have been shown to retain electrical conductivity up
to strains of 150% (Dalmas et al. 2005). Other results have proven the key role
that internal interfaces between nanotubes and the film matrix can play in deter-
mining properties. The sliding friction between carbon nanotubes and a soft latex
film can be adjusted with small dispersant molecules. It turn, it has a large impact
on the mechanical properties (Wang et al. 2008). Interface design therefore is
expected to be a particularly worthwhile pursuit.
To date, most of the emphasis has been on making hybrid materials and ordered
structures, but there has been far less emphasis on the relationship between the nano-
and mesostructure and properties. For colloidal nanocomposites to be developed for
demanding applications, greater effort must be made. Nevertheless, greater control
of the properties of nanocomposites based on latex films or polymer colloid particles
is resulting in new applications. Recent examples include thermoelectric devices
(Yu et al. 2008, 2009), scaffolds for cell cultures (Firkowska et al. 2006), biocata-
lytic coatings with controlled pore structure (Lyngberg et al. 2005), and pressure-
sensitive adhesives (Deplace et al. 2009, Wang et al. 2009).

6000 100
4
3
log (S/m)

5000 2
1 80
Electrical conductivity, (S/m)

0
Thermopower, S (V/K)

4000 -1
-2 60
0.1 1 10
3000 CNT wt%

40
2000

20
1000

S
0 0
0 5 10 15 20
CNT wt%

Fig. 8.4 Electrical conductivity (circles) and thermopower (squares) for a nanocomposite of a
poly(vinyl acetate) emulsion and carbon nanotubes. The thermoelectric figure of merit for a
nanocomposite with 20 wt.% CNTs is up to six times greater than found previously for
polymers. (Reproduced with permission from Yu et al. 2008; copyright 2008 American
Chemical Society)
8.5 Interdiffusion of Polymers in Multiphase Particles 267

In thermoelectric applications, in which temperature differentials are used to


generate electrical currents, nanocomposites of latex and carbon nanotubes offer a
desirable combination of low thermal conductivity (0.34 W/mK) coupled with
high electrical conductivity (4800 S/m) (Yu et al. 2008). Fig. 8.4 reports some of
the promising results.
In adhesive applications, core-shell latex particles have been used to create a
three-dimensional crosslinked network. The shells of the particles were
crosslinked together, and the particle cores remained viscous, resulting in the
desired nonlinear elasticity (Deplace et al. 2009). In other work, the addition of
clay-encapsulated hybrid particles at low concentrations in a soft latex film has
increased its adhesion energy through an effect on the balance of viscoelastic
properties. The adhesion energy is 70% greater in the nanocomposite compared to
a reference latex film. Nanostructured particles are required, as the effect is not as
pronounced if the polymer particles and laponite are added separately, as demon-
strated in Fig. 8.5 (Wang et al. 2009).

Fig. 8.5 A comparison of the probe-tack stress-strain curves for an acrylic adhesive containing
supracolloidal particles with a laponite clay shell (right) and a blend of the corresponding
polymer particles and laponite. (Reprinted from Wang et al. 2009, http:// dx.doi.org/
10.1039/B904740A; reproduced by permission of The Royal Society of Chemistry)

8.5 Interdiffusion of Polymers in Multiphase Particles

The use of hybrid particles has the advantage of avoiding the depletion-induced
phase separation that is inherent in the use of large and small (or hard and soft)
particle blends. In the case of hybrid particles consisting of hard inorganic and soft
polymer phases, the topic of interparticle diffusion has been neglected both
experimentally and theoretically. Conceptually, the strength development should
268 8 Future Directions and Challenges

be invariant to the presence of a hard component in the centre of each particle. A


guess of reptation over the distance of the polymers radius of gyration to achieve
full mechanical strength seems reasonable. This result however remains pure
speculation and requires verification. In addition, the effect of the hard component
will become more dominant as its volume fraction is increased. At higher load-
ings, it might significantly hinder the interdiffusion. This effect implies that there
has to be an optimal hard component loading, as a trade-off between achieving
high film hardness while also being able to achieve film formation.

8.6 Templating Film Topography

The film height profile is a complex function of the drying process. As outlined in
Chapter 3, the flows in drying films can be induced by horizontal drying fronts
pulling fluid towards the film edge, and surfactant-induced flows driving fluid
towards regions of lower surfactant concentration, typically the centre. The result
of the two flows in competition with each other allows a range of film profiles to
be created. A ridge around the film, a few mm from the edge, is evidence of the

Evaporation occurs uniformly from film surface

Flow instability caused by


surfactant at interface
Surfactant desorbed from
compacting particles

Flow from region


of high surfactant
concentration
Fluid region
Solid
region

Flow from fluid to replace


evaporated water in solid
region

Fig. 8.6 The three different flows associated with a drying film. Flow from the bulk to the edge
replaces solvent lost due to evaporation. Desorption of surfactant from coalescing particles can
cause a flow from the edge to the bulk and free surfactant in the bulk can cause flow instabilities
and film patterning.
8.7 Resolving the Film Formation Dilemma 269

balance of the two flows. The surfactant-induced flow can also result in instabili-
ties such as reported by Gundabala et al. (2008). The different flows are sketched
in Fig. 8.6 and, depending on the relative magnitude of each, different profiles will
be obtained. The use of these three types of flow to template desired film topogra-
phy will require careful control over the initial film profile and the drying condi-
tions. Yet, the possibility remains of creating surface structures with controlled
undulations in the micrometer size range.
An elegant manipulation of film topography has been achieved recently by a
technique called evaporative lithography (Harris et al. 2007). This technique
uses a mask essentially a plate with a pattern of holes to cause evaporation to
occur faster only in certain regions across a film surface. Flows are created in the
underlying film because of non-uniformities in particle concentration. Particles are
carried with the flow and hence accumulate in the fast-evaporating regions, which
are consequently higher when drying is complete. There are a number of parame-
ters that allow control over the extent of particle transport. These include the
initial particle volume fraction, the distance between the mask and the film, and
the dimensions of the masks patterns. Although evaporative lithography has been
demonstrated in a latex film, there is the scope for extending the application of the
technique. Coupling the use of masks with the creation of surfactant-induced
topography, as seen by Gundabala et al. (2008), will allow control over film
profiles over two discrete length scales.

8.7 Resolving the Film Formation Dilemma

Through this book, there has been discussion on the dilemma faced when trying to
create a hard coating via latex film formation. One option is to employ a hard
(high Tg) polymer and to achieve film formation at room temperature through the
use of plasticising additives or coalescent. But, this method is not favourable
because of its VOC emission. A second option is to use a hard polymer and then
carry out the film formation at temperatures well above the polymers Tg. This
method poses the risk of film cracking as capillary pressure develops in the weak
particle network (Tirumkudulu and Russel 2005, Lee and Routh 2004). At
elevated temperatures, particles will become sufficiently deformable to undergo
film formation by wet sintering or capillary deformation. Cracking, such as is seen
in Fig. 8.7, will be avoided (Lee and Routh 2006). However, the film formation of
industrial coatings through the use of ovens is an energy-intensive process that
adds to the carbon footprint of coatings. The future of the coatings industry,
along with many others, is green. There is a need to find ways to resolve the film
formation dilemma.
The environmental drive to reduce VOC levels will continue. There is a con-
tinuing need to find ways to soften latex particles to enable film formation,
without sacrificing the hardness level of the final coating. A possible method to
solve the problem is the use of polymers that are plasticised by water. The
270 8 Future Directions and Challenges

softened latex will film-form in the presence of water, but the polymer will then
harden as the water evaporates. A drawback is that such polymers will also be
hydrophilic and subject to water uptake in use as a coating. A second method is
the blending of hard and soft particles, as discussed in Chapter 7. The soft
particles create a continuous film whereas the hard particles increase the elastic
modulus and the hardness. Crosslinking of a soft polymer after film formation is
yet another way to achieve a hard film (Taylor and Winnik 2004) when film-
forming at room temperature.
A newly reported way to reduce the use of coalescent (and hence the VOC
content) in a latex blend, is to use a low-molecular weight (oligomeric) copolymer
latex to reduce the glass transition temperature of the same copolymer with a high-
molecular weight (Tomba et al. 2008). The oligomers, which have a lower glass
transition temperature, can diffuse readily into the high-molecular weight particles
and soften them. The rates of coalescence and diffusion can then be enhanced
without the introduction of a foreign molecule. The high molecular weight
component provides the necessary mechanical strength.

Fig. 8.7 Example of a film cast on glass substrates from acrylic latex with a glass transition
temperature of approximately 40C. When cast at room temperature, an opaque, cracked film
results. (Photograph courtesy of A. Georgiadis, University of Surrey)
8.7 Resolving the Film Formation Dilemma 271

A recently proposed method for reducing the use of coalescent, called De-
signed DiffusionTM and developed at the Rohm and Haas Company, relies on
blends of two types of polymer particles (Fu et al. 2009). A hard polymer (Tg
greater than room temperature) is the majority phase, and a soft (low Tg) polymer
is blended to achieve percolation. The key idea of the technology is to select a
coalescent that will partition within the hard polymer when the latex blend is in
the wet state. The softened particles will be able to undergo particle deformation
and coalescence. However, during the film formation process, the solubility of the
coalescent switches so that it partitions primarily in the soft phase. The hardness
of the high-Tg polymer then returns. The evaporation of the coalescent from the
low-Tg polymer phase is fast, so that desirable properties of film hardness, such as
blocking resistance and low dirt pick-up, develop quickly. Fig. 8.8 illustrates the
process. The rate of property development is significantly faster than when a
coalescent is added to hard latex. Furthermore, the concentration of VOCs in a
Designed DiffusionTM formulation has been reported to be 30% lower than in a
conventional formulation with similar final film properties.

Fig. 8.8 a In the wet latex, the coalescent molecules (identified with x are mainly partitioned in
the high Tg polymer particles. b The plasticized hard particles are able to be deformed during
film formation. c When all water has left the film, the coalescent molecules are transferred into
the low Tg polymer phase. d The coalescent molecules can diffuse quickly through the soft
polymer phase and evaporate from the film.
272 8 Future Directions and Challenges

As described so far, the technology sounds somewhat mysterious. There are three
key requirements for the hard and soft polymer phases that yield the designed
results. Each requirement stems from a scientific concept, as will be explained.
1. The Tg of the softer polymer should preferably be less than 5C. Not only
does this ensure rapid coalescence, but this requirement ensures a high free
volume within the polymer. The diffusion of the coalescent through the poly-
mer, and hence the rate at which it leaves the film during film formation, in-
creases as the free volume increases (Vrentas et al. 1985). A more open
polymer allows faster transport of the VOCs.
2. To ensure there is a pathway for the coalescent to escape from the soft polymer
phase, there should be percolation of the soft particles. Yet, the volume fraction
of the soft polymer phase should be as low as possible so that the properties of
the hard polymer phase are dominant and not compromised. The percolation
threshold of the soft phase can be lowered by using a smaller particle size in
comparison to the hard particles.
3. The soft polymer phase should ideally be more hydrophilic than the hard
polymer phase. This requirement ensures that the coalescent partitions mainly
in the hard polymer phase when the latex is in the wet state. When the water
evaporates from the soft polymer phase, the partitioning is shifted such that the
coalescent is more soluble in this phase.
The Rohm and Haas Company has reported that several types of polymers can
be successfully used as the soft polymer phase, including acrylic and styrene-
acrylic copolymers and polyurethane dispersions (Fu et al. 2009). The use of
standard coalescent, such as n-methylpyrrolidone and TexanolTM (Eastman
Chemical Company), has been demonstrated. This approach presents an exciting
way to resolve the film formation dilemma. Future research might find other ways.

References

Arnold C. et al. (2009) Mapping the route from wet to dry. European Coatings
Journal 11: 28-32
Casagrande C., Fabre P., Raphael E., Veyssie M. (1989) Janus beads Realisation
and behaviour at water oil interfaces. Europhys Lett 9: 252-255
Dalmas F., Chazeau L., Gauthier C., Masenelli-Varlot, K., Dendievel R., Cavaill
J.Y., Forr L. (2005) Multiwalled Carbon Nanotube/Polymer Nanocomposites:
Processing and Properties. J Polym Science Part B Polym Phys 43: 1186-1197
Deplace F., Rabjohns M.A., Yamaguchi T., Foster A.B., Carelli C., Lei C.-H.,
Ouzineb K., Keddie J.L., Lovell P.A., Creton C. (2009) Deformation and adhe-
sion of a soft-soft nanocomposite designed with structured polymer colloid
particles. Soft Matter 5: 1440-1447.
References 273

Firkowska I., Olek M., Pazos-Perz N., Rojas-Chapana, J., and Giersig M. (2006)
Highly Ordered MWNT-Based Matrixes: Topography at the Nanoscale Con-
ceived for Tissue Engineering. Langmuir 22: 5427-5434
Fu Z., Hejl A., and Swartz A. (2009) Designed diffusion technology: A paradigm
shift in film formation. European Coatings Journal Issue 6: 26-33
Glotzer S.C. and Solomon, M.J. (2007) Anistropy of building blocks and their
assembly into complex structures. Nature Materials 6: 557-562.
Gundabala V.R., Lei C.-H., Ouzineb K., Dupont O., Keddie J.L., and Routh A.F.
(2008) Lateral surface non-uniformities in drying latex films. AIChE J 54(12):
3092-3105
Hagen R., Salmen L., Karlsson O., Wesslen B. (1996) Viscoelastic properties and
film morphology of heterogeneous styrene-butadiene latexes. J Appl Polym Sci
62: 1067-1078
Haley J.C., Liu Y., Winnik M.A., Demmer D., Haslett T., Lau W. (2007) Tracking
polymer diffusion in a wet latex film with fluorescence resonance energy trans-
fer. Rev Sci Instrum 78: 084101
Haley J.C., Liu Y., Winnik M.A., Lau W. (2008) The onset of polymer diffusion
in a drying acrylate latex: how water initially retards coalescence but ultimately
enhances diffusion. J Coat Technol Res 5(2): 157-168
Harris D.J., Hu H., Conrad J.C., Lewis J.A. (2007) Patterning colloidal films via
evaporative lithography. Physical Review Letters 98: 148301.
Hong L., Cacciuto A., Luijten E., Granick S. (2006) Clusters of charged Janus
spheres. Nano Letters 6: 2510-2514.s
Joannopoulos J.D., Villeneuve P.R., Fan S. (1997) Photonic crystals: putting a
new twist on light. Nature 386: 143-149
Keddie J.L. (1997) Film formation of latex. Mat Sci Eng R: Reports R21(3): 101-
169
Knig A.M., Weerakkody T.G., Keddie J.L., Johannsmann D. (2008) Heterogene-
ous drying of colloidal polymer films: Dependence on added salt. Langmuir 24:
75807589
Lee W.P., Routh A.F. (2004) Why do drying films crack? Langmuir 20: 9885-
9888
Lee W.P., Routh A.F. (2006) Temperature dependence of crack spacing in drying
latex films. Ind. & Eng. Chem. Res. 45: 6996-7001
Lyngberg O.K., Solheid C., Charaniya, S., Ma, Y., Thiagarajan V., Scriven L.E.,
Flickinger M.C. (2005) Permeability and reactivity of Thermotoga maritime in
latex bimodal blend coatings at 80 C: a model high temperature biocatalytic
ccoating. Extremophiles 9: 197-207.
Ludwig I., Schabel W., Kind M., Castaing J.-C., Ferlin P. (2007) Drying and film
formation of industrial waterborne latices. AIChE J 53: 549-560
Mohraz A. and Solomon M.J. (2005) Direct visualisation of colloidal rod assem-
bly by confocal microscopy. Langmuir 21: 5298-5306
Pfau A., Sander R., Kirsch S. (2002) Orientational ordering of structured poly-
meric nanoparticles at interfaces. Langmuir 18: 2880-2887
274 8 Future Directions and Challenges

Roh K.-H., Martin D.C., Lahann J. (2005) Biphasic Janus particles with nanoscale
anisotropy. Nature Materials 4: 759-763
Ruhl T., Spahn P., Hermann C., Jamois C., Hess O. (2006) Double-inverse-opal
photonic crystals: The route to photonic bandgap switching. Adv Funct Mat 16:
885-890
Stebe K.J., Lewandowski E., Ghosh M. (2009) Oriented assembly of metamateri-
als. Science 325(5937) 159-60.
Taylor J.W., Winnik M.A. (2004) Functional latex and thermoset latex films. JCT
Research 1(3): 163-190
Tomba J.P., Ye X., Oh J., Lau W., Winnik M.A. (2008) Polymer blend latex
films: Miscibility and polymer diffusion studied by energy transfer. Polymer
49: 2055-2064
Tirumkudulu M.S., Russel W.B. (2005) Cracking in drying latex films. Langmuir
21: 4938-4948
Vrentas J.S., Duda J.L., Ling H.C., Hou A.C. (1985) Free-volume theories for
self-diffusion in polymersolvent system. II Predictive capabilities. J Polym
Sci, Polym Phys Ed 23: 289304
Wang T., Dalton A.B., Keddie J.L. (2008) Importance of molecular friction in a
soft polymer-nanotube nanocomposite. Macromolecules 41: 7656 7661
Wang T., Colver P.J., Bon. S.A.F., and Keddie J.L. (2009) Soft Polymer and
Nano-Clay Supracolloidal Particles in Adhesives: Synergistic Effects on Me-
chanical Properties. Soft Matter, 5: 3842-3849
Wohlleben W., Bartels F.W., Altmann S., Leyrer R.J. (2007) Mechano-optical
octave- tunable elastic colloidal crystals made from core-shell polymer beads
with self-assembly techniques. Langmuir 23: 2961-2969
Yu, C., Kim Y.S., Kim D., Grunlan J.C. (2008) Thermoelectric behaviour of
segregated-network polymer nanocomposites. Nano Letters 8: 4428-4432.
Yu, C., Kim Y.S., Kim D., Grunlan J.C. (2009) Thermoelectric behaviour of
segregated-network polymer nanocomposites. Nano Letters 9: 1283-1283.
Index

acomustic waves 34 atom transfer radical


acrylates 2 polymerization 220
acrylic acid groups 174 autocorrelation 45
acrylic copolymers 3 autohesion 151
adhesion, effect of surfactants 190 barrier resistance
adhesion energy 159 effect of surfmers 206
adsorption isotherms 1923 in nanocomposites 216
AFM see atomic force microscopy beam bending 3234
aggregation, definition of 20 blocking 159, 169, 216, 245
alkyd film 74 boundary layer 96
anisotropic particles 25961 Braggs law 232
anisotropy 259 brittle fracture 293
anthracene 77, 81 brittleness 215
Arrhenius equation 166 Brown, Robert 1
aspect ratio 240, 246 Brownian dynamics simulations
atomic force microscopy 628, 144 of drying 106
cantilever 62, 68 Brownian motion 1, 44
experimental parameters 65 applications of 50
height artefacts 64 Brown mechanism 125
indentation depth 64 capillary deformation 1245, 135
intermittent contact 63 experimental evidence 142
microtomed cross-sections 67 capillary length 110, 111
particle deformation 67 capillary pressure 1113, 124, 230
phase imaging 66 effect on cracking 116
contrast in 67 capillary waves 157
set point ratio 65 carbon nanotube 221, 234, 246,
TappingModeTM 63 263
tip 69 carboxylic acid groups 173
contamination 68 carpet backings 6
302 Index

chain critical micelle concentration 191


branching 1645 critical stress intensity factor 293
entanglement 159 critical volume fraction 234
length 249 crosslinking 58, 734, 175
pull-out 158 autoxidative 74
scission 159 control parameter 179
chalking 245 molecular weight effects 178
chemical patterning 231 two-pack 175
Clausius-Mossotti equation 51 two-pack in one pot 175
clay cryogenic electron microscopy see
exfoliation 221 electron microscopy
intercalation 221 currant-bun particle 221
close packing, random 10, 23, 100 dangling chains 178
cloudy-clear transition 29, 143 Darcy flow 112
coalescent reduction 268 Darcys law 104
coalescing aid 1745 Debye length 18, 114
effect on Tg 175 deformation map 1334, 139
selection of 175 depletion interactions 17, 20
coffee rings 110 Designed DiffusionTM 269
Col.9 245 desorption of surfactant 199
colloidal crystal 23, 232 deuterium 44
classification 238 dew point 283
growth 231 dialysed latex 189
colloidal stability, effect on drying diffraction limit 49
114 diffusing wave spectroscopy 46, 263
colloid dispersion 1 diffusion 10, 151
colloid science 1723 activation energy for 166
complex longitudinal modulus 35 competition with crosslinking
confocal microscopy 4950 175
laser scanning 50 effect of chain branching 164
confocal Raman microscopy 52, 74 effect of coalescing aids 174
construction materials 6 effect of membranes 173
convection of surfactant 194 effect of molecular weight 1645
core-shell particle see particle effect of particle size 172
crack healing 152, 294 effect of reduced mobility 171
cracking 1167 effect of temperature 165
in nanocomposites 235 in gel 177
relaxation mechanism 117 near Tg 167
crack point 29 of core shell particles 172
crack spacing 117 particle shell effects 164
creaming 275 scaling prediction 165
creeping flow 22, 273 scaling relations 157
critical coagulation concentration shift factor 168
115 surfactant 195
critical energy release rate 2956 tortuosity effects 169
Index 303

diffusion coefficient 22, 153, 166 wet STEM 412


dirt pick-up 189, 245 electron paramagnetic resonance 60
DLVO theory 17, 19 electron scattering 40
double cantilever beam 294 electrostatic repulsion 17, 1819
drag coefficient 22 ellipsometry 50, 52, 143
dry bulb temperature 282 emulsion polymerisation 2
drying 10, 95117 emulsion polymers, market for 9
effect of Peclet number 104, 106 encapsulated particle 221
effect of salt 106, 115 entanglement molecular weight
effect of surfactant 114 155, 160, 178
horizontal 107114 enthalpy of air 283
factors that affect 112 environmental (gaseous) detector
fronts 108, 109 38
MRI of 113 environmental legislation 1516
importance of 95 environmental scanning electron
particle distribution during 99 microscopy see electron microscopy
three-stage process 98 EU Directive 2004/42/EC 15
two-stage process 98 evanescent wave 49
vertical 99107 evaporation
factors affecting 102 effects on 97
drying fronts 15 rate 96, 296
dry sintering see sintering evaporative cooling 32, 96
dwell time in MR profiling 277 evaporative lithography 267
dynamic speckle 48 face-centred cubic 225
elastic particles 127 Fickian diffusion 153
elastic spheres 128 filler particles 168, 171
electrical conductivity 36, 216 effect on diffusion 81, 170
electrical impedance 36 film formation
electric force microscopy 6970 mechanical probe 32
electron beam damage 40 stages of 10, 11
electron microscopy 3642 film formation paradox 174
cryogenic scanning 37, 104, 108 film scratching 32
cryogenic transmission 125 film topography 267
dark field 41 flame retardancy 214
environmental, pump down 41 flammability 214
environmental scanning 36, flocculation, definition of 20
3740, 145 flow, particle in Newtonian fluid
design 39 276
scanning 36, 723 flow instabilities 266
backscattering electron images fluorescence decay curves 80, 81
73 fluorescence resonance energy
scanning transmission 36 transfer 61, 76
transmission 41, 712 simulations 79
freeze-fracture 72 forced Rayleigh scattering 58, 59
staining 72 Forster radius 77
304 Index

Forster relation 76 interaction potentials 17


fraction of mixing after interdiffusion 152
interdiffusion 79 effects on 80
fracture energy 159, 296 techniques to study 74
effect of diffusion 160 interdiffusion distance 162
time dependence 162 interfacial chain density 162
fracture strength 159, 296 interfacial strength 247
fracture toughness 159, 2934 interfacial width 75, 152
free radicals 40 interparticle interference 51
Frenkel theory 128 interpenetration distance 75, 157
FTIR spectroscopy 73 interphase 215
further gradual coalescence 151 interstitial space between latex
GARField 5658, 2779 particles 169
experimental design 57 inverse micro-Raman spectroscopy
experimental profiles 105 53, 263
gel point 35 iridescence 232
glass transition temperature, Janus particles 260
definition of 2 Johnson, Kendall and Roberts 127
gloss, effect of surfactant 188 Kelvin probe force microscopy
Graham, Thomas 1 6970
gravimetry 32 knife point 29
Guinier plot 75 Krieger-Dougherty expression 23
Halpin-Tsai equations 214 Langmuir isotherm 193
Hamaker constant 18 laponite 264
Hertz theory 127 laponite clay 228
hetero-flocculation 2234 lapping time 111
homogeneous particles 2134 latex
honeycomb 13 blends 213
horizontal drying see drying definition of 1
humidity 28192 dialysed 189
definition of 95, 281 gloves 8
relative, definition of 2812 market for 9
hybrid 213, 224 natural 8
types of 21725 sensitisation 8
hydrophobicity 245 latex film formation 10
ideal gas equation 282 publications on 16
industrial coater 6 latex foam structures 173
infrared microscopy 53, 146 light scattering 44, 83
infrared spectroscopy 52 dynamic 45
inisurfs 205, 207 in nanocomposites 234
inks 6 magnetic resonance imaging 55
inorganic nanocomposite particles magnetic resonance profiling and
219 particle deformation 140
inorganic nanoparticles 245 magnetogyric ration 54
Institute Laue Langevin 43 Marangoni flows 199, 202
Index 305

Marangoni instabilities 200 nanoparticle


mass transfer coefficient 97 dispersion 2334
mass transfer resistance 98 encapsulated 222
melt compression 232 hybrid 224
membrane bending 34 Navier-Stokes equation 22, 2735
membranes 1723 Newtonian fluid 22
meniscus 124, 125, 230 NMR see nuclear magnetic
MFFT see minimum film formation resonance
temperature non-adsorbing polymer 20
micelle 191 non-radiative energy transfer 58, 61
microrheology 45 nuclear magnetic resonance 54
miniemulsion polymerisation 217 MOUSE 56
minimum film formation spectroscopy 74, 202
temperature occupational exposure limits 15
and particle size ratio 236 oligomers 268
definition of 14 opal structure 232
effect of particle size 30 double-inverse 262
effect of surfactant 191 inverse 262
interpretation 30 open time 107, 111
MFFT bar 2931, 139, 143 optical cantilever see beam bending
for studying deformation 138 optical clarity front 113
standard for 29 optical stethoscopy 70
time effects 30 optical transmission 143
modern art 189 optical transparency 14
moist sintering see sintering packing, face-centered cubic 12
molecular mobility 171 paints, formulation of 4
molecular weight 165 paper coatings 6
Monte Carlo simulation of drying parameter map 131
106 partial pressure 296
MRI see magnetic resonance particle
imaging core-shell 21810, 226
multispeckle 46 film formation 227, 264
nanocomposites 21349 half moon 218
classification 213 lobed 218
conductivity 216 occluded structures 218
cracking in 235 particle assembly 225, 261
failure mechanism 247 particle blends
in paints 216 advantages of 233
light scattering in 234 film formation 234
properties 214 hard-soft 243
silica 227 particle compressibility 102
soft-soft 242 particle deformation 10, 12
stiffness of 214 atomic force microscopy 144
toughness of 215 driving forces 121, 122
viscoelasticity 215 effect of particle size 139
306 Index

effect of temperature 1379 quadrature spin-echo sequence 277


MFFT bar 143 quality factor 64
scaling argument 135 quantum efficiency of energy
particle deposition methods 230 transfer 78
particle interfacial area 122 quartz crystal microbalance 73
particle packing 12, 260 radiolysis 40
effect of surfactant 191 radius of gyration 157, 170, 172
front 109 compared to diffusion distance
size ratio effects 2356 170
particle spacing 51 Raman spectroscopy 52
patterned substrate 231 surfactant analysis 202
peak-to-valley height 144, 145 random coil 172
Peclet number 101, 195 raspberry particles 222
effect on drying 104, 106 Rayleigh theory 51
peel strength 190 reactive surfactant 205
pendular rings 126 refractive index 143
percolation 23842 measurement of 52
effect on properties 241 replicas, transmission electron
model 239 microscopy 71
of rods 240 reptation 14, 152, 154
thresholds 239 reptation time 156
phase separation 234 Reynolds number 274
in particles 261 rheology modifiers 4
phenanthrene 77, 81 rhombic dodecahedron 13, 122
photoacoustic spectroscopy 73 root mean square displacement of
photon correlation spectroscopy 45 chains 156
photonic crystals 262 Rouse entanglement time 156
Pickering emulsion polymerisation Rouse relaxation time 156
222 Routh and Russel film deformation
plane strain 293 model 130
plane stress 293 Rutherford backscattering
plasticisation 16, 81 spectrometry, surfactant analysis
by surfactant 187 202
plasticisers 174 saturated vapour pressure 296
plastic zone 295 scanning electric potential
Plateau borders 146 microscopy 6970
Poissons ratio 127 scanning electron microscopy see
poly-condensation 217 electron microscopy
poly(dimethyl siloxane) 245 scanning near-field optical
Porod law 756 microscopy 50, 701
pressure-limiting apertures 38 scanning transmission electron
pressure sensitive adhesives 190 microscopy see electron
application of 5 microscopy
psychrometric chart 283 scattering angle 43
pulse gap in MR profiling 279 scattering techniques 4252
Index 307

scratch resistance 216 Stokes-Einstein diffusion coefficient


secondary ion mass spectrometry 22, 44, 101
201, 204 Stokes flow 22, 275
sedimentation 275 stray-field imaging 55
sedimentation coefficient 102 strength 214
sedimentation velocity 276 stress relaxation modulus 131
seeded emulsion polymerisation styrene-acrylic copolymers 3
219 surface patterns 230
shear force microscopy 701 surface roughness 144
shear modulus 158 surfactant 185207
Sheetz deformation 126, 136 anionic 185
silica cationic 185
nanocomposites 227 classification of 185
nanoparticles 244 convection of 194
particles 169, 172 desorption 187, 199
sintering exudation 187
dry 1234, 136 cause of 192
theory 129 effect of surfmers 206
moist 126 effect of Tg 199
wet 123, 135 fate of 1867
skin formation 58, 107, 141, 146 gloss effect 188
experimental evidence 142 non-ionic 185
study of 59 plasticisation by 187
skin layer 55, 115, 146 segregation 198
small-angle neutron scattering solubility in polymer 187
424, 145 surfactant-free emulsion
parameters for 43 polymerisation 185
surfactant analysis 202 surfactant-induced flow 267
to study interdiffusion 75 surfmer 2057
small-angle X-ray scattering temperature, effect on particle
424 deformation 1379
sodium dodecyl sulphate 187 templates for drying 231
soft-soft nanocomposites 242 tensile strength 160
sorptive capacity 107 TexanolTM 174
specific volume 282 textile backings 6
speckle thermal conductivity 216
commercial instrument 49 thermoelectric applications 2634
interferometry 48 thin film analyser 146
spectrophotometry 83 time-temperature superposition 167
specular reflection 188 tortuosity 169, 172
spin-casting 228 toughness 215
spin-spin relaxation time 55, 58, 74 transmission electron microscopy
star polymer 165 see electron microscopy
steric stabilisation 234 transmission spectrophotometry 50
stick-slip 116 transport coefficient 104
308 Index

transurfs 205, 207 water


tube model 155 adsorption 190
turbidity 83 diffusion coefficient of
ultramicroscopy 50 vapour 96
ultrasonic reflection 3435, 73 distribution profiles 141
van der Waals attraction 17, 128 surface tension 125
van der Waals forces 115 water whitening 191
varnishes, formulation of 4 wavevector 43
vertical deposition 2289 wet bulb temperature 282
vertical drying profiles see drying wet sintering see sintering
viscoelastic particles 122, 130 wet STEM see electron microscopy
viscosity wetting 152, 157
dependence on volume Williams-Landel-Ferry equation
fraction 23 167
measurement of 32 Winnik, M.A., 76
viscous flow of particles 128 X-ray photoelectron spectroscopy
VOC see volatile organic 201
compounds X-ray scattering 44
volatile organic compounds 15, 138 Youngs modulus 241

You might also like