See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/225756343

Miniemulsion Polymerization

Chapter  in  Advances in Polymer Science · March 2005

DOI: 10.1007/b100115

CITATIONS READS
381 3,753

6 authors, including:

F. Joseph Schork Yingwu Luo

Georgia Institute of Technology Zhejiang University
145 PUBLICATIONS   4,647 CITATIONS    122 PUBLICATIONS   4,584 CITATIONS   

SEE PROFILE SEE PROFILE

Alessandro Butte
ETH Zurich
86 PUBLICATIONS   1,901 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

flexible transparent electrodes View project

Polyesters View project

All content following this page was uploaded by Yingwu Luo on 20 May 2014.

The user has requested enhancement of the downloaded file.

175
Advances in Polymer Science

Editorial Board:
A. Abe · A.-C. Albertsson · R. Duncan · K. Dušek · W. H. de Jeu
J. F. Joanny · H.-H. Kausch · S. Kobayashi · K.-S. Lee · L. Leibler
T. E. Long · I. Manners · M. Möller · O. Nuyken · E. M. Terentjev
B. Voit · G. Wegner
Advances in Polymer Science
Recently Published and Forthcoming Volumes

Inorganic Polymeric Nanocomposites Polymers and Light

and Membranes Volume Editor: Lippert, T. K.
Vol. 179, 2005 Vol. 168, 2004

Polymeric and Inorganic Fibres New Synthetic Methods

Vol. 178, 2005 Vol. 167, 2004

Poly(arylene Ethynylenes) Polyelectrolytes with Defined

From Synthesis to Application Molecular Architecture II
Volume Editor: Weder, C. Volume Editor: Schmidt, M.
Vol. 177, 2005 Vol. 166, 2004

Polyelectrolytes with Defined

Metathesis Polymerization
Molecular Architecture I
Volume Editor: Buchmeiser, M.
Volume Editor: Schmidt, M.
Vol. 176, 2005
Vol. 165, 2004
Polymer Particles Filler-Reinforeced Elastomers ·
Volume Editor: Okubo, M. Scanning Force Microscopy
Vol. 175, 2005 Vol. 164, 2003
Neutron Spin Echo in Polymer Systems Liquid Chromatography ·
Authors: Richter, D., Monkenbusch, M., FTIR Microspectroscopy · Microwave
Arbe, A., Colmenero, J. Assisted Synthesis
Vol. 174, 2005 Vol. 163, 2003

Advanced Computer Simulation Radiation Effects on Polymers

Approaches for Soft Matter Sciences I for Biological Use
Volume Editors: Holm, C., Kremer, K. Volume Editor: Kausch, H.
Vol. 173, 2005 Vol. 162, 2003

Microlithography · Molecular Imprinting Polymers for Photonics

Vol. 172, 2005 Applications II
Nonlinear Optical, Photorefractive and
Polymer Synthesis Two-Photon Absorption Polymers
Vol. 171, 2004 Volume Editor: Lee, K.-S.
Vol. 161, 2003
NMR · Coordination Polymerization ·
Photopolymerization Filled Elastomers · Drug Delivery Systems
Vol. 170, 2004 Vol. 160, 2002

Statistical, Gradient, Block

Long-Term Properties of Polyolefins
and Graft Copolymers by Controlled/
Volume Editor: Albertsson, A.-C.
Living Radical Polymerizations
Vol. 169, 2004
Authors: Davis, K. A., Matyjaszewski, K.
Vol. 159, 2002
Polymer Particles

Volume Editor: Masayoshi Okubo

With contributions by
A. Butté · P. Y. Chow · J. M. DeSimone · K. Fontenot · L. M. Gan
K. Ito · S. Kawaguchi · K. A. Kennedy · Y. Luo · M. Nomura
G. W. Roberts · J. P. Russum · F. J. Schork · W. Smulders · K. Suzuki
H. Tobita

2 3
The series presents critical reviews of the present and future trends in polymer and biopolymer science
including chemistry, physical chemistry, physics and material science. It is addressed to all scientists at
universities and in industry who wish to keep abreast of advances in the topics covered.

As a rule, contributions are specially commissioned. The editors and publishers will, however, always be
pleased to receive suggestions and supplementary information. Papers are accepted for “Advances in Polymer
Science” in English.

In references Advances in Polymer Science is abbreviated Adv Polym Sci and is cited as a journal.

The electronic content of APS may be found springerlink.com

Library of Congress Control Number: 2004110358

ISSN 0065-3195
ISBN 3-540-22923-X Springer Berlin Heidelberg New York
DOI 10.1007/b14102

This work is subject to copyright. All rights are reserved, whether the whole or part of the material is
concerned, specifically the rights of translation, re-printing, re-use of illustrations, recitation, broadcasting,
reproduction on microfilms or in any other ways, and storage in data banks. Duplication of this publication
or parts thereof is only permitted under the provisions of the German Copyright Law of September 9, 1965,
in its current version, and permission for use must always be obtained from Springer-Verlag. Violations are
liable to prosecution under the German Copyright Law.

Springer is a part of Springer Science+Business Media

springeronline.com
© Springer-Verlag Berlin Heidelberg 2005
Printed in The Netherlands
The use of registered names, trademarks, etc. in this publication does not imply, even in the absence of a
specific statement, that such names are exempt from the relevant protective laws and regulations and there-
fore free for general use.
Cover design: KünkelLopka GmbH, Heidelberg/design & production GmbH, Heidelberg
Typesetting: Fotosatz-Service Köhler GmbH, Würzburg
Printed on acid-free paper 02/3141 xv – 5 4 3 2 1 0
Volume Editor
Prof. Masayoshi Okubo
Kobe University
Fac. Engineering Dept.
Rokko-dai 1-1, Nada-ku
657-8501 Kobe, Japan
[email protected]

Editorial Board
Prof. Akihiro Abe Dutch Polymer Institute
Department of Industrial Chemistry Eindhoven University of Technology
Tokyo Institute of Polytechnics PO Box 513
1583 Iiyama, Atsugi-shi 243-02, Japan 5600 MB Eindhoven, The Netherlands
[email protected]
Prof. Jean-François Joanny
Prof. A.-C. Albertsson Physicochimie Curie
Department of Polymer Technology Institut Curie section recherche
The Royal Institute of Technology 26 rue d’Ulm
S-10044 Stockholm, Sweden F-75248 Paris cedex 05, France
[email protected] [email protected]

Prof. Ruth Duncan Prof. Hans-Henning Kausch

Welsh School of Pharmacy EPFL SB ISIC GGEC
Cardiff University J2 492 Bâtiment CH
Redwood Building Station 6
King Edward VII Avenue CH-1015 Lausanne, Switzerland
Cardiff CF 10 3XF [email protected]
United Kingdom
[email protected] Prof. S. Kobayashi
Department of Materials Chemistry
Prof. Karel Dušek Graduate School of Engineering
Institute of Macromolecular Chemistry, Kyoto University
Czech Kyoto 615-8510, Japan
Academy of Sciences of the [email protected]
Czech Republic
Heyrovský Sq. 2 Prof. Kwang-Sup Lee
16206 Prague 6, Czech Republic Department of Polymer Science &
[email protected] Engineering
Hannam University
Prof. Dr. W. H. de Jeu 133 Ojung-Dong
FOM-Institute AMOLF Daejeon 306-791, Korea
Kruislaan 407 [email protected]
1098 SJ Amsterdam, The Netherlands
[email protected]
and
VI Editorial Board

Prof. L. Leibler Prof. Oskar Nuyken

Matière Molle et Chimie Lehrstuhl für Makromolekulare Stoffe
Ecole Supèrieure de Physique TU München
et Chimie Industrielles (ESPCI) Lichtenbergstr. 4
10 rue Vauquelin 85747 Garching, Germany
75231 Paris Cedex 05, France [email protected]
[email protected]
Dr. E. M. Terentjev
Prof. Timothy E. Long Cavendish Laboratory
Department of Chemistry Madingley Road
and Research Institute Cambridge CB 3 OHE
Virginia Tech United Kingdom
2110 Hahn Hall (0344) [email protected]
Blacksburg, VA 24061, USA
[email protected] Prof. Brigitte Voit
Institut für Polymerforschung Dresden
Prof. Ian Manners Hohe Straße 6
Department of Chemistry 01069 Dresden, Germany
University of Toronto [email protected]
80 St. George St.
M5S 3H6 Ontario, Canada Prof. Gerhard Wegner
[email protected] Max-Planck-Institut
für Polymerforschung
Prof. Dr. Martin Möller Ackermannweg 10
Deutsches Wollforschungsinstitut Postfach 3148
an der RWTH Aachen e.V. 55128 Mainz, Germany
Pauwelsstraße 8 [email protected]
52056 Aachen, Germany
[email protected]
Advances in Polymer Science
Also Available Electronically

For all customers who have a standing order to Advances in Polymer Science,
we offer the electronic version via SpringerLink free of charge. Please contact
your librarian who can receive a password for free access to the full articles by
registering at:

springerlink.com

If you do not have a subscription, you can still view the tables of contents of the
volumes and the abstract of each article by going to the SpringerLink Home-
page, clicking on “Browse by Online Libraries”, then “Chemical Sciences”, and
finally choose Advances in Polymer Science.

You will find information about the

– Editorial Board
– Aims and Scope
– Instructions for Authors
– Sample Contribution
at springeronline.com using the search function.
Preface

Preface

In this special volume on polymer particles, recent trends and developments

in the synthesis of nano- to micron-sized polymer particles by radical poly-
merization of vinyl monomers in environmentally friendly heterogeneous
aqueous and supercritical carbon dioxide fluid media are reviewed by promi-
nent worldwide researchers. Polymer particles are prepared extensively as
synthetic emulsions and latexes, which are applied as binders in the industri-
al fields of paint, paper and inks, and films such as adhesives and coating
materials. Considerable attention has recently been directed towards aqueous
dispersed systems due to the increased awareness of environmental issues.
Moreover, such polymer particles have already been applied to more advanced
fields such as bio-, information, and electronic technologies. In addition to the
obvious commercial importance of these techniques, it is of fundamental sci-
entific interest to completely elucidate the mechanistic details of macromole-
cule synthesis in the “microreactors” that the polymer particles in these het-
erogeneous systems constitute.
In the first chapter, Professor Nomura et al. review features of emulsion
polymerization, which is applied for the synthesis of submicron-sized poly-
mer particles, with particular emphasis on particle nucleation and growth and
polymer structure development. In the second chapter, Professor Schork
describes the basic features of miniemulsion polymerization, which is deeply
related to emulsion polymerization, but offers advantages for the synthesis of
hybrid particles in which hydrophobic substances are included. This chapter
also covers controlled/living radical polymerization, which has been devel-
oped over the past 10 years. It is of great importance both from an academic
and industrial perspective to make controlled/living radical polymerization
compatible with heterogeneous systems. In the third chapter, Professsor Gan
and coworkers review microemulsion polymerization and its applications for
the synthesis of polymer nanoparticles and nanocomposites of polymeric/
inorganic substances. In the fourth chapter, Professors Kawaguchi and Ito
review various features of dispersion polymerization, which is a useful tech-
nique for the synthesis of micron-sized monodisperse polymer particles,
focusing on the preparation of novel functional particles and the control of
particle size. In the final chapter, heterogeneous polymerization of fluoro-
olefins in supercritical carbon dioxide fluids is reviewed. One of the authors,
X Preface

Professor DeSimone, is the pioneer of heterogeneous polymerizations in

supercritical carbon dioxide.
As the editor of this Special Volume on “Polymer Particles”, I would like to
thank all of the authors who made valuable contributions in spite of their
undoubtedly busy schedules. I believe this work will be of great use for scien-
tists in both academia and industry.

Kobe, November 2004 Masayoshi Okubo

Contents

Emulsion Polymerization: Kinetic and Mechanistic Aspects

M. Nomura · H. Tobita · K. Suzuki . . . . . . . . . . . . . . . . . . . . . . 1

Miniemulsion Polymerization
F. J. Schork · Y. Luo · W. Smulders · J. P. Russum · A. Butté · K. Fontenot . . 129

Microemulsion Polymerizations and Reactions

P. Y. Chow · L. M. Gan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257

Dispersion Polymerization
S. Kawaguchi · K. Ito . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299

Heterogeneous Polymerization of Fluoroolefins

in Supercritical Carbon Dioxide
K. A. Kennedy · G. W. Roberts · J. M. DeSimone . . . . . . . . . . . . . . . 329

Author Index Volumes 101–175 . . . . . . . . . . . . . . . . . . . . . . . 347

Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 367

Adv Polym Sci (2005) 175: 1–128
DOI 10.1007/b100116
© Springer-Verlag Berlin Heidelberg 2005

Emulsion Polymerization:
Kinetic and Mechanistic Aspects
Mamoru Nomura ( ) · Hidetaka Tobita · Kiyoshi Suzuki
Department of Materials Science and Engineering, Fukui University, Fukui, Japan
[email protected], [email protected], [email protected]

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2 Emulsion Polymerization Kinetics . . . . . . . . . . . . . . . . . . . . . . 4

2.1 Generally Accepted Kinetics Scheme . . . . . . . . . . . . . . . . . . . . . 4
2.2 Summary of the Smith-Ewart Theory . . . . . . . . . . . . . . . . . . . . . 6

3 Kinetics and Mechanisms of Emulsion Polymerization . . . . . . . . . . . 7

3.1 Radical Entry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.1.1 Diffusion-Controlled Entry . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.1.2 Propagation-Controlled Entry . . . . . . . . . . . . . . . . . . . . . . . . 11
3.1.3 Miscellaneous Kinetic Problems in Radical Entry . . . . . . . . . . . . . . 13
3.2 Radical Desorption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.2.1 Desorption in Homopolymer Systems . . . . . . . . . . . . . . . . . . . . 16
3.2.2 Desorption in Copolymer Systems . . . . . . . . . . . . . . . . . . . . . . 19
3.2.3 Miscellaneous Kinetics Problems in Radical Desorption . . . . . . . . . . 21
3.3 Particle Formation and Growth . . . . . . . . . . . . . . . . . . . . . . . . 22
3.3.1 Particle Formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.3.2 Particle Growth in Homopolymer Systems . . . . . . . . . . . . . . . . . . 36
3.3.3 Particle Growth in Copolymer Systems . . . . . . . . . . . . . . . . . . . . 42
3.3.4 Monomer Concentration in Polymer Particles . . . . . . . . . . . . . . . . 47
3.3.5 Reaction Calorimetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.4 Effect of Initiator Type . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3.5 Effect of Additives and Impurities . . . . . . . . . . . . . . . . . . . . . . 66
3.6 Effects of Other Important Factors . . . . . . . . . . . . . . . . . . . . . . 74

4 Kinetic Aspects in Polymer Structure Development . . . . . . . . . . . . . 81

4.1 Molecular Weight Distribution (MWD) . . . . . . . . . . . . . . . . . . . . 81
4.1.1 Monte Carlo (MC) Simulation Method . . . . . . . . . . . . . . . . . . . . 81
4.1.2 Instantaneous Molecular Weight Distribution . . . . . . . . . . . . . . . . 83
4.1.3 Effect of Chain-Length-Dependent Bimolecular Termination . . . . . . . . 89
4.1.4 Accumulated Molecular Weight Distribution . . . . . . . . . . . . . . . . . 91
4.1.5 Determination of Monomer Transfer Constants from MWD . . . . . . . . 92
4.2 Branched and Crosslinked Polymer Formation . . . . . . . . . . . . . . . 94
4.2.1 Long-Chain Branched Polymers . . . . . . . . . . . . . . . . . . . . . . . 94
4.2.2 Crosslinked Polymers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

5 Continuous Emulsion Polymerization . . . . . . . . . . . . . . . . . . . . 108

6 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
2 M. Nomura et al.

Abstract The current understanding of the kinetics and mechanisms of batch and continu-
ous emulsion polymerizations is summarized from the viewpoints of particle formation and
growth and polymer structure development. There are numerous factors that affect these
processes; among them, studies on the radical transfer and monomer partitioning between
phases, which are key factors for particle formation and growth, are reviewed and discussed.
Attention is also focused on the effects of initiator type, additives and impurities in the recipe
ingredients, and agitation, each of which sometimes exert crucial influences on the processes
of particle formation and growth. In relation to polymer structure development, important
aspects of the molecular weight distribution and branched/crosslinked polymer formation
are highlighted.

Keywords Emulsion polymerization · Kinetics · Particle nucleation · Particle growth ·

Molecular weight distribution · Nonlinear polymers

Abbreviations
AA acrylic acid
AAm acrylamide
AN acrylonitrile
APS ammonium persulfate
AIBN 2,2¢-azobis-isobutyronitrile
BA butyl acrylate
Bu butadiene
CCTVFR continuous Couette-Taylor vortex flow reactor
CLTR continuous loop-tubular reactor
CMC critical micellar concentration
CSTR continuous stirred tank reactor
CTR continuous tubular reactor
DVB divinylbenzene
E ethylene
KPS potassium persulfate
MA methyl acrylate
MAA methacrylic acid
MC Monte Carlo
MMA methyl methacrylate
MWD molecular weight distribution
NaLS sodium laurylsulfate
n-BA n-butyl methacrylate
PSD particle size distribution
PSPC(R) pulsed sieve plate column (reactor)
PT(R) pulsed tubular (reactor)
S-E Smith and Ewart
SEC size exclusion chromatography
St styrene
VAc vinyl acetate
VCl vinyl chloride
Am total surface area of micelles per unit volume of water
Ap total surface area of polymer particles per unit volume of water
as surface area occupied by a unit amount of emulsifier
dm diameter of a micelle
dp diameter of a polymer particle
Dw diffusion coefficient for radicals in the aqueous phase
Emulsion Polymerization: Kinetic and Mechanistic Aspects 3

Dp diffusion coefficient for radicals inside a polymer particle

E(t) residence time distribution function
F absorption efficiency factor defined in Eq. 8
f initiator efficiency
fi fraction of i-radicals in the polymer particle phase
I0, [I0] initial initiator concentration
kd rate constant for initiator decomposition
kem mass transfer coefficient for micelles defined by Eq. 7
kep mass transfer coefficient for polymer particles defined by Eq. 7
kf rate coefficient for radical desorption per particle
kmf chain transfer rate constant to monomer
ktw rate coefficient for bimolecular radical termination in the aqueous phase
kTf chain transfer rate constants to chain transfer agent (CTA)
md partition coefficient for monomer (monomeric radicals) between particle and
aqueous phases defined by md=[M]p/[M]m
Mm aggregation number of emulsifier molecules per micelle
M0 initial monomer concentration
[M]pc constant monomer concentration in polymer particles at saturation swelling
[M]w monomer concentration in the aqueous phase
–
n average number of A-radicals per polymer particle
A
Nn number of polymer particles containing n radicals
Rp rate of polymerization
[R*w] radical concentration in the aqueous phase
S0 initial emulsifier concentration
Sm concentration of emulsifier forming micelles
vp volume of a polymer particle
XMc critical monomer conversion where monomer droplets disappear from the
aqueous phase
aw nondimensional parameter defined by Çwvp/ktpNT
e defined by (kep/kem)Mm in Eq. 37
l partition coefficient for radicals between particle and water phase
m volumetric growth rate per polymer particle
q mean residence time
Ç radical entry rate per polymer particle defined by Eq. 12
Çe overall rate of radical entry into polymer particles
Çp polymer density
Çw rate of radical generation per unit volume of water

1
Introduction

There are four main types of liquid-phase heterogeneous free-radical poly-

merization; microemulsion polymerization, emulsion polymerization, mini-
emulsion polymerization and dispersion polymerization, all of which can
produce nano- to micron-sized polymeric particles. Emulsion polymerization
is sometimes called macroemulsion polymerization. In recent years, these het-
erophase polymerization reactions have become more and more important
4 M. Nomura et al.

technologically and commercially, not only as methods for producing high-

performance polymeric materials, but also from an environmental point of
view. It is well known that microemulsion, miniemulsion and dispersion poly-
merizations bare many similarities to emulsion polymerization in the kinetics
of particle nucleation and growth and in polymer structure development.
Therefore, for optimal design and operation of these heterophase free radical
polymerizations, it is important to have detailed knowledge of the kinetics and
mechanisms of emulsion polymerization. In this article, recent developments
in emulsion polymerization are reviewed from kinetic and mechanistic per-
spectives.
Between 1995 and 1997, three excellent books on emulsion polymerization
were published and provide extensive reviews of the subject up to 1995 [1–3].
Therefore, this review article will focus on research work that has appeared
since ~1996.We will also include historically important work from before 1995
in this review article.

2
Emulsion Polymerization Kinetics

2.1
Generally Accepted Kinetics Scheme

Emulsion polymerization takes place over a number of steps, where various

chemical and physical events take place simultaneously during the process of
particle formation and growth. Figure 1 depicts the generally accepted scheme
for the kinetics of emulsion polymerization.
Three major mechanisms for particle formation have been proposed to date.
Figure 1a shows the proposed scheme for particle formation in emulsion poly-
merization initiated by water-soluble initiators. Particle formation is consid-
ered to take place when either: (1) a free radical in the aqueous phase enters a
monomer-swollen emulsifier micelle and propagation proceeds therein (mi-
cellar nucleation); (2) the chain length of a free radical growing in the aqueous
phase exceeds its solubility limit and precipitates to form a particle nucleus
(homogeneous nucleation), or; (3) a free radical growing in the aqueous phase
enters a monomer droplet and propagation proceeds therein (droplet nucle-
ation). However, if the resultant polymer particles are not stable enough, the
final number of polymer particles produced, regardless of the mechanism of
particle formation, is determined by coagulation between the existing particles
(coagulative nucleation).
In the process of particle growth, various chemical and physical events oc-
cur in both the aqueous and particle phases, as illustrated in Fig. 1b [1].We now
know that the polymerization takes place exclusively in the resultant polymer
particle phase, wherever the free radicals are generated. Smith and Ewart [4]
were the first to establish a quantitative description of the processes of parti-
Emulsion Polymerization: Kinetic and Mechanistic Aspects 5

(a)

(b)

Fig. 1 (a) Three major mechanisms of particle formation, and (b) various chemical and
physical events that occur during the process of particle growth in an emulsion polymer-
ization
6 M. Nomura et al.

cle formation and growth in emulsion polymerization on the basis of the

achievements made by Harkins et al. [5]. This is now called the Smith-Ewart
theory. It is not an exaggeration to say that almost all of the theoretical devel-
opments in emulsion polymerization that have been made so far are based on
the Smith-Ewart theory.

2.2
Summary of the Smith-Ewart Theory

The Smith and Ewart theory (the S-E theory) describes the basic concept of
emulsion polymerization. Its main points are briefly reviewed here. Smith and
Ewart showed that the rate of emulsion polymerization, which proceeds
exclusively in the polymer particles, is given by
R = k [M ] nN– (1)
p p p T

where kp is the propagation rate constant, [M]p is the monomer concentration

in the monomer-swollen polymer particles, NT is the number of monomer-
swollen polymer particles per unit volume of water and n – is the average num-
ber of radicals per particle, defined as

• • •
– = ∑ nN
n ∑ Nn = ∑ nNn NT (2)
n
n=1 0 n=1

where Nn is the number of polymer particles containing n free radicals. Nn

is described by the following balance equation that takes into account three rate
processes: (1) radical entry into, (2) radical desorption (exit) from, and (3)
bimolecular radical termination inside the polymer particle
dNn/dt = (Çe /NT)Nn–1 + kf (n + 1)Nn+1 + ktp [(n + 2)(n + 1)/vp] (3)
– Nn {Çe /NT + kf n + ktp [(n(n – 1)/vp]} = 0
where kf is the rate coefficient for radical desorption per particle, vp is the
volume of a polymer particle, ktp is the rate coefficient for bimolecular radical
termination inside the polymer particles, and Çe is the overall rate of radical
entry into polymer particles, defined by
Ç = Ç + k nN– – 2k [R*]2 (4)
e w f T tw w

where Çw is the rate of radical production per unit volume of water, ktw is the
rate coefficient for bimolecular radical termination in the aqueous phase, and
[R*w] is the radical concentration in the aqueous phase.
On the other hand, they derived an expression that predicts the number
of polymer particles produced, NT, assuming that (i) a monomer-swollen
emulsifier micelle is transformed into a polymer particle by capturing a free
radical from the aqueous phase, (ii) the volumetric growth rate per particle m
is constant, at least during particle formation, and (iii) free radical activity does
not transfer out of a growing particle
Emulsion Polymerization: Kinetic and Mechanistic Aspects 7

NT = k(Çw /m)0.4 (as S0)0.6 (5)

where k is a constant between 0.37 and 0.53, as is the surface area occupied
by a unit amount of emulsifier, S0 is the initial emulsifier concentration (the
concentration of emulsifier forming micelles), and Çw is the rate of radical
generation per unit volume of water, given by
Çw = 2kd f [I0] (6)
where kd is the rate constant for initiator decomposition, f is the initiator
efficiency, and [I0] is the initial initiator concentration. Since the appearance of
the S-E theory, much effort has been directed into investigating the physical
meanings of various parameters such as Çe, kf and ktp, and the effects of these
parameters on the three key factors of emulsion polymerization, [M]p, n – and N .
T

3
Kinetics and Mechanisms of Emulsion Polymerization

3.1
Radical Entry

One of the most important parameters in the S-E theory is the rate coefficient
for radical entry. When a water-soluble initiator such as potassium persulfate
(KPS) is used in emulsion polymerization, the initiating free radicals are gen-
erated entirely in the aqueous phase. Since the polymerization proceeds ex-
clusively inside the polymer particles, the free radical activity must be trans-
ferred from the aqueous phase into the interiors of the polymer particles, which
are the major loci of polymerization. Radical entry is defined as the transfer of
free radical activity from the aqueous phase into the interiors of the polymer
particles, whatever the mechanism is. It is believed that the radical entry event
consists of several chemical and physical steps. In order for an initiator-derived
radical to enter a particle, it must first become hydrophobic by the addition
of several monomer units in the aqueous phase. The hydrophobic oligomer
radical produced in this way arrives at the surface of a polymer particle by mol-
ecular diffusion. It can then diffuse (enter) into the polymer particle, or its
radical activity can be transferred into the polymer particle via a propagation
reaction at its penetrated active site with monomer in the particle surface layer,
while it stays adsorbed on the particle surface.A number of entry models have
been proposed: (1) the surfactant displacement model; (2) the collisional
model; (3) the diffusion-controlled model; (4) the colloidal entry model, and;
(5) the propagation-controlled model. The dependence of each entry model on
particle diameter is shown in Table 1 [12].
However, some of these models have been refuted, and two major entry
models are currently widely accepted. One is the diffusion-controlled model,
which assumes that the diffusion of radicals from the bulk phase to the surface
8 M. Nomura et al.

Table 1 Dependence of entry rate coefficient on particle diameter, as predicted by different

models [12]

Entry model Dependence on dp

Surfactant displacement [7] none

Collisional [8] dp2
Diffusional [9] dp
Colloidal [10] dp
Propagational [6, 11] no dependence

of a polymer particle is the rate-controlling step. The other is the propagation-

controlled model, which assumes that since only z-mer radicals can enter the
polymer particles very rapidly, the generation of z-mer radicals from (z–1)-mer
radicals by a propagation reaction in the aqueous phase is the rate-controlling
step.

3.1.1
Diffusion-Controlled Entry

Smith and Ewart [4] first proposed that the transfer of free radical activity into
the interior of a polymer particle takes place by the direct entry of a free rad-
ical into a polymer particle. They pointed out that the rate of radical entry into
a polymer particle is given by the rate of diffusion of free radicals from an in-
finite medium of concentration [R*w] into a particle of diameter dp with zero
radical concentration,
Çe /NT = 2p Dw dp[R*w] = kep [R*w] (7)
where Dw is the diffusion coefficient for the radicals in the water phase and kep
the mass transfer coefficient for radical entry into a particle. However, for
simplicity, they actually used a rate coefficient that is proportional to the square
of the diameter (the surface area). Since then, most researchers have treated the
problem of particle formation by assuming that the rate of radical entry into
a micelle and a polymer particle is proportional to the surface area (the colli-
sional entry model) [8, 13].
On the other hand, Nomura and Harada [14] proposed a kinetic model for
the emulsion polymerization of styrene (St), where they used Eq. 7 to predict
the rate of radical entry into both polymer particles and monomer-swollen
micelles. In their kinetic model, the ratio of the mass-transfer coefficient for
radical entry into a polymer particle kep to that into a micelle kem, kep/kem, was
the only one unknown parameter (Eq. 37). They determined the value of kep/kem
to be about 103 by comparing the model’s predictions with experimental re-
sults. However, the observed value of kep/kem was at least two orders of magni-
tude greater than that predicted by Eq. 7, because kep/kem=dp/dm (dm is the
Emulsion Polymerization: Kinetic and Mechanistic Aspects 9

diameter of a micelle) according to Eq. 7 and the value of dp/dm would be 10 at

the most during particle formation. This was considered to indicate that the
radical capture efficiency of a micelle is a factor of about 100 less than that of
a particle. Taking this into consideration, they implicitly introduced a concept
called the “radical capture efficiency of a micelle relative to a polymer particle”
to adjust for this disagreement and pointed out two possible reasons for the
lower radical capture efficiency of a micelle. One is that the energy barrier
against the entry of charged radicals into micelles may be higher than that into
polymer particles. The other is that an oligomeric radical, having entered a mi-
celle, may pass through the micelle without adding at least one extra monomer
unit because the volume of the micelle is so small that the mean residence time
of the radical in the micelle is too short for the radical to add another unit.
The concept of “radical capture efficiency” was further elaborated on by
Hansen et al. [15–17]. By applying the theory of mass transfer with simultane-
ous chemical reactions, they proposed the following expression to represent the
net rate of radical absorption by a particle, introducing an “absorption effi-
ciency factor” F into Eq. 8
Çe /NT = 2p Dw dp[R*w]F = kep [R*w]. (8)
Therefore, F represents a factor that describes the degree to which absorption
is lowered compared to irreversible diffusion, and is given by

 
1 Dw –1
(9)
3 = 7 (X coth X – 1) + W¢
F l Dp
where X=(dp/2){(kp[M]p+nktp/vp)/Dp}1/2 and l is the equilibrium partition co-
efficient between particles and water for radicals, W¢ is the potential energy

Fig. 2 Capture efficiency F as a function of particle size rp for different values of the parti-
tion coefficient, l, and the number of radicals in a polymer particle for St polymerization;
(a) n=0 and (b) n=1
10 M. Nomura et al.

barrier analogous to Fuchs’ stability factor, Dp is the diffusion coefficient for the
radicals inside a particle, kp is the propagation rate constant, [M]p is the
monomer concentration in particles, n is the number of radicals in a particle,
NA is Avogadro’s number and vp is the particle volume. It should be noted that
radicals are captured inside the particles only if they react therein; otherwise
they will eventually diffuse out and back to the water phase. Figure 2 shows
an example of F versus particle radius rp, calculated from Eq. 9. The important
conclusion of Eq. 9 is, as is clear from Fig. 2, that the value of F for a particle con-
taining radicals is higher than that for a particle containing no radicals.
According to Fig. 2a, it approximately holds that Fµdp2, and hence this model
gives kep/kem=(dp /dm)(Fp/Fm)=(dp /dm)3 @103, the value of which is in good
agreement with the result obtained by Nomura et al. in the emulsion polymer-
ization of St [14].
A much simpler model for the radical capture (absorption) efficiency F can
be derived by introducing the concept of radical desorption from a polymer
particle, developed in Section 3.2.1. The probability F for a radical to be cap-
tured inside a particle containing n radicals by any chemical reaction (propa-
gation or termination) is given by

kp [M]p + ktp(n/vp)
F = 99993 (10)
Ko + kp [M]p + ktp(n/vp)

where Ko is the overall radical desorption rate constant for a particle, defined
by Eq. 19 and shown later in Section 3.2.1. For simplicity, no distinction is made
here between radicals with and without initiator fragments at their ends. In the
case where Kopkp[M]p, ktp(n/vp), substitution of Eq. 19 into Eq. 10 leads to

  
kp [M]p md yDw
F = kp [M]p /Ko = 99 1 + 82 dp2 µ dp2 (11)
12Dw mdDp

The result of Eq. 11 agrees with Fµdp2 obtained above by Ugelstad and Hansen
[15]. Therefore, both Eq. 9 developed by Hansen and Ugelstad and Eq. 11
developed here can explain the value of kep/kem@103 found experimentally
by Nomura et al. [14], although no direct experimental confirmation of the
validity of these radical capture models have been reported yet.
Unzueta et al. [18] derived a kinetic model for the emulsion copolymeriza-
tion of methyl methacrylate (MMA) and butyl acrylate (BA) employing both
the micellar and homogeneous nucleation mechanisms and introducing the
radical absorption efficiency factor for micelles, Fm, and that for particles, Fp.
They compared experimental results with model predictions, where they em-
ployed the values of Fp=10–4 and Fm=10–5, respectively, as adjustable param-
eters. However, they did not explain the reason why the value of Fm is an order
of magnitude smaller than the value of FP. Sayer et al. [19] proposed a kinetic
model for continuous vinyl acetate (VAc) emulsion polymerization in a pulsed
Emulsion Polymerization: Kinetic and Mechanistic Aspects 11

sieve plate column reactor, where they assumed that both micellar and homo-
geneous nucleation takes place, and introduced the radical absorption efficiency
factor Fm for micelles and Fp for polymer particles, respectively. They could
explain the experimental results by employing Fm=1.0¥10–5 and Fp=3.3¥10–3 in
the model predictions, indicating that kep/kem=330. This value agrees fairly well
with the value of 100 found for the St system [14], but is 30 times less than
kep/kem=104 found for the VAc system [20].Araújo et al. [21] developed a detail-
ed dynamic mathematical model that describes the evolution of particle size
distributions (PSDs) during the emulsion copolymerization of VAc and
Veova10 in a continuous loop-tubular reactor and compared results from it
with their experimental data. They could describe the process of micellar par-
ticle formation by introducing radical absorption efficiency factors for micelles
of Fm=1.5¥10–4 and for particles of Fp=1.5¥10–3, respectively, although they also
did not provide a reason why the value of Fm is 1/10 of the value of Fp. This gives
kep/kem=(dp/dm)(Fp/Fm)@102 if one assumes that dp/dm@10. On the other hand,
Herrera-Ordóñez et al. [22] also developed a mathematical model for St emul-
sion polymerization employing Eq. 8 as the radical capture rate coefficient,
where the expression for the capture of monomeric radicals is that used by
Hansen and Ugelstad [15, 17], while a more detailed modification was made for
the entry of initiator-derived radicals.

3.1.2
Propagation-Controlled Entry

Maxwell et al. [11] proposed a radical entry model for the initiator-derived
radicals on the basis of the following scheme and assumptions. The major
assumptions made in this model are as follows: An aqueous-phase free radical
will irreversibly enter a polymer particle only when it adds a critical number
z of monomer units. The entrance rate is so rapid that the z-mer radicals can
survive the termination reaction with any other free radicals in the aqueous
phase, and so the generation of z-mer radicals from (z–1)-mer radicals by the
propagation reaction is the rate-controlling step for radical entry. Therefore,
based on the generation rate of z-mer radicals from (z–1)-mer radicals by prop-
agation reaction in the aqueous phase, they considered that the radical entry
rate per polymer particle, Ç (Ç=Çe/NT) is given by
Ç = kpw[IM*z–1][M]w NT (12)
where kpw is the propagation rate constant in the aqueous phase and [M]w
is the monomer concentration in the aqueous phase. By substituting the steady-
state concentration of (z–1)-mer radicals [IM*z–1] into Eq. 12, the approximate
expressions for Ç and the initiator efficiency, fentry are derived, respectively, as

 
2kd[I] d9kd[I] k41
t,w
1–z 2kd[I]
Ç = 92 926 + 1 = 721 fentry (13)
Np kp,w[M]w Np
12 M. Nomura et al.

 
d9kd[I] k41
t,w
1–z
fentry = 926 + 1 (14)
kp,w[M]w

There has been discussion on the value of z. Maxwell et al. [11] proposed a
semi-empirical thermodynamic model to predict the value of z for persulfate-
derived oligomeric radicals, which is given by
z @ 1 + int(–23 kJmol–1/{RT ln[Msat]w}) (15)
where the integer function (int) rounds down the quantity in parentheses to the
nearest integer value and [Msat]w is the saturation solubility of the monomer in
mol dm–3. On the other hand, Sundberg et al. [23] proposed a thermodynamic
method for estimating the critical chain length z of entry radicals with a
hydrophilic end group (such as SO4–) using a simple two-layer lattice model. The
values of z calculated by both Sundberg et al. and by Maxwell et al. (Eq. 15) are
listed in Table 2.
Several research articles have been published that deal with the methodology
for determining the radical entry rate Ç, the initiator efficiency fentry and the
actual values of z. Hawkett et al. [24] developed a method for determining the
value of Ç along with the desorption rate coefficient kf, termed the slope-and-
intercept method. This method is experimentally simple, but has several draw-
backs [25]. For example, it is only applicable to the so-called zero-one system
– ≤0.5) with negligible radical termination in the aqueous phase. It is usually
(n
very difficult to judge whether or not the radical termination in the aqueous
phase is negligible. Moreover, it gives a large error if an induction period caused
by any trace of impurity exists. Marestin et al. [26] proposed an experimental
method for directly determining the entry rate of a critical size MMA oligomer
into the polymer particle using the seeded emulsion polymerization of MMA

Table 2 Predicted Z values for persulfate-derived radicals

Monomer Z value

Maxwell et al., Sundberg et al.,

at 50 °C [1] at 25 °C [23]

2-EHA – 1
Styrene (St) 2–3 2
Butyl methacrylate (BMA) 3 2
Butyl acrylate (BA) 2–3 2
Butadiene (Bu) 3 2
Ethyl acrylate (EA) – 4
Methyl methacrylate (MMA) 4–5 4
Vinyl acetate (VAc) – 5
Methyl acrylate (MA) – 8
Acrylonitrile (AN) – >10 (estimate: 12)
Emulsion Polymerization: Kinetic and Mechanistic Aspects 13

initiated by KPS. The initial seed latex used was synthesized so as to have
radical traps (TEMPO) covalently bound onto the particle surface. When an
aqueous phase-propagating radical entered a seed particle, the nitroxide
moiety led to the formation of a stable alkoxyamine. Therefore, the kinetics of
radical entry into the seed particles was followed by monitoring the decay of
the ESR signal from the nitroxide in the samples withdrawn from the reactor.
They obtained fentry=0.36 for KPS at 70 °C and fentry=0.33 for V-50 at 70 °C,
respectively. Maxwell et al. [11] obtained the value of z@2 by comparing the
model predictions with the experimental results in the emulsion polymeriza-
tion of St. Schoonbrood et al. [27] reported z@18 for a 80:20/St:MA emulsion
copolymerization system. However, there is an example where it is difficult to
explain the kinetic behavior of the emulsion copolymerization of St and AAm
without assuming z=(one St monomer unit), as shown later in Section 3.2.3.

3.1.3
Miscellaneous Kinetic Problems in Radical Entry

Two major entry models – the diffusion-controlled and propagation-controlled

models – are widely used at present. However, Liotta et al. [28] claim that the
collision entry is more probable. They developed a dynamic competitive
growth model to understand the particle growth process and used it to simu-
late the growth of two monodisperse polystyrene populations (bidisperse sys-
tem) at 50 °C.Validation of the model with on-line density and on-line particle
diameter measurements demonstrated that radical entry into polymer particles
is more likely to occur by a collision mechanism than by either a propagation or
diffusion mechanism.
One of the most basic and important problems to be clarified in emulsion
polymerization is what kind of radicals can enter the polymer particles.A pro-
posal now widely accepted is that in the emulsion polymerization of St initiated
by, for example, potassium persulfate (KPS), an entering radical must be sur-
face active, so it must be a charged oligomer such as ·MzSO4– [11]. Tauer et al.
[29] carried out MALDI-TOF-MS investigations of the polymer inside the
particles at the end of a persulfate-initiated emulsifier-free emulsion polymer-
ization of St, and showed that besides sulfate end groups, a variety of different
end groups for the polymer molecules exist almost independently of the buffer
concentration employed during the polymerization. These results support the
existence of the following end group combinations: H–H, H–OH, K+–OO–OH,
K+ –OO–K+–OO, OH–OH, K+–OSO3–H, K+–OSO3–OH, K+–OSO3–K+–OSO3. They
concluded from these experimental results that the entering radicals can be
either corresponding primary radicals (H, OH, or O3SO radicals) or oligomer
radicals, and therefore surface activity is not a prerequisite condition for the en-
tering radicals.
Another important problem that has been debated for a long time is whether
or not the electric charges and the emulsifier layers on the surfaces of the poly-
mer particles affect the radical entry rate of a charged radical (Ç). It is now con-
14 M. Nomura et al.

firmed that for conventional emulsifiers such as sodium lauryl sulfate (NaLS,
electrostatic stabilizer), the extent of the surface coverage on a particle and the
ionic strength have no effect on radical entry, within experimental error mar-
gins [10, 30]. Adams et al. [30] found that the radical entry rate is independent
not only of the surface coverage and the ionic strength, but also of differences
in the kind of charges between the particle surface and the entering free radi-
cal. Furthermore, they refuted, on the basis of experimental data, the proposals
that the rate of radical entry is controlled by surfactant displacement [7] and
by collision of free radicals with the polymer particles [31]. Penbos et al. [10] car-
ried out the seeded emulsion polymerization of St using a PSt seed latex with
negative surface charge in combination with three different initiators: persul-
fate anions (negatively charged radicals), hydrogen peroxide/iron (II) (neutral
radicals) and 2,2¢-azobis(2-amidinopropane) (V-50, cationic free radicals).
They found that the rate of radical entry showed no significant dependence on
the kind of charges of these initiating radicals. El-Aasser et al. [25] examined the
effect of an adsorbed layer of the nonionic emulsifier (steric stabilizer) TritonX-
405 (octylphenoxypolyethoxy ethanol with an average number of 40 ethylene
oxide units) on the rate of radical entry by changing the ratio of anionic (NaLS)
and nonionic emulsifiers in the seeded emulsion polymerization of St initiated
by KPS. They concluded that any variation in surface coverage by the nonionic
emulsifier did not significantly affect the value of Ç within experimental error,
suggesting that the adsorbed layer of the nonionic emulsifier on the particle
surface would not act as a barrier to radical entry. Contrary to the observations
of El-Aasser et al. [25], Kuster et al. [32] observed a large decrease in the des-
orption rate coefficient for monomeric radicals in the emulsion polymerization
of St conducted using poly(ethylene oxide) nonylphenol surfactant with an
average number of 30 ethylene oxide units as the steric stabilizer. They con-
cluded from this observation that a “hairy” layer near the particle surface would
also act as a barrier for the entry of such uncharged radicals because re-entry
of desorbed radicals is the reverse process of desorption. Wang et al. [33] con-
ducted the seeded emulsion polymerization of St using the reactive surfactant
sodium dodecyl allyl sulfosuccinate (TREM LF-40) and its polymeric counter-
part, poy(TREM), and only observed an increase in the PSt molecular weight
when poly(TREM) was used. This was considered to be a consequence of the de-
crease in the rate of bimolecular radical termination resulting from the decrease
in the rate of radical entry into the polymer particles due to the diffusion bar-
rier of the hairy layer near the particle surface to the diffusing radicals.
On the other hand, several reports have been published that point out that
when a polymeric surfactant acting as an electrosteric stabilizer is used, the rate
of radical entry into a polymer particle should decrease due to a diffusion
barrier of the hairy layer built up by the polymeric surfactant adsorbed on the
surface of the polymer particles [34–36]. Coen et al. [34] found that in the
seeded emulsion polymerization of St using a PSt seed latex stabilized elec-
trosterically by a copolymer of acrylic acid (AA) and St, the electrosteric stabi-
lizer greatly reduced the radical entry rate Ç compared to the same seed latex
Emulsion Polymerization: Kinetic and Mechanistic Aspects 15

with a conventional electrostatic stabilizer. The explanation was that the “hairy”
layer on the surface of the polymer particle acted as a diffusion barrier to
the entering radicals. Kim et al. [35] also observed the decrease in the value of
Ç in the seeded emulsifier-free emulsion polymerization of MMA conducted
using KPS as the initiator and a PSt seed latex electrosterically coated by a
copolymer of St and styrenesulfonate (NaSS), which constitutes a thick hairly
layer on the particle surface.Vorwerg et al. [36] observed a decrease in the value
of Ç, although it was dependent on the pH, in the seeded emulsion polymer-
ization of St using seed latexes electrosterically stabilized with poly(acrylic
acid). Leeman et al. [37], on the other hand, synthesized amphiphilic water-
soluble polyelectrolytes, a poly(methyl methacrylate-block-sulfonated glycidyl
methacrylate) block copolymer (PMMA-b-SPGMA; polyanionic block copoly-
mer) and a poly(methyl methacrylate-block-quaternized N,N¢-dimethylamino-
ethyl methacrylate) block copolymer (PMMA-b-QPDMAEMA; polycationic
block copolymer). They carried out the emulsion polymerization of MMA at
60 °C in the presence of these two copolymers as surfactants and with the
following four types of initiators: KPS producing anionic radicals, H2O2 and
AIBN both producing neutral radicals and AAP (2,2¢-azobis(2-amidinopropane))
which generates cationic radicals in acid medium and neutral radicals in basic
medium. There was no major difference in the rate of polymerization Rp when
the initiators producing neutral radicals were combined with either of these
two oppositely charged copolymer surfactants. But a large decrease in Rp was
observed when the charge of the entering radical was different to the charge of
the block polyelectrolyte surfactant. Considering the experimental results
of El-Aasser et al. [25], the former finding is acceptable, but the latter one is
contrary to our expectations and the conclusions given in the articles [34–36].
Therefore, clarifying whether or not the adsorbed layer of polymeric surfac-
tants on the particle surface act as a diffusion barrier to the entering radicals
is still an important problem that needs to be solved in the future.
In the emulsion copolymerization of the water-soluble monomer acryl-
amide (AAm) and the sparingly water-soluble monomer St using KPS as ini-
tiator, Kawaguchi et al. [38] observed odd kinetic behavior. Regardless of the
fraction of AAm in the monomer feed, the polymerization of both monomers
started from the beginning of the reaction, but soon the AAm polymerization
slowed down and finally stopped, while the polymerization of St continued very
smoothly to the end. Nomura et al. [39] examined the reason for this abnormal
kinetic behavior and ascribed the reason to unusual radical entry behavior.
They studied the seeded emulsion copolymerization of St and AAm at 50 °C
using PSt particles as the seed and KPS as the initiator, and found that the change
in the number of PSt seed particles Ns caused a drastic change in the kinetic
behavior of this emulsion copolymerization system.When the number of seed
particles was less than a certain critical value Nc (~2.5¥1012 particles/cm3-
water), both St and AAm started polymerizing as soon as the initiator was added.
However, when the number of seed particles was higher than Nc, an apparent
induction period suddenly appeared for AAm polymerization; in other words,
16 M. Nomura et al.

although more than 99% of the initially charged AAm existed in the aqueous
phase, the AAm polymerization did not start until the St conversion exceeded
around 75%. Therefore, the apparent induction period was the time necessary
for the St conversion to reach 75%. On the other hand, the St polymerization
started and continued very smoothly from the beginning to the end of the
reaction, independently of the AAm polymerization.We have not yet succeeded
in explaining this interesting phenomenon quantitatively, but the reason may
be explained qualitatively as follows [40]. Using values for the monomer reac-
tivity ratio and the concentrations of St and AAm in the aqueous phase, it is
clear that the addition rate of St radicals to AAm is about 100 times faster than
to St and the addition rate of AAm radicals to AAm is about 5 times faster than
to St in the aqueous phase. Therefore, once an AAm radical is formed in the
aqueous phase, this AAm radical preferentially adds AAm monomer in the
aqueous phase without entering the polymer particles due to its hydrophilicity.
When the number of initially charged PSt seed particles is higher than Nc, no
AAm polymerization occurs in the aqueous phase although all free radicals are
produced in the aqueous phase and almost all AAm monomer units exist in the
aqueous phase. This implies that KPS radicals preferentially add St monomer
units as soon as they are generated in the aqueous phase and enter the seed
particles before the resultant St radicals add one AAm monomer unit. This is
because the mean residence time of the St radicals in the aqueous phase before
they enter the seed particles is shorter than the average time necessary for these
radicals to add one AAm monomer unit. On the other hand, when the number
of seed particles is less than Nc, the mean residence time of most of the St rad-
icals would become longer than the average time necessary for a St radical to
add one AAm monomer unit, and so these St radicals are mostly transformed
into AAm radicals, although some of these radicals can enter the seed particles
and continue St polymerization inside the particles. Therefore, the St radicals
produced continue AAm polymerization in the aqueous phase. The mechanism
of radical entry proposed in this copolymerization system contradicts the con-
clusions of Maxwell et al. [11], who claimed that KPS radicals need to add at
least 2–3 St monomer units before entering the monomer-swollen polymer
particles.

3.2
Radical Desorption

3.2.1
Desorption in Homopolymer Systems

The desorption (exit) of free radicals from polymer particles into the aqueous
phase is an important kinetic process in emulsion polymerization. Smith and
Ewart [4] included the desorption rate terms into the balance equation for Nn
particles, defining the rate of radical desorption from the polymer particles
containing n free radicals in Eq. 3 as kf nNn. However, they did not give any
Emulsion Polymerization: Kinetic and Mechanistic Aspects 17

detailed discussion on radical desorption. Ugelstad et al. [41] pointed out that
the rate coefficient for radical desorption (the desorption rate coefficient) could
be a function of particle size, the rate of chain transfer to monomer, the rate
of polymerization and the diffusion coefficients involved in the transport
processes leading to desorption of radicals, and suggested that, in the emulsion
polymerization of VCl, the desorption rate coefficient, kf might be expressed in
the form
kf = kvp–2/3 = k¢dp–2 (16)
On the other hand, Nomura and Harada [14, 42–45] pointed out that radical
desorption from the polymer particles and micelles plays a decisive role in par-
ticle formation and growth, and further that there are many examples of kinetic
deviations from the S-E theory that are attributable to radical desorption. First,
they theoretically derived the desorption rate coefficient from both stochastic
[42] and deterministic approaches, [14, 42, 43] based on a scheme consisting of
the following three consecutive steps: (1) chain-transfer of a polymeric radical
to a monomer molecule or a species like CTA in a polymer particle, followed
by (2) diffusional transportation of the resulting low molecular weight radical
to the particle-water interface, and (3) successive diffusion into the bulk water
phase through a stagnant film adjacent to the surface of the particle. In mod-
eling the rate coefficient for radical desorption, the following assumptions were
made:
A1. Polymer particles contain at most one radical (zero-one system).
A2. An oligomeric radical with no more than s monomer units can desorb
from and re-enter into a polymer particle with the same rate, irrespective
of its chain length.
A3. Instantaneous termination takes place when another radical enters a
particle already containing a radical.
A4. No distinction is made between radicals with or without an initiator
fragment on its end.
A5. Water-phase reactions such as propagation, termination, and chain-trans-
fer to a monomer are negligible for the desorbed radicals. This means that
all of the desorbed radicals would re-enter particles and the loss of these
radicals occurs only through the event given by the assumption A3).
A6. The physical and chemical properties of chain transfer agent (CTA) rad-
icals are approximately equal to those of monomer radicals.
Based on these assumptions, they derived the desorption rate coefficient kf as

Çw (1 – – Çw (1 – n– )ki
   
n) kmf kTf [T]p
kf = KoI 9948 + K + +
KoI n– + ki[M]p n– Np kp kp [M]p KoI n– + ki[M]p kp n– Np
o 6 661 99401

 
S j
kp [M]p
· ∑ 9941– (17)
j=1 Kon + kp[M]p
18 M. Nomura et al.

When it is assumed that initiator-derived radicals do not exit and only mono-
meric and CTA radicals produced by chain transfer to a monomer and/or a
CTA can desorb (s=1), Eq. 17 can be simplified as

  
kp [M]p kmf kTf [T]p
kf = Ko 986
– 6 + 661 (18)
Kon + kp[M]p kp kp [M]p

where kmf and kTf are the chain transfer rate constants to monomer and to CTA,
respectively, and Ko is the overall desorption rate constant per particle for
monomeric (or CTA) radicals, which is approximately given by [42–44]

    
2pDw dp yDw –1 12Dw d
Ko = 95 1 + 91 = 93 (19)
mdvp mdDp mddp2

where d=(1+yDw/mdDp)–1 denotes the ratio of the aqueous-side to overall

diffusion resistance, y is a numerical constant between 1 and 6 that depends
on the mass-transfer coefficient employed (y=1 if Eq. 7 is assumed to be ap-
plicable to the mass-transfer process inside the polymer particles) [42–45], md
is the partition coefficient for monomer radicals between the polymer parti-
cle and aqueous phases, defined by md=[M]p/[M]w, and the term, yDw/mdDp is
the ratio of the particle-side to water-side diffusion resistance. Ugelstad et al. [9]
obtained the same expression as Eq. 19 with y=1. On the other hand, Casey
et al. [1, 46] also derived the same expression as Eq. 19 with d=1. However,
it must be noted that when the diffusion resistance inside the particle is far
greater than that in the aqueous-side effective diffusion film, that is, Dw/mdDpp1,
one gets

   = 9
d 
–1
2pDw dp yDw 12Dp
Ko = 95 1 + 91 2
(19a)
mdvp mdDp p

For the St emulsion polymerization, for example, in the absence of CTA, it

usually holds that Ko n– Ok [M] , at least in the range where monomer droplets
p p
exist. Then, Eq. 18 is reduced to

      
kmf 12Dwd kmf 12DwCmd –2
kf = Ko 6 = 63111
2 6 = 6311131 d p (20)
kp mddp kp md

where Cm is the monomer chain transfer constant. The validity of Eq. 20 has
been demonstrated by Nomura et al. [20, 42, 44, 47] and Adams et al. [48].
Equation 18 was derived under the assumption that the physical and chemical
properties of a CTA radical are approximately equal to those of a monomeric
radical. However, if it is necessary to take into account the differences in the
physical and chemical properties between monomeric and CTA radicals, Eq. 18
can be modified approximately as
Emulsion Polymerization: Kinetic and Mechanistic Aspects 19

   
kmf [M]p kTf [T]p
kf = Kom 9862– + KoT 666152
– (18a)
Kom n + kp[M]p KoT n + kiT[M]p

where Kom and KoT are the overall desorption rate constants per particle for
monomeric and CTA radicals, respectively.
Ugelstad et al. [9] later derived almost the same desorption rate coefficient
as Eq. 19 given by

    = 61212
m d  k¢ 
–1
12Dw kmf Dw 12Dwd kmf
kf = 64 2 61 1 + 6121 2
(21)
mddp k¢p mdDp
6
d p p

where k¢p is the rate constant for the reaction between a monomeric radical
formed by chain transfer and a monomer, the value of which may be different
from the value of kp, the propagation rate constant.
Asua et al. [49–51] modified Eq. 18 in the absence of a CTA to include more
general cases, taking into account the fate of the desorbed radicals (both chem-
ical reactions in the water phase and re-entry) as

 
Ko
kf = kmf [M]p 6121941 (22)
Kob + kp[M]p

where b stands for the fraction of desorbed radicals that cannot re-enter
because of the aqueous phase termination or propagation, and is given by

kp[M]w + ktw[T*]w
b = 999492 (23)
kp[M]w + ktw[T*]w + kaNT

where [T*]w is the total radical concentration in the water phase and ka is the
mass-transfer coefficient for radical entry, and [M]w is the monomer concen-
tration in the water phase.
On the other hand, Casey and Morrison et al. [52, 96] derived the desorption
rate coefficient for several limiting cases in combination with their radical
entry model, which assumes that the aqueous phase propagation is the rate-
controlling step for entry of initiator-derived free radicals. Kim et al. [53] also
discussed the desorption and re-entry processes after Asua et al. [49] and
Maxwell et al. [11] and proposed some modifications. Fang et al. [54] discussed
the behavior of free-radical transfer between the aqueous and particle phases
(entry and desorption) in the seeded emulsion polymerization of St using KPS
as initiator.

3.2.2
Desorption in Copolymer Systems

As we discuss later in Section 3.3.3, Nomura et al. [45, 47] first derived the rate
coefficient for radical desorption in an emulsion copolymerization system by
20 M. Nomura et al.

extending the approach developed for emulsion homopolymerization under

the same assumptions as A1–A6 given in Section 3.2.1. This methodology
is now termed the “pseudo-homopolymerization approach”. According to this
approach, the average rate coefficient for radical desorption, defined, for ex-
ample, in a binary emulsion copolymerization system with monomers A and
–
B, kf , is given by
– – /n
– – –
kf = kfA(n A t) + kfB(nB/nt) = kfA fA + kfB fB (24)

where kfA is the desorption rate coefficient for A-monomeric radicals, – nt is

the average number of total radicals per particle, – nA the average number of
A-radicals per particle (n–t = n–A + n–B), and fA is the fraction of A-radicals in the
particle phase and is expressed, at steady-state, as a function of the propaga-
tion rate constant kp, the monomer reactivity ratio r¢, and the monomer con-
centration in the polymer particles [M]p, in the following form.
n–t = n–A + n–B (25)

In the case where all the desorbed A-monomeric radicals reenter the polymer
particles, the desorption rate coefficient for A-monomeric radicals kfA is given
by

 
CmAArA[MA]p + CmBA[MB]p
kfA = KoA 99998641 (26)
rB {(KoA –
nt/kpAA) + [MB]p} + [MA]p

where KoA is the overall desorption rate constant per particle for A-monomeric
radicals given by Eq. 19 in Section 3.2.1 and CmBA is the chain transfer constant
for a B-radical to A-monomer.
López et al. [55] investigated the kinetics of the seeded emulsion copoly-
merization of St and BA in experiments where the diameter and number of seed
particles, and the concentration of initiator were widely varied. The experi-
mental data were fitted with a mathematical model in which they used the
desorption rate coefficient developed by Forcada et al. [56] for a copolymer-
ization system. The desorption rate coefficient for the A-monomeric radical
that they used was a modification of Eq. 22 and Eq. 23, and is given by

 
K0A
kfA fA = (kmf,AA fA + kmf,BA fB)[MA]p 9999862 (27)
bAK0A + k¢pAA[MA]p + k¢pAB[MB]p

where kmf,AB denotes the chain transfer constant of the A-radical to the B-mono-
mer and bA is the fraction of the desorbed A-monomeric radicals that cannot
reenter the polymer particles because of the aqueous phase termination or
propagation. Barudio et al. [57], on the other hand, developed a simulation model
for emulsion copolymerization based on the pseudo-homopolymerization
approach, where they used the average rate coefficient for radical desorption
– – –
given by kde = (12Dw z/mddp2)(kmf /kp). Saldivar et al. have presented a survey of
Emulsion Polymerization: Kinetic and Mechanistic Aspects 21

emulsion copolymerization models that have been published in the literature,

and a comprehensive mathematical model for emulsion copolymerization [58],
along with its experimental verification [59, 60]. In Appendix A of the former
paper, they present a detailed discussion on the average rate coefficient for
radical desorption, which is applicable to a multimonomer system.
Only a few experimental studies have been published that aim to demonstrate
the validity of the average rate coefficient for radical desorption given by Eqs. 24
to 27 [45, 47] directly. Vega et al. [61] modeled the batch emulsion copolymer-
ization of AN and Bu in order to simulate an industrial process and improve
the final polymer quality. The mathematical model they used was an extended
version of that developed by Guliotta et al. [62] for the continuous emulsion
polymerization of St and Bu. Due to the high solubility of AN in water, the
effect of the desorption of AN radicals was taken into consideration in the
model. The average rate coefficient for radical desorption used was given by
Eqs. 24 and 27. Barandiaran et al. [63] proposed a method to estimate the rate
coefficient for radical desorption in emulsion copolymerization and gave
the values of this parameter for the MA-VAc and MMA-BA emulsion copoly-
merization systems.

3.2.3
Miscellaneous Kinetics Problems in Radical Desorption

The rate coefficient for radical desorption was derived by assuming that the
adsorbed layer of conventional or polymeric surfactant on the surface of the
polymer particle does not act as an interfacial diffusion barrier to the desorb-
ing neutral monomeric radicals. However, Kusters et al. [32] studied the kinet-
ics of particle growth in emulsion polymerization systems with a surface-active
initiator (an “inisurf ”). The inisurf employed was the diester of 4,4¢-azobis(4-
cyanopentanoic acid), the initiator moiety, with poly(ethylene oxide) nonyl-
phenol, the surfactant moiety. They observed a large decrease (one order of
magnitude) in the desorption rate coefficient for monomeric radicals in the
emulsion polymerization of St. They ascribed the reason for the decrease in the
rate of radical desorption to a “hairy” layer of the polymeric surfactant, which
would play the role of a diffusion barrier. Coen et al. [34] also reported that
in the seeded emulsion polymerization of St using a PSt seed latex stabilized
electrosterically by a copolymer of AA and St, the electrosteric stabilizer greatly
reduced the rate of radical desorption compared to the same seed latex with an
electrostatic stabilizer. They interpreted the reason for the decrease in the rate
of radical desorption by assuming that the aqueous-phase diffusion of mono-
meric radicals is slower in the hairy layer. Recently, Vorwerg et al. [36] carried
out a kinetic study of the seeded emulsion polymerization of St using PSt seed
lattices electrosterically stabilized with poly(acrylic acid) (pAA). They found
that seed lattices with a high-coverage of pAA (above 50 mC cm–2) exhibited a
significant reduction in radical desorption (by a factor of ~3 compared to the
ionically stabilized seed) at low pH.
22 M. Nomura et al.

3.3
Particle Formation and Growth

3.3.1
Particle Formation

As we mentioned in Section 2.1 (Fig. 1a), there are three major models for par-
ticle formation in emulsion polymerization. According to these models, poly-
mer particles are formed:
1. When a free radical in the aqueous phase enters a monomer-swollen emul-
sifier micelle and polymerization proceeds therein (micellar nucleation).
2. When the chain length of a free radical growing in the aqueous phase
exceeds its solubility limit and precipitates to form a particle nucleus
(homogeneous nucleation).
3. When a free radical growing in the aqueous phase enters a monomer droplet
and polymerization proceeds therein (droplet nucleation).
However, when the resultant polymer particles become unstable and coagulate,
then whatever the mechanism of particle formation is, the final number of
polymer particles produced is determined by a limited coagulation between
existing polymer particles (coagulative nucleation).
Smith and Ewart [4] derived an expression that can predict the number of
polymer particles produced, by assuming that:
1. A monomer-swollen emulsifier micelle is transformed into a polymer par-
ticle by capturing a free radical from the aqueous phase [4, 5].
2. The volumetric growth rate per particle m is constant, at least during parti-
cle formation (m=dnp/dt=constant).
3. Free radical activity does not transfer out of a growing particle (kf@0).
4. The amount of emulsifier that dissolves in the water phase without forming
micelles and adsorbs on the surface of emulsified monomer droplets may be
neglected.
Based on these assumptions, two limiting cases were discussed.
Case A: The rate of radical entry into micelles that results in the formation
of new particles is approximately equal to the rate of radical generation in the
water phase (Çw), as long as emulsifier micelles are present; in other words,
dNT
7 = Çw (28)
dt
Particle formation stops at the time tc, when the emulsifier micelles have just
disappeared because all of the emulsifier molecules comprising the emulsifier
micelles have been transferred to the surfaces of growing polymer particles
for adsorption. The volume vp,c at time tc of a particle formed at time t is
vp,c=m(tc–t), and so the surface area ap,c of this particle at time tc is given by
Emulsion Polymerization: Kinetic and Mechanistic Aspects 23

ap,c=s(tc–t)2/3 where s =[(4p)1/23m]2/3. Therefore, the total surface area Ap,c

of all the polymer particles present at time tc is given by the integral Ap,c=
tc
∫ s (tc – t)2/3 Çwdt = 3/5 sÇw tc5/3 . No micelles exist (Am=0) at time tc, and so all
0
of the charged emulsifier molecules are adsorbed onto the surfaces of polymer
particles present. Therefore, it holds that Ap,c=(3/5)sÇw tc5/3=asS0. In this case, the
number of polymer particles produced (NT) can be obtained by substituting tc
into NT=Çwtc as
NT = 0.53(Çw /m)0.4 (asS0)0.6 (29)
where Am and Ap are the total surface area of the micelles and the total surface
area of the polymer particles per unit volume of water, respectively, as is the
surface area occupied by a unit amount of emulsifier, and S0 is the amount of
initially charged emulsifier per unit volume of water (the initial emulsifier con-
centration).
Case B: Radicals enter both micelles and polymer particles at rates that are
proportional to their surface areas (collision theory), so that the rate of new
particle formation is given by

 
Am Çw
dNT/dt = Çw 77 = 08 (30)
Am + Ap 1 + Ap /Am
Then, it follows that
NT = 0.37(Çw /m)0.4 (asS0)0.6 (31)
On the other hand, Nomura et al. [14] proposed a different approach for pre-
dicting the number of polymer particles produced, where the new concept of
“radical capture efficiency” of a micelle relative to a polymer particle was pro-
posed. The assumptions employed were almost the same as those of Smith and
Ewart, except that the volumetric growth rate m of a polymer particle was not
considered to be constant. It was also assumed that all of the radicals formed
in the aqueous phase enter either micelles or polymer particles with negligible
termination in the aqueous phase. In this approach, the following elementary
reactions and their respective rates were defined.
(1) Particle formation by radical entry into a micelle
R* + ms Æ N*, kemms [R*w] (32)
(2) Formation of a dead particle by radical entry into an active particle con-
taining a radical
R* + N* Æ N0, kep N*[R*w] (33)
(3) Formation of an active particle by radical entry into a dead particle con-
taining no radical
R* + N0 Æ N*, kep N0[R*w] (34)
24 M. Nomura et al.

where ms is the number of monomer-swollen micelles, [R*w] the concentration

of free radicals in the aqueous phase, N* the number of active particles
containing a radical, N0 the number of dead particles containing no radical, kem
the rate constant for radical entry into micelles, and kep the rate constant for
radical entry into particles. Using these rate expressions, the following equa-
tions, describing the balance of radicals in the aqueous phase and the rate of
particle formation, were obtained:

d[R*w]
91 = Çw – kemms [R*w] – kep NT [R*w] (35)
dt

dNT
71 = kemms [R*w] (36)
dt

where NT is the total number of polymer particles produced (NT=N*+N0).

Introducing the aqueous phase concentration [R*w], obtained by applying the
steady state assumption to Eq. 35 into Eq. 36, and rearranging leads to

dNT Çw Çw
71 = kemms [R*w] = 7788 = 77241 (37)
dt 1 + (kep NT /kemms ) 1 + (eNT /Sm)

where kepNT/kemms denotes the ratio of the rate of radical entry into polymer
particles to that into micelles and is rewritten as eNT/Sm, where e=(kep/kem)Mm
and e is the one unknown parameter, which affects the number of polymer par-
ticles produced. Here, Sm is the total number of emulsifier molecules forming
micelles, and Mm is the aggregation number of emulsifier molecules per micelle,
defined by Mm=Sm/ms. By solving a set of simultaneous differential equations
describing NT, N*, the monomer conversion XM, and using the balance equation
for the number of emulsifier micelles ms , the number of polymer particles pro-
duced NT can be predicted with respect to the initial emulsifier (S0) and initia-
tor concentrations (I0) (or Çw=2kd f [I0]) as shown by NTµÇ0.3 0.7
w S 0 . In the case
of VAc emulsion polymerization [20], the authors took into account radical
desorption from the polymer particles, yielding the following expression in
place of Eq. 37.

dNT Çw + kde –
nNT Çw + kde –
nNT
= k m
em s [R*
w ] = = (38)
dt 1 + (kepNT/kemms ) 1 + (e NT/Sm)
52 7788 772411

In this derivation, it was assumed that only monomer-transferred radicals could

desorb from the polymer particles, and that radical termination in the aqueous
phase was negligible. The calculated result was correlated by NTµÇw0.04S 00.94, from
which it was found that radical desorption from the polymer particles de-
creases the value of the exponent of the initiator dependence and increases
the value of the exponent of the emulsifier dependence. To obtain agreement
between the predicted and experimental values of NT, however, it was necessary
Emulsion Polymerization: Kinetic and Mechanistic Aspects 25

to introduce a value of e that is far greater than that predicted by using the
diffusion theory given by Eq. 7 (kep/kem=dp/dm). A value of e=1.28¥105 was
necessary for St emulsion polymerization [14], and a value of e=1.2¥107 for
VAc emulsion polymerization [20], while the value of e predicted by diffusion
theory (Eq. 7) was ~1000 in both systems because Mm≈100 and dp/dm≈10 hold
(roughly) in Interval I of particle formation, as already discussed in Sect. 3.1.1.
The authors, therefore, proposed the concept of the “radical capture efficiency”
of a micelle relative to a polymer particle to correct for this disagreement. The
same phenomenon has been encountered by several researchers [18, 19, 21]
(Sect. 3.1.1). However, the reason for the disagreement between the predicted
and experimental values of e has not been found yet.
As we discussed in Sect. 3.1.1, Hansen et al. [15] made significant improve-
ments to the concept of the radical capture efficiency proposed by Nomura
et al. [14]. Taking this concept into consideration, they examined the effect of
radical desorption on micellar particle formation in emulsion polymerization
[65]. Assuming that radical entry is proportional to the xth power of the micelle
radius and the polymer particle radius, they proposed the following general
expression for the rate of particle formation:
Çw + kf –
 
x
dNT – dcemsd m nNT
52 = (Çw + kf nNT) 5886 = (39)
  
dt x x
dcemsd m + NTd p
59941
NT dp
1 + 53 5
dcems dm
where dce is the radical capture efficiency of a micelle relative to a polymer par-
ticle, and is related to e by dce=(Mm/e)(dp/dm)x. The condition x=1 corresponds
to the diffusional entry model, while the condition x=2 corresponds to the col-
lisional entry model. Using x=1, they calculated the effect of radical desorption
in the emulsion polymerizations of St, MMA, VAC, and VCl on the number
of polymer particles produced, and demonstrated that the following general
rule for initiator and emulsifier exponents, which was first found by Nomura
et al. [20, 43], could also be applied to the emulsion polymerizations of these
monomers.
NT µ I01–zS 0z (40)
where 0.6<z<1.0. The value of z increases from 0.6 (a common value for St) to
1.0 (a common value for VAc) with increasing radical desorption.
Particle formation below the critical micellar concentration (CMC) in
emulsion polymerization is now accepted to take place according to the ho-
mogeneous nucleation mechanism. Among several quantitative treatments of
homogeneous particle formation in emulsion polymerization, the best-known
model was that proposed by Fitch and co-workers [66]. Their model is based
on the assumption that when the chain length of a free radical growing in the
aqueous phase reaches its solubility limit (critical chain length), it precipitates
to form a primary particle, and that particle formation will be hindered if these
growing oligomers are absorbed in polymer particles formed earlier. Hansen
26 M. Nomura et al.

et al. [67] made significant improvements on the Fitch model [the HUFT
(Hansen-Ugelstad-Fitch-Tsai) model]. According to Hansen et al, the rate of
particle formation is given by
dNT
52 = kpw Mw (RIjIcr + RMjMcr) (41)
dt
where kpw and Mw are the propagation rate constant and the monomer con-
centration in the aqueous phase, respectively, and RIjcr and RMjcr are the aque-
ous phase concentrations of oligomer radicals with critical chain length derived
from initiator and monomer radicals, respectively. This equation means that
oligomers stemming from initiator and monomer radicals precipitate as par-
ticles when they propagate beyond their respective critical degree of polymer-
ization, jIcr and jMcr. The authors derived the steady-state expressions for RIj and
RMj and obtained a general equation for homogeneous particle formation by
nserting them into Eq. 41. Furthermore, in order to simplify it, they neglected
particle formation from oligomers stemming from the desorbed monomeric
radicals, along with several other assumptions, and obtained
dNT – –j
(42)
52 = Çw (1 + ktwRw /kpw Mw + kc NT /kpw Mw) Icr
dt
–
where kc is the average rate coefficient for radical entry into polymer particles,
and Rw is the total radical concentration in the aqueous phase. Assuming that,
as an approximation, radical absorption by particles may be neglected in the
calculation of Rw, one gets

NT (t) = {[k1Çw jcr t + (k2 + 1) jcr]1/jcr – k2 – 1}/k1 (43)

–
where Rw=(Çw/ktw)1/2 in this case, and k1 = kc/kpwMw and k2=(ktwÇw)1/2/kpwMw.
On the other hand, Tauer et al. [68] developed a framework for modeling
particle formation in emulsion polymerization on the basis of a combination
of classical nucleation theory with radical polymerization kinetics and the
Flory-Huggins theory of polymer solutions. The basic assumption adopted was
that water-borne oligomers form stable nuclei under critical conditions. The
only adjustable model parameter was the activation energy of nucleation. The
model allows us to calculate the chain length of the nucleating oligomers, the
number of chains forming one nucleus, the diameter of the nucleus, the total
number of nuclei formed, and the rate of nucleation. Further, they experimen-
tally studied particle formation in the very early stages of the emulsifier-free
emulsion polymerization of St by monitoring the optical transmission and the
conductivity of the reaction mixture on-line [69].
Usually particle formation by initiation in the monomer droplets (droplet nu-
cleation) is not considered important in conventional emulsion polymerization.
This is because of the low absorption rate of radicals into the monomer droplets,
relative to the other particle formation rates. However, when the monomer
Emulsion Polymerization: Kinetic and Mechanistic Aspects 27

droplets are very small, they become an important source of particle formation
because the monomer droplets can compete for aqueous phase free radicals with
emulsifier micelles. This mode of heterogeneous polymerization is now called
“miniemulsion polymerization” and is reviewed in another chapter.
It is accepted that particle formation below the CMC in emulsion polymer-
ization takes place by homogeneous nucleation. However, there have been claims
that homogeneous nucleation is the main particle formation mechanism, even
above the CMC. Lichti et al. [70] investigated the mechanism of particle forma-
tion in the emulsion polymerization of St using sodium dodecyl sulfate (SDS)
as an emulsifier, and proposed the concept of “coagulative nucleation”. They
measured the full PSDs by electron microscopy at consecutive times soon after
the cessation of particle formation, and found that the PSDs obtained during
particle formation (Interval I) were positively skewed, confirming the role of co-
agulation, even above the CMC. Based on this phenomenon, they concluded that
the particle formation process does not occur by either simple micellar entry
or homogeneous nucleation mechanisms. Therefore, they suggested a mecha-
nism for particle formation where the homogeneous nucleation of oligomers
in the aqueous phase creates small primary polymer particles, and these
primary particles coagulate to produce polymer particles. On the basis of this
experimental finding, Feeny et al. [71] proposed a detailed theory for coagula-
tive nucleation and the PSDs in emulsion polymerization. The theory com-
bined and extended Müller-Smoluchowski coagulation kinetics with the DLVO
theory. Expressions were provided for the time evolutions of the nucleation
rate, particle number, and PSD. They showed that with physically reasonable
values for the parameters of the coagulation kinetics, agreement was obtained
with experimental data for the St emulsion polymerization system. Richards et
al. [72] developed a mathematical model for emulsion copolymerization. The
model combined the theory of coagulative nucleation of homogeneously nu-
cleated precursors with detailed species material and energy balances to calcu-
late the time evolution of the concentration, size and colloidal characteristics of
polymer particles, the monomer conversion, the copolymer composition, and
the molecular weight in an emulsion copolymerization system.
Although it is now accepted that particle formation below the CMC in emul-
sion polymerization takes place according to the homogeneous nucleation
mechanism, there has been debate as to whether homogeneous nucleation is
still operative even above the CMC, especially when relatively water-soluble
monomers are polymerized in emulsion in the presence of emulsifier micelles.
To date, most investigators believe that in the emulsion polymerization of
partially water-soluble monomers such as MMA and methyl acrylate (MA),
polymer particles are generated not by a micellar mechanism, but by homo-
geneous nucleation even in the presence of emulsifier micelles. This is because
the emulsion polymerization involving these monomers does not follow the
Smith-Ewart theory, and moreover, because an inflection point cannot be seen
around the CMC of the emulsifier on the particle number versus initial emul-
sifier concentration curve, where an abrupt and sharp decrease in the number
28 M. Nomura et al.

of polymer particles produced is usually observed if particle formation occurs

by the micellar mechanism [73, 74]. Nomura et al. [75] carried out the emulsion
polymerization of vinylidene chloride (VDC) at 50 °C using NaLS as the emul-
0.7I 0.3M 0, where S
sifier and KPS as the initiator. They found that NTµSm 0 0 m
is the initial concentration of emulsifier forming micelles, I0 is the initial
initiator concentration and M0 is the amount of monomer initially charged
per unit volume of water. Although the solubility of VDC in water at 25 °C is
2.5¥10–2 mol/dm3-water (~0.25 wt%), which is about ten times more water-sol-
uble than St, but is about 1/10 of the water-solubility of MMA, an inflection
point was definitely observed around the CMC on the particle number versus
initial emulsifier concentration curve. This seems to indicate that micellar nu-
cleation occurred, although Gilbert et al. [70, 71, 76] refuted micellar nucleation
even in the emulsion polymerization of sparingly water-soluble monomer St in
the presence of SDS micelles. Sajjadi et al. [77] investigated the kinetic features
of the batch emulsion polymerization of BA using SDS as the emulsifier and
KPS as the initiator. They observed that the number of polymer particles pro-
duced was proportional to the 0.54th power of emulsifier concentration, to the
0.39th power of initiator concentration, and was practically independent of
the monomer/water ratio. Particle formation was found to occur even during
Interval III, when undissociated micelles existed.
Experimental investigations that deal in detail with particle formation in
emulsion copolymerization are scarce. Nomura et al. [78] studied the kinetics
of particle formation and growth in the emulsion copolymerization of VDC
and MMA using NaLS as the emulsifier and KPS as the initiator. The number
of polymer particles produced was determined using particle diameters mea-
sured by both electron microscopy (TEM) and dynamic light scattering (DLS)
for comparison. They found that NTµS 01.0I 00.3F0, where S0 and I0 are the initial
emulsifier and initiator concentrations, respectively, and F is the weight frac-
tion of MMA in the initial monomer feed. It was also found that the particle
number determined via the DLS particle diameter was always about 1/2~1/3
of that determined by TEM. This is due to the difference between the average
particle diameters determined from TEM and DLS.Yuan et al. [79] carried out
the emulsion terpolymerization of St, butadiene (Bu) and AA, and found that
the mechanism of water-soluble oligomer formation during the emulsion poly-
merization differed depending on whether the SDS emulsifier concentration
was above or below the CMC. This may demonstrate that particle formation is
also closely connected to the presence of emulsifier micelles.
Herrera-Ordóñez et al. [22, 80] discussed particle formation during the
emulsion polymerization of St above the CMC of the emulsifier used (SDS),
based on their detailed mathematical model for emulsion polymerization.
By comparing the model predictions with experimental data, they concluded
that micellar nucleation dominates over the homogeneous nucleation above
the CMC, and that coagulation is not significant, even if it does take place. Fur-
thermore, they [81] concluded that particle formation by micellar nucleation
is at all times at least ten orders of magnitude greater than that by homoge-
Emulsion Polymerization: Kinetic and Mechanistic Aspects 29

neous nucleation, even in the emulsion polymerization of relatively water-sol-

uble monomer MMA. Recently, however,Varela de la Rosa et al. [82–84] carried
out detailed experimental studies on the emulsion polymerization of St at
50 °C, using SDS as the emulsifier and KPS as the initiator, and proposed the
modified description shown by the following, which is very different from the
widely accepted classical description of St emulsion polymerization.
1. In Interval I, the rate of polymerization and the number of polymer parti-
cles produced increase. Particle formation takes place predominantly by
micellar nucleation. Micelles disappear between 5% and 10% conversion,
marking the end of this interval.
2. In Stage II (referred to as Stage II to differentiate it from the classical Inter-
val II), the rate of polymerization and the number of polymer particles
continue to increase but at a slower rate. Polymer particles are formed by
homogeneous nucleation as long as monomer droplets and enough emul-
sifier (>0.05 mM) are present in the system. The end of this stage is marked
by the disappearance of monomer droplets, but particle nucleation may or
may not end at this time.
3. The number of polymer particles produced is proportional to the 0.36th
power of the initial monomer concentration.
Therefore, the mechanism of particle formation is still anything but a settled
question, even in the emulsion polymerization of St.
Only a few papers [85–88] have been published so far that discuss method-
ologies that could be used to discriminate experimentally between micellar and
homogeneous nucleations. Nomura et al. [85, 86] proposed an experimental
method to gauge to what extent the homogeneous and micellar nucleations are
operative in a given emulsion polymerization system. The method involves the
emulsion copolymerization of the sparingly water-soluble monomer St with
partially water-soluble monomers such as MMA or MA, followed by measure-
ment of the composition of the copolymers produced at the very beginning of
the reaction, including the interval of particle formation. This method is based
on the fact that the composition of the copolymer produced at the very begin-
ning of the reaction reflects the comonomer composition at the locus of parti-
cle formation. In other words, the copolymer composition serves as a probe of
the locus of particle formation. They carried out the emulsion copolymeriza-
tions of St-MMA and St-MA, where the weight ratio of two monomers was 1:1.
The copolymer compositions observed at the very beginning of the reactions
(far less than 1% monomer conversion) in the presence of emulsifier (NaLS)
micelles were definitely very different from those observed when the emulsi-
fier micelles were absent, reflecting the comonomer composition in the locus
of particle nucleation. These experimental results revealed that the micellar
nucleation is the main particle formation mechanism, even in the emulsion
polymerization of moderately water-soluble MMA and MA in the presence of
emulsifier micelles. Recently, Chern et al. [87, 88] proposed a novel method in
which a water-insoluble dye was used as a probe to study the particle nucleation
30 M. Nomura et al.

mechanism. The rationale behind the method was that if particle formation
took place by homogeneous nucleation, the resultant polymer particles would
contain negligible amounts of dye because the transport of the dye species
from the monomer droplet phase to the resultant polymer particles could be
neglected due to the insolubility of the dye in the aqueous phase. If, on the
other hand, particle formation took place by micellar nucleation, the resultant
polymer particles would contain an amount of dye corresponding to that sol-
ubilized in the micelles. They carried out the semibatch emulsion polymeriza-
tion of St and of MMA in the presence of the dye. In the semibatch emulsion
polymerization of MMA, for example, the experimental results showed that
when the emulsifier (SDS) concentration is above its CMC, mixed mode parti-
cle nucleation (micellar and homogeneous nucleation) was the predominant
mechanism. However, a question raised for this study is that if the transport of
the dye species from the monomer droplets to the resultant polymer particles
can be neglected, how is the dye transported from the monomer droplet phase,
where the dye is dissolved, to the monomer-swollen micelle phase where the
dye is solubilized.
Semibatch seeded emulsion polymerizations are quite common in industrial
operations. One of the most important problems in semibatch seeded emulsion
polymerization is how to control secondary particle formation. It is well known
that the amount of emulsifier must be carefully fed during starved-fed semi-
batch seeded emulsion polymerization. Too little emulsifier leads to emulsion
instability and hence coagulation, while too much emulsifier leads to secondary
particle formation by the micellar mechanism. Wang et al. [89] developed a
method for controlling the emulsifier level in starved-fed emulsion polymer-
ization. Morrison et al. [90] studied the conditions for secondary particle
formation in emulsion polymerization systems where the amount of added
emulsifier was below the CMC. They advanced their discussion based on the
HUFT model (Eq. 41), incorporating their reaction-controlled entry model, and
deduced a simple means for determining conditions for the onset and extent
of secondary particle formation. Coen et al. [91] further extended the work
of Morrison et al. [90] and proposed an extensive model for the PSD, particle
number, particle size and amount of secondary particle nucleation in emulsion
polymerization. Prescott et al. [92] proposed a simplified model for particle
formation, which is particularly useful for exploring the conditions required
for the growth of large particles, while avoiding secondary particle formation.
Butucea et al. [93] studied the seeded emulsion polymerization of VCl to es-
tablish the conditions needed to avoid the formation of new polymer particles
(secondary nucleation), and proposed new parameters: (1) MSA, the minimum
surface area of seed particles necessary to capture all initiator-derived (ionic)
radicals generated in the aqueous phase at a given initiator concentration; (2)
MCCI, the maximum critical concentration of initiator per unit surface of seed
particles under which the formation of new polymer particles is avoided; (3)
PVR1, the polymer volume per active growing radical necessary for the radical
to be within the particle for one second.
Emulsion Polymerization: Kinetic and Mechanistic Aspects 31

It has been reported that both the surface charge density and the degree of
surface coverage by emulsifier on the seed particles affect the behavior of sec-
ondary particle nucleation in seeded emulsion polymerization because these
factors control the rate of radical entry into seed particles. Vorwerg et al. [36]
carried out a kinetic study of seeded emulsion polymerization using PSt seed
particles electrosterically stabilized with poly(acrylic acid), and studied the
effect of the degree of surface coverage by poly(acrylic acid) on both radical
entry (Ç) and desorption (kf), through which secondary particle nucleation is
influenced under the condition of a fixed number of seed particles. The be-
havior of Ç and kf for the low-coverage particles was the same as that of the PSt
seed stabilized by initiator fragments and adsorbed emulsifier. The high-cov-
erage particles, on the other hand, exhibited strongly reduced kf values (by a
factor of three) at low pH, but Ç was only slightly lower than for the ionically
stabilized seed, while at high and neutral pH, secondary particle nucleation and
a decreased polymerization rate was observed with increasing pH (despite
an increase in particle number), indicating a reduced Ç value. Therefore, an
extensive electrosteric stabilizer reduced the rate coefficient for radical entry
(and for radical desorption), inducing secondary particle nucleation. Cheong
et al. [35, 94] investigated the effects of surface charge density on the kinetics
of secondary particle formation. They carried out three emulsifier-free seeded
emulsion polymerizations of MMA using monodispersed seed particles with
different surface charge densities, prepared from the St and NaSS comonomers
using the two-stage shot-growth process. In the case of the highest surface
charge density (72.7 mC/cm2), secondary particle nucleation was observed.
They ascribed the reason for this to the reduced rate of radical entry into the
seed particles resulting from electrical repulsion between seed particles and
entering oligmeric radicals [35]. They [94] proposed a mathematical model
to explain the effects of seed particle surface charge density on secondary
particle nucleation by introducing a modified Smolchowski equation and the
DLVO theory.
Sajjadi [95] examined the conditions for secondary particle formation and
coagulation in the seeded semibatch emulsion polymerization of BA under
monomer-starved conditions. They arrived at very interesting conclusions: (1)
particle coagulation occurred if the particle surface coverage (qcr1) dropped
below qcr1=0.25±0.05; (2) secondary particle formation occurred above a
critical surface coverage of qcr2=0.55±0.05, indicating that the presence of mi-
celles in the reaction vessel is not the only prerequisite for micellar nucleation
to occur; (3) the number of polymer particles remained approximately con-
stant if the critical surface coverage was within (qcr1=0.25)<q<(qcr2=0.55), and;
(4) this surface coverage band is equivalent to the surface tension band of
42.50±5.0 dyne/cm that is required to avoid particle formation and coagulation
in the course of polymerization. Sajjadi [96] also carried out an experimental
investigation on particle formation under monomer-starved conditions in the
semibatch emulsion polymerization of St. They observed that the number of
polymer particles formed increased with a decreasing monomer feed rate, and
32 M. Nomura et al.

that a much larger number of particles (1–2 orders of magnitude greater) than
that generally expected from a conventional batch emulsion polymerization
was obtained. It is clear from Eq. 29 that any variation in the formulation or
process variables that results in a reduction in the volumetric growth rate per
particle m will enhance particle formation as long as polymer particles are
generated from emulsifier micelles. A depressed rate of particle growth can
be achieved either by the starvation of polymer particles or by a reduction in
the capability of polymer particles to swell monomer, with the exception of
impurities (see Sect. 3.5) and radical desorption (see Sect. 3.2). In this case,
the former was the reason for an increase in particle formation. The latter
could be achieved, for example, if a small amount of crosslinking agent was
used in the formulation. It is well known that crosslinking will decrease the
extent of particle swelling by a monomer and thereby reduce the rate of parti-
cle growth [97]. Sajjadi [98] also investigated particle nucleation in the seeded
emulsion polymerization of St in the presence of Aerosol-MA emulsifier mi-
celles and in the absence of monomer droplets (Interval III). A larger number
of polymer particles were found to form in Interval III than in the corre-
sponding seeded batch operation in the presence of monomer droplets. The in-
crease in the number of polymer particles could be attributed to the reduced
rate of growth of new particles, which retarded the depletion of the emulsifier
micelles.
There are an enormous variety of commercial emulsifiers that are employed
in emulsion polymerization. Emulsifiers are generally categorized into four
major classes: anionic, cationic, nonionic and zwitterionic (amphoteric). The an-
ionic and nonionic emulsifiers are the most widely used. In addition, mixtures
of emulsifiers are also often used. Since the effects of the molecular structure
and chemical and physical properties of an emulsifier on particle formation are
still far from being well understood, numerous experimental investigations on
particle formation have been carried out to date with various nonionic emulsi-
fiers [99–102], mixed emulsifiers (ionic and nonionic emulsifiers) [18, 103–106]
and reactive surfactants [33, 107–110]. Recently, polymeric surfactants have
become widely used and studied in emulsion polymerizations [111–116]. A
general review of polymeric surfactants was published in 1992 by Piirma [117].
Recently, emulsion polymerization stabilized by nonionic and mixed (ionic and
nonionic) emulsifiers was reviewed by Capek [118].
Özdeğer et al. studied the role of the nonionic emulsifier Triton X-405 (octyl-
phenoxy polyethoxy ethanol) in the emulsion homopolymerization of St [99]
and n-butyl acrylate (n-BA) [100], and in the emulsion copolymerization of St
and n-BA [101]. In the emulsion homopolymerization of St, they noted two
separate nucleation periods, resulting in bimodal PSDs.Although the total con-
centration of the emulsifier was maintained at a level above its CMC based on
the water phase in the recipe, the portion of the emulsifier initially present in
the aqueous phase was below the CMC due to partitioning between the oil and
aqueous phases. Due to the nature of this emulsifier, the first of the two nucle-
ation periods was attributed to homogeneous nucleation, while the second was
Emulsion Polymerization: Kinetic and Mechanistic Aspects 33

attributed to micellar nucleation. In the case of n-BA, contrary to the case of St,
the emulsifier was found to partition primarily into the aqueous phase, leading
to nucleation in the presence of micelles and unimodal PSDs. Particle formation
was accompanied by limited aggregation in the early stages of the reaction with
particles being formed past 50% monomer conversion in some cases. In the
emulsion copolymerization of St and n-BA, unimodal PSDs were observed at
the lowest (4.2 mM) and the highest (12.5 mM, 16.2 mM) levels of the emulsi-
fier, while bimodal PSDs were produced at intermediate levels. These results
were also attributed to emulsifier partitioning between the oil and aqueous
phases. Lin et al. [102] also studied the kinetics of the emulsion polymerization
of St in the presence of the nonionic emulsifier NP40 (nonylphenol poly-
ethoxylate with an average of 40 oxyethelene units per molecule). The initially
charged emulsifier concentration was well above its CMC. The number of poly-
mer particles produced was proportional to the 2.4th power of the total emul-
sifier concentration. This deviation of particle formation kinetics from the
S-E theory (the 0.6th power) was attributed to the low water solubility of the
emulsifier (higher solubility in the monomer droplets), the increased agglom-
eration of polymer particles for the system with the lower amount of emulsi-
fier, and the increased contribution of miniemulsion polymerization kinetics
to the system with a higher emulsifier level.
Anionic emulsifiers provide electrostatic stability to the resultant polymer
particles. The efficiency of anionic emulsifiers is dependent on various para-
meters, such as ionic strength and pH, but this can be a major drawback in
terms of the stability of the resultant polymer particles. On the other hand,
nonionic polymeric emulsifiers provide the steric stabilization provided by the
thermodynamically-favored steric repulsion between particles. It is therefore
practical to use mixtures of anionic and nonionic emulsifiers in emulsion poly-
merization to take advantage of these two different stabilization mechanisms.
Chern et al. [104, 105] studied the CMC of the mixed emulsifier SDS/NP40 for
various compositions at 25 °C and at 80 °C by performing surface tension mea-
surements. They examined the effect of the mixed emulsifier SDS/NP40 on par-
ticle formation in the emulsion polymerization of St at 80 °C. They found that
adding only a small amount of SDS to NP40 dramatically increased the number
of polymer particles produced, and that the emulsion polymerization of St with
the mixed emulsifier SDS/NP40 did not follow the conventional S-E theory. The
number of polymer particles produced, NT, was described by the expression
NTµS0a, where S0 is the concentration of mixed emulsifier, and the values of a
were 0.60, 0.76, 1.3 and 1.1 for experiments with molar concentrations of NP40
of 0%, 40%, 70% and 100%, respectively. Colombié et al. [106] investigated the
role of mixed anionic-nonionic emulsifier systems in particle formation in the
emulsion polymerization of St. They carried out the emulsion polymerization
of St using a mixture of SDS and Triton X-405 at 70 °C. They found that adding
just 1 mM SDS to 6.4 mM Triton X-405 produced a dramatic increase in the rate
of polymerization and the number of polymer particles produced. The increase
in the number of polymer particles was 17 times that without SDS. When in-
34 M. Nomura et al.

creased amounts of SDS (3 and 5 mM) were used in combination with 6.4 mM
Triton X-405, the number of polymer particles produced increased. No sec-
ondary particle nucleation was noted upon the disappearance of the droplets,
and the resulting latexes were stable. They attributed this behavior to the change
in the partitioning of the Triton X-405 between the oil and aqueous phases by
changing the amount of SDS added. When no SDS was used, 95% of the Triton
X-405 was associated with the oil phase as opposed to 78% when 5 mM SDS was
used. Therefore, increasing the amounts of Triton X-405 in the water phase
by increasing the amount of SDS added allowed the formation of mixed emul-
sifier micelles, resulting in an increase in the number of polymer particles
produced and the rate of polymerization. Unzueta et al. [103] carried out the
semicontinuous emulsion copolymerization of MMA and n-BA using mixed
emulsifier systems (anionic sodium lauryl sulfate and nonionic polyethylene
oxide lauryl ether (Brij35)) and found that narrower PSDs with larger average
particle sizes were obtained with mixed emulsifier systems than those obtained
with single anionic systems. Furthermore, they developed a mathematical
model for the emulsion copolymerization of MMA and n-BA with mixed
anionic/nonionic emulsifier systems, where the CMC and micelle composition
of mixed emulsifiers was predicted using the thermodynamics of nonideal
mixtures [18].
Reactive surfactants have also been used in emulsion polymerization [33,
107–110, 114]. This is because the disadvantages of the surfactants that are
typically used in emulsion polymerization, such as instability of the latex and
surfactant migration during film formation, can be overcome in theory by
using a reactive surfactant. Studies of the use of reactive surfactants in het-
erophase polymerizations (up to 1997) have already been extensively reviewed
[107]. Amalvy et al. [108] investigated the particle formation process in the
emulsion polymerizations of St, MMA and VAc stabilized by sodium dodecyl
sulfopropyl maleate, a polymerizable surfactant (surfmer), focusing their at-
tention on whether the reactivity of the surfmer with the main monomer(s)
and the polymerization locus play critical roles in the particle formation
in emulsion polymerization systems. The results obtained suggested that the
presence of the surfmer did not affect the particle formation mechanism. They
concluded from the shape of the logNT vs logS0 (surfmer concentration) plot
that the polymer particles were formed by micellar nucleation in the case of St
and by homogeneous nucleation in the case of MMA and VAc. Wang et al. [33,
109, 110, 114] studied the emulsion polymerization of St using the reactive sur-
factant sodium dodecyl allyl sulfosuccinate (TREM LF-40) and its polymeric
counterpart (poly(TREM)) as anionic polymeric emulsifiers in terms of the
polymerization kinetics. The use of TREM LF-40 gave NTµS 00.5–0.6 and RpµN T0.7
at constant initiator concentration. The reasons for the unusual kinetics com-
pared to those with SDS (RpµN T1.0) were ascribed to chain transfer to TREM
LF-40, copolymerization of St with TREM LF-40, and the influence of the ho-
mopolymer TREM LF-40 (poly(TREM)) and/or the copolymer (poly(TREM-
co-St)) on the entry and the exit rates of the free radicals. In contrast, by
Emulsion Polymerization: Kinetic and Mechanistic Aspects 35

varying the initiator concentration, the kinetics were found to have the same
dependencies as the conventional emulsifier (RpµN T1.0µI 00.4). In the case of poly-
(TREM), the dependencies of Rp and NT on S0 and I0 varied depending on
experimental conditions (RpµN T1.0µS 00.2–0.4, and RpµN T1.0µI 0.6–0.8
0 ). It was inferred
that homogeneous nucleation was dominant when using poly(TREM), even at
concentrations exceeding its CMC. This was different from the monomeric
TREM LF-40 emulsifier.
Recently, polymeric surfactants have received considerable attention in
industry. They provide the steric repulsion between interacting particles,
which gives the latex excellent stability against high electrolyte concentration,
freeze-thaw cycling and high shear rates. Cochin et al. [111] carried out a
comparative study of the emulsion polymerization of St using conventional,
polymerizable and polymeric emulsifiers. Ayoub et al. [112] investigated the
emulsion polymerization of St with amphiphatic copolymers [of VAc and
methoxy polyoxyethlene (PVAc-b-MPOE)(35:65, 27:73, 19:81 wt/wt) prepared
with a macroradical initiator in the presence of benzoyl peroxide] as the emul-
sifier. The experimental results for the number of polymer particles produced
(NT) versus emulsifier concentration (S0) were as follows: NTµS01.82 (65%),
NTµS 02.1 (73%) NTµS01.66 (81%). They [113] also studied the emulsion polymer-
izations of VAc and St using the polymeric emulsifier prepared from poly-
oxyethylene methylether (POE, 66%) and St (34%). They did not measure the
number of polymer particles produced, but the rate of polymerization was
found to be proportional to the 0.9th and 0.76th powers of the initiator (KPS)
concentrations, and to the 0.77th and 0.66th powers of the emulsifier concen-
trations for VAc and St monomers, respectively. Kato et al. [115] investigated the
emulsion polymerization of St using poly(methyl methacrylate (MMA)-co-
methacrylic acid (MAA)) with different copolymer compositions as polymeric
emulsifiers. They examined the effect of the copolymer compositions, molec-
ular weights and MAA contents of the polymeric emulsifiers on the number of
polymer particles produced. They found that the number of polymer particles
produced showed a slight dependence on the copolymer molecular weight,
having a maximum when the molecular weight was in the range 5,000–10,000,
and decreasing monotonously with the content of MMA in the copolymer.
Cheong et al. [116] studied the kinetics of particle nucleation and growth in the
emulsion polymerization of St using water-soluble polyurethane resins (PUR)
as the emulsifier. They found that the number of polymer particles produced
became constant in the early stage of polymerization when the concentration
of the initially charged PUR was lower. However, the monomer conversion
where the particle number became constant increased with increasing the
initial PUR concentration. The constant particle number observed (NT) was
correlated as NTµ[PUR]00.6–0.7[KPS]00.4 (where [PUR]0 and [KPS]0 are the PUR
and KPS concentrations, respectively). These dependencies are almost the same
as those predicted by the S-E theory.
36 M. Nomura et al.

3.3.2
Particle Growth in Homopolymer Systems

As is clear from Eq. 1, the rate of particle growth (Rp/NT) is proportional to the
monomer concentration, [M]p and the average number of radicals per particle,
–
n, respectively. Thus, –n is one of the basic parameters that characterize the
kinetic behavior of particle growth in an emulsion polymerization system.
Early researchers devoted their efforts to deriving a quantitative description of
–
n by solving Eq. 3 for –n defined by Eq. 2 [4, 119, 120].
Smith and Ewart [4] did not obtain a general solution to Eq. 3, but rather
solved it for three limiting cases at steady-state conditions, that is, dNn /dt=0.

Case 1. The number of radicals per particle is smaller than unity.

In this case, it holds that, Çe/NT Okf, (Çe/NT)N0@kfN1, and N0@NT. Furthermore,
a. When radical termination in the water phase is dominant; in other words,
Çw@2ktw[Rw*]2, then
–
n = (Ç /2k )1/2m v O 0.5 (44)
w tw d p

where ktw is the termination rate constant in the water phase, Çw is the rate of
radical generation per unit volume of water, [Rw*] is the concentration of radi-
cals in the water phase, and md is the partition coefficient of radicals between
the water and the polymer particle phases. However,
b. When termination in the polymer particles is dominant,
–
n = (Ç /2k N )1/2 O 0.5 (45)
w f T

The requirement for this condition is obtained as (4p2Dw2dp2NT/kf)oktw from ad-

ditional assumptions that Çe=2pDwdp[Rw*]NT and 2ktw[Rw*]2O2(Çe/NT)N1, where
Dw is the diffusion coefficient for the radicals in the water phase and dp the
diameter of the particles.

Case 2. The number of radicals per particle is approximately equal to 0.5.

The requirements for this case are given as kf O Çe/NT<ktp/vp. Then we have
n– = 0.5 (46)
Equation 46 usually holds in St emulsion polymerization under normal con-
ditions and is generally well known as the S-E theory.

Case 3. The number of radicals per particle is larger than unity.

This situation will prevail when the average time interval between successive
entries of radicals into a polymer particle is much smaller than the average time
for two radicals in the same particle to coexist without mutual termination; in
other words, Çe/NTpktp/vp.
–
n = (Ç v /2k )1/2 p 0.5 (47)
e p tp
Emulsion Polymerization: Kinetic and Mechanistic Aspects 37

Moreover, when both radical termination in the water phase and radical de-
sorption from the particles are negligible, Eq. 47 is reduced to
–
n = (Ç v /2k N )1/2 p 0.5 (48)
w p tp T

In this case, the kinetic behavior is quite similar to that of suspension poly-
merization, except that the polymer particles are supplied with free radicals
from the external water phase. When the polymerization proceeds according
to Eq. 48, the system is sometimes referred to as obeying “pseudo-bulk” kinet-
ics.
A general solution to Eq. 3 was provided by Stockmayer [121] with minor
corrections by O’Toole [119]. On the other hand, Ugelstad et al. [120] proposed
the most useful and widely applicable expression for – n given by
Im (a) 2a
n– = (a/4) 921 = (1/2) 99966
Im –1(a) 2a
m + 9977 (49)
2a
m + 1 + 9171
m+2+…

where Im(a) is the modified Bessel function of the first kind, m=kfvp/ktp , and
a=a2/8=Çevp/ktpNT. On the other hand, the radical balance in the water phase
(Eq. 4) leads to the following relationship using the non-dimensional parame-
ters, a, aw, m and Y.
a = a + mn – – Ya2 (50)
w

where aw=Çwvp/ktpNT and Y=2ktwktp/ka2NTvp. They solved the simultaneous

equations, Eq. 49 and Eq. 50 for –
n, and plotted the calculated value of –
n against
the value of aw that consisted of variables of known values with m varied as a
parameter for various fixed values of Y. Figure 3(a) shows an example of the
– versus loga for varying m and Y=0 [120].
plot of logn w
On the other hand, Nomura et al. [122a] provided a semi-theoretical ex-
pression for –
n corresponding to Y=0, and compared it with the experimental
data shown in Fig. 3(b) [122b].

       
2 1/2 1/2
– 1 aw aw aw 1 aw 1
n=3 aw + 5 + 2 aw + 5 – aw + 5 + 3 + 5 – 3 (51)
2 m m m 4 2 2

The values predicted by Eq. 51 agree well with those predicted by Eq. 49 within
less than 4%. This type of plot is called a “Ugelstad plot” and has been applied
as a criterion to determine whether a system under consideration obeys either
zero-one kinetics (n – ≤0.5) or pseudo-bulk kinetics (n
– >0.5).
Nomura et al. [42, 43, 64] showed that when the value of the term ktp/vp is
very large (the rate of bimolecular termination in the polymer particles is very
rapid), –n is expressed by
– C 2 + lll
n = (– C + dllll 2C)/2 (52)
38 M. Nomura et al.

Fig. 3 (a) Average number of radicals per particle –

n as predicted by Eqs. 49 and 50, and pre-
sented as a function of the parameters aw and m for Y=0, (b) comparison between predicted
and observed values of n– .

where C=Çw/kfNT. In Fig. 4, they plotted the value of –

n against the value of C for
the monomers of VCl, VAc, MMA, n-BMA and St obtained from the literature.
It was found that the experimental values of –
n are in fairly good agreement with
those predicted by Eq. 52, although the values of the parameters used in this
comparison may not necessarily be exact.
An example of a successful application of the Ugelstad plot to determine
some of the rate coefficients involved in emulsion polymerization was pre-
Emulsion Polymerization: Kinetic and Mechanistic Aspects 39

Fig. 4 Comparison of experimental values of –

n with those predicted by Eq. 52

sented by Nomura et al. [123]. Although this approach is rather laborious and
time-consuming, they could successfully determine the propagation rate con-
stant kp and the value of m, from which the desorption rate coefficient kf could
then be deduced, in seeded and unseeded emulsion polymerization of VDC at
50 °C. On the other hand, Hawkett et al. [24] developed a method for deter-
mining Ç and kf termed the slope-and-intercept method. This method is sim-
ple and straightforward, but has several drawbacks (as stated in Sect. 3.1.2), so
care must be taken when this method is used. Asua et al. [12, 124] proposed a
new approach for the estimation of kinetic parameters such as the entry and
desorption rate coefficients, the termination rate constant in the aqueous
phase, the rate coefficient for initiator decomposition and the propagation rate
constant in emulsion homopolymerization systems under zero-one conditions.
40 M. Nomura et al.

They claim that accurate values for the parameters are obtainable with this
approach provided that a sufficient number of experiments with a minimum
range of variation are available.
Recently, several modeling papers have been published which are useful for
the design and operation of emulsion homopolymerization processes [22, 80, 81,
125–127]. Mendoza et al. [125] developed a mathematical model that could pre-
dict the monomer conversion, particle diameter, number of polymer particles
produced, and the number-average and weight-average molecular weights in the
unseeded emulsion polymerization of St using n-dodecyl mercaptan as CTA.
This model was validated by fitting the experimental data to the model’s pre-
dictions. Kiparissides et al. [126] proposed a comprehensive mathematical
model to quantify the effect of the oxygen concentration on the polymerization
rate and PSD in the unseeded emulsion polymerization of VCl. Particle forma-
tion was assumed to proceed by both the homogeneous and micellar nucleation
mechanisms.Asua et al. [127] developed a mathematical model for seeded emul-
sion polymerization stabilized with polymerizable surfactants (surfmers). The
model included the most distinctive features of surfmer polymerization, in-
cluding partitioning of the unreacted surfmer between the surface of the poly-
mer particles and the aqueous phase, and the surfmer burying itself inside the
polymer particles. The model also included the possibility of having radical
concentration profiles in the polymer particles. Herrera-Ordóñez et al. [22, 80,
81] proposed a detailed mathematical model of the kinetics of St emulsion
polymerization, which was a modification and adaptation of previous works
reported in the literature. By comparing model predictions with experimental
results, they arrived at the conclusion that initiator-derived radicals with only
one monomeric unit also make a significant contribution to the rate of radical
capture by polymer particles, which contradicts the conclusion obtained by the
Sydney Group [11, 91]. They applied the model to the emulsion polymerization
of MMA above the CMC of the emulsifier to discuss the mechanism of particle
formation and growth in this system, and concluded that particle formation by
micellar nucleation is at all times at least ten orders of magnitude greater than
the homogeneous one, although MMA is moderately water-soluble [81].
Although the emulsion polymerization of VAc is already one of the most
studied systems, research articles on this topic are still being published [128–
132]. Gilmore et al. [128, 129] presented a mathematical model for particle
formation and growth in the isothermal semibatch emulsion polymerization
of VAc stabilized with poly(vinyl alcohol) (PVA). The model accommodated
grafting onto the PVA backbone during particle formation, and polymeric sta-
bilization.When the emulsion recipe, process conditions and kinetic parameters
are supplied, the model can predict the various species concentrations along
with the monomer conversion and particle size and number profiles. In Part II
of a series of papers [129], model predictions were compared with semibatch
and batch experimental results. Budhiall et al. [130] investigated the role of
grafting in particle formation and growth during the emulsion polymerization
of VAc with partially hydrolyzed PVA as the emulsifier and KPS as the initiator.
Emulsion Polymerization: Kinetic and Mechanistic Aspects 41

They found that: (1) the number of polymer particles produced was dependent
on the PVA blockiness; (2) the PVA with the higher degree of blockiness led
to the formation and stabilization of more polymer particles; (3) particle for-
mation continued to high conversions for the more random PVA (Poval 217),
whereas it appeared to stop at intermediate conversions for the blockier PVA
(PVA217EE); (4) all systems exhibited limited aggregation of the polymer par-
ticles during the polymerization process, and; (5) the greatest amount of graft-
ing of the PVA stabilizer onto the polymer particles occurred early in the
reactions (XM<25%), presumably contributing primarily to the stabilization of
the particles. Shaffie et al. [131] studied the kinetics of the emulsion polymer-
ization of VAc initiated by redox initiation systems of different persulfate cations
such as KPS, sodium persulfate (NaPS), and ammonium persulfate (APS); each
of them was coupled with a developed acetone sodium bisulfate adduct as the
reducing agent. In emulsion polymerization, the exhaustion of the separate
monomer droplet phase (the onset of Interval III) is usually followed by a
decrease in the rate of polymerization due to the decrease in monomer concen-
tration in the polymer particles. This is not the case for the emulsion polymer-
ization of VAc, where the rate of polymerization remains constant throughout
most of Interval III (ca. 20~90% conversion). In the case of VAc, Interval III starts
from around 20% conversion [20, 64]. In order to explain the reason for the
independence of the polymerization rate from the monomer concentration,
Nomura et al. [20] developed a model that takes into account the particles con-
taining at most two free radicals, where no instantaneous termination inside
the particles is assumed. Based on this model, they ascribed the reason to
an increase in the value of – n due to the gel effect, which compensates for the
decrease in monomer concentration with conversion. The decrease in the
termination rate constant inside the polymer particles due to an increase in
viscosity with conversion (the gel effect) prolongs the coexisting time of two
free radicals in the same particle, thereby increasing the value of –n. Chern et al.
[133] also developed a model that includes particles containing at most two free
radicals. However, they attributed the reason to a decrease in the rate coefficient
for radical desorption due to an increase in viscosity with conversion, which
results in an increase in the value of –
n. On the other hand, Bruyn et al. [132] pro-
posed a kinetic model that considers a zero-one system with instantaneous rad-
ical termination inside the particles. The model assumes that radical loss is by
transfer to a monomeric species which is very slow to propagate and whose
radical activity is lost by desorption and termination, either in the aqueous
phase or when it enters a particle containing a radical. Since the transfer step
is rate determining, the rate of this process is proportional to monomer con-
centration, which then cancels the dependence on the monomer concentration
in the overall polymerization rate expression. This model also predicts that the
radical loss rate coefficient should be either ktrCp (Limit 2a [1]) or 2ktrCp
(Limit 2b [1]), where ktr is the rate coefficient for transfer to monomer and Cp
is the monomer concentration in the particles. They [134] also studied the
kinetics and mechanisms of the emulsion polymerization of vinyl neo-decanete
42 M. Nomura et al.

(VnD), a practically water-insoluble monomer (its water solubility@4¥10–5 M

at 50 °C) at 50 °C using sodium persulfate (NaPS) as the initiator and SDS as the
emulsifier. They found that the polymerization rate was nearly independent
of particle number for a given initiator concentration (approximately inde-
pendent of particle size). They regarded this as consistent with the Limit 2b
zero-one kinetics of emulsion polymerization, whereby a monomeric radical
resulting from transfer to a monomer goes from particle to particle by desorp-
tion and reentry until it eventually enters a particle containing a growing radi-
cal, whereupon it undergoes very rapid bimolecular termination. Therefore, they
explained the kinetic behavior of this system with the same mechanisms and
model applied to the VAc system [132]. The most important claim raised in
these articles is that, contrary to the conclusion of Nomura et al. [43, 64], the rate
coefficient of radical desorption is independent of the water solubility of a
desorbing monomeric radical. Therefore, the validity of their claim is still open
for discussion and further studies are needed for its final solution.
Matsumoto et al. [135] studied the kinetics of the emulsion crosslinking
polymerization and copolymerization of allyl methacrylate (AMA) with MMA,
BMA and ethylene dimethacryllate (EDMA).

3.3.3
Particle Growth in Copolymer Systems

Ballard et al. [136] presented an extended S-E theory that provides a description
of the emulsion copolymerization system during Interval II and III and sug-
gested the possibility of using an “average” rate coefficient to treat the copoly-
merization system. On the other hand, Nomura et al. [45, 47, 137] first developed
an approach to generalize the S-E theory for emulsion homopolymerization to
emulsion copolymerization by introducing “average (or mean) rate coefficients”
for propagation, termination and radical desorption. This methodology was
termed the “pseudo-homopolymerization approach” [138] or the “pseudo-kinetic
rate constant method” [139] and is now widely applied, not only to emulsion
copolymerization systems, but also to other homogeneous free radical copoly-
merization systems. Nomura et al. [47, 122a, 140] demonstrated that the equa-
tions derived so far for emulsion homopolymerization can also be applied with-
out any modification to a binary emulsion copolymerization system with
monomers A and B by substituting the following “average rate coefficients” for
the corresponding rate constants for emulsion homopolymerization.
The polymerization rate for the A-monomer is expressed as
–
r = k [M ] –
pA pA nN
A p t T (53)
where [MA]p is the concentration of A-monomer in the polymer particles and
– – =–
nt is the average number of total radicals per particle (n nA + –
nB). The over-
t
all rate of copolymerization is defined by
–– –
R = k [M] –
p p n N = r + r = (k [M ] + k [M ] )n– N
p t T pA pB pA A p pB B p t T (54)
Emulsion Polymerization: Kinetic and Mechanistic Aspects 43
–
where kp is the overall propagation rate coefficient defined by Eq. 54 and is a
function of the propagation rate constant, monomer reactivity ratio and mole
fraction of each monomer in the polymer particles. In the case of a binary
emulsion copolymerization system, for example, the average rate coefficients
are defined as follows:
–
1. The average rate coefficient for the propagation of the A-monomer, kpA, is
given by
–
kpA = kpAA fA + (kpBB/rB)fB (55)

Here
–
nA kpBBrA[MA]p
fA = 5
– = 999953 (25)
nt kpAArB[MB]p + kpBBrA[MA]p

where kpBA denotes the rate constant for the propagation of a B-radical to an A-
monomer, fA is the fraction of A-radicals in the particle phase (fA+fB=1), and
rB is the B-monomer reactivity ratio.
–
2. The average rate coefficient for radical desorption, kf , is defined using the
equations
– /n
kf = kfA(n – – –
A t) + kfB(nB/nt) = kfA fA + kfB fB (24)
and

 
CmAArA[MA]p + CmBA[MA]p
kfA = KoA 99998641 (26)
rB {(KoA –
nt/kpAA) + [MB]p} + [MA]

where kfA is the desorption rate coefficient for the A-monomeric radicals, CmBA
is the chain transfer constant of a B-radical to an A-monomer, and KoA is the
desorption rate constant for A-monomeric radicals, given by Eq. 19 in
Sect. 3.2.1.
–
3. The average rate coefficient for radical termination in the particle phase, ktp,
is defined by
–
ktp = ktpAA f A2 + 2ktpAB fA f B + ktpBB f B2 (56)

where ktpAB is the bimolecular radical termination rate constant between A- and
B-radicals. Other average coefficients, such as the average chain transfer coef-
ficient, can be defined, if necessary, using the same principle. This approach can
be easily extended to multimonomer emulsion polymerization systems [138,
141]. Based on an exact mathematical treatment, Giannetti [142] concluded that
the pseudo-homopolymerization approach represents a suitable approximation
for most copolymerization systems of practical interest, except for very special
cases.
44 M. Nomura et al.

Since the appearance of the pseudo-homopolymerization approach, a wide

variety of mathematical models have been developed for emulsion copolymer-
ization systems using this approach, in order to thoroughly understand the
mechanisms involved in particle formation, growth processes, and to predict the
copolymerization rate, the properties of the copolymer obtained (molecular
weight and copolymer composition), and colloidal characteristics (the particle
number and PDS) [18, 27, 55, 58–63, 122(a), 140, 143–147]. Nomura et al. [122(a)]
first proposed a kinetic model that introduced the pseudo-homopolymerization
approach. They showed that the model could fairly accurately predict the mono-
mer conversion versus time histories observed in the emulsion copolymerization
of St and MMA. Moreover, they showed that the model could be successfully ap-
plied to the MMA-VAc system to predict the propagation rate constant and
monomer reactivity ratio for each monomer, respectively. Furthermore, they pre-
sented both experimental and modeling work for the unseeded emulsion copoly-
merization of St and MMA, including the particle formation process [140]. The
model was an extension of that used for simulating the kinetic behavior of the
emulsion homopolymerization of MMA [74]. This model describes both the
number of polymer particles produced and the monomer conversion versus
time histories observed in the emulsion copolymerization of St and MMA con-
ducted at 50 °C using KPS as the initiator and NaLS as the emulsifier.
Barandiaran et al. [63] also developed a mathematical model based on the
pseudo-homopolymerization approach. Furthermore, they proposed a method
to predict the kinetic parameters (kd f, kep, kf) in emulsion copolymerization
using only calorimetric measurements, and gave the values of these parameters
for the MA-VAc, MMA-BA emulsion copolymerization systems. Schoonbrood
et al. [27] carried out a kinetic study of the seeded emulsion copolymerization
of St with the relatively water-soluble monomer MA to investigate the mecha-
nisms of radical entry into particles, radical desorption from particles, and the
fate of radical species in the aqueous phase. For this purpose, they extended
their propagation-controlled entry model to an emulsion copolymerimeriza-
tion system by applying the pseudo-homopolymerization approach. López et
al. [55] used calorimetric measurements to study the kinetics of the seeded
emulsion copolymerization of St and BA. They varied the diameters of the seed
particles, the number of initially charged seed particles, and the initial initiator
concentration. A mathematical model was used to fit the experimental data for
conversion versus time using the entry and desorption coefficients as adjustable
parameters. Martinet et al. [145] carried out the emulsion copolymerization of
a-methyl styrene (aMSt) and MMA at various temperatures (60, 70, 85 °C) in or-
der to study the kinetic behavior, investigating the conversion, particle size, and
the average number of radicals per particle, as well as the copolymer composi-
tion, microstructure, molecular weight distributions (MWDs), and the glass
transition temperature (Tg). Unzueta et al. [18] proposed a mathematical model
for emulsion copolymerization with mixed emulsifier systems, and carried out
the seeded and unseeded emulsion copolymerizations of MMA and BA. Good
agreement was found between the experimental results in batch and semicon-
Emulsion Polymerization: Kinetic and Mechanistic Aspects 45

tinous reactors and the corrsponding model predictions. Vega et al. [61, 147]
and Dubé et al. [144] both developed mathematical models for the emulsion
copolymerization of AN and Bu initiated by a redox initiator system, with the
aim of simulating an industrial process and of improving the final polymer
quality. Due to the high solubility of AN in water, the following effects were
included: (a) the homopolymerization of AN in the aqueous phase; (b) the
desorption of AN radicals from the polymer particles, and; (c) homogeneous
particle nucleation. Saldívar et al. [58–60, 143] carried out extensive investiga-
tions on emulsion copolymerization. In the first paper, they presented a de-
tailed mathematical modeling of emulsion copolymerization reactors along
with comprehensive reviews of earlier models [58]. Then, they validated their
model with experimental results obtained in the emulsion copolymerization of
St and MMA, and demonstrated the generality of the model by applying it to
three illustrative problems [143]. Furthermore, they performed a systematic
experimental study of ten binary and three ternary emulsion copolymerization
systems involving St, MMA, BA, Bu,VAc,AA and E [59]. The predictions for the
evolution of conversion and average particle diameter in batch emulsion
copolymerizations from the model were compared with experimental data for
four emulsion copolymerizations of St with the comonomers MMA, BA, Bu and
AA.After data fitting for unknown or uncertain parameters, the model was ca-
pable of quantitatively explaining the experimental observations for conversion
evolution, but could only qualitatively explain the particle size evolution [59].
In industrial emulsion polymerization processes, a small amount of water-
soluble carboxylic monomer (such as AA) is often added to improve the col-
loidal stability and surface properties of the resulting latex particles. Therefore,
numerous studies have been carried out to date to clarify the influence of the AA
monomer on the kinetic behavior of the emulsion copolymerization of St and
AA [148–152] and of emulsion terpolymerizations including AA [79, 153, 154].
Shoaf et al. [148] presented a kinetic model that describes the reaction behavior
of emulsion copolymerization systems where significant polymerization occurs
in both the particle and aqueous phases. The model was applied to two seeded
carboxylated emulsion copolymerization systems, AA-St and methacrylic acid
(MAA)-St. They observed that the reaction behavior is greatly affected by the
type of acid monomer used, the partitioning of the monomer between the
various phases, and the locus of polymerization, and furthermore that the
mechanism for the AA-St system is more complicated than that for the MAA-
St system. They suggested that the primary reaction locus in the AA-St system
shifts from the particles to the aqueous phase after the hydrophobic monomer,
St, has been consumed. Yang et al. [149, 150] studied the effects of the initial
comonomer (AA) concentration on the monomer conversion and particle
number (NT) in the emulsifier-free emulsion copolymerization of St and AA.
They proposed an end-chain extension model to explain the experimental
results where the monomer conversion to the power of 2/3 is proportional to
the reaction time and log(NT)=36.00+9.44[AA]w. Slawinski et al. [151] investi-
gated the influence of AA on the particle growth process, with pH as the main
46 M. Nomura et al.

parameter in the seeded emulsion copolymerization of St and AA. To avoid the

effect of pH on the decomposition of KPS [153, 155], they carried out the
copolymerization at pH 2.5 (complete protonation) or pH 7 (neutralization). It
was found that pH was the dominating factor in the incorporation of AA onto
the particle surface.Wang et al. [152] examined the effect of AA and MAA sep-
arately on the total monomer conversion and the distributions of the carboxylic
groups at different positions (the surfaces and cores of particles, and in the
aqueous phase) in the emulsifier-free emulsion copolymerization of St with
AA or MAA. On the other hand, Santos et al. [153] carried out a batch emulsion
terpolymerization of St, BA and AA or MAA to study the effect of pH on the
polymerization rate, monomer conversion, and glass transition temperature of
the polymers produced, as well as the distributions of the carboxylic groups at
different positions (the surfaces and cores of particles, and in the aqueous
phase).Yan et al. [154] studied the kinetics and mechanisms of an emulsifier-free
emulsion terpolymerization of St, MMA and AA. They found that the rates of
particle formation and copolymerization increased with increasing concentra-
tions of AA and APS and polymerization temperature. Yuan et al .[79] carried
out the emulsion terpolymerization of St, Bu and AA in order to understand the
roles of the water-soluble oligomers produced. It was found that increasing the
AA concentration in the recipe increased the water-soluble oligomer concen-
tration and the number of polymer particles, thereby increasing the rate of
polymerization.
On the other hand, Xu et al. [156] studied the emulsifier-free emulsion
terpolymerization of St, BA and the cationic monomer N-dimethyl, N-butyl,
N-ethyl metacrylate ammonium bromide (DBMA) using oil-soluble azobis
(isobutyl-amidine hydrochloride) (AIBA) as the initiator. They found that
with increasing DBMA and AIBA concentrations, the number of oligomeric
radicals increased, resulting in an increased polymerization rate, as shown by
Rpµ[DBMA]0.64[AIBA]0.67. Fang et al. [157] investigated the kinetics and the
colloidal properties of the resulting polymer latexes in the emulsifier-free emul-
sion copolymerization of St and the nonionic water-soluble comonomer AAm,
using an amphoteric water-soluble initiator, 2,2¢-azobis[N-(2-carboxyethyl)-
2–2-methylpropionamidine]-hydrate (VA057). They found that the rate of poly-
merization at 20% conversion was proportional to the initiator concentration
to the power of 0.52.
Kostov et al. [158, 159] carried out a kinetic and mechanistic investigation of
tetrafluoroethylene and propylene with a redox system containing tert-butyl-
perbenzoate (TBPB). They found that RpµI 00.54S 00.42, where Rp is the rate of poly-
merization, I0 is the initial TBPB concentration, and S0 is the initial emulsifier
(C7F15 COONH4) concentration. Noël et al. [160] studied the effect of water sol-
ubility of the monomers on the copolymer composition drift in the emulsion
copolymerization of MA and vinyl ester combinations. Urretabizkaia et al. [161]
investigated the kinetics of the high solids content semicontinuous emulsion
terpolymerization of VAc, MMA and BA. The effects of operating variables
(feed flow rate, total amount of emulsifier, concentration of initiator and so on)
Emulsion Polymerization: Kinetic and Mechanistic Aspects 47

on the time evolution of the conversion, terpolymer composition, and the to-
tal number of polymer particles were investigated. The experimental results
were analyzed by means of a mathematical model that incorporated the main
features of the system. Ge et al. [162] studied the inverse emulsion copolymer-
ization of (2-methacryloyloxyethyl) trimethyl ammonium chloride and AAm
initiated with KPS. Aqueous monomer solutions were emulsified in kerosene
with a blend of two emulsifers (Span80 and OP10). Particle formation was
supposed to take place by monomer droplet nucleation. The observed rate of
polymerization is represented by RpµI 00.52S 00.38M 01.50, where M0 is the monomer
concentration.

3.3.4
Monomer Concentration in Polymer Particles

It is clear from Eq. 1 that the monomer concentration in a polymer particle is

one of the three key factors that control the particle growth rate, and accord-
ingly, the rate of polymerization. In emulsion polymerization, the course of
emulsion polymerization is usually divided into three stages, namely, Inter-
vals I, II and III. In Intervals I and II of emulsion homopolymerization, the
monomer concentration in the polymer particles is assumed to be approxi-
mately constant. In Interval III, it decreases with reaction time. Two methods
are now used to predict the monomer concentration in the polymer particles
in emulsion homopolymerization: empirical and thermodynamic methods.
According to the empirical method [14, 20, 163], the monomer concentration
in Intervals I and II can be expressed as
[M]p = [M]pc (57)
Interval III begins when the monomer droplets disappear from the system at
the monomer conversion XMc. The monomer concentration in this interval
(XM>XMc) is approximately given by

 
1 – XM
[M]p = [M]pc 94 (58)
1 – XMc

where [M]pc is the constant monomer concentration at saturation swelling.

On the other hand, several researchers [164–167] have tried to thermody-
namically describe the swelling behavior of polymer particles by one
monomer. The thermodynamic approach now used is based on the so-called
Morton equation given by

 
DFip 1 2
2Vmgj p1/3
7 = ln(1 – jp) + jp 1 – 31
– + cj p + 97 = 0 (59)
RT Pn r0RT

where DFip is the partial molar free energy of the monomer in the polymer
–
particles, jp is the volume fraction of polymer in the polymer particles, Pn is
48 M. Nomura et al.

the number-average degree of polymerization, c is the Flory-Huggins interac-

tion parameter, r0 is the unswollen radius of the particle, R is the gas constant,
T is the temperature, Vm is the partial molar volume of the monomer, and g is
the interfacial tension between the particles and the aqueous phase. Since the
– –
value of Pn is usually very large, the term 1/Pn can be neglected. Given values
of c and r0, Eq. 59 can be solved iteratively to yield jp. Then, by introducing the
value of jp into the following equation, one can get the saturation monomer
concentration in the polymer particles.
1 – jp
[M]pc = 91 (60)
Vm

Maxwell et al. [166] discussed the effects of several factors on the saturation and
partial swelling of polymer particles by monomers using Eq. 59 and the Vanzo
equation [168] that deals with the partial swelling of polymer particles in In-
terval III. By comparing theory and experiments for the MA and poly(MA-co-
St) system, the authors showed that the monomer partitioning was insensitive
to temperature, particle radius, copolymer composition, polymer molecular
weight, polymer cross-linking, the value of the Flory-Huggins interaction pa-
rameter, and the particle-water interfacial tension, and that the conformational
entropy of mixing of monomer and polymer was the significant term in de-
termining the degree of partial particle swelling by the monomer. Contrary to
Maxwell et al. [166],Antonietti et al. [167] observed a pronounced dependence
of the swelling ratio on particle size where absolute values of swelling were
much lower than those described by the classical Morton equation. In order
to explain this phenomenon, the authors presented a modified description that
considered size-relevant effects (such as the Kelvin pressure and depletion)
using an additional osmotic pressure term, which increases with the inverse of
the particle size. They also studied the effect of different types of covalently
bound surface stabilizing groups on the degree of swelling, and found that elec-
trically-stabilized particles resulted in higher swelling ratios and significant
lower values for the interfacial energy as compared to sterically stabilized
particles.
In an emulsion copolymerization, monomer partitioning between the mono-
mer droplet, polymer particle and aqueous phases plays a key role in deter-
mining the rate of copolymerization and the copolymer composition. Two
approaches (empirical and thermodynamic) have been proposed to predict the
monomer concentrations in the polymer particles in an emulsion copolymer-
ization system. In the emulsion copolymerization of St and MMA, Nomura et al.
[45, 122, 140] first proposed an empirical approach for predicting the saturated
concentration of each monomer in the polymer particles as a function of the
monomer composition in the monomer droplets, as shown by

1
[Mi]p = 9182 (61)
ai + bi /Wi,m
Emulsion Polymerization: Kinetic and Mechanistic Aspects 49

Fig. 5 Comparison of the observed saturation concentrations of St and MMA monomers in

polymer particles with those predicted by Eq. 61 [45]

where the subscript i denotes monomer i, ai and bi are the numerical constants
particular to monomer i, and Wi,m is the weight fraction of monomer i in the
monomer droplets. Figure 5 is an example that shows good agreement between
the predictions from Eq. 61 and experimental results.
The authors demonstrated experimentally that the saturation monomer
concentration in the polymer particles was insensitive to particle radius and
copolymer composition, and also that the weight fraction of the monomer i in
the polymer particles (Wi,m) was approximately equal to that in the monomer
droplets (Wi,p); in other words,
Wi,m = Wi,p (62)
Thermodynamic methods were developed based on the extended equation by
Ugelstad et al. [169], and have been further dealt with by various researchers
[170–181]. Maxwell et al. [170] worked on the partitioning of two monomers
between the polymer particle, monomer droplet and aqueous phases in an
50 M. Nomura et al.

emulsion copolymerization system, and proposed a thermodynamic approach

that could be easily extended to deal with systems of three or more solvents
and/or monomers. They derived the following equations for i and j monomers
by taking into account the partial molar free energy of mixing of the monomer
with polymer, the contribution of monomer to the interfacial free energy, the
partial molar free energy of the monomer in the monomer droplets and the
partial molar free energy of the monomer in the aqueous phase, respectively.

ln jpi + (1 – mij)jpj + jp + cij j pj2 + cip j p2 + jpj + jp (cij + cip – cjpmij)

(63a)
 
2Vmi gj p1/3 [Mi]w
+ 9521 = ln jdi + (1 – mij) jdj + cij j dj2 = ln 94
r0RT [Mi]w,sat

ln jpj + (1 – mji)jpi + jp + cij j pi2 + cjp j p2 + jpi + jp (cij + cjp – cipmji)

(63b)
 
2Vmjgj p1/3 [Mj]w
2 = ln
+ 9521 = ln jdj + (1 – mji) jdi + cij j di
r0RT [Mj]w,sat
94

where jp is the volume fraction of polymer in the latex particles, jpi, jdi, jpj
and jdj respectively represent the volume fractions of monomers i and j in the
polymer particles and monomer droplet phases, cij, cip and cjp are the Flory-
Huggins interaction parameters between each of the respective monomers i
and j and the polymer, mij is the ratio of the molar volumes of monomers i and
j (so mij=Vmi/Vmj, where Vmi and Vmj are the molar volumes of monomers i and
j, respectively), [Mi]w is the concentration of monomer i in the aqueous phase,
and [Mi]w,sat is its saturation concentration value if there are no other mono-
mers present. The derivations of Eq. 63a and Eq. 63b involve the reasonable
assumption that mip and mjp, the ratios of the respective molar volumes of
monomers i and j and the molar volume of polymer are negligible compared
to all other terms. Furthermore, they made the following three assumptions to
simplify Eqs. 63a and 63b.
1. For many pairs of monomers, the differences between the molar volumes of
the monomers is slight. If this is the case, the ratio of the molar volumes of
monomer i and j is well approximated by unity, so mij=mji=1.
2. The contribution to the partial molar free energy arising from the residual
(enthalpic and non-conformational entropic) partial molar free energy of
mixing of the two monomers is small relative to all other terms in the
monomer droplet phase.
3. The interaction parameters for each monomer with the same polymer are
equal (cip=cjp).
They finally obtained the following simple expressions for saturation swelling.
jpi jdi
= (64)
jpj 42
42 jdj
Emulsion Polymerization: Kinetic and Mechanistic Aspects 51

fpi = fdi (65a)

fpj = fdj (65b)

where fpi, fdi, fpj and fdj represent the volume fractions of monomers i and j in
the polymer particle and monomer droplet phases, respectively. This equation
relates the ratios of the volume fractions (or concentrations) of monomers i
and j in the monomer droplet and particle phases. Eq. 65 is basically the same
as Eq. 62 derived empirically by Nomura et al. [122(a)]. The validity of Eq. 65
was experimentally demonstrated in the St-MA, St-BA and MA-BA systems us-
ing seed polymer particles with different copolymer compositions and diam-
eters. Based on Eq. 65 and experimental results, they finally proposed the fol-
lowing simple empirical expressions that could predict the concentration of
monomers i and j in the polymer particles.
Ci = fdi [(Ci,m – Cj,m)fd,i + Cj,m] (66a)
Cj = fdj[(Cj,m – Ci,m)fd,j + Ci,m] (66b)
where Ci and Cj are the concentrations of monomers i and j in the polymer
particles, and Ci,m and Cj,m the maximum saturations of monomers i and j in
the polymer particles (the homo-monomer swelling concentrations in the par-
ticles). By comparing the predictions from Eqs. 64 and 65 with experimental
data from the St-MA, St-BA, MA-BA systems, they demonstrated that Eqs. 64
and 65 could provide adequate predictions for the monomer concentrations
in the polymer particles. In these discussions, the ratio of the molar volumes
of monomers i and j (mij) was assumed to be unity. On the other hand, Schoon-
brood et al. [171] examined the validity of the assumption mij=1, made by
Maxwell et al. [170], and demonstrated that this assumption could be used with
systems where mij deviated from 1, at least up to a value of 2.
Noël et al. [172] experimentally determined both the saturation and partial
swelling of MA-VAc copolymer latex particles by MA-VAc monomer mixtures.
Monomer partitioning at saturation swelling could be predicted using the
simplified relationships developed by Maxwell et al. [170]. On the basis of the
work by Maxwell et al. [166, 170], Noël et al. developed an extended thermo-
dynamic model for monomer partitioning at the partial swelling of latex par-
ticles by two monomers with limited water solubility in order to predict the
monomer concentrations and fractions within the different phases, and con-
firmed the validity of this model by showing that the model’s predictions were
in good agreement with the observed monomer partitioning.
On the other hand, Schoonbrood et al. [173] investigated multimonomer
partitioning in latex particles and derived simple equations describing mono-
mer partitioning among the latex particle, monomer droplet and aqueous
phases during Intervals II and III in emulsion copolymerization with any num-
ber of low to moderately water-soluble monomers, by extending the approaches
developed by Maxwell et al. [170] and Noël et al. [172]. They showed that it is
mainly the conformational entropy from mixing the monomer and polymer
52 M. Nomura et al.

that governs the partitioning behavior, and that other contributions to the free
energies of the monomers in the polymer particles are marginal. They con-
firmed that all of the assumptions made in this study were valid using experi-
mental results for St, MMA and MA, and confirmed that the simple equations
proposed describe the monomer partitioning with these three monomers in
Intervals II and III very well. In this approach, the only parameters needed to
calculate the monomer concentrations in all of the phases were the saturation
concentrations of each monomer in the polymer particles, and the saturation
concentrations of each monomer in the aqueous phase.
By combining thermodynamically-based monomer partitioning relation-
ships for saturation [170] and partial swelling [172] with mass balance equa-
tions, Noël et al. [174] proposed a model for saturation and a model for partial
swelling that could predict the mole fraction of a specific monomer i in the
polymer particles. They showed that the batch emulsion copolymerization
behavior predicted by the models presented in this article agreed adequately
with experimental results for MA-VAc and MA-Inden (Ind) systems. Karlsson
et al. [176] studied the monomer swelling kinetics at 80 °C in Interval III of
the seeded emulsion polymerization of isoprene with carboxylated PSt latex
particles as the seeds. The authors measured the variation of the isoprene sorp-
tion rate into the seed polymer particles with the volume fraction of polymer
in the latex particles, and discussed the sorption process of isoprene into the
seed polymer particles in Interval III in detail from a thermodynamic point of
view.
These thermodynamic equations provide the most complete description
of the swelling of polymer particles by monomers, but include a rather large
number of parameters whose accurate estimation requires extensive work.
Considering this, Gugliotta et al. [175] presented a criterion for choosing which
monomer partitioning models developed so far in the mathematical modeling
of emulsion copolymerization should be applied to a given system. In the math-
ematical simulations, the seeded emulsion copolymerization of four monomer
systems with a wide variety of reactivity ratios and water-solubilities were con-
sidered: BA-St, VAc-MA, VAc-BA and St-MAA. They investigated the effect of
the complexity of the monomer partitioning equations, the type of process, the
solid contents, and the amount of seed on the time evolution of the conversion
and copolymer composition, and tabulated a summary of recommended
monomer partitioning models.
In the industrial production of structured AN-Bu-St (ABS) latex particles,
the grafting copolymerization of AN and St on crosslinked polybutadiene (PB)
seed latex is carried out in emulsion polymerization. Therefore, information on
the effect of PB crosslinking density on the swelling of PB latex particles by a
St-AN monomer mixture is very important for the production of ABS copoly-
mers with desired properties. Mathew et al. [177] studied the effect of several
thermodynamic parameters, such as the crosslinking density, particle size and
monomer mixture composition on the swelling behavior of PB latex particles
by pure St and AN, and St-AN mixtures of various compositions. They reported
Emulsion Polymerization: Kinetic and Mechanistic Aspects 53

that, in the case of mixtures, the higher the AN concentration in the mixture,
the lower the maximum swelling by St, and the opposite effect was observed for
AN swelling. The parameters describing the interaction between the two
monomers were found to be functions of the composition in the initial mixture.
Liu and Nomura et al. [178–180] carried out a series of investigations on the
swelling behaviors of St-AN (SAN) and ABS latex particles by St-AN monomer
mixtures. In the first article, Nomura et al. [178] examined the effects of copoly-
mer composition and its compositional inhomogeneity in SAN latex particles
on their swelling behavior, and found that both the copolymer composition and
the compositional inhomogeneity in SAN latex particles had little or no influ-
ence on the swellability of SAN latex particles with a St-AN monomer mixture,
as long as the weight fraction of AN monomer units in SAN latex particles was
less than a certain value (between 0.6 and 0.8). Based on the experimental data,
they proposed semiempirical equations that could predict the saturation con-
centration of each monomer in the SAN latex particles as a function of the
comonomer composition in the monomer droplets and the overall copolymer
composition in the SAN latex particles. In the follow-on article, Liu et al. [179]
investigated the possibility of a thermodynamic correlation between both
the partial and saturation swelling of SAN latex particles by St-AN monomer
mixtures. First, they determined the three unknown Flory-Huggins interaction
parameters between each monomer and homopolymer, cA,PA, cA,PS, and cS,PS (S:
styrene, PA: polyacrylonitrile, PS: polystyrene) by fitting the thermodynamic
swelling equations to the experimentally observed monomer concentrations in
SAN latex particles. Then, they showed that the AN concentrations predicted
by using these interaction parameters agreed fairly well with those observed.
However, agreement between the predicted and observed St concentrations
was somewhat worse than that for the AN concentrations. On the basis of the
preceding studies, Liu et al. [180] further studied the saturation swelling of ABS
latex particles by a St-AN monomer mixture. In order to describe the observed
saturation swelling behavior, they proposed a two-phase swelling model
based on the assumptions that in ABS latex particles, St-AN (SAN) copolymer
domains were randomly dispersed in a continuous PB matrix, and further that
thermodynamic equilibrium was attained among the SAN copolymer domain,
PB matrix and monomer droplet phases. By using the proposed model, the
effects of various factors on the saturation concentrations of each monomer
in the ABS latex particles were experimentally and theoretically discussed.
The factors examined were the polymer crosslinking density M 3C, the interfa-
cial tension between the PB latex particle and aqueous phases g, the ratio of the
molar volumes of the St and AN monomers mab, the weight fraction of AN units
in the SAN copolymer domains HA, and the weight fraction of PB in the ABS
latex particles HPB. It was found that the saturation concentration of each
monomer in the SAN latex particles was insensitive to M 3C, g, mAB and HA, but
that HPB had a large influence on the saturation concentration of the AN
monomer but almost no influence on the saturation concentration of the St
monomer. They finally concluded that the two-phase swelling model developed
54 M. Nomura et al.

in this study could predict the saturation concentrations of St and AN mono-

mers in the ABS latex particles quite well.
On the other hand, Aerdts et al. [181] carried out partial and saturation
swelling experiments in latex particles of St-MMA (SMMA) copolymers, poly-
butadiene (PB) and composite particles containing PB with St and MMA grafted
on, and compared the results from them to predicted results from the semi-
empirical equations developed by Maxwell et al. [166, 170]. They showed that
the partitioning of MMA was independent of the type of polymer/SMMA
copolymers of different compositions and PB. Moreover, the partitioning of
MMA in PB was independent of particle size, polymer crosslinking density and
the presence of SMMA copolymer grafted onto PB. It was further shown that
the swelling of latex particles by St and MMA monomer mixtures was also in-
dependent of the polymer type of the latex particles and that the saturation
partitioning of monomers between the latex particle and monomer droplet
phases could be predicted by the simplified equations of Maxwell et al. (Eq. 66).
Said et al. [182] studied the effects of adding inorganic electrolytes on the
emulsion polymerization of St using three different ionic emulsifiers and potas-
sium and sodium chlorides as the inorganic electrolytes. They observed a sig-
nificant increase in the rate of polymerization in all cases as the concentration
level of the added electrolyte was increased. At the same time, they carried out
saturation swelling measurements and found a slight increase in the monomer
concentration inside the polymer particles as the level of added electrolyte
concentration was increased. They thought that one reason for an increase in
the rate of polymerization was the increase in the monomer concentration
inside the polymer particles as the level of added electrolyte concentration was
increased.
Tognacci et al. [183] discussed various methods for measuring the monomer
concentration in the polymer particles. The method proposed by the authors
is a direct estimation of the solvent activity by the GC (gas chromatography)
measurement of its partial pressure in the gas phase at equilibrium with the
polymer particle, monomer droplet (if any) and aqueous phase in the latex.
They proposed an original measuring technique and carried out measurements
for different monomers (St, MMA, and VAc) and polymeric matrices (PSt and
MMA-VAc copolymer), both above and below saturation conditions (corre-
sponding to Intervals II and III). They compared the experimental data with
that predicted by the monomer partitioning relationships derived by Maxwell
et al. [166, 170] and Noël et al. [172].

3.3.5
Reaction Calorimetry

Reaction calorimetry has been widely explored in studies of the kinetics of

heterophase polymerization in recent years [63, 82–84, 147, 184–191]. There are
several advantages of using a reaction calorimeter: (1) the rate of polymeriza-
tion is obtained directly, using the monomer conversion calculated from the
Emulsion Polymerization: Kinetic and Mechanistic Aspects 55

integral of the heat of reaction curve, and (2) nearly continuous information
is obtained. Using this information, a more detailed examination of the poly-
merization kinetics can be made, allowing the observation of important fea-
tures which cannot be seen using any other technique (such as gravimetry and
gas chromatography). Therefore, reaction calorimetry provides a powerful tool
for investigating heterophase polymerization [82]. Recently, many papers have
been published on reaction calorimetry studies into the kinetics of emulsion
polymerization. The kinetics of the emulsion polymerization of St was rein-
vestigated in detail [82–84, 184]. The effect of surface charge density on the
kinetics of the seeded emulsion polymerization of St was studied [185]. The de-
pendence of the reaction rate profiles on the water solubility of the monomers,
on the presence of CTA, on the types and concentrations of the stabilizer and
the initiator, and on the polymerization temperature was investigated [187].
The influence of oxygen on the kinetics of chemically initiated seeded emulsion
homopolymerization of St and the seeded emulsion copolymerization of St and
BA was investigated [188]. Reaction calorimetry has been used to estimate the
parameters in emulsion copolymerization systems [63], to control monomer
conversion and copolymer composition in semi-batch unseeded emulsion
copolymerization on-line [186, 189–191], and to monitor the copolymer com-
position, the average molecular weight, and the average degree of branching in
the semi-batch unseeded emulsion copolymerization of AN and Bu [147].
Most reaction calorimeters work according to heat-flow calorimetry prin-
ciples. The heat of reaction Qr evolved from a reaction mixture running at Tr
under isothermal conditions is transferred to the fluid in the cooling jacket
according to the equation
Qr = Rp (– DHp) = UAconst (Tr – Tj) (67)
where Rp is the reaction rate, DHp is the heat of reaction per unit amount of
reactant, U is the overall heat transfer coefficient, Aconst is the constant heat
transfer area, and Tj is the fluid temperature in the cooling jacket. The flow rate
of the cooling fluid is so high that the cooling fluid temperature at any position
in the jacket is considered to be, to a good approximation, equal to Tj. The tem-
perature of the cooling fluid Tj is adjusted to keep the reaction temperature
constant at Tr, and the heat of reaction Qr is found by calculating the value of
UAconst(Tr–Tj). Here, the heat transfer area Aconst and the value of U (determined
by calibration before and after the reaction) are treated as constant during
the reaction. However, the value of U is likely to change during the reaction
wherever the viscosity of the reaction mixture varies and/or the deposition of
scale on the heat transfer surface occurs during the reaction.
To avoid the drawbacks mentioned above, a novel calorimeter was developed
[192], as shown in Fig. 6, which can accurately measure the heat of reaction
independently of the variation of U during a reaction.
The working principle is as follows. A cooling fluid at temperature Ti at the
inlet is fed into the cooling jacket at a constant mass flow rate Fin, and the wet-
ted heat transfer area Avar in the jacket (which can be varied in this calorime-
56 M. Nomura et al.

Fig. 6 Schematic diagram of a novel calorimeter with variable heat-transfer area

ter) is controlled to keep the temperature of the reaction mixture at the desired
reaction temperature Tr by adjusting the fluid level (c) in the jacket, which is
achieved by regulating the flow rate of the cooling fluid at the outlet with a com-
puter-controlled throttle valve positioned there. The fluid temperature in the
cooling jacket is considered to be Tj throughout the jacket because of satisfac-
tory mixing due to the kinetic energy of the cooling fluid entering through a
nozzle. Therefore, the temperature of the cooling fluid flowing through the out-
let of the jacket is also Tj. To offset the heat loss from the reaction calorimeter
Qloss, a constant heat flux Qh(>Qloss) is passed into the reaction mixture through
an immersed electrical heater (f) in order to maintain the reaction mixture at
a constant temperature Tr even in the absence of the reaction. A reference run
is carried out with no reaction before measuring the heat of reaction. Then, the
steady-state heat balance for the reaction mixture in the calorimeter at tem-
perature Tr is given by

Qh = Qm0 + Qloss (68)

Qm0 = FinCp (Ti – Tj) = UAvar (Tr – Tj) (69)

where Qm0 is the heat flux transferred to the cooling fluid across the reactor wall
and Tj is the cooling fluid temperature at the outlet of the jacket. Since the
values of Fin, Ti and Tj are measurable, the heat flux Qm0 is obtained by calcu-
lating the value of FinCp(Ti–Tj), where Cp is the specific heat of the cooling fluid.
Therefore, this calculated value of Qm0 gives a base-line reading and is constant
Emulsion Polymerization: Kinetic and Mechanistic Aspects 57

as long as Qloss remains constant. On the other hand, when the reaction is
allowed to take place at the constant temperature Tr, one gets the steady-state
heat balances given by
Qh + Qr = Qm + Qloss (70)
Qm = FinCpi (Ti – Tj) = UAvar (Tr – Tj) (71)
where Qm is the heat flux corresponding to the new reading when the reaction
is taking place. As long as the value of Qloss is kept constant by maintaining
the temperature around the region of the calorimeter constant, one gets the
following expression from Eqs. 68 to 71.
Qr = Qm – Qm0 (72)
Therefore, one can derive the heat of reaction Qr independently of the value of
U from the difference between the new and base-line readings.

3.4
Effect of Initiator Type

There are two types of chemical initiator that can be used to initiate emulsion
polymerization. They are water-soluble initiators (such as KPS, hydrogen per-
oxide-iron (II) redox system) and oil-soluble initiators (like azobis-isobuty-
ronitrile (AIBN), benzoil peroxide (BPO), benzoil peroxide-N,N-dialkylaniline
redox system). Water-soluble initiators are more commonly used in emulsion
polymerization than oil-soluble initiators. However, oil-soluble initiators are
sometimes used when the fragments derived from ionic water-soluble initia-
tors are not desirable either in the latex serum or on the surface of the polymer
particles. Water-soluble initiators produce almost all of the free radicals in
the water phase, because the amount of initiator partitioned into the organic
phases is usually negligible. Contrary to water-soluble initiators, oil-soluble
initiators distribute among the four phases: monomer-swollen micelles, mono-
mer-swollen polymer particles, monomer droplets (if any), and the water
phase. In the case of oil-soluble initiators, only a small fraction of radicals are
produced in the water phase because the amount of initiator partitioned into
the water is usally very small. Therefore, it is useful to find out whether the
different principal initiator loci of polymerization systems with water-soluble
and oil-soluble initiators brings about any differences in the kinetics and mech-
anisms of polymerization between both initiator systems.
Several researchers have carried out experimental and/or theoretical inves-
tigations on emulsion polymerizations initiated with oil-soluble initiators and
reported that the kinetics of the emulsion polymerizations is basically similar
to that initiated with water-soluble initiators [193–202]. Breitenbach et al. [193]
carried out the emulsion polymerization of St initiated by BPO at 50 and 60 °C.
The authors interpreted the experimental results by assuming a relatively
rapid exchange of low molecular weight radicals between the micelle-polymer
58 M. Nomura et al.

particle and water phases. Van der Hoff [194] conducted the emulsion poly-
merization of St initiated by cumen hydroperoxide (CHP) and suggested three
possible mechanisms. One possibility is the entry of single radicals generated
from the fraction of the initiator dissolved in the water phase. A second possi-
bility is that single radicals are formed by desorption of one of a pair radicals
(that form within the particles or by a side reaction) into the water phase. An-
other possible mechanism is that pairs of radicals are produced in the emulsi-
fier layer, and only the organic radical (C9H11O·) enters the particle, while the
inorganic initiator fragment (OH·) remains in the water phase where it can
undergo further reaction. The author stated that there was no direct evidence
that any of these three mechanisms in fact came into play. Dunn et al. [195]
carried out the emulsion polymerization of St at 60 °C using octadecyl sulfate
as the emulsifier and AIBN as the initiator. They found that the number of
polymer particles produced varied approximately with the 0.4th power of the
initially charged initiator concentration. This behavior is quite similar to that
usually found in the emulsion polymerization of St initiated with KPS. They
ascribed this kinetic similarity to desorption of either of the primary radicals
that formed as a pair into the water phase, leaving a single radical inside the
polymer particle for initiation. It was also found that only 4% of the whole ini-
tially charged initiator was effective in the emulsion polymerization in contrast
with an efficiency of ~50% found in bulk or solution polymerizations. Barton
et al. [196] investigated the effect of an oil-soluble initiator (AIBN) on the
kinetics and mechanism of the emulsion polymerization of BMA at 60 °C in
the presence of the anionic emulsifier disodium dodecylphenoxybenzene
disulfonate. They compared the results obtained with the course of the emul-
sion polymerization of BMA initiated by KPS, and proposed that the radicals
produced by decomposition in the aqueous phase determine the kinetics of the
polymerization.
On the other hand, Il’menev et al. [197] carried out the emulsion polymer-
ization of St at 50 °C using oil-soluble initiators such as AIBN, BPO and lauryl
peroxide (LPO). The water-solubilities of AIBN, BPO and LPO at 20 °C are 3.6,
0.1 and 0.01 mol/dm3-water, respectively. The rate of polymerization conducted
at 50 °C with 0.025 mol/dm3-St of each initiator was found to be (fastest to
slowest): AIBN, BPO and LPO; in other words, in the order of decreasing
water-solubility. They estimated the average times for a primary radical to ter-
minate, propagate, and desorb into the aqueous phase, respectively, when a pair
of radicals are generated in a micelle and a polymer particle. They concluded
that the contribution to the polymerization (particle formation and growth)
from the free radicals that are produced in pairs in the micelles and polymer
particles is almost negligible, because they are very likely to cause rapid gem-
inate termination, and that the free-radicals generated in the monomer droplets
also play only a small part in the polymerization, because their desorption into
the water phase can be ignored. This view was strongly supported later by the
theoretical and experimental work of Nomura et al. [198–202]. Nomura et al.
[198] proposed a theoretical approach by which the effects of various factors
Emulsion Polymerization: Kinetic and Mechanistic Aspects 59

on the average number of radicals per particle – n could be predicted in seeded

emulsion polymerizations initiated by oil-soluble initiators. In their approach,
the following six kinetic events were considered, (i) the generation of a pair
of radicals inside the particles, (ii) radical entry into the particles from the
aqueous phase, (iii) overall radical desorption including both primary initia-
tor radicals and single-unit monomeric radicals produced by chain transfer to
monomer molecules, (iv) bimolecular termination of radicals in the particles,
(v) bimolecular termination of radicals in the aqueous phase, and (vi) genera-
tion of radicals in the aqueous phase by decomposition of the initiator dissolved
in the phase. Based on these events, they formulated a set of six differential equa-
tions describing the system in which particles containing more than six radicals
per particle could be neglected. The authors introduced a new parameter
K=Çw/Çp, where Çw is the rate of radical production per unit volume of water,
and Çp is the rate of radical production inside the particles per unit volume of
water. They solved a set of differential equations numerically and plotted the
calculated values of – n against ap=(Çpvp/ktpNT) for the range 0£K£•. Here,
the bimolecular termination of radicals in the water phase was assumed to be
negligible (Y=0).An advantage of this approach is that the time evolution of the
average number of radicals per particle can be evaluated. These plots were
found to be quite similar to those obtained for the case of KPS, except for K=0.
On the other hand, when K=0 (so the oil-soluble initiator employed is com-
pletely insoluble in water) no region of – n =0.5 was found regardless of the
values of ap and m (the parameter relating to the rate of radical desorption),
and the polymerization proceeded according to suspension polymerization
kinetics. Therefore, it was concluded that, kinetically, the similar behavior of
emulsion polymerization initiated by oil-soluble initiators to that initiated
by water-soluble initiators originated from the water-soluble portion of the
oil-soluble initiator rather than from the desorption of the initiator radicals
produced in the particles.
In order to delve deeper into the similarities and differences between the
kinetic behaviors of emulsion polymerization initiated by oil-soluble initiators
or water-soluble initiators, Nomura et al. [199–202] carried out extensive in-
vestigations into the kinetics and mechanisms of the unseeded and seeded
emulsion polymerizations of St at 50 °C using sodium lauryl sulfate (NaLS) as
the emulsifier and AIBN as the initiator, and obtained the following conclu-
sions:

1. The latex (polymer) particles are generated from the emulsifier micelles and
the number of latex particles produced is proportional to the 0.70th power
of the initial concentration of the emulsifier forming micelles and to the
0.30th power of the concentration of initially charged AIBN. This behavior
is very similar to that observed when the water-soluble initiator KPS is used.
2. The polymerization takes place both in the monomer droplets and in the
latex particles produced. The polymerization inside the monomer droplets
proceeds according to the kinetics of suspension polymerization until the
60 M. Nomura et al.

monomer droplets have disappeared from the reaction mixture due to

complete absorption by the resultant latex particles. On the other hand, the
polymerization in the latex particles proceeds according to emulsion poly-
merization kinetics, independently of the polymerization in the monomer
droplets. The total amount of polymer produced inside the monomer
droplets is only several percent of the whole polymer produced. Moreover,
the molar mass of the polymer produced in the monomer droplets is the
same as that produced by bulk polymerization under comparable conditions
and is only about one-hundredth of that produced in the latex particles.
3. The free radicals produced from the fraction of initiator dissolved in the wa-
ter phase are responsible for particle formation and growth in the emulsion
polymerization of St initiated by AIBN. The free radicals produced in pairs
in the polymer particles play almost no role in the polymerization inside the
polymer particles because pairs of radicals produced within a volume as
small as a monomer-swollen latex particle or a monomer-swollen micelle
are very likely to recombine.
4. A kinetic model developed for unseeded emulsion polymerization based on
the knowledge and conclusions obtained above could explain the progress
of polymerization inside both the monomer droplets and the latex particles
in the seeded emulsion polymerization of St initiated by AIBN at 50 °C.

Therefore, they showed both theoretically and experimentally that the kinetic
behavior of the emulsion polymerization of St initiated by AIBN is basically sim-
ilar to that initiated by KPS, and concluded that this similarity is mainly due to
the radicals produced from the water-soluble fraction of the initiator, because
the radicals produced pair-wise inside the small volume of a monomer-swollen
latex particle or a monomer-swollen micelle are very likely to recombine.
Several researchers have also experimentally and theoretically investigated
the reasons for this kinetic similarity [203–208]. Asua et al. [203] proposed a
mathematical model that can predict the average number of radicals per par-
ticle –n in seeded emulsion polymerization initiated by oil-soluble initiators.
Their model includes the parameter fw that denotes the fraction of the initia-
tor dissolved in the aqueous phase, and the following various kinetic events: (i)
generation of radicals inside the particles, (ii) desorption of primary initiator
radicals from the polymer particles before reacting with a monomer molecule,
(iii) termination of radicals by bimolecular reaction in the particles, (iv) de-
sorption of single-unit monomeric radicals produced by chain transfer to
monomer molecules, (v) absorption of radicals from the aqueous phase into the
particles, (vi) termination of radicals in the aqueous phase, and (vii) generation
of radicals in the aqueous phase by decomposition of the initiator dissolved in
that phase. They calculated the average number of radicals per particle in a typ-
ical example of the seeded emulsion polymerization of St, using their model that
distinguishes between desorption of primary initiator radicals and single-unit
monomeric radicals. The effect of increasing the water-soluble fraction fw of the
initiator from 0 to 0.1 was calculated for various particle diameters in the range
Emulsion Polymerization: Kinetic and Mechanistic Aspects 61

23–231 nm. For a fixed particle size, the value of – n was found to be essentially
independent of the fraction fw of the initiator present in the aqueous phase,
even for fw=0. Moreover, the plot of – n versus the seed particle diameter was
quite similar to that found for the emulsion polymerization initiated by KPS.
The authors therefore concluded that the kinetic similarity mainly originated
from desorption of the initiator radicals produced in the particles rather than
from decomposition of the initiator present in the aqueous phase. Mørk et al.
[208], however, pointed out that the almost identical values of – n found for a
completely water-insoluble initiator (fw=0) appeared to contradict calculations
performed by Mørk et al. [208] and Nomura et al. [198], and that calculations by
Asua et al. [203] could not be taken as evidence that a desorption mechanism is
the reason for the similarity. Alduncin et al. [204] studied the seeded emulsion
polymerization and the miniemulsion polymerization of St using an oil-soluble
initiator (AIBN) in an attempt to elucidate the main locus of radical formation
in emulsion polymerization initiated by an oil-soluble initiator. The monomer/
water weight ratio (M/W) was varied while keeping the monomer/initiator
ratio constant. They found that the average number of radicals per particle (n –)
increased as the M/W ratio increased. This was taken as evidence that the over-
all rate of radical entry into a particle increased when the M/W ratio increased.
They claimed that this phenomenon could only be explained by assuming that
the radicals responsible for emulsion polymerization initiated by oil-soluble
initiators are mainly those produced from the initiator partitioned into the
polymer particles, followed by desorption into the water phase. Mørk et al. [208]
concluded, on the basis of their calculations, that the argument provided by
Alduncin et al. [204] was not strong enough to resolve the issue of the similar-
ities between the kinetic behaviors of emulsion polymerization with oil-solu-
ble and water-soluble initiators.
Mørk et al. [206–208] recently published a series of theoretical works. In the
first article of this series [206], the authors aimed to develop expressions that
would allow easy and rapid calculation of the average number of radicals per
particle in emulsion polymerizations with a constant number of reaction loci
containing an oil-soluble initiator. Taking into account pairwise formation of
radicals in the particles, desorption and reabsorption, water phase termination,
solubility of the initiator in the water phase, and the possible formation of a
single radical species, they derived the recurrence relation that determined the
stationary state distribution of radicals in a particle. The calculation was based
on a probabilistic analysis leading to a third-order recurrence relation solved
using confluent, hypergeometric Kummer functions. The calculated results
confirmed the previous finding of Nomura et al. [198] that the kinetics of emul-
sion polymerizations carried out with oil-soluble initiators are quite similar to
those with water-soluble initiators, provided that the oil-soluble initiator is not
completely insoluble in the water phase. The main intention of the second
article [207] was to develop equations that make it relatively easy to assess
the effects of the most common experimental variables on the stationary state
average number of radicals per particle in a bidispersed seeded emulsion poly-
62 M. Nomura et al.

merization, and on the competitive growth of differently-sized seeded particles

in this system. In the third article [208], the authors extended the third-order
recurrence relation derived in the first article to the general case by including
single radical formation in the particles; for example, by a redox reaction along
with the formation of pairs of radicals by thermal decomposition. They carried
out calculations for the case where single radicals are generated inside the par-
ticles by an oil-soluble initiator, with or without the simultaneous formation of
pairs of radicals. The calculations showed that:
1. From a kinetic point of view, single radicals generated in the particles be-
haved quite similarly to radicals produced in the water phase.
2. However, at high rates of radical desorption, the effect of water phase ter-
mination on the average number of radicals per particle is much more
prominent when the radicals are produced in the water phase.
3. In the system where an oil-soluble initiator generates both single radicals
and pairs of radicals, the contribution of the latter to the average number of
radicals per particle is almost negligible.
4. When an oil-soluble initiator distributes between phases, the single radicals
that are responsible for the similar kinetic behavior observed with water-sol-
uble and oil-soluble initiators originate from the water-soluble fraction of
the initiator rather than from a desorption/reabsorption mechanism as
claimed by Asua et al. [203] and Alducin et al. [204].
Consequently, the authors supported the conclusion of Nomura et al. [198, 199]
that the reason for the similar kinetic behaviors observed for water-soluble and
oil-soluble initiators originates from the water-soluble fraction of the initiator.
Unlike in conventional emulsion polymerization, no monomer droplets ex-
ist in a microemulsion polymerization system, and hence, oil-soluble initiators
partition into the monomer-swollen micelles, the resultant polymer particles
and the water phase. Therefore, in microemulsion polymerization, the poly-
merization only proceeds in the monomer-swollen micelles and the resultant
polymer particles over the entire course of polymerization. Pairs of radicals
produced in volumes as small as monomer-swollen micelles and polymer par-
ticles may terminate as soon as they are generated. If so, it is expected that the
radicals responsible for the polymerization in the monomer-swollen micelles
and the resultant polymer particles would usually be those generated from the
fraction of the initiator dissolved in the water phase. In order to examine
whether this expectation is correct, oil-in-water (O/W) microemulsion poly-
merizations of St were carried out using four kinds of oil-soluble azo-type ini-
tiators with widely different water-solubilities [209]. It was found that the rates
of polymerization with these oil-soluble initiators were almost the same irre-
spective of their water-solubilities, when the polymerizations were carried out
with the same rate of radical production for the whole system for all of the
oil-soluble initiators used. Moreoever, the rate of polymerization with any of
these oil-soluble initiators was only about 1/3 of that with KPS at the same rate
of radical production. Considering that the rate of polymerization was pro-
Emulsion Polymerization: Kinetic and Mechanistic Aspects 63

portional to the 0.5th power of the initiator concentration regardless of whether

the initiator used was oil-soluble or water-soluble [210], the authors concluded
that the apparent efficiencies of these oil-soluble azo-type initiators were all
only 1/9 of that of KPS. This might suggest that although radicals were gen-
erated in the monomer-swollen micelles and polymer particles as well as the
water phase, only 1/9 of the radicals generated were active in the microemul-
sion polymerization of St, while the rest were lost somewhere (possibly in the
water phase) by bimolecular termination. These experimental results seem to
support the desorption/reabsorption mechanism proposed by Asua et al. [203],
although Candau et al. [205] also suggested that, in the case of AIBN, the radi-
cals that initiate the polymerization are not those from the initiator localized
within the monomer-swollen micelles, but from the initiator dissolved in the wa-
ter phase. Therefore, the role of oil-soluble initiators in the kinetics of hetero-
geneous polymerizations such as emulsion and microemulsion polymerizations
is still unanswered, and further studies are needed for its final elucidation. A
recent review article [211] refers to the role of oil-soluble initiators in hetero-
geneous polymerizations including emulsion polymerization.
Conventional emulsion polymerizations are usually initiated by chemical ini-
tiators. However, there are some disadvantages in the use of chemical initiators.
For example, some of the products produced by termination in the aqueous
phase may undergo subsequent reactions, resulting in discoloration of the final
latex, or any residual initiator present after the polymerization may act as an
undesirable contaminant. To avoid these problems, alternative techniques for
radical initiation that are safe and inexpensive have been explored. Ultrasound
has been increasingly used to realize novel chemical reactions and enhance the
reaction rate; this emerging field is called “sonochemistry”. Relatively recently,
several researchers have investigated the possibilities of using ultrasonic irra-
diation as a way to initiate free radical species in emulsion polymerizations of
various monomers [212–219]. Biggs et al. [212] conducted pioneering work
on the ultrasonically-initiated emulsion polymerization of St at 30 °C (±5 °C)
using NaLS as the emulsifier and a 20 kHz horn sonifier as a ultrasound gen-
erator. From experiments carried out at a fixed ultrasound intensity, they con-
cluded that: (1) radicals produced as a consequence of the cavitation process
were sufficient to cause polymerization; (2) the rate of polymerization increased
to a maximum at about 30% conversion before decreasing, showing no constant
region; (3) the rate of polymerization increased with increasing concentration
of initially charged NaLS (negligible polymerization without NaLS); (4) the
diameters of final latex particles were very small (around 50 nm), and the PSt
molecular weights were high (>106); (5) there was continuous formation of
polymer particles, and; (6) the small particle sizes, high polymerization rates,
and continuous nucleation of polymer particles were postulated to be due to
the continuous formation of very small monomer droplets in the ultrasonic
field, which could efficiently scavenge the radicals formed during the cavitation
process. The authors concluded from these experimental results that this poly-
merization system had many similarities with microemulsion polymerization
64 M. Nomura et al.

but at a considerably reduced emulsifier level. They extended this study, mainly
to clarify the effects of varying the input ultrasound intensity [213], and found
that: (1) a marked increase in the rate of polymerization was seen as the input
power was increased; (2) despite the increase in the rate of polymerization, the
increasing intensity did not affect the resultant polymer particle sizes, which
were in all cases 40–50 nm, and; (3) increases in both the concentration of NaLS
and the reaction temperature resulted in an increased rate of polymerization
at a fixed input intensity, but the particle sizes were invariant.
Cooper et al. [214] carried out the emulsion homopolymerization of BA and
of VAc, and also the emulsion copolymerization of BA and VAc at 30 °C (±5 °C)
using 20-kHz ultrasound as the initiator with SDS and Aerosol AT as the emul-
sifiers, respectively. The homopolymerization rate of VAc (10 wt%) was much
lower than that of BA (10 wt%), and interestingly, lower rates of BA emulsion
homopolymerization were observed at higher temperatures. The reason for
such a large difference in the rate of polymerization between the BA and VAc
systems was explained by the greater evaporation of the more volatile VAc
monomer into the cavity, suppressing cavitation and thereby reducing the rate
of radical production. The average particle sizes produced in the BA system and
the copolymerization system with 50:50 wt% BA and VAc were very small;
around 15–20 nm, respectively. But the average particle sizes produced in the
emulsion homopolymerization of VAc were much larger, showing a size of
around 300 nm. The reason for producing smaller polymer particles in both the
BA homopolymer and BA-VAc copolymer systems than in the VAc homopoly-
mer system even at low emulsifier concentrations was attributed to a high rate
of particle formation due to a large number of very small monomer-emulsion
droplets that were to be transformed into polymer particles.
Grieser et al. [215] investigated the kinetics and mechanisms of the emulsion
polymerization of MMA and of BA at 30 °C (±5 °C) using ultrasonic irradiation
(20-kHz horn sonifier) and a cationic emulsifier, dodecyltrimetylammonium
chloride (DTAC). They observed the formation of stable dispersions with par-
ticle diameters in the range of 40–150 nm and with polymer molecular weights
greater than 106 g mol–1. In the case of MMA, the average particle size was found
to be constant throughout the reaction time (sonication time) and independent
of the initial DTAC concentration. The final particle size decreased as the initial
DTAC concentration was decreased, but the rate of polymerization was ap-
proximately the same over the concentration range of DTAC examined. In the
case of BA, the kinetic behavior was basically the same as that of MMA, except
that the average particle size was constant (~30 nm) up to 50 min of sonication,
after which a dramatic increase in size (100–140 nm) was observed when the
initial TDAC concentration was comparatively low. Based on their experimen-
tal data, the authors proposed the kinetics and mechanisms of the ultrasonic
(sonochemical) initiation in this polymerization process. When ultrasound
is applied in a liquid medium, the cavitation event that occurs as ultrasound
travels through the liquid medium, producing microbubbles in the solution.
When the microbubbles rapidly collapse, this leads to high local temperatures
Emulsion Polymerization: Kinetic and Mechanistic Aspects 65

of the order of 4000–5000 K within the bubble and at least 1250 K in the liquid
immediately surrounding the interfacial region. In an aqueous medium, such
high temperatures lead to the homolysis of water, creating hydroxyl (·OH) and
hydrogen (·H) radicals. The authors assumed that primary organic radicals
produced from MMA and BA were unlikely to play a major role at 20 kHz even
though MMA and BA are volatile and could enter cavitation bubbles to produce
a variety of primary organic radicals by thermal decomposition. The hydroxyl
and hydrogen radicals generated in the aqueous phase add several monomer
units and then enter the miniemulsion droplets produced by ultrasonication,
initiating polymerization. Therefore, the results obtained strongly support a
polymerization process involving a miniemulsion polymerization system,
where continuous formation of polymer particles takes place throughout the
polymerization.
Chou and Stoffer [216, 217] carried out the ultrasonically-initiated free rad-
ical emulsion polymerization of MMA at ambient temperature using NaLS as
the emulsifier, and published two articles on this topic. In the first article [216],
the authors studied: (1) the nature and source of the free radicals for the initia-
tion process; (2) the effects of different types of cavitations, and; (3) the depen-
dence of the polymerization rate, the number of polymer particles generated,
and the polymer molecular weight on the acoustic intensity, argon gas flow
rate, surfactant concentration, and the initial monomer concentration. They
found that, in the absence of argon gas flow, no polymerization took place, and
that, contrary to Grieser et al. [215], the source of the free radicals for the ini-
tiation process came from the degradation of the NaLS, presumably in the
aqueous phase. The molecular weight of the poly(MMA) obtained varied from
(2.5–3.5)¥106 g mol–1, and the monomer conversion was up to 70%. The rate of
polymerization was found to be proportional to the acoustic intensity to the
power of 0.98, to the argon gas flow to the power of 0.086, and to the emulsifier
concentration to the power of 0.08 in the emulsifier concentration range of
0.035–0.139 M. The number of polymer particles was found to be proportional
to the acoustic intensity to the power of 1.23, to the argon gas flow to the power
of 0.16, and to the emulsifier concentration to the power of 0.3 in the emulsifier
concentration range of 0.035–0.139 M. In the second article [217], the radical
generation process was studied. Based on this experimental study, the authors
tried to explain the kinetic data obtained in the previous work. In this study, rad-
ical trapping experiments were used to investigate the effects of acoustic inten-
sity, argon gas flow rate, and NaLS concentration on the extent of free radical
generation in aqueous NaSL solutions. Aqueous solutions of NaLS were ul-
trasonically irradiated in the presence of a radical scavenger. The NaLS mol-
ecules then decomposed by ultrasound to form free radicals in the aqueous
phase. It was found that the extent of free radical generation increased as: (1)
the 0.6th power of the acoustic intensity, (2) the 0.44th power of the argon gas
flow rate, (3) the 0.35th power of the emulsifier concentration in the emulsifier
concentration range of 0.035–0.139 M. These experimental results were found
to explain the effects of acoustic intensity, argon gas flow rate, surfactant con-
66 M. Nomura et al.

centration on the rate of polymerization and the number of polymer particles

generated.
Recently,Wang et al. [218, 219] carried out the ultrasonically-initiated emul-
sion polymerization of MMA using a 20 kHz ultrasonic generator and NaLS as
the emulsifier, respectively, in order to find a way to reach a high monomer con-
version. It was found that, with increasing NaLS concentration, the monomer
conversion increased significantly, but in the absence of NaLS, monomer con-
version remained nearly zero. Therefore, the NaLS emulsifier played a key role
and appeared to serve as an initiator. They observed that (1) an increase in the
reaction temperature resulted in an increase in the monomer conversion, (2) an
appropriate increase in the N2 purging rate also increased the monomer con-
version, and (3) the polymer particles prepared were nanosized, even with a
small amount of emulsifier. Optimized reaction conditions were obtained using
these experimental results, and so a high monomer conversion of about 67%
and high molecular weight polymers of several million could be obtained in a
period of about 30 min. They [219] also studied the ultrasonically-initiated
emulsion polymerization of n-BA in order to investigate the factors that affect
the induction period and the rate of polymerization, and proposed a mecha-
nism for ultrasonically-initiated emulsion polymerization. Increasing the N2
flow rate, temperature, NaLS concentration and power input, and decreasing
the monomer concentration resulted in further decreases in the induction
period and increased the rate of polymerization. Under optimized reaction
conditions, the conversion of BA reached 92% in 11 min. In addition, they
carried out a feasibility study on semicontinuous and continuous ultrasoni-
cally-initiated emulsion polymerization.

3.5
Effect of Additives and Impurities

Most kinetic studies on emulsion polymerization carried out in universities

and industrial research laboratories have been done under extremely clean
conditions. The polymerization is conducted in a high-purity nitrogen atmos-
phere with any remaining oxygen in the reaction system removed by degassing.
High purity initiators and emulsifiers are used, and the commercially-available
monomers are purified (by, say, distillation) to remove any inhibitors used
during storage as well as any other reactive organic impurities that may act as
radical scavengers or CTAs. In industry, however, it is usually impractical to
purify the monomers, initiators, emulsifiers, water, and so on to remove reac-
tive impurities from them. Moreover, the polymerization is usually carried
out in an industrial-grade low-purity nitrogen atmosphere containing a trace
of oxygen. The presence of inhibitors in the reaction mixture will affect both
particle formation and growth processes. Therefore, it is very important to
understand the effects of any impurities present in the starting materials when
attempting the optimum design and operation of emulsion polymerization
processes.
Emulsion Polymerization: Kinetic and Mechanistic Aspects 67

It has been recognized that the presence of oxygen during emulsion poly-
merization can have detrimental effects on the course of a reaction, causing
inhibition periods and retarding the reaction rate. Relatively few publications
have addressed the issue of the effects of oxygen in emulsion polymerization
[126, 188, 220–223]. With the intention of clarifying the effect of stirring on
emulsion polymerization, Nomura et al. [220] carried out St emulsion poly-
merization under three nitrogen atmospheres with different purities (contain-
ing a trace of oxygen) and at different stirring speeds. They observed that the
faster the stirring speed, the longer the retardation period. They attributed the
result to the diffusion limited transfer of oxygen from the headspace into the
water phase through the liquid surface, which was controlled by stirring. Fur-
thermore, they found that the polymerization rate following a long retardation
period was often greater than that after a shorter retardation period, indicating
that the final number of polymer particles produced with a long retardation
period was higher than that with a shorter retardation period. The reason for
this is discussed later. The same trend was also observed by other researchers
[188, 222]. Cunningham et al. [222] examined the effects of oxygen on the in-
duction period, conversion kinetics, molecular weight and particle size during
the emulsion polymerization of St, by varying the initial dissolved oxygen con-
centration in the aqueous phase. They found that the length of the induction
period did not vary linearly with the initial oxygen level, suggesting diffusion
from the reactor headspace to the aqueous phase could have a significant impact
on rates of particle formation and growth. Furthermore, the higher the initial
dissolved oxygen level, the longer the induction period and the smaller the
average diameter of polymer particles in the final latex product, which indicates
that the longer the induction period, the greater the number of polymer parti-
cles produced. Their experimental results suggested that, during the induction
and retardation period, the oxygen molecules in the reactor headspace were
continuously transferred into the aqueous phase, and some of them are con-
sumed by the radicals in the aqueous phase, but most of them diffuse further
into both the monomer-swollen micelles and polymer particles. Therefore, the
oxygen molecules that have diffused into the monomer-swollen micelles and
polymer particles inhibit the growth of radicals within them, thereby reducing
the volumetric growth rate per particle, m and resulting in an increase in the
number of polymer particles produced according to Eq. 29.
Arbina et al. [188] investigated the influence of oxygen on the kinetics of the
chemically-initiated seeded emulsion homopolymerization of St and the seeded
emulsion copolymerization of St and BA using reaction calorimetry. They dis-
cussed whether oxygen behaved kinetically as an ideal inhibitor. In the experi-
ments, they observed that oxygen not only caused an inhibition period, but also
behaved like a retarder by reducing the polymerization rate. Their explanation
for this seemingly contradictory behavior was the existence of mass-transfer
limitations from the reactor headspace to the latex, resulting in a gradual and
continuous flow of oxygen into the aqueous phase. Their own experiments
showed that this induction period decreased with increasing initiator concen-
68 M. Nomura et al.

tration. When the headspace to aqueous phase ratio was decreased, the induc-
tion period was reduced and the polymerization rate increased. The length of
the retardation period caused by oxygen in the seeded emulsion polymerization
is known to depend on the kind of monomer. In seeded emulsion polymeriza-
tions, the inhibition period may be followed by a retardation period during
which the polymerization rate increases to a steady state value. The retardation
period observed in the seeded emulsion polymerization of VAc is unusually long
compared to that of St or MMA. Bruyn et al. [223] tried to quantitatively explain
the reason for this unusually long retardation in terms of the initiator effi-
ciency, fentry, proposed by Maxwell et al. [11]. They argue that this unusually
long retardation is due to the high radical entry efficiency of the aqueous phase
oligomeric radicals, which allows latex particles to compete with dissolved oxy-
gen for these initiating radicals. In the case of VAc, this is due to the high value
of the product of the propagation rate constant and the water solubility.As oxy-
gen is consumed, the competition increasingly favors the entry of initiating rad-
icals into polymer particles and the rate of polymerization gradually increases.
In another case, Kiparissides et al. [126] considered the different effects of the
presence of oxygen on the kinetics and PSD in the emulsifier-free emulsion
polymerization of VCl. Oxygen is capable of reacting with primary initiator rad-
icals and the resulting oligomeric radicals in the aqueous phase to produce vinyl
polyperoxides. Taking this into account, they developed a mathematical model
into which the combined role of oxygen as an inhibitor and a radical generator,
through the formation and subsequent decomposition of vinyl polyperoxides,
was incorporated.
CTAs are used not only to reduce the molecular weight of the polymer pro-
duced, but also to limit the extent of the branching and crosslinking of the poly-
mer produced in diene-polymerization. It is well known that an ideal CTA is able
to reduce the molecular weight of the polymer produced in a homogeneous bulk
or solution free radical polymerization, without affecting the overall rate of
polymerization. For a long time it has been known that even an ideal CTA such
as mercaptan could affect not only the molecular weight of the polymer pro-
duced in emulsion polymerization, but also the rate of the polymerization [224],
but details about the effects of CTAs (including mercaptans) on the kinetics of
heterophase radical polymerizations, like emulsion polymerizations, have only
been revealed recently [225–229].
Whang et al. [225] observed that the rate of polymerization decreased in the
seeded emulsion polymerization of St in the presence of carbon tetrabromide,
which functions as an ideal CTA in the homogeneous bulk or solution free
radical polymerization of St. They pointed out that this appeared to result from
the enhanced rate of radical desorption of the free radicals from the polymer
particles. Also, a substance which behaves as an ideal CTA in homogeneous
polymerizations might apparently function as a retarder in heterogeneous free
radical polymerization. Lichti et al. [227] advanced the discussion by Whang
et al. [225] and argued that the increase in the rate of radical desorption which
brought about the decrease in the rate of polymerization paralleled the increase
Emulsion Polymerization: Kinetic and Mechanistic Aspects 69

in the chain transfer constant for the additives: CBr4>CCl4>St. The efficiency
of desorption of free radicals formed by chain transfer from the latex particles
followed the inverse order CBr4<CCl4<St, and this reflected the reactivities of
the low-molecular weight free radicals (formed by atom abstraction) with the
monomer.
At almost the same time, Nomura et al. [226] carried out an extensive ex-
perimental study of the effect of typical CTAs such as CCl4, CBr4 b, and four pri-
mary mercaptans (C2, n-C4, n-C7 and n-C12) on particle formation and growth
processes in the unseeded emulsion polymerization of St. They found that these
CTAs, which had almost no effect on the rate of bulk polymerization of St, de-
creased the rate of polymerization per particle m and so increased the number
of polymer particles produced (see Eq. 29). From a theoretical point of view,
they suggested that these effects could be enhanced by increasing the rate of
radical desorption from the polymer particles by adding a CTA with a higher
value of the chain transfer constant and/or with higher water-solubility.
Nomura et al. [20, 43] pointed out that the number of polymer particles pro-
duced NT could be expressed as NTµS 0zI 01–z in an emulsion polymerization sys-
tem following micellar particle formation. They also showed that, by increasing
the rate of radical desorption from the polymer particles in the interval of par-
ticle formation with the help of CTA, the emulsifier dependence exponent, z,
would increase from about 0.6 to 1.0, thereby decreasing the initiator depen-
dence exponent from about 0.4 to 0. This was also confirmed experimentally.
Therefore, they demonstrated that the effect of CTA on particle formation and
growth in the emulsion polymerization of St could be explained in terms of
desorptions of chain-transferred radicals from the polymer particles.
Maxwell et al. [228] discussed the effect of CTA, such as mercaptan, on ini-
tiator efficiency and extended their quantitative model for initiator efficiency
to take into account the effect of adding CTA. They assumed the following
model. The effect of the CTA on the entry rate occurs by facilitating the pro-
duction of aqueous-phase free radical species (CTA radicals) via transfer be-
tween species such as ·MnSO4– (where M is a monomer entity and n<z) and the
CTA in the aqueous phase. The CTA radicals will be formed at a reasonable rate
provided that the CTA is not too water-insoluble (as in C12H25SH) and the
resultant CTA radical is able to enter the latex particles rapidly because of this
relative insolubility in water. If the monomer-derived ·MnSO4– tends to suffer
aqueous-phase termination rather than entry, the overall rate of entry (and
hence initiator efficiency) will increase. They claimed that this mechanism
could explain the accelerating (promoting) effect of intermediate molecular
weight CTAs (C10–C12) on the emulsion polymerization of monomers such
as Bu, where the z value is large and so initiator efficiency is very low in the
absence of CTA, because most of the ·MnSO4– undergoes termination rather
than entry into the latex particles. However,Weerts et al. [229] studied the “pro-
moting effect” in the emulsion polymerization of Bu using SDS, potassium
stearate and potassium oleate as the emulsifiers and sodium or potassium per-
oxodisulfate, 4,4¢-azobis-(4-cyanopentanoic acid and AIBN as the initiators, and
70 M. Nomura et al.

concluded that the promoting effect appears to be related to impurities present

in the emulsifier, because it was found to be completely absent in emulsifier-free
polymerizations. They also demonstrated that a simple redox reaction between
a sulfate radical anion and thiol could not provide a satisfactory explanation for
the promoting effect. Therefore, the promoting effect of thiol in the emulsion
polymerization of diene-hydrocarbons is still poorly understood.
Commercially-available monomers usually contain an inhibitor such as
4-tert-butylcatechol (TBC) or hydroquinone (HDQ) to prevent undesired au-
topolymerization before their use in polymerizations. In industry, however,
distillation of the monomer to remove inhibitor is rarely carried out. Therefore,
a good understanding of the effect of inhibitor on the kinetics of emulsion
polymerization is important when designing and operating an industrial emul-
sion polymerization process. Huo et al. [230] investigated the effects of HDQ
and TBC on the kinetics of the emulsion polymerization of St, where HDQ is
a water-soluble inhibitor and TBC is an oil-soluble inhibitor, respectively. They
found that HDQ produced an induction period proportional to the amount
of HDQ present, but did not produce any significant difference in the rate of
polymerization regardless of the different HDQ levels in the systems. Moreover,
the final particle number, average particle diameter, molecular weights, and
latex viscosities were identical to within experimental error for all four runs
conducted. In contrast to HDQ, the monomer-soluble inhibitor TBC showed
more complex effects. They found that the higher the level of TBC, the larger
the number of polymer particles produced, and a longer induction time and an
increased retardation of the initial rate of polymerization with increasing TBC
level was evident, but the rate of polymerization at intermediate conversion in-
creased for smaller TBC level runs because of the increased number of polymer
particles produced. However, at higher TBC levels (200 ppm), TBC was never
fully depleted before the polymerization was completed, and so in spite of the
much larger number of polymer particles generated, the overall rate of poly-
merization never exceeded that for the purified monomer. The viscosity of the
latex produced increased with the TBC level in the system, which was a direct
consequence of the decrease in particle diameter. The average molecular weight
of polymer formed decreased somewhat, and the polydispersity of the polymer
produced increased from 1.5 to 2.0 with the TBC level, indicating that the dead
polymer was predominantly formed by reactions with the impurity (TBC),
which acted as a CTA as well as a retarder. These results were well predicted by
the modification of an existing model for Case II emulsion polymerization
which incorporated the effects of impurities. Penlidis et al. [221] also made nec-
essary modifications to their Case I model for batch and continuous emulsion
polymerization to account for the effects of reactive monomer-soluble impu-
rities (TBC).A detailed mechanism was proposed through which TBC affected
the emulsion polymerization of St [231]. TBC would oxidize into 4-tert-butyl-
1,2-benzochinon (TBBC) during the storage of St and also during the sulfate-
initiated emulsion polymerization of St. TBBC is soluble in St and is therefore
readily absorbed by the polymer particles produced. TBBC is known to act
Emulsion Polymerization: Kinetic and Mechanistic Aspects 71

as a radical accepter, so radical transfer from a growing radical to TBBC may

occur in the polymer particle. Because the TBBC radicals are relatively stable
due to conjugation, probably most of the TBBC can be converted into TBBC rad-
icals. Because of this radical transfer to TBBC, the chain growth in the particles
stagnates and so the rate of polymerization per particle (m) decreases, thereby
increasing the number of polymer particles produced. The TBBC radicals, which
desorb from the particles because of their charge, can react in different ways.
The desorbed TBBC radicals react with persulfate radicals in the aqueous
phase, converting the TBBC radicals back to TBBC and consuming persulfate
radicals (the inhibition mechanism). These TBBC molecules may be taken up
by the particles once more and transformed into TBBC radicals again, and so
on. Of course, the desorbed TBBC radicals can also contribute to the emulsion
polymerization in the usual way by reacting with monomer as initiating
radicals.
Barton et al. [232] conducted the emulsion polymerization of the sparingly
water-soluble monomer St and of the fairly water-soluble monomer MMA in
the presence and absence of the water-soluble inhibitor, potassium nitrosodi-
sulfonate (Fremy’s salt). By using oil-soluble dibenzoyl peroxide (DBP) and
water-soluble ammonium peroxodisulfate (APS) as the free-radical initiators,
they examined the effect of the location of the initiator on the kinetics of these
emulsion polymerizations with SDS as the emulsifier. Figure 7 shows an ex-
ample of the experimental results.
When the emulsion polymerization of St was carried out in the presence of
Fremy’s salt and ammonium peroxodisulfate in the aqueous phase, a distinct

Fig. 7 A typical example of the effect of water-soluble (APS) and oil-soluble (DBP) initiators
on the progress of the emulsion polymerization of St in the presence of a water-soluble
radical inhibitor (Fremy’s salt, FS); for [APS]=5¥10–4 mol/dm3, empty circles indicate [FS]=0,
filled circles indicate [FS]=10–4 mol/dm3; for [DBP]=5¥10–4 mol/dm3, empty squares indicate
[FS]=0, filled squares indicate [FS]=10–4 mol/dm3
72 M. Nomura et al.

inhibition period was observed, but after the end of the inhibition period the
conversion versus time curve was almost the same as that encountered in the
absence of Fremy’s salt. On the other hand, when the emulsion polymerization
of St was initiated by the oil soluble initiator DBP, the monomer conversion
versus time curve observed in the presence of Fremy’s salt was identical to that
seen in the absence of Fremy’s salt. In the case of MMA, the results were basi-
cally the same as those with St. Also, the rate of emulsion polymerization ini-
tiated by DBP was almost the same as the rate of bulk polymerization initiated
by DBP. This indicated that in the emulsion polymerization initiated by DBP,
the polymerization did not proceed in the monomer-swollen micelles and the
resultant polymer particles according to emulsion polymerization kinetics, but
in the monomer droplets according to bulk kinetics. This implies that neither
member of a radical-pair generated in a monomer-swollen micelle initiates
polymerization, either because geminate termination took place before either
of the pair of radicals desorbs from the micelle into the aqueous phase, or
because both of the radicals desorb as soon as they are generated and are scav-
enged in the aqueous phase. This finding is closely related to the claim [200]
that in the emulsion polymerization of St initiated by oil-soluble initiators, the
free radicals generated from the fraction of the initiator dissolved in the aque-
ous phase mainly participate in particle formation from the monomer-swollen
micelles. Barton et al. claimed that, although many studies had been performed
to clarify the effect of a variety of water-soluble and oil-soluble inhibitors on
emulsion polymerization, no unequivocal results had been obtained for their
effects on its kinetics, and specifically for their effect on particle formation, and
that this problem is still open for discussion.
Ignoring their side effects, chain transfer agents (CTAs) were originally
used in emulsion polymerization as additives to regulate the molecular weight
distribution of the resultant polymers and to limit the extent of branching and
crosslinking of the polymer produced in diene-polymerization. Recently, sev-
eral investigations based on this point of view have been published [57, 125,
233–238]. Barudio et al. [57] studied the effect of CTAs (tert-butanethiol and
n-dodecanethiol) on the microstructures of copolymers (the molecular weight
distribution (MWD) and glass transition temperature (Tg)) and the diameters
of polymer particles produced in the batch and semibatch emulsion copoly-
merizations of St and BA. The experimental results were interpreted in terms
of enhanced radical desorption and diffusion limitations of CTA between the
monomer droplet and particle phases. They proposed a kinetic model that was
able to successfully compute the kinetic constants, the number of radicals per
particle, the GPC/SEC diagram and the DSC thermogram related to the MWD
and Tg, respectively. Salazar et al. [234] developed a mathematical model that
included the effect of CTA (tert-nonyl mercaptan) on particle formation and
the average molecular weights in the batch and monomer-starved emulsion
polymerizations of St. Asua and co-workers [125, 233, 235–237] published
several reports on the effects of CTAs on the kinetics and the microstructures
of the resultant polymers in seeded and unseeded emulsion homo- and co-
Emulsion Polymerization: Kinetic and Mechanistic Aspects 73

polymerizations. Echevarria et al. [233] developed a closed-loop control strat-

egy based on on-line gas chromatographic measurements of both monomer
and CTA concentrations to obtain emulsion polymers of well-defined MWD.
The control strategy was experimentally assessed, producing widely different
MWDs in the emulsion polymerization of St using carbon tetrachloride (CCl4)
as CTA. Mendoza et al. [125] studied the effect of a CTA (n-dodecyl mercaptan)
on the MWD in the emulsion polymerization of St. It was found that the CTA
had no effect on the rate of polymerization but substantially affected the MWD
of the resultant polymers, and that the efficiency of the CTA in reducing the
MWD was lowered by mass-transfer limitations. The rate-controlling step for
CTA mass-transfer was the diffusion of the CTA from the surface of monomer
droplets to the aqueous phase, as already pointed out by Nomura et al. [239].
Mendoza et al. examined the process variables affecting CTA mass-transfer and
developed a mathematical model that could predict monomer conversion,
particle diameter, number of polymer particles, and number-average and
weight-average molecular weights. Plessis et al. [237] investigated the effect of
a CTA (dodecane-1-thiol) on the kinetics, gel fraction, level of branches and sol
molecular weight distribution in the seeded semibatch emulsion polymeriza-
tion of n-BA. They found that the gel fraction was strongly affected by the CTA
concentration, and that the sol weight-average molecular weight decreased
with increasing CTA concentration, whereas no effect on the kinetics and the
level of branches was observed. Their proposed mathematical model was able
to explain the effect of the process variables fairly well. Sayer et al. [235, 236]
published two papers. One [236] discussed the effect of a CTA (dodecanethiol)
on the kinetics and MWD of the semicontinuous emulsion copolymerization
of MMA and BA. It was found that the CTA had only a slight effect on the
reaction rate, but it significantly affected the secondary particle formation.
Moreover, the effects of the CTA concentration on the gel formation and the
mass-transfer limitations of the CTA were discussed. The other [235] dealt with
the effects of different strategies for copolymer composition control on the
MWD and gel fraction in the starved and semistarved seeded emulsion copoly-
merization of MMA and BA in the presence of dodecanethiol (CTA). It was
shown that simultaneous control of the copolymer composition and the MWD
was feasible. When the monomers were fed following the optimal semistarved
strategy, the MWD was controlled by employing dodecanethiol as the CTA.
Gugliotta et al. [238] studied the control of polymer molecular weight using
n-nonyl mercaptan (nNM) as the CTA in the emulsion polymerization of St,
with the aim of producing PSt latex particles of low molecular weight polydis-
persity at high conversion and in short reaction times. They claimed it was
preferable to use nNM instead of other CTAs like tert-dodecyl mercaptan or
CCl4.
Okaya et al. [240] investigated the effect of additives such as alcohol (iso-
propyl alcohol) on the initial stage of the emulsion polymerization of MMA
(1 wt%) initiated by APS in the presence of PVA (1%). They found that 90%
of MMA and 60% PVA were grafted and that stable polymer particles with an
74 M. Nomura et al.

average diameter of 80 nm were produced. However, the addition of alcohols

such as isopropyl alcohol to the system decreased the grafting to a great extent,
resulting in an increase in particle size. This was attributed to decreased hy-
drogen abstraction from PVA by sulfate radicals, due to the competing hydro-
gen abstraction from the low molecular weight alcohol.

3.6
Effects of Other Important Factors

Batch, semi-batch and continuous emulsion polymerizations are usually carried

out in stirred tank reactors, where agitation by a stirrer is necessary. The type
of stirrer chosen and its stirring speed can often affect the rate of polymeriza-
tion, the number of polymer particles and their size distribution (PSD), and the
molecular weight of the polymer produced. However, the effect of stirring on
emulsion polymerization has never been the main research parameter in re-
search programs [241]. This is probably due to the conflicting results obtained
so far by various researchers.
Shunmukham et al. [242] studied the effect of stirring on the emulsion poly-
merization of St, and concluded that violent agitation decreased the rate of
polymerization, as shown in Fig. 8.
Schoot et al. [243], on the other hand, criticized Shunmukham’s conclusion,
stating that this strange effect of agitation observed by Shunmukham might
have been due to the absorption of contaminant oxygen into the reaction mix-

Fig. 8 A typical example of the effect of stirring on the progress of the emulsion polymer-
ization of St in the presence of a typical inhibitor oxygen (40 °C, initiator: H2O2)
Emulsion Polymerization: Kinetic and Mechanistic Aspects 75

ture from nitrogen atmosphere in which the polymerization was carried out.
Evans et al. [244] carried out the emulsion polymerization of VDC at 36±1 °C
using NaLS as the emulsifier and APS as the initiator, and found that: (1) the
first stage polymerization rate decreased with increasing stirring speed; (2) the
second stage polymerization rate increased with increasing stirring speed, and;
(3) the third stage polymerization rate was independent of stirring speed. To
explain their results, they suggested two factors through which stirring affected
the polymerization rate. The first factor was the reduced levels of effective
emulsifier available for the formation of polymer particles, caused by the
adsorption of emulsifier molecules onto monomer droplets finely dispersed by
the stirring (in the first stage). The other factor was the effect of monomer-
transport from the monomer droplets to the polymer particles where the poly-
merization proceeded (in the second stage) upon the rate. Omi et al. [245] came
to the contrary conclusion that when the monomer used was St, stirring did not
influence emulsion polymerization as long as initial emulsification conditions
were not changed. They considered that stirring affected the polymerization
only through the former of the two factors suggested by Evans et al. [244]. Later,
Nomura et al. [75] carried out a kinetic study of the batch emulsion polymer-
ization of VDC at 50 °C with KPS as the initiator and NaLS as the emulsifier,
aiming to derive a quantitative explanation for the effect of stirring observed
by Evans et al. [244]. Unfortunately, however, the polymerization proceeded
very smoothly and the abnormal kinetic behavior caused by a change in the
stirring speed was not observed. Therefore, they concluded that the effect of
stirring observed by Evans et al. [244] in the VDC emulsion polymerization
must be a very special case. Therefore, the reasons for such abnormal kinetic
behavior remain a mystery.
These results illustrate that investigations into the effects of stirring on
emulsion polymerization, have produced inconsistent results and conclusions.
Further research was therefore needed to elucidate the effect of stirring in more
detail. Nomura et al. [220] carried out an extensive investigation into the effect
of stirring on the emulsion polymerization of St initiated by KPS at 50 °C with
NaLS as the emulsifier, with the intention of explaining the effects of stirring
quantitatively by showing how agitation affects emulsion polymerization, what
steps of the polymerization are affected by stirring, and whether a suitable
range of agitation exists in emulsion polymerization. The reactor used was
a cylindrical glass vessel with a dished bottom, fitted with four baffle plates
located at 90° intervals, and a four-bladed paddle type impeller. They concluded
that the effect of stirring on emulsion polymerization appears through the
following four factors:

1. When the polymerization is carried out in the presence of an imperfectly

purified nitrogen atmosphere, the retardation period is prolonged with
agitation due to an increase in the absorption rate of oxygen from the ni-
trogen atmosphere into the reaction mixture through the gas-liquid inter-
face. It has often been observed that the polymerization rate after a longer
76 M. Nomura et al.

retardation period is higher than that after a shorter retardation period.

The reason for this can be explained according to the S-E theory as follows.
If the polymerization is retarded by oxygen during particle formation, the
volumetric growth rate per particle (m) decreases, and so the number of
polymer particles produced (NT) would increase according to Eq. 29. When
the contaminant oxygen molecules in the nitrogen atmosphere are almost
completely consumed and so the supply of them into the reaction mixture
is not sufficient to restrain the polymerization appreciably, the polymeriza-
tion rate increases in proportion to the increase in NT.
2. In a pure nitrogen atmosphere, there is an optimum range of stirring speeds
where emulsion polymerization is not affected by agitation. If the stirring
speed is higher than the above-mentioned optimum range, the number of
polymer particles decreases by coagulation during the course of polymer-
ization, and so the polymerization rate also decreases.
3. At lower stirring speeds, on the other hand, stirring controls the rate of
monomer transport from the monomer droplets to the polymer particles,
thereby controlling the rate of polymerization. The rate-determining step is
usually the monomer-transport step from the monomer droplets to the aque-
ous phase, because the monomer-transport step from the aqueous phase to
the polymer particles is much faster than the former step due to the much
greater total surface area of the polymer particles compared to that of the
monomer droplets.
4. At low emulsifier concentrations near the CMC, an increase in the degree of
agitation results in a reduction of the emulsifier used for the formation
of polymer particles (like micelles). This is because the monomer droplets
become smaller as the degree of agitation is increased, and so the amount
of emulsifier adsorbed onto the surfaces of the monomer droplets increases
in proportion to the increased surface area of the monomer droplets. This
brings about a decrease in the number of polymer particles produced, and
so a decrease in the rate of polymerization.

On the other hand, Weert et al. [246] investigated the effects of stirring on the
kinetics of the emulsion polymerization of Bu at 60 °C using NaLS as the emul-
sifier. They carried out the polymerizations in a 2.3-liter reactor fitted with four
baffle plates located at 90° intervals and a twelve-flat-bladed turbine impeller.
In all of the experiments, the system was pre-emulsified by stirring for a few
minutes at 400 rpm, before adjusting the stirring speed n to the desired level.
The number of polymer particles produced, NT, was found to remain constant
within experimental error beyond a sufficiently high value of n, while a discon-
tinuous increase in NT became apparent towards lower values of n. The change
in NT was significant, especially at low n. They ascribed the reason for this
change to the fact that the level of emulsifier available for particle formation and
stabilization was influenced by the degree of agitation, as already pointed out by
Nomura et al. [220]. They discussed the effect of stirring in connection with the
flow conditions in the reactor, which are closely related to the stirring speed, and
Emulsion Polymerization: Kinetic and Mechanistic Aspects 77

finally arrived at the conclusions that the stirring speed influenced this poly-
merization system by reducing the effective emulsifier concentration available
for particle formation and stabilization at higher n, and by limiting the diffu-
sion of monomer to the polymer particles at low n. Arai et al. [247] studied the
effect of agitation on the kinetics of the soapless emulsion polymerization
of MMA in water at 65 °C. As expected, the stirring influenced the monomer
conversion versus time history, the molecular weight of the polymer produced,
and the number of polymer particles produced versus time. They found that
agitation was an important factor that affected the rate of monomer-transport
from the monomer droplets to the water phase. The authors proposed a quan-
titative kinetic model that could predict the effect of monomer transport on
the rate of polymerization. Kostov et al. [158] also studied the effects of poly-
merization conditions, including stirring, on the emulsion copolymerization of
tetrafluoroethlene (TFE) and propylene (P) with ammonium perfluorooc-
tanoate, initiated by a redox initiator system containing tert-butylperbenzoate.
They found that both the rate of copolymerization and the molecular weight
of the copolymer produced increased as the stirring speed increased up to
450 rpm, but then became independent of the speed above 450 rpm. The ex-
planation for this was that the stirring affected the rate of mass-transfer of TFE
and P from the gaseous to aqueous phases, which was the rate-controlling step
for speeds less than 450 rpm. Kim et al. [248] also reported the importance of
agitation in the semi-batch emulsion polymerization of TFE carried out using
a chemical initiator (APS) and a fluorinated surfactant (FC-143). The rate of
polymerization was found to increase linearly with the stirring speed. Based on
the experimental findings, they concluded that the diffusion or dissolution
of TFE into the aqueous phase was the rate-determining step through which
agitation affected the polymerization. Özdeģer et al. [249] investigated the ef-
fect of stirring speed and impeller type (axial and radial flow impellers) on the
kinetics of the emulsion copolymerization of St and n-BA using Triton X-405
(octylphenoxy polyethoxy etanol) as the emulsifier.At low solids content (30%),
the impeller type and speed did not have any significant effect on both the fi-
nal number of polymer particles and the overall rate of polymerization. The
PSDs were unimodal in all cases. For high solids content (50%), the rate of poly-
merization carried out with the axial flow impeller was slower, indicating that
fewer polymer particles were produced. Bimodal PSDs were obtained for both
cases. These differences were attributed to the partitioning of the emulsifier.
The axial flow impeller created more shear than the radial flow impeller. This
resulted in more monomer droplets being formed, leading to more of the emul-
sifier being associated with them. This also resulted in fewer emulsifier micelles
being available for particle formation, thereby leading to the lower number of
polymer particles produced.
In industrial emulsion polymerizations, CTAs like mercaptan are often used
to regulate the molecular weight of the polymer produced. In some cases, other
ingredients that directly participate in the polymerization reaction are used
to modify the properties of the polymer latex produced. In these cases, these
78 M. Nomura et al.

reacting species must be transported from one phase, for example, the monomer
droplets, via the aqueous phase, to the monomer-swollen polymer particles
where the reaction takes place. Therefore, when designing a latex product to
have particular properties, it is important to quantitatively elucidate the diffu-
sional behavior of these reacting species when they move between the phases
in an emulsion polymerization system. However, only a few researchers [125,
234, 239, 250–255] have presented quantitative discussions on the mass-trans-
fer problem involved in emulsion polymerization. Brooks [250] discussed the
monomer diffusion rate in an emulsion polymerization system. The author
calculated the maximum diffusion rate from monomer droplets to a polymer
particle via the water phase by using a simple diffusion equation and showed
that this rate was usually far greater than the rate of polymerization per parti-
cle. He also suggested that an adsorbed emulsifier layer on the surface of the
polymer particles would not impede monomer transfer to the particles. Finally,
he concluded that in most systems the diffusional processes that occurred in
the water phase would not affect the course of the polymerization. However,
Nomura et al. [220] pointed out a possibility that the monomer-transport step
from the monomer droplets to the water phase could control the polymeriza-
tion rate when the intensity of agitation was comparatively low.
Nomura et al. [239] discussed the mass-transfer problem in more detail for
the seeded emulsion polymerization of St initiated by KPS at 50 °C using NaLS
as the emulsifier and CTAs as the diffusing species. They carried out the seeded
emulsion polymerization of St using five normal aliphatic mercaptans with dif-
ferent molecular weights (n-C7, n-C8, n-C9, n-C10, and n-C12) under the condi-
tions of 400 rpm (stirring speed), NT=1.4¥1014 particles/cm3-water (the num-
ber of seed polymer particles), dpo=48 nm (the average diameter of seed
particles), and Tmo=8.44¥10–6 mol/cm3-water (the concentration of initially
charged mercaptan per unit volume of water), and measured the consumption
rate of each mercaptan. They found that the consumption rate decreased dras-
tically with the molecular weight of mercaptan, although the consumption rate
of n-C8 was not so different from that of n-C7. They analyzed these experimen-
tal data using a proposed diffusion model derived on the basis of the so-called
two-films theory, and concluded that the concentration of the CTA in the poly-
mer particles during the polymerization dropped to a value much lower than
the one that would be attained if thermodynamic equilibrium for the CTA were
reached between the monomer droplets and the polymer particles. This was
due mainly to the CTA molecules’ resistance to transfer across the diffusion
film at the interface between the monomer droplet and water phases. The au-
thors suggested that the proposed model can also be used to predict the rate of
mass-transfer of any sparingly water-soluble reacting species from monomer
droplets to the polymer particles where this species participate directly in the
polymerization (this may include the monomer itself).
It is well known that emulsion polymerizations of highly water-insoluble
monomers such as octadecyl methacrylate (OM), dodecyl methacrylate (DM),
and stearyl acrylate (SA) are generally not feasible using traditional surfactant
Emulsion Polymerization: Kinetic and Mechanistic Aspects 79

systems. This is because monomer transport from the monomer droplets to the
water phase is diffusion limited [239]. However, almost simultaneously, Rimmer
et al. [256, 257], Leyer et al. [258] and Lau et al. [259] reported that these highly
water-insoluble monomers could be emulsion polymerized in the presence of
b-cyclodextrin (b-CD). These studies uncovered a very interesting phenome-
non. Rimmer et al. [256, 257] successfully conducted the emulsion polymeriza-
tion of OM and DM at 70 °C using KPS as the initiator and Dowfax 2A1 as the
emulsifier. They claimed that the reason for this successful polymerization was
that the use of b-CD appeared to aid monomer transport from the monomer
droplets to the polymer particles across the aqueous phase by increasing the
apparent water-solubility of these monomers, because the CDs apparently
solubilize the hydrophobic compounds in aqueous media. On the other hand,
Leyrer et al. [258] reported that they also succeeded in emulsion-polymerizing
SA in the presence of methyl-b-CD, and claimed that in this case the CD served
as a phase transfer agent, because the ability of the CD to form a water-soluble
complex with hydrophobic molecules made it easier for the SA molecules to
leave the monomer droplets and to be released from the complex after arriv-
ing at the surfaces of the growing polymer particles. The authors reported that
only 5 wt% of CD was necessary to polymerize almost 100% of SA.
Recently, Soares and Hamielec [252] presented a review article on the study
of transport phenomena in emulsion polymerization and introduced a case
study on how to increase the amount of ethylene (E) content in the copolymer
produced by the emulsion copolymerization of E and VAc under the conditions
of a mass-transfer-controlled polymerization rate. Zubitur et al. [255] studied
the effect of agitation on the batch and semicontinuous emulsion polymeriza-
tions of St in a reactor equipped with a four-paddle type stirrer and dodecyl
mercaptan as the CTA. Here, the CTA mass-transfer from the monomer droplets
to the aqueous phase was the rate-controlling step. They showed that the mole-
cular weights of polymers decreased as the stirring speed was increased because
of the improvement in the CTA mass transfer from the monomer droplets to the
aqueous phase due to the improved emulsification of the monomer droplets. For
semicontinuous polymerization, the instantaneous conversion increased as the
stirring speed was increased for speeds less than 150 rpm because the system
was monomer diffusion controlled, whereas at stirring speeds higher than
150 rpm, the agitation was strong enough for the polymerization rate to be
kinetically controlled. They [251] further studied the effect of agitation on the
monomer and CTA transport step from the monomer droplets to the polymer
particles in the semicontinuous emulsion polymerization of St and BA with
KPS as the initiator and NaLS as the emulsifier, respectively. Polymerizations
were carried out in a 2 dm3 glass reactor fitted with a stainless-steel anchor-
type stirrer. It was found that when neat monomer addition was used, a mild
degree of agitation (0.1 kW/m3) was needed to overcome monomer mass trans-
fer limitations. However, a moderate degree of agitation (0.3 kW/m3) was not
enough to avoid mass transfer limitations when dodecyl mercaptan (CTA) was
present. Preemulsification of the feed was used to minimize the mass transfer
80 M. Nomura et al.

limitations of both the monomer and the CTA, even for a gentle degree of
agitation (0.01 kW/m3). The molecular weights of the polymers produced de-
pended on the presence of the CTA. In the presence of the CTA, the molecular
weights decreased with the stirring speed, whereas they increased in the absence
of the CTA.
Salazar et al. [234] investigated how the molecular weight could be controlled
in a starved emulsion polymerization of St using tert-dodecyl mercaptan and
tert-nonyl mercaptan (more water-soluble) as the CTAs. The authors showed
that in a starved polymerization with tert-dodecyl mercaptan, a mass-transfer
resistance to the mercaptan was required to fit the observed PSt molecular
weights to the model predictions, but this extra mass-transfer resistance could
be neglected in the case of the more water-soluble tert-nonyl mercaptan. Cun-
ningham et al. [253, 254] investigated the seeded emulsion polymerization of St
in order to study the effects of n-dodecly mercaptan on the polymer molecu-
lar weight distribution. In the emulsion polymerization of St with n-dodecly
mercaptan as the CTA, the transport of the CTA from the monomer droplets
to the polymer particles is diffusion limited, meaning that it was difficult to
calculate molecular weights, except perhaps by using empirical approaches.
They developed a methodology that used the mass transfer model developed
by Nomura et al. [239], which allowed the CTA concentration within the poly-
mer particles to be determined, regardless of whether or not the CTA was at
its equilibrium value, and validated the essential correctness of the approach
by comparing experimental molecular weight distributions with the model’s
predictions. The authors further suggested that the methodology used might
be amenable to online applications. Mendoza et al. [125] carried out a study of
the kinetics of St emulsion polymerization using n-dodecly mercaptan as the
CTA. In this study, it was found that the CTA had no effect on the polymeriza-
tion rate, but had a substantial effect on the molecular weight distribution
(MWD). The efficiency of the CTA in reducing the MWD was lowered by the
mass transfer limitations. The process variables affecting the CTA mass-trans-
fer were also investigated. For example, the average molecular weight of the
polymers produced was found to decrease with increasing stirring speed. The
authors developed a mathematical model to predict monomer conversion, the
number of polymer particles, and the number-average molecular weights, and
then validated the proposed model by fitting it to the experimental data.
During an emulsion polymerization, one often encounters the formation
of coagulum, which is sometimes fatal for products such as paints. Moreover,
it may prevent the scale-up of commercially-acceptable latex. Therefore, the
formation of coagulum during an emulsion polymerization is an important
industrial problem and may be closely related to agitation. Although some
researchers [260–262] postulated that coagulum may form during emulsion
polymerization due to tangential or shear stresses, which originate in the re-
action mixture due to agitation, the literature has very little information on any
quantitative experimental data in this field. Vanderhoff [260] discussed this
problem and proposed two mechanisms for the formation of coagulum in
Emulsion Polymerization: Kinetic and Mechanistic Aspects 81

the emulsion polymerization process: (i) a failure of the stability of the latex,
giving rise to flocculation and growth of the aggregates to macroscopic size
(lumps), and (ii) a different mechanism of polymerization, for example poly-
merization in large monomer droplets or a separate monomer layer in the
vapor space above the latex and on the reactor surfaces. Lowry et al. [261] stud-
ied the phenomenon of shear-induced coagulation in emulsion polymerization
carried out in a stirred tank reactor. Here, the authors correlated the coagulum
formation for different emulsion polymerizations to various agitation parame-
ters. For a low Reynolds number, it was shown that the stirring speed is impor-
tant, whereas, for a high Reynolds number, power consumption is the important
parameter. Matejicek et al. [262] studied the influence of agitation on the
creation of coagulum during the semicontinuous emulsion terpolymerization
of St-BA-AA carried out in 25 dm3 and 5 m3 reactors, respectively, and gave the
relationship between the amount of coagulum formed and the intensity of agita-
tion. The authors found that the amount of coagulum formed (Y%) was cor-
related to the specific impeller power input ei introduced by agitation, showing
that the dependence of the coagulum content in the dispersion on impeller
speed passes through a minimum.

4
Kinetic Aspects in Polymer Structure Development

4.1
Molecular Weight Distribution (MWD)

4.1.1
Monte Carlo (MC) Simulation Method

Polymerization rate represents the instantaneous status of reaction locus, but

the whole history of polymerization is engraved within the molecular weight
distribution (MWD). Recently, a new simulation tool that uses the Monte Carlo
(MC) method to estimate the whole reaction history, for both linear [263–265]
and nonlinear polymerization [266–273], has been proposed. So far, this tech-
nique has been applied to investigate the kinetic behavior after the nucleation
period, where the overall picture of the kinetics is well understood. However,
the versatility of the MC method could be used to solve the complex problems
of nucleation kinetics.
The MC method is a powerful technique for investigating complicated
phenomena that are difficult to solve by the conventional differential equation
approach. In the MC approach, all one needs are the individual probabilities
of various kinetic events. It is easy to understand the advantages of applying
the MC method to emulsion polymerization if we note that it is possible to
simulate the formation processes of all polymer molecules in each polymer
particle directly because the volume of the reaction locus is very small. One
82 M. Nomura et al.

unique characteristic of emulsion polymerization is that high molecular weight

polymers are produced without decreasing the polymerization rate, and this is
due to the compartmentalization of polymerization reactions inside the polymer
particles, resulting in the isolation of macroradicals. Usually, the loci of poly-
merization is made up of 1016 to 1018 polymer particles per liter. As a conse-
quence, the number of monomeric units in each reaction locus is limited to about
105 to 108. For example, suppose the diameter of a polymer particle at 100% con-
version is 0.1 mm, the density of the polymer is 1 g/cm3, and the molecular weight
of a monomeric unit is 100. In this case, the total number of monomeric units in
this polymer particle is 3¥106. If the number average chain length is 104, which
is not unusually large in emulsion polymers, each polymer particle consists of
only 300 polymer molecules at 100% conversion. For such a small number of
polymer molecules, one can readily simulate the formation processes of all poly-
mer molecules using an MC method in a straightforward manner.
In this section we discuss unique MWDs formed via linear emulsion poly-
merization, while the kinetics of branched and crosslinked polymer formation
are considered in Sect. 4.2.
Before the MC simulation method was proposed, theoretical analyses of
the MWDs from linear emulsion polymerizations had been conducted on the
basis of kinetic population balance equations [274–277] and Markovian sta-
tistics [278–280]. These approaches have clarified that the MWD of polymer
molecules formed in emulsion polymerization is fundamentally different to
that from corresponding bulk polymerization. However, due to the complex
heterogeneous nature of the polymerization system, in which entry and des-
orption of oligomeric radicals are involved, applying these methods to real sys-
tems is not straightforward, and analytical solutions are limited to very special
cases. Often, the effect of radical desorption on the MWD is taken into account
using the first-order chain stoppage reaction [276–280], which does not reflect
the real kinetics. Although some approximate methods have been proposed
[277, 281], it appears to be a formidable task to correctly account for the chain-
length dependence of radical desorption in the conventional approaches.
In MC simulation, any kinetic event can be accounted for, as long as the prob-
ability of each kinetic event is represented explicitly. Chain length dependent
kinetics can be accounted for in a straightforward manner if the functional form
is provided. In conventional MC simulations of molecular build-up processes,
the monomeric units are added to each growing polymer molecule one-by-one;
therefore, a multitude of random numbers and calculations are required to sim-
ulate the formation of each polymer molecule. To get around this problem,
a new concept, the competition technique, was proposed in order to drastically
reduce the amount of calculation required for the simulation [263, 264].
Figure 9 illustrates a case where two polymer radicals exist in a polymer
particle.
In this technique, the imaginary time (or equivalently, the imaginary chain
length, given by P=kp[M]pt) for a certain event to occur is calculated by using the
appropriate probability distribution for each type of event. If the given process
Emulsion Polymerization: Kinetic and Mechanistic Aspects 83

Fig. 9 Schematic drawing illustrating the simulation method based on the competition tech-
nique

is considered random and independent of chain-length, one can simply estimate

the time for each event to occur by generating one random number that follows
the most probable distribution [263, 264].After calculating the imaginary chain
lengths, a kind of event competition is considered, and the shortest “imaginary
chain length” is chosen as the “real event”. In the figure, Pf2 is chosen as a real
event. If this chain length is 1000, one can add 1000 monomeric units to both
polymer radicals by using only four random numbers, not 2000 random num-
bers as in the conventional type of MC simulations. It is often claimed that MC
methods are time-consuming, however, by using a well-designed method, the
calculation time can be reduced significantly. In addition, virtually any type of
information can be obtained from a set of MC simulations, and significant in-
sights into the complex reaction system can be obtained in a straightforward
manner. The MC method is useful for investigating emulsion polymerization
kinetics that involve various types of simultaneous kinetic events.

4.1.2
Instantaneous Molecular Weight Distribution

In this part, the instantaneous MWD formed over a very small time interval
is considered. This is equivalent to considering the distribution of polymer
84 M. Nomura et al.

chains formed in a certain fixed environment. Note that, because the polymer/
monomer ratio is kept approximately constant during Interval II, the instanta-
neous MWD may be a reasonable approximation for the linear polymers formed
during Interval II. This is not the case, however, for nonlinear polymer forma-
tion, as discussed later.
Assuming that the polymer particles have a uniform size, a single statistical
polymer particle that is representative of the whole population of particles can
be considered. This polymer particle is a kind of imaginary micro-reactor that
does not modify either the volume or the polymer/monomer ratio, even after
polymer chains are produced, so the reaction environment is kept constant
except that the number and chain length distribution of the macroradicals are
changed stochastically. By producing a large number of polymer molecules
consecutively with this imaginary polymer particle, the instantaneous MWD
can be determined.
After the nucleation period, three types of kinetic processes determine the
kinetics of emulsion polymerization: radical entry, radical desorption, and
polymer chain formation in the polymer particles. The kinetics of emulsion
polymerization are fully described by the following five dimensionless para-
meters:
1. Radical entry
Çe 1
e = 961 = 941 (73)
kp[M]pNT kp[M]p –
te

where Çe is the rate of radical entry into polymer particles, including the
reentry of desorped radicals, NT is the number of polymer particles, and –
te is
the average time interval between radical entry.
Assuming a random entry of radicals to all polymer particles, the imagi-
nary chain length Pe shown in Fig. 9 follows the most probable distribution,
and can be determined by using a random number between 0 and 1, y, as fol-
lows [273]:
Pe = (1/e) ln(1/y) (74)
2. Chain transfer
[CTA]p
Cf = Cm + CfCTA 93 (75)
[M]p
where Cm is the monomer transfer constant, and CfCTA is the constant of transfer
to chain transfer agents.
In the same way as in Eq. 74, the imaginary chain lengths Pf1 and Pf2 shown
in Fig. 9 can be determined by using a random number between 0 and 1, y, as
follows:
Pf = (1/Cf) ln(1/y) (76)
Emulsion Polymerization: Kinetic and Mechanistic Aspects 85

3. Bimolecular termination
2ktp
x = 93251 (77)
kp[M]p vpNA

where ktp is the bimolecular termination rate constant, in which the termina-
tion rate is represented by Rt=2ktp[R*]2, not Rt=ktp[R*]2 used in [263–273]. NA
is Avogadro’s number, and vp is the volume of a swollen polymer particle. Note
that the bimolecular termination rate depends on the particle size in emulsion
polymerization, and is larger for smaller polymer particles.
Assuming that the bimolecular terminations are independent of chain-
length, the imaginary chain length P12 shown in Fig. 9 can be determined by:
P12 = (1/x) ln(1/y) (78)
Note that Eq. 78 must be considered for all possible radical pairs.
4. Radical desorption
K0
d = 93 (79)
kp[M]p
where K0 is the desorption rate coefficient for an oligomeric radical.Any chain-
length-dependent radical desorption can be accounted for using the MC
method, in principle. However, it is often reasonable to assume that only
monomeric radicals can exit. For such cases, the average desorption rate coef-
ficient for all radicals in polymer particles, kf, which appears in the Smith-Ewart
equation [4] can be approximated by [43, 44, 122] kf @K0Cf. In this article, the
simulated results for such simple cases are shown.
If a chain transfer reaction is the actual event, as shown in Fig. 9, a mono-
meric or a CTA radical is formed. Neglecting the difference between mono-
meric and CTA radicals, the probability of radical exit when the chain transfer
reaction occurs is given by:
d
Pdes = 99961 (80)
1 + Cf + (n – 1) x + d
where n is the number of radicals in the polymer particle.
5. Type of bimolecular termination
ktp,c
jc = 99 (81)
ktp,c + ktp,d
where ktp,c and ktp,d are the bimolecular termination rate constants by combi-
nation and by disproportionation, respectively.
If bimolecular termination is the actual event, the probability that the
termination is by combination is equal to jc.
86 M. Nomura et al.

The MC simulation was performed for 1¥104 polymer molecules in each

condition [264]. It was shown that both the average number of radicals – n and
the distribution of the number of radicals per polymer particle agree com-
pletely with the analytical solution derived by O’Toole [119], and it was con-
firmed that the present MC simulation can be conducted without significant
statistical errors [264].
To clarify the unique characteristics of the MWD of emulsion polymers,
one of the simplest and most common cases [264], where neither chain trans-
fer (Cf=0) nor radical desorption (d=0) occurs, is considered here. The magni-
tude of bimolecular termination (x) is changed with a constant radical entry
frequency (e=2¥10–4). Figure 10 shows the calculated number- and weight-
– –
average chain lengths and the polydispersity index (PDI= Pw/Pn) as a func-
tion of –
n.

Fig. 10 Calculated number- and weight-average chain lengths as a function of the average
number of radicals in a polymer particle, –
n, with e =2¥10–4 and Cf =d =0
Emulsion Polymerization: Kinetic and Mechanistic Aspects 87

In the present investigation, because the value of – n increases when the

termination rate x is decreased, the average chain length becomes larger as – n
increases. As clarified in earlier theoretical investigations [276, 277], the PDI is
the largest when –n =0.5, which could be considered a typical example of emul-
sion polymerization. At – n =0.5, PDI=4 if polymer chains are formed solely via
bimolecular termination by disproportionation, and PDI=2 when formed by
combination. In homogeneous polymerizations, it is well known that PDI=2
when polymer chains are entirely formed by disproportionation, and PDI=1.5
when they are formed solely by combination. Therefore, the PDI of the in-
stantaneous MWD is larger for emulsion polymerization, and as the average
number of radicals per polymer particle – n increases, the PDI approaches that
for homogeneous polymerization.
The reason for the broader distribution in a typical emulsion polymerization
(the so-called zero-one system with – n =0.5) can be rationalized from the point
of view of a very fast termination reaction in a polymer particle. When an
oligomeric radical (a monomeric radical is assumed in the present simulation)
enters a polymer particle that contains a macroradical, the bimolecular termi-
nation occurs before the oligomeric radical grows to a sufficient chain length.
An investigation of the MWD produced makes it easier to understand the
origin of this behavior. Figure 11 shows the chain length distribution formed
by disproportionation (solid curve) and by combination (broken curve).

Fig. 11 Calculated chain length distribution on a number and weight basis, for –
n=0.500
88 M. Nomura et al.

The independent variable is the logarithm of chain length (log10P), which is

usually employed in size exclusion chromatography (SEC) analysis. The upper
figure shows the chain length distribution on a number basis, while the lower
figure shows that on a weight basis. The chain length distribution formed
by disproportionation is the same as the distribution of radicals that cause
bimolecular termination. When polymer chains are formed by bimolecular
termination from disproportionation, very large peaks appear at smaller chain
lengths in the number fraction distribution (N(log10P)). These discrete peaks cor-
respond to integral numbers of monomer units in the chain lengths: 1, 2,
3, ... In the present case, half of the polymers formed by disproportionation are
oligomeric chains from very fast bimolecular termination.When polymers are
formed via bimolecular termination by combination, on the other hand, the
chain length of the radical that has just entered is so small that the dead poly-
mer distribution obtained is almost the same as the macroradical distribution
in the polymer particles, which is given by the most probable distribution
(PDI=2).

Fig. 12 Instantaneous chain length distribution on a number and weight basis, where
dead chains are formed by disproportionation termination. The value of –
n is increased by
decreasing the bimolecular termination rate
Emulsion Polymerization: Kinetic and Mechanistic Aspects 89

In addition, because the weight fraction of the oligomeric chains formed by

disproportionation is very small, differences between the termination modes
cannot be found in the weight fraction distributions obtained via usual SEC
techniques, as shown in the lower figure of Fig. 11. Therefore, one needs to pay
careful attention to the measurement of the MWD, especially when the effect
of bimolecular termination by disproportionation cannot be neglected. If the
oligomeric peaks cannot be determined accurately, the obtained PDI drops
from 4 to 2, and it is expected that these oligomeric molecules are neglected in
usual SEC analysis.
As –n increases, the unique characteristics of the MWD formed in emulsion
polymerization are lost, as shown in Fig. 12.
The smaller and larger peaks in the N(log10P) formed via bimolecular
termination by disproportionation merge as – n increases. On the other hand,
such drastic MWD change cannot be observed in the weight based distributions
W(log10P) that are usually measured by SEC. When – n =2, the MWD formed has
already become very close to that for homogeneous polymerization, and in
terms of the MWD, pseudobulk polymerization kinetics would be a reasonable
approximation. Because – n increases with particle size, the MWD changes with
the particle size. This would be of special interest in cases with broad particle
size distributions, as in the case of a continuous emulsion polymerization
using a stirred tank reactor.
In our illustrative calculated results, chain transfer reactions are neglected in
order to highlight unique characteristics of emulsion polymerization. However,
the radical entry rate into a polymer particle is often much smaller than the
chain transfer frequency; e OCf in emulsion polymerization usually. In such
cases, dead polymer chain formation is dominated by chain transfer reactions,
and the instantaneous weight fraction distribution is given by the following
most probable distribution:
N(P) = Cf exp (– Cf P) (82a)
W(P) = C 2fP exp (– Cf P) (82b)

4.1.3
Effect of Chain-Length-Dependent Bimolecular Termination

The bimolecular termination reaction in free-radical polymerization is a

typical example of a diffusion controlled reaction, and is chain-length-depen-
dent [282–288]. When pseudobulk kinetics applies, the MWD formed can be
approximated by that resulting from bulk polymerization, and it can be solved
numerically [289–291]. As in the other extreme case where no polymer parti-
cle contains more than one radical, the so-called zero-one system, the bimole-
cular termination reactions occur immediately after the entrance of second
radical, so unique features of chain-length-dependence cannot be found.
Assuming that the average time interval between radical entries is the same for
all particles and that the weight contribution from oligomeric chains formed
90 M. Nomura et al.

via disproportionation termination can be neglected, the weight fraction

distribution is given by the following most probable distribution:
W(P) = (Cf + e )2P exp {– (Cf + e )P} (83)
On the other hand, however, it is not straightforward to calculate the MWDs
for intermediate cases using the conventional approach. A notable advantage
of using an MC simulation technique is that it can be applied to virtually any
type of emulsion polymerization, and can account for the chain-length-de-
pendent bimolecular termination reactions in a straightforward manner [265].
Sample simulation results for instantaneous MWDs were shown [265] that were
obtained using parameters for styrene polymerization that were reported by
Russell [289].
When the bimolecular terminations are highly diffusion controlled, the
termination reactions are dominated by interactions between radicals with
short and long chain lengths even in bulk polymerization, and the MWD of the
longer polymer radicals tends to follow the most probable distribution [287,
292]. Under such conditions, oligomeric chains that can be observed only in the
number fraction distribution may be formed via disproportionation termina-
tion irrespective of particle size. Figure 13 shows the effect of particle size on
the instantaneous chain length distribution where the bimolecular termina-
tions are from disproportionation [265].

Fig. 13 Effect of particle size on the instantaneous chain length distribution, where bi-
molecular terminations are chain-length dependent and are by disproportionation
Emulsion Polymerization: Kinetic and Mechanistic Aspects 91

As the particle size increases, the bimolecular termination rate decreases. For
cases with chain-length independent bimolecular termination, the oligomeric
peak in the number fraction distribution moves toward larger chain length, as
shown in Fig. 12. On the other hand, if the bimolecular terminations are highly
diffusion controlled, the oligomeric peak location does not move, but the peak
height becomes smaller due to an increased amount of dead polymer formation
from chain transfer to the monomer, which is accounted for in this simulation.
Note that we should not expect these oligomeric peaks to be detected via usual
SEC analysis, represented on the basis of the weight fraction distribution, as
shown in the lower panel of Fig. 13. The termination mode may not be distin-
guished from the SEC data.
The chain-length-dependence of bimolecular termination reactions needs
to be taken into account in order to be able to accurately estimate the MWD
formed, except when very small polymer particles are formed and/or chain
transfer reactions dominate over dead polymer chain formation.

4.1.4
Accumulated Molecular Weight Distribution

In a deterministic approach, the instantaneous MWDs are calculated first, and

then the accumulated MWD is obtained by integrating the instantaneous
MWDs. However, in MC simulations, we can follow the reaction history of each
polymer particle directly, and the full MWD is obtained by simulating a large
number of polymer particles [263]. Therefore, highly complex reaction kinetics
can be simulated directly in a straightforward manner.
During emulsion polymerization, the polymer concentration in the polymer
particle that is the locus of polymer chain formation is larger than that in cor-
responding bulk polymerization. Therefore, the possibility of chain transfer re-
actions to the polymer occurring is higher, even when the monomer conversion
to polymer is not very large. Besides these branching reactions (which will be
discussed in Sect. 4.2), another accidental branching may occur during emul-
sion polymerization. In emulsion polymerization, the time interval between
radical entries is usually large, and chain transfer to monomer tends to be the
dominant chain termination mode, in the absence of other chain transfer
agents. Depending on the mechanism of the chain transfer reaction, active
terminal double bonds may be formed by the monomer transfer reaction,
which may lead to terminal double bond polymerization (TDBP) [263]. Active
terminal double bonds may not be formed during styrene polymerization
[293]; however, in order to investigate the potential importance of the TDBP
in general, a simulation that incorporates TDBP was conducted using kinetic
parameters for the styrene polymerization.
Figure 14 shows the development of the weight fraction distribution with
and without TDBP [263].
The parameter K shown in the figure represents the reactivity of the termi-
nal double bonds, defined by K=k¢p /kp, and k¢p is the rate constant of TDBP. The
92 M. Nomura et al.

Fig. 14 Accumulated weight fraction distribution development with and without terminal
double bond polymerization

average time interval between entries for radicals generated in the water phase
(excluding reentry of desorped radicals) is 50 s. The value of np is the total num-
ber of monomeric units bound into polymer chains in a polymer particle, and
therefore, np increases as the reaction proceeds. In the simulation, the poly-
mer/monomer ratio in the polymer particle is kept constant. Both with and
without TDBP, the MWD profiles do not change significantly during polymer-
ization (as long as monomer droplets exist); however, much larger polymer
molecules can be formed by the TDBP and this can change the MWD profile
significantly. In emulsion polymerization, the enhanced accidental branching
caused by a higher polymer concentration cannot be neglected, even for reac-
tion systems in which the branching reactions are not significant in the bulk
polymerization.

4.1.5
Determination of Monomer Transfer Constants from MWD

The traditional method of determining the monomer transfer constant Cm is

the Mayo method [294, 295], where the inverse of the number average chain
–
length Pn is extrapolated to zero polymerization rate. To obtain reliable Cm
–
values, one needs to measure rather large Pn values to high precision that can
then be extrapolated to zero polymerization rate. In addition, linear extrapo-
lation is not guaranteed if bimolecular termination reactions are chain-length-
dependent [296].
A simple alternative method was proposed by Gilbert et al. [296, 297] to
determine the chain transfer constants based on the chain length distribution
(CLD). If the dominant chain termination mechanism is chain transfer to
monomer, the instantaneous numerical MWD (the number fraction distribu-
tion) is given by:
Emulsion Polymerization: Kinetic and Mechanistic Aspects 93

N(P) = Cm exp (– CmP) (84)

In many emulsion polymerizations, the monomer/polymer ratio is kept con-
stant during Interval II, and the accumulated MWD is approximately equal to
the instantaneous distribution. Equation 84 shows that the Cm value can be de-
termined from the slope of the lnN(P) versus P plot.
In a zero-one system in which a radical has just entered a polymer particle
containing one polymer radical, and is terminated instantaneously, the num-
ber fraction distribution in the absence of a polymer transfer reaction is given
by:
N(P) = (Cm + e ) exp {– (Cm + e )P} (85)

where e =1/(kp[M]p – te), as given by Eq. 73.

It is evident from Eq. 85 that the condition e =1/(kp[M]p–
te) O Cm is needed to
apply the CLD method to emulsion polymerization. Note that the radical entry
rate may be increased through the radical exit. Even when these conditions
are satisfied, a higher polymer concentration than for the corresponding bulk
polymerization may result in more occurrences of the polymer transfer reac-
tion.
The effect of the polymer transfer reaction on the applicability of the CLD
method can be examined by applying the MC simulation method [298].
Figure 15 shows the MC simulation results for the condition Cm=5¥10–5, e =
1/(kp[M]p– te) =5¥10–7, and for the polymer transfer constant Cfp=5¥10–5, when
the total number of monomeric units bound into polymer chains in a polymer
particle np=1¥106.
The upper panel of Fig. 15 shows the weight fraction distribution obtained by
taking the logarithm of chain length as an independent variable, as in an SEC
analysis. The distribution is clearly much broader than without the polymer
transfer reactions, as shown in [298]. The lower panel shows the plot of ln(N(P))
versus P, which is clearly curved although a straight regression line could be
drawn around the peak region of the W(logP) curve. The slope obtained for this
case is –5.102¥10–5, and therefore e+Cm=5.102¥10–5. If we use e=5¥10–7, we
obtain Cm=5.052¥10–5, which is sufficiently close to the true monomer transfer
constant, 5¥10–5. Note that the MC simulation inevitably involves a small
amount of statistical error, and the slope changes slightly if the same simula-
tion is repeated. However, an important conclusion from these kinds of simu-
lations [298] is that, although the MWD is affected significantly by the polymer
transfer reactions, the CLD method is still considered applicable as a reason-
able approximation for many systems. The Cm value can be estimated reason-
ably well from the plot of ln(N(P)) versus P, by taking the slope around the
peak region of the W(logP) curve, as long as the polymer transfer constant is
not too large.
The MC simulation method can be used to find the experimental conditions
where the CLD method can be used to determine the monomer transfer con-
stant.
94 M. Nomura et al.

Fig. 15 Monte Carlo simulation results for emulsion polymerization that involves polymer
transfer reactions, under the conditions Cm=5¥10–5, e =5¥10–7 and Cfp=5¥10–5, without
radical desorption

4.2
Branched and Crosslinked Polymer Formation

4.2.1
Long-Chain Branched Polymers

Nonlinear polymer formation in emulsion polymerization is a challenging

topic. Reaction mechanisms that form long-chain branching in free-radical
polymerizations include chain transfer to the polymer and terminal double
bond polymerization. Polymerization reactions that involve multifunctional
monomers such as vinyl/divinyl copolymerization reactions are discussed
separately in Sect. 4.2.2. For simplicity, in this section we assume that both
the radicals and the polymer molecules that formed are distributed homoge-
neously inside the polymer particle.
In nonlinear emulsion polymerization, a comprehensive mathematical model
must account for the following unique characteristics of emulsion polymeriza-
Emulsion Polymerization: Kinetic and Mechanistic Aspects 95

tion: (a) compartmentalization of radicals, (b) higher polymer concentration

effects, and (c) limited space effects.

4.2.1.1
Compartmentalization of Radicals

Compartmentalization of radicals into polymer particles may yield a unique

MWD for the linear chains, as discussed in Sect. 3.1, except when the dominant
chain termination mode is the chain transfer reaction. Branched polymer
molecules are assemblies of linear polymer chains (called primary chains), and
compartmentalization effects on the primary chain length distribution must be
properly accounted for.
The compartmentalization of radicals may produce another important effect
when large-sized branched polymer molecules are formed by chain transfer to
polymer plus combination termination. As clarified in Sect. 4.1, when the – n
value is small, the frequency of bimolecular termination reactions between
large polymer radicals drops significantly compared to models that do not ac-
count for compartmentalization of radicals. From this fact, it is easy to see that
the size of branched polymer molecule is smaller than that calculated without
considering compartmentalization effects [281].

4.2.1.2
Higher Polymer Concentration Effects

The mechanism of emulsion polymerization ensures that the polymer con-

centration at the polymerization locus is semidilute or concentrated, which
results in a greater probability of branch chain formation [299, 300]. This
effect produces unique average branching densities and unique distributions
of branching densities, that are significantly different from corresponding bulk
polymerization [266, 301].
Figure 16a shows the development of average branching density in emulsion
polymerization and in a corresponding bulk polymerization, both of which
involve the polymer transfer reactions [301].
In bulk polymerization, the average branching density increases with the as
the polymerization progresses, while it is fairly high even from a very early stage
of the emulsion polymerization. This difference in behavior can be explained by
the different polymer concentrations at the locus of each type of polymeriza-
tion. In bulk polymerization, the polymer concentration at low conversions is
very low. Because the branches are formed by the reaction between a polymer
molecule and a polymer radical, a lower polymer concentration results in a
lower frequency of branch formation and the branching density increases as
the polymer concentration increases. On the other hand, in emulsion poly-
merization, the polymer concentration at the locus of polymerization (in the
polymer particle) is high, even just after the formation of polymer particles.
A higher polymer concentration results in a higher branching reaction rate. If
96 M. Nomura et al.

Fig. 16 Development of the average branching density a and the branching density distri-
bution b for emulsion and bulk polymerizations. The conversion at which monomer droplets
disappear in emulsion polymerization is xc=0.5

the polymer concentration stays the same during the lifetimes of the monomer
droplets, the ratio of the branching and propagation reaction rates is kept
constant, resulting in a constant average branching density until the deple-
tion of the monomer droplets, as shown in Fig. 16a. Emulsion polymerizations
enhance the frequency of branching and crosslinking reactions due to the
higher polymer concentration associated with the existence of monomer
droplets.
Figure 16b shows the development of the branching density distribution,
where Çb(q,y) shows the expected branching density of the primary polymer
molecule born at conversion x=q, when the conversion at the present time
is x=y. The primary chains formed in the early stages of polymerization are
subjected to branching reactions for a longer period of time, and therefore, the
expected branching density is higher than those chains formed in the later
Emulsion Polymerization: Kinetic and Mechanistic Aspects 97

stages of polymerization. This is the reason that we obtain decreasing functions

for both emulsion and bulk polymerization. However, the heterogeneity of the
branched structure is more significant for emulsion polymers, as shown in
Fig. 16b. Note that the fact that the average branching density does not change
as long as monomer droplets exist, as shown in Fig. 16a, does not mean that
a homogeneous branched structure is formed during that period.

4.2.1.3
Limited Space Effects

Considerations of “radical compartmentalization” and “higher polymer con-

centration effects” are not sufficient to describe the processes that build branched
polymer molecules in emulsion polymerization, and the effects of limited space
must be properly taken into account [266–269].
The locus of polymerization is confined to a very small space, which not
only limits the highest molecular weight attainable but can also change the
whole MWD profile significantly. The MC simulation technique [266–269]
is the only method that can currently take the effects of limited space into
account. If the effects of limited space are ignored, the calculated molecular
weight may exceed the molecular weight of a whole polymer particle. For ex-
ample, when the crosslinks – the bridges that connect chains – are formed, con-
ventional deterministic models that do not account for the effects of limited
space may predict that the second order moment of the MWD goes to infinity,
which is clearly wrong, because the maximum molecular weight is limited by
the particle size. In general, those models that do not include the particle size
as a parameter when describing the formation of branched chains are illogical.
The limited space effects present rather difficult problems to account for in
the conventional deterministic approach. The problem can be highlighted by
considering the following hypothetical example. Suppose that there are three
polymer chains and three radicals (R·), and that polymer transfer reactions are
about to occur [269]. Within these polymer chains, suppose that one chain is
much larger than the other two, as shown in Fig. 17.

Fig. 17 Hypothetical example that illustrates the differences between the polymer transfer
reactions that occur in bulk polymerization a and in emulsion polymerization b
98 M. Nomura et al.

If all of these species exist in the same reaction locus as in Fig. 17a, it would
be highly probable that all of the radicals would attack the largest chain. In
other words, the chain transfer rate of the polymer chain with chain length P,
vfp,Pp, is proportional to its chain length:
vfp,PP = kfp[R*] r[PP] (86)
where kfp is the rate constant for chain transfer to polymer, [R·] is the total
radical concentration, and [PP] is the concentration of polymer molecules with
chain length P in the whole reaction mixture.
On the other hand, suppose that each of these polymer molecules is isolated
into different particles, and that each particle contains one radical, as shown
in Fig. 17b. If the radical causes the polymer transfer reaction, the partner
must be the polymer molecule that happens to exist in the same particle (so it
cannot partner a larger polymer molecule that exists in a different polymer
particle).As a consequence, the expected size of the polymer molecule attacked
by a radical is smaller for emulsion systems than for the homogeneous model
shown in Fig. 17a.
Equation 86 is commonly used for homogeneous reaction systems, but it is
not exact in emulsion polymerization. The value of [PP] is different for each
polymer particle, and the value obtained for [PP] when all of the particles are
combined cannot be used either. Strictly, one needs to determine a discrete
distribution function of polymer molecules in each polymer particle.
Figure 18 shows the simulated MWD profiles that clearly demonstrate the
effects of limited space [273].
Because the total number of monomeric units bound into polymer molecules
is np=4¥105 for the present calculation, the high molecular tail can never exceed
4¥105. Figure 18 shows that the model that does not account for the effects of

Fig. 18 Comparison of the calculated weight fraction distribution with Cf=Cfp=5¥10–4 and
xc=0.5. For the emulsion polymerization model, the total number of polymerized monomeric
units in a polymer particle np=4¥105, which is equal to the size of a dried polymer particle
Emulsion Polymerization: Kinetic and Mechanistic Aspects 99

limited space (the macro-reactor model) is incorrect, because the formation of

polymer molecules that are too large to fit in a polymer particle is predicted.
More details on the effects of limited space can be found in [268, 269].
The MC simulation method can account for the effects of limited space in
a straightforward manner, because the kinetics of polymer formation inside
each polymer particle is simulated directly in the MC method. The analytical
solution for the development of the weight-average DP was also derived for
a simpler case [268], but for more detailed information one needs to resort to
the MC method. In MC simulations, one can investigate the structure of each
polymer molecule directly, so that highly detailed information can be ob-
tained. Figure 19 shows an example of the branched structure formed during
emulsion polymerization that involves chain transfer to polymer, where the
primary chains follow the most probable distribution with a number average
of 1000.
As we can see, a rather large number of smaller branches exist, which is ob-
viously not what we might expect from the term long-chain branches! The 3-D

Fig. 19 Example of a branched polymer molecule formed in a model emulsion polymer-

ization. The probability that the chain end is connected to a backbone chain is Pb=0.7. The
primary chains follow the most probable distribution with a number-average chain length
of 1000
100 M. Nomura et al.

structure of each nonlinear polymer molecule can be estimated, most simply

in a q solvent, using the structural information shown in Fig. 19. By determin-
ing the hydrodynamic size of each polymer molecule, we can also estimate the
size exclusion chromatography (SEC) elution curve [302–306].

4.2.1.4
Formation of the Bimodal Molecular Weight Distribution

The bimodal MWDs of emulsion-polymerized polyethylenes, which are sig-

nificantly different from those formed in bulk polymerization, have been
reported experimentally [307–309]. An MC simulation was conducted for the
experimental conditions reported in [307], and an example is shown in Fig. 20
[310].
The kinetic parameters used were mostly taken from the literature, and an
important assumption made was that the particle diameter is about 80 nm
(=1.4¥108 g/mol in molecular weight, which is shown by an arrow in Fig. 20). In
[307], it was reported that (1) the particle size is about 50 nm, and that (2) the
weight of such a particle is close to the molecular weight of the high molecular
weight, narrow distribution component. However, the weight of a low-density
polyethylene particle 50 nm in diameter is 6¥10–17 g, which is 3.6¥107 g/mol in
molecular weight. This molecular weight is too small to contain the polymers in
the high molecular weight tail, whose molecular weight can be as large as
8¥107 g/mol, and so the value of 1.4¥108 g/mol (≈80 nm in diameter) was used
in the MC simulation.
According to the MC simulation, the high molecular weight, narrow distri-
bution component consists of the largest polymer molecule in each polymer
particle, and the bimodal MWD is formed because of the limited space effects.
Assuming a simple zero-one system during Interval II, a model analysis was
conducted to clarify the conditions needed to form bimodal MWD through the
effects of limited space [311].

Fig. 20 MWD of emulsion-polymerized polyethylene. The experimental conditions are

discussed in detail in [307], and the simulation method is described in [310]
Emulsion Polymerization: Kinetic and Mechanistic Aspects 101

The instantaneous MWD of the primary chains formed during Interval II in

a zero-one system (assuming combination termination) is given by the most
probable distribution, whose number fraction distribution is given by:

Npc(P) = t exp (– tP) (87)

xc [CTA]p 1
t = Cm + Cfp 93 + CfCTA + 75 + 753 (88)
(1 – xc) [M]p kp[M]p–
te

where xc is the weight fraction of polymer in the polymer particle, which is

kept approximately constant during Interval II, and is often equal to XMc.
The probability that a newly-formed primary chain starts growing from
a radical center on a backbone chain (in other words, the probability that a
primary chain end is connected to a backbone chain), Pb, is given by:

Cfp xc /(1 – xc )
Pb = 9999999998 (89)
Cm + Cfp xc /(1 – xc ) + CfCTA[CTA]p /[M]p + 1/(kp[M]p–
te)

A model analysis was conducted, which assumed that (1) both t and Pb do not
change during polymerization, (2) all polymer particles are formed instanta-
neously, and (3) the number of primary chains in a polymer particle, npc, is
the same for all particles. Important conclusions were that (i) bimodal MWDs
(represented in terms of W(logP)) are formed if Pb is larger than 0.5, and (ii) for
Pb>0.5, the weight-average molecular weight increases without limits over the
whole course of polymerization. Note that the second conclusion does not
indicate that gelation occurs. Figure 21 shows the calculated development of
the weight-average chain length [311].
For Pb<0.5, the weight-average chain length reaches a constant value that
– – –
is given by Pw = Pwp /(1 – 2Pb), where Pwp is the weight-average chain length of
the primary chains. On the other hand, the weight-average molecular weight
increases without limit for Pb>0.5, but it increases very gradually and it takes
an infinitely long time to reach an infinitely large polymer molecule. It was
found [313] that the formed MWD is a power-law distribution [312] that pos-
sesses fractal characteristics is formed.
For the emulsion-polymerized polyethylene shown in Fig. 20, which clearly
shows a bimodal MWD, Pb=0.813>0.5. For another example investigated in
[311], Pb=0.711>0.5, and the MWD is bimodal.
On the other hand, for the emulsion polymerization of vinyl acetate, Friis
et al. reported that (i) the weight-average molecular weight does not increase
significantly until the monomer droplets are depleted, and that (ii) the MWDs
are unimodal [314, 315]. They considered the TDBP in addition to the polymer
transfer reactions; however, the contribution of the TDBP is minor and a qual-
itative discussion on the MWD shape could be made without needing to involve
the TDBP. The parameters they used are Cm=2.32¥10–4, Cfp=3.98¥10–4 and
xc=0.2, which gives Pb=0.3.With Pb=0.3, the theoretical analysis in [311] showed
 102

Fig. 21 Calculated development of the weight-average chain length during the model emulsion polymerization
M. Nomura et al.
Emulsion Polymerization: Kinetic and Mechanistic Aspects 103

that (i) the weight-average chain length reaches a steady state value rather
quickly, and that (ii) the MWDs formed are unimodal, both of which agree with
experimental observations.
Although the kinetic behavior during Interval III was not considered in
[311], Pb=0.5 is an important consideration when we look at the possibility of
forming bimodal MWDs in emulsion polymerization that involves chain trans-
fer to polymer.

4.2.2
Crosslinked Polymers

4.2.2.1
Crosslinked Structure

Assuming that classical chemical kinetics are valid and that the crosslinking
reaction rate is proportional to the concentrations of polymer radicals and
pendant double bonds, it was shown theoretically that the crosslinked polymer
formation in emulsion polymerization differs significantly from that in cor-
responding bulk systems [270, 316]. To simplify the discussion, it is assumed
here that the comonomer composition in the polymer particles is the same as
the overall composition in the reactor, and that the weight fraction of polymer
in the polymer particle is constant as long as the monomer droplets exist. These
conditions may be considered a reasonable approximation to many systems, as
shown both theoretically [316] and experimentally [271, 317]. First, consider
Flory’s simplifying assumptions for vinyl/divinyl copolymerization [318]; that:
(1) the reactivities of all types of double bonds are equal, (2) all double bonds

Fig. 22 Crosslinking density distribution development during bulk and emulsion poly-
merization under Flory’s simplifying assumptions, where the initial mole fraction of divinyl
monomer is 0.01, and xc=0.4 for emulsion polymerization
104 M. Nomura et al.

react independently, and (3) there are no cyclization reactions in finite mole-
cules. Under these simplifying assumptions, a completely homogeneous net-
work is formed in homogeneous batch polymerization [319].
Figure 22 shows the calculated results for the expected crosslinking density
of primary chains formed at conversion q when the conversion at the present
time is y, under Flory’s simplifying assumptions for bulk and emulsion copoly-
merization.
The expected crosslinking density is the same for all primary chains at any
stage of polymerization in a bulk system, indicating that a statistically homo-
geneous network is formed. On the other hand, the crosslinking density dis-
tribution is completely different in emulsion polymerization. In particular,
the crosslinking densities of the primary chains formed in the earlier stages
of polymerization are very high, and the variance of the crosslinking density
distribution is significant in emulsion polymerization. The bending at q=0.4 oc-
curs because the conversion at which monomer droplets disappear is assumed
to be 0.4 in these calculations. Homogeneous networks cannot be formed, even
under Flory’s simplifying assumptions in emulsion crosslinking copolymer-
ization.
Figure 23 shows the average crosslinking density development under Flory’s
simplifying assumptions.
The average crosslinking density is higher for emulsion polymerization due
to the higher polymer concentration effects inside the polymer particles, which
are the loci of polymerization. The fact that the average crosslinking density
is high even from a very early stage of polymerization has been confirmed
experimentally, and this agrees with theoretical calculations adequately [271,

Fig. 23 Average crosslinking density development under Flory’s simplifying assumptions

Emulsion Polymerization: Kinetic and Mechanistic Aspects 105

272, 317]. The fact that the average crosslinking densities in some emulsion
polymerizations may not change much during emulsion polymerization, as
shown in Fig. 23, is sometimes misunderstood as producing a homogeneous
polymer network [320]. However, this argument may not be true, as shown in
the crosslinking density distribution profile shown in Fig. 22.

4.2.2.2
Unique Molecular Weight Distributions

It has long been recognized that microgels formed in emulsion polymerization

possess only supermolecular size and weight [321]. It is known that a reaction
system with divinyl monomer makes the particle size smaller compared to the
cases without crosslinker [97, 322].Assuming that the final particle diameter is
50 nm, the molecular weight of this particle is about 4¥107 g/mol. Such micro-
gels could be soluble in a good solvent, and the MWD could be determined by

Fig. 24 Experimental results for the weight-average molecular weight development during
the emulsion crosslinking copolymerization of styrene (St)/divinylbenzene (DVB) and
methyl methacrylate (MMA)/ethylene glycol dimethacrylate (EGDMA). Data from [323]
106 M. Nomura et al.

using SEC even when each polymer particle forms a single crosslinked polymer
molecule.
In a typical recipe for microgel formation, a sufficient amount of crosslinker
is used. Because the crosslinking density tends to be high from a very early
stage of polymerization, the MC simulation results showed [270, 271] that each
polymer particle tends to consist of a single large crosslinked polymer mole-
cule even from the early stage of polymerization. In such cases, it was shown
[270, 271] that (1) the weight-average molecular weight grows linearly with con-
version, and (2) the MWD formed is narrow and the distribution shifts to larger
molecular weights, preserving the narrow profile, as polymerization proceeds.
Experimental data [323–325] in which the MWDs are measured by SEC with
on-line multiangular laser light scattering (MALLS) show the above trend [326].
Figure 24 shows some examples of the weight-average molecular weight de-
velopment, which show linear increases with respect to conversion.
The y-intercept might approximately correspond to the weight-average
molecular weight of primary chains. It was further confirmed [326] that the
MWD of polymer molecules and that of whole polymer particles are the same,
showing that each polymer particle essentially consists of one crosslinked
polymer molecule. Figure 25 shows such an example, and the MWD formed
agrees reasonably well with the particle size distribution represented in terms
of molecular weights.
On the basis of the MC simulation results [270], it is expected that if the
amount of divinyl monomer is reduced to the level of, say, several crosslinkers
per primary chain, a bimodal MWD (as shown in Fig. 26) may result.
The arrows indicate the molecular weight of a dried polymer particle, which
gives an upper limit to the molecular weight of the polymers attainable. Qual-
itatively speaking, the high molecular weight peaks are formed because the
crosslinked polymer molecules want to grow further due to the higher chain

Fig. 25 MWD of polymer molecules and dried polymer particles in the emulsion copoly-
merization of St and DVB (20 mol%) [326]
Emulsion Polymerization: Kinetic and Mechanistic Aspects 107

Fig. 26 MC simulation results that show bimodal MWDs in the emulsion copolymerization
of vinyl/divinyl monomers [270]

Fig. 27 Experimentally obtained MWD developments of polymer molecules and particles

for the emulsion copolymerization of St and DVB [327]
108 M. Nomura et al.

connectivity, but they cannot because of the limitation of a small particle size.
Experimentally, such unique bimodal distributions were obtained in the
styrene/divinyl benzene system [327] by setting the number of divinylbenzene
molecules per chain to ~5. An example of a bimodal MWD, together with the
particle size distribution [327], is shown in Fig. 27.
It was found that the locations of these peaks can be controlled indepen-
dently. The location of a high molecular weight peak is mainly controlled by the
particle size, and the location of a low molecular weight peak is controlled by
the chain lengths of the primary polymer molecules.
The MC simulation method is particularly suitable for investigating emul-
sion polymerization that involves various simultaneous kinetic events with a
very small locus of polymerization. The MC simulation method will become a
standard mathematical tool for the analysis of complex reaction kinetics, both
for linear and nonlinear emulsion (co)polymerization.

5
Continuous Emulsion Polymerization

Continuous emulsion polymerization processes are industrially important for

the large-scale production of synthetic polymer latexes, and have been used
particularly where the solid polymer is to be recovered by coagulating the
polymer latex. St-Bu rubber latex was one of the earliest latex products manu-
factured using continuous emulsion polymerization processes consisting of a
number of stirred-tank reactors in series (CSTRs). Since the 1940s, continuous
emulsion polymerization processes have been developed for a variety of prod-
ucts and with different reactor configurations [328]. This is because these con-
tinuous reactor systems have several advantages, such as [329]:
– Economical production of large-volume or closely-related products
– Uniform product quality
– Full utilization of heat transfer capability
– Fewer problems with wall polymer build-up and coagulation.
However, these systems have also potential disadvantages, such as [329]:
– Less flexibility in terms of the operation and control of product character-
istics
– Possible production of off-specification material during start-up or product
change-overs
– Difficulties with the direct development of continuous processes based on
the information from batch and semi-batch R&D.
Reflecting the importance of continuous emulsion polymerization processes,
numerous investigations have been carried out to date, which are categorized
into three groups: (1) studies on the reactor configuration (stirred-tank reactors,
tubular type reactors such as a simple tubular reactors, pulsed tubular reactors
Emulsion Polymerization: Kinetic and Mechanistic Aspects 109

and loop-tubular reactors, pulsed packed column reactors, Couett-Taylor vor-

tex flow reactors and combinations of these reactors); (2) studies on operational
techniques (pre-reactor concept [330] and split-feed operation [331–333]), and;
(3) studies on the kinetics and mechanisms of particle formation and growth in
a given reactor system. Research work carried out before ~1990 was introduced
in a compact but excellent review article [328], which also included those cited
in a review article published in 1977 [334]. The present review, therefore, mainly
refers to the research work that has been published since ~1990, but also in-
cludes any historically important works reported before ~1990.
The stirred-tank reactor and the tubular reactor are two basic reactors used
for continuous processes, so much of the experimental and theoretical studies
published to date on continuous emulsion polymerization have been conducted
using these reactors. The most important elements in the theory of continuous
emulsion polymerization in a stirred-tank reactor or in stirred-tank reactor
trains were presented by Gershberg and Longfield [330]. They started with the
S-E theory for particle formation (Case B), employing the same assumptions as
stated in Sect. 3.3, and proposed the balance equation describing the steady-state
number of polymer particles produced as:
dNT/dt = Çw[Am/(Am + Ap)] – NT/q = 0 (90)
where q is the mean residence time of a single CSTR. Introducing the relation,
Am+Ap=asS0 into Eq. 90 yields
NT = Çw[1 – (Ap/asS0)] (91)
where S0 is the emulsifier concentration in the feed. The residence time distri-
bution for polymer particles in a perfectly mixed CSTR is given by E(t)dt=
dNT/NT=(1/q)exp(–t/q)dt. The total surface area of polymer particles per unit
volume of water Ap is obtained by the integration
NT NT ∞
Ap = ∫ apdNT = (36p)1/3 ∫ v p2/3dNT = (36p)1/3NT ∫ (mt)2/3E(t) dt (92)
0 0 0

where ap is the surface area of a polymer particle and E(t) is the residence time
distribution function. Combining Eqs. 91 and 92 gives
Çwq
NT = 99984 (93)
1 + (4.36 Çw m 2/3q 5/3asS0)
Thus, the rate of polymerization Rp can be predicted using Eqs. 1, 51 and 93.
Omi et al. [335] and Nomura et al. [163] pointed out that Eq. 93 suggests the
existence of a maximum number of polymer particles NT,max at the optimum
residence time qmax, which is given by
qmax = 0.53(asS0/Çwm 2/3)3/5 (94)
NT,max = 0.21(Çw /m )0.4(aSS0)0.6 (95)
110 M. Nomura et al.

By comparing Eq. 31 with Eq. 95, they also pointed out that NT,max is only 58%
of that produced in a batch operation with the same recipe and temperature,
and also that
qmax = 0.83tc (96)
where tc is the time at which particle formation stops due to complete deple-
tion of micelles in the water phase, in a batch operation with the same recipe
as that used in a continuous operation.
Gershberg and Longfield [330] carried out the continuous emulsion poly-
merization of St at 70 and 100 °C in a train of three CSTRs, and compared the
experimental results with theoretical predictions. They presented the following
conclusions and suggestions;
1. The theoretical equation NTµRpµÇw0S 01.0q–2/3, derived from Eqs. 1 and 93,
correctly predicts the effect of the operating variables upon the rate of poly-
merization in the stirred-tank reactors at any stage at 70 °C for large q
values.
2. The number of polymer particles produced at 70 °C (and hence, the rate of
polymerization) was higher than that at 100 °C, as Eq. 93 predicted.
3. The theory and the experimental data from this study demonstrates that in
a train of CSTRs, essentially all of the particles form in the first reactor.
Therefore, it is possible to maximize the monomer conversion in the latex
leaving the first reactor by keeping the temperature and the residence time
at the first reactor as low as possible in order to produce the maximum num-
ber of polymer particles and so increase the rate of polymerization in the
succeeding stages. This is the so-called “pre-reactor” concept.
4. NT,max certainly occurs at low q, as the theory predicts.
5. Sustained oscillation can take place in the monomer conversion.
6. The residence time in the first stage should be long enough to overcome the
retarding effects of traces of oxygen and/or impurities in the feed-stream
upon particle formation and growth.
Poehlein and Degraff [336] extended the derivation of Gershberg and Long-
field [330] to the calculation of both molecular weight and particle size distri-
bution in the continuous emulsion polymerization of St in a CSTR. On the
other hand, Nomura et al. [163] carried out the continuous emulsion poly-
merization of St in a cascade of two CSTRs and developed a novel model for the
system by incorporating their batch model [14], which introduced the concept
that the radical capture efficiency of a micelle relative to a polymer particle
was much lower than that predicted by the diffusion entry model (Ç µd1.0).
The assumptions employed were almost the same as those of Smith and
Ewart (Sect. 3.3), except that the model did not assume a constant value of m.
The elementary reactions and their rate expressions employed in the first stage
are as follows:
Emulsion Polymerization: Kinetic and Mechanistic Aspects 111

1. Particle formation by radical entry into a micelle (Eq. 32):

R* + ms Æ N*, kemms [R*w] (32)
2. Formation of dead particles by entry of a radical into an active particle con-
taining a radical (Eq. 33):
R* + N* Æ N0, kep N*[R*w] (33)
3. Formation of active particles by entry of a radical into a dead particle con-
taining no radical (Eq. 34):
R* + N0 Æ N*, kep N0[R*w] (34)
where ms is the concentration of monomer-swollen micelles, [Rw*] is the con-
centration of free radicals in the water phase, N* is the concentration of active
particles containing a radical, N0 is the concentration of dead particles that con-
tain no radicals, kem is the rate constant for the entry of radicals into micelles,
and kep is the rate constant for the entry of radicals into particles. If these rate
expressions are used, then the balance equation describing the steady-state
number of polymer particles produced becomes
dNT/dt = Çw[k1msR*/(k1msR* + k2NT R*)] – NT/q = 0 (97)
where NT=N* + N0. Then Eq. 97 is rearranged to give
NT = Çwq/[1 + (k2NT/k1ms)] = Çwq/(1 + eNT/Sm) (98)
where e=(k2/k1)Mm; Mm is the aggregation number per micelle and Sm is the
concentration of emulsifier forming micelles, given by Sm=S0–(Ap/as). Rear-
ranging Eq. 98 along with the steady-state balance equations for the monomer
and the number of active particles finally yields
eNT2/(Çwq – NT) = S0 – kv (KM0 /2)[Çwq/(Çwq – NT/2)]2/3NTq 2/3 (99)
where kv and K are the constants associated with the adsorption area of the
emulsifier on the particles’ surface and the particle growth rate, respectively,
and M0 is the concentration of the monomer in the feed. Equation 99 can be
used to calculate the relationship between NT and q, and that between NT,max
and qmax.A comparison of the experimental data and theoretical predictions for
the dependence of NT on q is presented in Fig. 28 [163, 331].
The experimental data showed much better agreement with the results
predicted using Eq. 99 than the results calculated using the Gershberg and
Longfield model (Eq. 93), which gave a higher value of NT,max and a smaller
value of qmax. The reason for this may be that the rate of radical entry into a
micelle must be less than that predicted by the collision entry model, given by
Çw[d m2 /(dm2 + dp2)].
Gerrens and Kuchner [337] investigated the continuous emulsion polymer-
ization kinetics in a cascade of three CSTRs, and with St and MA as monomers
with different solubilities in water. They showed that the experimental results
obtained with St agreed with the predictions from the Gershberg and Longfield
112 M. Nomura et al.

Fig. 28 Effect of the mean residence time in the first reactor q on the number of polymer
particles produced (S0 (NaLS)=12.5 g/dm3-water, I0 (KPS)=1.25 g/dm3-water, M0 (St)=
500 g/dm3-water; 50 °C. Experimental data: empty circles, first reactor; filled circles, second
reactor

model, but for MA only the particle number varied in the manner predicted by
the Gershberg and Longfield model. This showed that a relatively water-solu-
ble monomer does not correlate with the model based on S-E theory. The
existence of multiple steady-states for the isothermal operation of continuous
emulsion polymerization in a CSTR was first demonstrated by Gerrens et al.
[338, 339]. It is now well understood from a kinetic point of view that this
phenomenon takes place as a consequence of the so-called “gel-effect” (the
Trommsdorff-Norrish effect).
Oscillations in the number of polymer particles, the monomer conversion,
and the molecular weight of the polymers produced, which are mainly observed
in a CSTR, have attracted considerable interest. Therefore, many experimental
and theoretical studies dealing with these oscillations have been published
[328]. Recently, Nomura et al. [340] conducted an extensive experimental study
on the oscillatory behavior of the continuous emulsion polymerization of VAc
in a single CSTR. Several researchers have proposed mathematical models
that quantitatively describe complete kinetics, including oscillatory behavior
[341–343]. Tauer and Müller [344] proposed a simple mathematical model for
the continuous emulsion polymerization of VCl to explain the sustained oscil-
lations observed. Their numerical analysis showed that the oscillations depend
on the rates of particle growth and coalescence. However, it still seems to be
difficult to quantitatively describe the kinetic behavior (including oscillations)
of the continuous emulsion polymerization of monomers, especially those with
relatively high solubility in water. This is mainly because the kinetics and mech-
Emulsion Polymerization: Kinetic and Mechanistic Aspects 113

anisms of particle formation in systems containing such monomers are not

completely understood yet. It is now known that oscillations in the monomer
conversion and the molecular weight of the polymer produced occur due to
oscillations in the number of polymer particles. Since oscillatory behavior is
undesirable in practice from the standpoint of stable operation and product
quality problems, a variety of techniques have been proposed to eliminate
the oscillations in the number of polymer particles. The simplest but most
effective technique among them would be to add a small continuous tubular
type reactor – a pre-reactor with plug flow, operating as a seed generator (a
seeding reactor), located upstream of the main CSTRs [331, 345].
To meet the growing demand for the large-scale production of latex with
narrow PSD and consistent quality, it was initially recognized that continuous
emulsion polymerization in a tubular reactor might be advantageous from the
standpoint of higher performance (compared to a CSTR), greater heat trans-
fer, better quality control and lower equipment costs [346]. Nevertheless, very
little research work has been done with tubular reactors compared to CSTRs.
Feldon et al. [346] successfully performed the continuous emulsion copoly-
merization of St and Bu in a continuous tubular reactor. It was originally
thought that the polymerization needed to be conducted in the turbulent flow
regime in order to obtain satisfactory heat transfer and mixing. However, it was
found that this turbulent flow gave rise to the formation of a pre-coagulum,
which resulted in an accumulation of polymer particles on the reactor wall
which eventually plugged the reactor. Therefore, the stable and long-term
operation of a continuous tubular reactor (CTR) is usually a very difficult task
[346]. In addition, when the rate of polymerization is low, as it often is in such
reactors, a very long tube is necessary to achieve high monomer conversion.
This is not practical because it is difficult to transport a highly viscous latex
product through a long tube.
In order to make full use of the advantages of a tubular reactor while avoid-
ing its drawbacks, the use of a continuous loop-tubular reactor (CLTR) has
been proposed. Background on this discovery, as well as a discussion of the
main characteristics of the continuous loop-tubular reactor, was first provided
by Geddes [347]. He studied the changes in properties during emulsion copoly-
merization in a continuous loop reactor, with particular reference to particle
size distribution. Cycling of properties similar to that observed in CSTRs was
found [348]. Bataille and co-workers [349, 350] carried out the emulsion ho-
mopolymerizations of St and VAc using a batch loop-tubular reactor. They
found that a limiting conversion, lower than that obtained in a stirred-tank
batch reactor, occurred for all cases, and that the value of the limiting conver-
sion was a maximum at the laminar-turbulent transition point. Lee et al.
[351–352] carried out the continuous emulsion polymerization of St in a con-
tinuous loop-tubular reactor with recycling. The effects of the emulsifier, ini-
tiator, and monomer concentrations, temperature, and mean residence time on
the monomer conversion, average particle diameter, the number of polymer
particles produced and, in some cases, the average molecular weight, were ex-
114 M. Nomura et al.

amined to clarify the characteristics of a CLTR. Overshoots in both monomer

conversion and the number of polymer particles were observed, but these di-
minished at about three times the mean residence time. Neither oscillations nor
multiple steady-states were observed. The steady-state number of polymer par-
ticles was independent of the mean residence time. Asua and co-workers [21,
353–356] carried out the redox-initiated emulsion copolymerization of VAc and
Veoba 10 in a CLTR. They studied the effect of the flow rate on shear-induced
coagulation and the effect of the start-up strategy on the smoothness of the op-
eration and on the amount of off-specification product, as well as proposing a
mathematical model for the process [353, 354]. They compared the perfor-
mance of a CLTR with that of a CSTR operating under similar conditions, and
concluded that, in most cases, both reactors showed a similar performance, but
under exigent conditions, the CSTR was prone to thermal runaway, whereas the
CLTR was much safer [355]. Moreover, the effects on the reactor performance
of both macromixing (residence time distribution) and micromixing (the de-
gree of monomer distribution in polymer particles) in the reaction mixture
were examined and the macromixing was characterized by means of tracer re-
sponse experiments [356]. They also showed that when the recycle ratio (the ra-
tio of the flow rate inside the reactor to the feed flow rate) was increased, the
behavior of the reactor approached that of a CSTR, and that when the feed was
pre-emulsified, the state of micromixing was substantially improved because
the rate of monomer diffusion from the monomer droplets into the polymer
particles was enhanced. Araújo et al. [21] developed a detailed dynamic math-
ematical model that described the evolution of PSDs during the emulsion
copolymerization of VAc and Veova 10 in a CLTR and compared the calculated
results with their experimental data.
With the aim of improving the performance of a continuous tubular reactor
by decreasing its backmixing, a pulsed tubular (PT) reactor [357], a pulsed
packed column (PPC) reactor [358], and a pulsed sieve plate column (PSPC) re-
actor [19] have been proposed as continuous reactors with near-plug flow. These
reactors are considered to be modified versions of a tubular reactor, but with less
backmixing than a simple tubular reactor. Paquet and Ray [357] successfully
conducted the continuous emulsion polymerization of MMA in a PT reactor.
When they performed the first four runs without any pulsation, only one of the
runs was successful; the other three plugged. They found, therefore, that the use
of a pulsation source eliminated the reactor fouling and plugging problem that
has frequently occurred in continuous tubular reactors. Also, no oscillatory
behavior was observed. The exit conversion remained constant for three resi-
dence times and was close to that observed in a corresponding batch reactor. No
clear dependence of the pulsation rate on monomer conversion was revealed.
Mayer et al. [358] investigated the performance of a PPC reactor in the con-
tinuous emulsion polymerization of St. They found that the number of poly-
mer particles produced in the PPC reactor depended strongly on the residence
time distribution (RTD) – in other words, on the pulsation conditions – and that
it had a value between those recorded for the batch and the CSTR processes.
Emulsion Polymerization: Kinetic and Mechanistic Aspects 115

They developed a mathematical model based on a micellar nucleation hypothesis

and plug flow with axial dispersion, and showed that there was a good agreement
between the model predictions and the experimental results. On the other hand,
Scholtens et al. [359] extended the Mayer model and predicted the effect of RTD
on the intermolecular chemical composition distribution (CCD) of copolymers
produced in the continuous emulsion polymerization of St and MA in a PPC
reactor with side feed streams. They concluded that the PPC reactor was a good
continuously operated alternative to a semi-batch process that produces latexes
with a predicted CCD. Sayer et al. [19] utilized a PSPC reactor for the continuous
emulsion polymerization of VAc and developed a dynamic mathematical model
to simulate the experimental results, showing that the continuous PSPC reactor
could be described by an axial dispersion model that covered the entire range
between the plug flow and perfectly mixed stirred-tank reactors.Another possi-
ble way of producing a plug flow is to use a reactor system with a cascade of CSTR
trains. Certainly, the larger the number of stages, the closer the flow pattern will
approach plug flow. However, an increase in the number of stages is not neces-
sarily desirable because of the probability of shear-induced latex coagulation,
caused by the increased exposure of latex particles to stirring.
As an alternative to a cascade of CSTR trains, a novel continuous reactor
with a Couette-Taylor vortex flow (CTVF) has been proposed, which can
realize any flow pattern between plug and perfectly mixed flows [361–366]. A
continuous Couette-Taylor vortex flow reactor (CCTVFR) consists of two con-
centric cylinders with the inner cylinder rotating and with the outer cylinder
at rest. Figure 29 shows a typical flow pattern caused by the rotation of the
inner cylinder.
Nomura et al. [360, 364] first utilized a Couette-Taylor vortex flow reactor
(CTVFR) for the continuous emulsion polymerization of St to clarify its char-

Fig. 29 Taylor vortices between two concentric cylinders (Couette-Taylor vortices): inner
cylinder rotating, outer cylinder at rest
116 M. Nomura et al.

acteristics as a continuous emulsion polymerization reactor, and developed a

model for the continuous emulsion polymerization of St in a single CCTVFR
that incorporated their batch model. The flow pattern is governed by the
dimensionless number called the Taylor number, Ta, defined by

  
wbRi b
Ta = 9 4 (100)
v Ri

where Ri is the inner cylinder radius, b is the radial clearance between two con-
centric cylinders, n is the kinematic viscosity, and w is the angular velocity of
the inner cylinder.When the Taylor number exceeds a certain value between 46
and 60, called the critical Taylor number, Tac, a transition occurs from pure
Couette flow to a flow regime in which toroidal vortices are regularly spaced
along the cylinder axis, as shown in Fig. 29, which is the so-called “Couette-Tay-
lor vortex flow”. They adopted the tank-in-series model that has only one
parameter, N, the number of tanks in series, to characterize the deviation of the
flow in the CCTVFR from plug flow. The relationship between the number
of tanks N and the Taylor number Ta was determined by using a stimulus-
response method. Figure 30 shows an interesting example of the monomer con-

Fig. 30 Effect of rotational speed of inner cylinder on steady-state monomer conversion

(S0 (NaLS)=6.25 g/dm3-water, I0 (KPS)=1.25 g/dm3-water, M0 (St)=100 g/dm3-water; 50 °C)
Emulsion Polymerization: Kinetic and Mechanistic Aspects 117

version versus reaction time, where the steady-state monomer conversion was
increased from 27 to 35% simply by decreasing the rotational speed of the
inner cylinder from 145 to 30 rpm.
This can be explained by the fact that the flow in the CCTVFR became closer
to plug flow as the Taylor number was dropped closer to Tac . Therefore, the
steady-state particle number and the steady-state monomer conversion could
be arbitrarily varied by simply varying the rotational speed of the inner cylin-
der. Moreover, no oscillations were observed, and the rotational speed of the
inner cylinder could be kept low, so that the possibility of shear-induced co-
agulation could be decreased. Therefore, a CCTVFR with these characteristics
is considered to be highly suitable as a pre-reactor for a continuous emulsion
polymerization process. In the case of the continuous emulsion polymerization
of VAc carried out with the same CCTVFR, however, the situation was quite
different [365]. Oscillations in monomer conversion were observed, and almost
no appreciable increase in steady-state monomer conversion occurred even
when the rotational speed of the inner cylinder was decreased to a value close
to Tac . Why the kinetic behavior with VAc is so different to that with St cannot
be explained at present.
Ohmura et al. [361, 362] carried out the continuous emulsion polymeriza-
tion of St and of VAc in a single CCTVFR from the standpoint of controlling
the particle size distribution of the latex produced. They also observed no
oscillations in the monomer conversion for the St system, but did observe
them in the system with VAc. They concluded that a self-sustained oscillation
would be useful for controlling the size of latex particles and for raising the
monomer conversion. On the other hand, Schmidt et al. [363] studied the con-
tinuous emulsion polymerization of n-BA in a single CCTVFR. The special
flow pattern and residence time distributions of the reactor were investigated
at different flow conditions. Both the hydrodynamics and the monomer con-
version of the CCTVFR were modeled using computational fluid dynamics
simulations. The model predictions were in good agreement with experi-
mental results.
In order to increase the performance and the productivity of the whole
continuous reactor system, it has been suggested that a pre-reactor operating
as a seed generator (a seeding reactor) should be placed upstream of a reactor
train. Even when it is optimized to have the mean residence time as qmax, the
steady-state number of polymer particles in a continuous stirred-tank pre-re-
actor, reaches only 58% of the number produced in an ideal plug flow reactor
or in a batch reactor using the same recipe and temperature conditions [163,
335]. Therefore, it is generally desirable to place a pre-reactor with plug flow
behavior (such as tubular, PSPC, PPC and CTVF reactors) upstream of the main
reactors, because this can produce a similar number of polymer particles to
that produced in a batch reactor if properly operated. Instead of incorporating
a pre-reactor, Nomura et al. [331, 332] have devised an operational technique
called “split-feed operation”, by which the number of polymer particles pro-
duced could be increased to much higher than that in a batch reactor under
118 M. Nomura et al.

Fig. 31 Schematic diagram of split-feed operation in continuous emulsion polymerization

the same recipe conditions. Figure 31 indicates the general concept of the split-
feed operation.
In a batch emulsion polymerization of St, they found that the number of
polymer particles produced increased when the amount of initially charged
monomer was decreased below a critical value, Mc, as shown in Fig. 32, where
the solid line shows the values predicted by
NT = (a3sÇ2p /36p)S03 M 0–2I 00 (101)
where Çp is the polymer density, and S0, M0 and I0 are the emulsifier, monomer
and initiator concentrations, respectively.
When a monomer split-feed operation based on the experimental result
shown in Fig. 32 was applied, for example, to a continuous tubular pre-reactor
with some backmixing, the number of polymer particles increased by about
30% at MF1=0.02 g/cm3-water, compared to the number produced in a batch re-
actor, as shown in Fig. 33.
The monomer split-feed operation was also shown to work in a continuous
stirred-tank pre-reactor. When certain conditions are fulfilled, the split-feeds

Fig. 32 Effect of lowering the initial monomer concentration on particle formation in

the batch emulsion polymerization of St (S0 (NaLS)=6.25 g/dm3-water, I0 (KPS)=1.25 g/dm3-
water, M0 (St)=variable; 50 °C)
Emulsion Polymerization: Kinetic and Mechanistic Aspects 119

Fig. 33 Effect of monomer concentration MF1 fed to the first tubular seeding reactor with
back mixing on the number of polymer particles produced (S0 (NaLS)=6.25 g/dm3-water,
I0 (KPS)=1.25 g/dm3-water, MF1 (St)=variable; 50 °C. Experimental data: empty circles, par-
ticle number observed at t=40 min in a batch reactor; filled circles, steady-state particle num-
ber observed in the first tubular seeding reactor operated with mean residence time
t =40 min)

of emulsifier, initiator and water could effectively increase the number of poly-
mer particles produced. The split-feed operation has additional advantages in
that the volume of a pre-reactor can be made smaller, and oscillations can be
eliminated when it is applied to a continuous stirred-tank pre-reactor. Penlides
et al. [333] have also discussed the advantages of the split-feed operation.
Despite the industrial importance, very little work on the kinetics of con-
tinuous emulsion copolymerization has been reported. Poehlein et al. [366, 367]
carried out the continuous emulsion copolymerization of St-AM and St-AN us-
ing a continuous reactor system comprised of a tubular reactor followed by one
or two stirred-tank reactors, and developed a steady-state model by employing
the kinetic model proposed by Nomura et al. [14] for batch emulsion copoly-
merization. Their model could predict the latex particle size distribution,
average number of radicals per particle, rate of copolymerization, and copoly-
mer composition. They also investigated the continuous emulsion copolymer-
ization of a moderately water-soluble monomer, ethyl acrylate, with a com-
pletely water-soluble monomer, methacrylic acid. Several continuous processes
involving a tubular reactor and/or a CSTR were designed and utilized so as to
produce a latex product with properties similar to the batch product [368, 369].
Nomura et al. [370] carried out the continuous emulsion copolymerization of
a sparingly water-soluble monomer, St, with a moderately water-soluble
monomer, MMA, in a single CSTR, in order to experimentally elucidate how the
kinetic behavior of the continuous emulsion homopolymerization of a spar-
ingly water-soluble monomer changes when it is copolymerized with a mod-
erately water-soluble monomer.
120 M. Nomura et al.

6
Concluding Remarks

The kinetics and mechanisms of particle growth and polymer structure de-
velopment are comparatively well understood compared to those of particle
nucleation. Therefore, the rate of polymerization and the properties of the poly-
mer produced can be (roughly) estimated as long as the number of polymer
particles produced is known (for example, in seeded emulsion polymerization).
However, the prediction of the number of polymer particles produced is still
far from being an established technique. Therefore, further efforts are needed
to qualitatively and quantitatively clarify the effects of numerous factors that
affect the process of particle formation in order to gain a more quantitative
understanding of emulsion polymerization.

References

1. Gilbert RG (1995) Emulsion polymerization: A mechanistic approach. Academic,

London
2. Lovell PA, El-Aasser MS (eds)(1997) Emulsion polymerization and emulsion polymers.
Wiley, New York
3. Fitch RM (1997) Polymer colloids: A comprehensive introduction. Academic, London
4. Smith WV, Ewart RH (1948) J Chem Phys 16:592
5. Harkins WD (1947) J Am Chem Soc 69:1428
6. Leslie GL, Napper DH, Gilbert RG (1992) Aust J Chem 45:2057
7. Yeliseeva VI (1982) In: Piirma I (ed) Emulsion polymerization. Academic, New York,
p 247
8. Gardon JL (1968) J Polym Sci A1 6:623
9. Ugelstad J, Hansen FK (1976) Rubber Chem Technol 49:536
10. Penboss IA, Napper DH, Gilbert RG (1983) J Chem Soc Farad T 1 79:1257
11. Maxwell IA, Morrison BR, Napper DH, Gilbert RG (1991) Macromolecules 24:1629
12. López de Arbina L, Barandiaran MJ, Gugliotta LM, Asua JM (1996) Polymer 37:5907
13. Gardon JL (1968) J Polym Sci A1 6:643
14. Harada M, Nomura M, Kojima H, Eguchi W, Nagata S (1972) J Appl Polym Sci 16:811
15. Hansen FK, Ugelstad J (1978) J Polym Sci Pol Chem 16:1953
16. Hansen FK (1992) ACS Sym Ser 492:12
17. Hansen FK (1993) Chem Eng Sci 48:437
18. Unzueta E, Forcada J (1997) J Appl Polym Sci 66:445
19. Sayer C, Palma M, Giudici R (2002) Ind Eng Chem Res 41:1733
20. Nomura M, Harada M, Eguchi W, Nagata S (1976) ACS Sym Ser 24:102
21. Araujo PHH, de la Cal JC, Asua JM, Pinto JC (2001) Macromol Theor Simul 10:769
22. Herrera-Ordóñez J, Olayo R (2000) J Polym Sci Pol Chem 38:2201
23. Dong Y, Sundberg DC (2002) Macromolecules 35:8185
24. Hawkett BS, Gilbert RG, Napper DH (1980) J Chem Soc Farad T 1 76:1323
25. Colombie D, Sudol ED, El-Aasser MS (2000) Macromolecules 33:4347
26. Marestin C, Guyot A, Claverie J (1998) Macromolecules 31:1686
27. Schoonbrood HAS, German AL, Gilbert RG (1995) Macromolecules 28:34
28. Liotta V, Georgakis C, Sudol ED, El-Aasser MS (1997) Ind Eng Chem Res 36:3252
Emulsion Polymerization: Kinetic and Mechanistic Aspects 121

29. Tauer K, Deckwer R (1998) Acta Polym 49:411

30. Adams ME, Trau M, Gilbert RG, Napper DH, Sangster DF (1988) Aust J Chem 41:1799
31. Penboss IA, Gilbert RG, Napper DH (1986) J Chem Soc Farad T 1 82:2247
32. Kusters JMH, Napper DH, Gilbert RG (1992) Macromolecules 25:7043
33. Wang X, Boya B, Sudol ED, El-Aasser MS (2001) Macromolecules 34:8907
34. Coen EM, Lyons RA, Gilbert RG (1996) Macromolecules 29:5128
35. Cheong IW, Kim JH (1997) Colloid Polym Sci 275:736
36. Vorwerg L, Gilbert RG (2000) Macromolecules 33:6693
37. Leemans L, Jerome R, Teyssie P (1998) Macromolecules 31:5565
38. Ohtsuka Y, Kawaguchi H, Sugi Y (1981) J Appl Polym Sci 26:1637
39. Nomura M, Ichikawa H, Fujita K, Okaya T (1997) J Polym Sci Pol Chem 35:2689
40. Nomura M, Suzuki K (2004) Prog Coll Pol Sci 124:7
41. Ugelstad J, Mørk PC, Dahl P, Rangnes P (1969) J Polym Sci C27:49
42. Nomura M, Harada M, Eguchi W, Nagata S (1971) J Chem Eng Jpn 4:54
43. Nomura M (1982) In: Piirma I (ed) Emulsion polymerization.Academic, New York, p 191
44. Nomura M, Harada M (1981) J Appl Polym Sci 26:17
45. Nomura M, Yamamoto K, Horie I, Fujita K, Harada M (1982) J Appl Polym Sci 27:2483
46. Casey BS, Morrison BR, Maxwell IA, Gilbert RG, Napper DH (1994) J Polym Sci Pol
Chem 32:605
47. Nomura M, Kubo M, Fujita K (1983) J Appl Polym Sci 28:2767
48. Adams M, Napper DH, Gilbert RG, Sangster DF (1986) J Chem Soc Farad T 1 82:1979
49. Asua JM, Sudol ED, EL-Aasser MS (1989) J Polym Sci Pol Chem 27:3903
50. Barandiaran MJ, Asua JM (1996) J Polym Sci Pol Chem 34:309
51. Asua JM (1998) Polymer 39:2061
52. Morrison BR, Casey BS, Lacik I, Leslie GL, Sangster DF, Gilbert RG, Napper DH (1994)
J Polym Sci Pol Chem 32:631
53. Kim JU, Lee HH (1996) Polymer 37:1941
54. Fang S-J, Wang K, Pan Z-R (2003) Polymer 44:1385
55. López de Arbina L, Barandiaran MJ, Gugliotta LM, Asua JM (1997) Polymer 38:143
56. Forcada J, Asua JM (1990) J Polym Sci Pol Chem 28:987
57. Barudio I, Guillot, Fevotte G (1998) J Polym Sci Pol Chem 36:157
58. Saldivar E, Dafniotis P, Ray WH (1998) J Macromol Sci R M C 38:207
59. Saldiver E, Araujo O, Giudici R, López-Barrón C (2001) J Appl Polym Sci 79:2380
60. Araujo O, Giudici R, Saldivar E, Ray WH (2001) J Appl Polym Sci 79:2360
61. Vega JR, Gugliotta LM, Bielsa RO, Brandolini MC, Meira GR (1997) Ind Eng Chem Res
36:1238
62. Gugliotta LM, Brandolini MC,Vega JR, Iturralde EO, Azum JL, Meira GR (1995) Polym
React Eng 3:201
63. Barandiaran MJ, López de Arbina L, de la Cal JC, Gugliotta LM, Asua JM (1995) J Appl
Polym Sci 55:231
64. Nomura M, Harada M, Nakagawara K, Eguchi W, Nagata S (1971) J Chem Eng Jpn 4:160
65. Hansen FK, Ugelstad J (1979) Makromol Chem 180:2423
66. Fitch RM (1981) ACS Sym Ser 165:1
67. Hansen FK, Ugelstad J (1982) In: Piirma I (ed) Emulsion polymerization. Academic,
New York, p 51
68. Tauer K, Kuhn I (1995) Macromolecules 28:2236
69. Kuhn I, Tauer K (1995) Macromolecules 28:8122
70. Lichti G, Gilbert RG, Napper DH (1983) J Polym Sci Pol Chem 21:269
71. Feeney PJ, Napper DH, Gilbert RG (1984) Macromolecules 17:2520
72. Richards JR, Congalidis JP (1989) J Appl Polym Sci 37:2727
122 M. Nomura et al.

73. Sütterlin N (1980) In: Fitch RM (ed) Polymer colloids II. Plenum, New York, p 583
74. Nomura M, Fujita K (1994) Polym React Eng 2:317
75. Nomura M, Kodani T, Ojima J, Kihara Y, Fujita K (1998) J Polym Sci Pol Chem 36:1919
76. Morrison BR, Maxwell IA, Gilbert RG, Napper DH (1992) In: Daniels ES, Sudol ED,
El-Aasser MS (eds) Polymer Latexes. ACS Symp Ser. 492. Am Chem Soc, Washington
DC, p 28
77. Sajjadi S, Brooks BW (1999) J Polym Sci Pol Chem 37:3957
78. Nomura M, Sakai H, Kihara Y, Fujita K (2002) J Polym Sci 40:1275
79. Yuan X-Y, Dimonie VL, Sudol ED, Roberts JE, El-Aasser MS (2002) Macromolecules
35:8356
80. Herrera-Ordóñez J, Olayo R (2000) J Polym Sci Pol Chem 38:2219
81. Herrera-Ordóñez J, Olayo R (2001) J Polym Sci Pol Chem 39:2547
82. Varela de la Rosa L, Sudol ED, EL-Aasser MS, Klein A (1996) J Polym Sci Pol Chem
34:461
83. Varela de la Rosa L, Sudol ED, EI-Aasser MS, Klein A (1999) J Polym Sci Pol Chem
37:4066
84. Varela de la Rosa L, Sudol ED, EI-Aasser MS, Klein A (1999) J Polym Sci Pol Chem
37:4073
85. Nomura M, Satpathy US, Kouno Y, Fujita K (1988) J Polym Sci Pol Lett 26:385
86. Nomura M, Takahashi K, Fujita K (1990) Makromol Chem–M Symp 35/36:13
87. Chern C-S, Lin C-H (1998) Polymer 40:139
88. Chern C-S, Lin C-H (2000) Polymer 41:4473
89. Wang Z, Paine AJ, Rudin A (1995) J Polym Sci Pol Chem 33:1597
90. Morrison BR, Gilbert RG (1995) Macromol Symp 92:13
91. Coen EM, Gilbert RG, Morrison BR, Leube H, Peach S (1998) Polymer 39:7099
92. Prescott SW, Fellows CM, Gilbert RG (2002) Macromol Theor Simul 11:163
93. Butucea V, Sarbu A, Georgescu C (1998) Angew Makromol Chem 255:37
94. Cheong I-W, Kim J-H (1998) Macromol Theor Simul 7:49
95. Sajjadi S (2000) J Polym Sci Pol Chem 38:3612
96. Sajjadi S (2001) J Polym Sci Pol Chem 39:3940
97. Nomura M, Fujita K (1993) Polym Int 30:483
98. Sajjadi S (2002) J Polym Sci Pol Chem 40:1652
99. Ozdeger E, Sudol ED, El-Aasser MS, Klein A (1997) J Polym Sci Pol Chem 35:3813
100. Ozdeger E, Sudol ED, El-Aasser MS, Klein A (1997) J Polym Sci Pol Chem 35:3827
101. Ozdeger E, Sudol ED, El-Aasser MS, Klein A (1997) J Polym Sci Pol Chem 35:3837
102. Lin S-Y, Capek I, Hsu T-J, Chern C-S (1999) J Polym Sci Pol Chem 37:4422
103. Unzueta E, Forcada J (1995) Polymer 36:4301
104. Chen L-J, Lin S-Y, Chern C-S, Wu S-C (1997) Colloid Surface A 122:161
105. Lin S-Y, Capek I, Hsu T-J, Chern C-S (2000) Polym J 32:932
106. Colombie D, Sudol ED, El-Aasser MS (2000) Macromolecules 33:7283
107. Asua JM, Schoonbrood HAS (1998) Acta Polym 49:671
108. Amalvy JI, Unzue MJ, Schoonbrood HAS, Asua JM (1998) Macromolecules 31:5631
109. Wang X, Sudol ED, El-Aasser MS (2001) J Polym Sci Pol Chem 39:3093
110. Wang X, Sudol ED, El-Aasser MS (2001) Macromolecules 34:7715
111. Cochin D, Laschewsky A, Nallet F (1997) Macromolecules 30:2278
112. Ayoub MMH, Nasr HE, Rozik NN (1998) J Macromol Sci Pure 35:1415
113. Ayoub MMH (1998) J Elastom Plast 30:207
114. Wang X, Sudol ED, El-Aasser MS (2001) Langmuir 17:6865
115. Kato S, Nomura M (1999) Colloid Surface A 153:127
116. Cheong I-W, Nomura M, Kim J-H (2001) Macromol Chem Phys 202:2454
Emulsion Polymerization: Kinetic and Mechanistic Aspects 123

117. Piirma I (1993) Polymer surfactants (Surfactant Science Series No 42). Marcel Dekker,
New York
118. Capek I (2002) Adv Colloid Interfac 99:77
119. O’Toole JT (1965) J Appl Polym Sci 9:1291
120. Ugelstad J, Mørk PC, Aasen JO (1967) J Polym Sci A1 5:2281
121. Stockmayer WH (1957) J Polym Sci 24:314
122. (a) Nomura M, Fujita K (1985) Makromol Chem Suppl 10/11:25, (b) Nomura M (1987)
Kobunshi Kagaku 36:680
123. Sakai H, Kihara Y, Fujita K, Kodani T, Nomura M (2001) J Polym Sci Pol Chem 39:1005
124. Asua JM, Adams ME, Sudol ED (1990) J Appl Polym Sci 39:1183
125. Mendoza J, de la Cal JC, Asua JM (2000) J Polym Sci Pol Chem 38:4490
126. Kiparissides C, Achilias CDS, Frantzikinakis CE (2002) Ind Eng Chem Res 41:3097
127. de la Cal JC, Asua JM (2001) J Polym Sci Pol Chem l39:585
128. Gilmore CM, Poehlein GW, Schork FJ (1993) J Appl Polym Sci 48:1449
129. Gilmore CM, Poehlein GW, Schork FJ (1993) J Appl Polym Sci 48:1461
130. Budhlall BM, Sudol ED, Dimonie VL, Klein A, El-Aasser MS (2001) J Polym Sci Pol Chem
39:3633
131. Shaffie KA, Moustafa AB, Mohamed ES, Badran AS (1997) J Polym Sci Pol Chem 35:3141
132. Bruyn HD, Gilbert RG, Ballard MJ (1996) Macromolecules 29:8666
133. Chern C-S, Poehlein GW (1987) J Appl Polym Sci 33:2117
134. Bruyn HD, Miller CM, Bassett DR, Gilbert RG (2002) Macromolecules 35:8371
135. Matsumoto A, Kodama K, Aota H, Capek I (1999) Eur Polym J 35:1509
136. Ballard MJ, Napper DH, Gilbert RG (1981) J Polym Sci Pol Chem 19:939
137. Nomura M, Kubo M, Fujita K (1981) Mem Fac Eng Fukui Univ 29:167
138. Storti G, Carra S, Morbidelli M, Vita G (1989) J Appl Polym Sci 37:2443
139. Tobita H, Hamielec AE (1989) Macromolecules 22:3098
140. Nomura M, Horie I, Kubo M, Fujita K (1989) J Appl Polym Sci 37:1029
141. Chen S-A, Wu K-W (1988) J Polym Sci Pol Chem 26:1487
142. Giannetti E (1989) Macromolecules 22:2094
143. Saldivar E, Ray WH (1997) Ind Eng Chem Res 36:1322
144. Dube MA, Penlidis A, Mutha RK, Cluett WR (1996) Ind Eng Chem Res 35:4434
145. Martinet F, Guillot J (1999) J Appl Polym Sci 72:1627
146. Vicente M, Leiza JR, Asua JM (2001) AIChE J 47:1594
147. Vega JR, Gugliotta LM, Meira GR (2002) Polym React Eng 10:59
148. Shoaf GL, Poehlein GW (1991) J Appl Polym Sci 42:1213
149. Yang B-Z, Chen L-W, Chiu W-Y (1997) Polym J 29:744
150. Yang B-Z, Chen L-W, Chiu W-Y (1997) Polym J 29:737
151. Slawinski M, Schellekens MAJ, Meuldijk J, Herk AMV, German AL (2000) J Appl Polym
Sci 76:1186
152. Wang PH, Pan C-Y (2001) Colloid Polym Sci 279:98
153. Santos AMD, Mckenna TF, Guillot J (1997) J Appl Polym Sci 65:2343
154. Yan C, Cheng S, Feng L (1999) J Polym Sci Pol Chem 37:2649
155. Henton DE, Powell C, Reim RE (1997) J Appl Polym Sci 64:591
156. Xu Z, Yi C, Cheng S, Zhang J (1997) J Appl Polym Sci 66:1
157. Fang S-J, Fujimoto K, Kondo S, Shiraki K, Kawaguchi H (2000) Colloid Polym Sci
278:864
158. Kostov GK, Petrov PC (1994) J Polym Sci Pol Chem 32:2229
159. Petrov PC, Kostov GK (1994) J Polym Sci Pol Chem 32:2235
160. Noel LFJ, Altveer JLV, Timmermans MDF, German AL (1996) J Polym Sci Pol Chem
34:1763
124 M. Nomura et al.

161. Urretabizkaia A, Asua JM (1994) J Polym Sci Pol Chem 32:1761

162. Ge X, Ye Q, Xu X, Chu G, Zhang Z (1998) J Appl Polym Sci 67:1005
163. Nomura M, Kojima H, Harada M, Eguchi W, Nagata S (1971) J Appl Polym Sci 15:675
164. Morton M, Kaizerman S, Altier MW (1954) J Colloid Sci 9:300
165. Gardon JL (1968) J Polym Sci A1 6:2859
166. Maxwell IA, Kurja J, Doremaele GHV, German AL, Morrison BR (1992) Makromol
Chem 193:2049
167. Antonietti M, Kasper H, Tauer K (1996) Langmuir 12:6211
168. Vanzo E, Marchessault RH, Stannett V (1965) J Colloid Sci 20:62
169. Ugelstad J, Mørk PC, Mfutakamba HR, Soleimany E, Nordhuus I, Schmid R, Berge A,
Ellingsen T, Aune O, Nustad K (1983) In: Pohelein GW, Ottewill RH, Goodwin JW (eds)
Science and technology of polymer colloids, vol 1 (NATO ASI Ser E:67). Martinus
Nijhoff, Boston, MA, p 51
170. Maxwell IA, Kurja J, van Doremaele GHJ, German AL (1992) Makromol Chem 193:2065
171. Schoonbrood HAS, German AL (1994) Macromol Rapid Commun 15:259
172. Noël LFJ, Maxwell IA, German AL (1993) Macromolecules 26:2911
173. Schoonbrood HAS, Boom MATVD, German AL, Hutovic J (1994) J Polym Sci Pol Chem
32:2311
174. Noël LFJ, van Zon JMAM, Maxwell IA, German AL (1994) J Polym Sci Pol Chem 32:1009
175. Gugliotta LM, Arzamendi G, Asua JM (1995) J Appl Polym Sci 55:1017
176. Karlsson O, Wesslen B (1998) J Appl Polym Sci 70:2041
177. Mathey P, Guillot J (1991) Polymer 32:934
178. Nomura M, Liu X, Ishitani K, Fujita K (1994) J Polym Sci Pol Phys 32:2491
179. Liu X, Nomura M, Fujita K (1997) J Appl Polym Sci 64:931
180. Liu X, Nomura M, Liu Y-H, Ishitani K, Fujita K (1997) Ind Eng Chem Res 36:1218
181. Aerdts AM, Boei MWA, German AL (1993) Polymer 34:574
182. Said ZFM, Fataftah ZA (1996) Polym Int 40:307
183. Tognacci R, Storti G, Bertucco A (1996) J Appl Polym Sci 62:2341
184. Varela de la Rosa L, Sudol ED, EI-Aasser MS, Klein A (1999) J Polym Sci Pol Chem
37:4054
185. Cheong I-W, Kim J-H (1999) Colloid Surface A 153:137
186. BenAmor S, Colombie D, McKenna T (2002) Ind Eng Chem Res 41:4233
187. Tauer K, Muller H, Schellenberg C, Rosengarten L (1999) Colloid Surfuce A 153:143
188. López de Arbina L, Gugliotta LM, Barandiaran MJ, Asua JM (1998) Polymer 39:4047
189. de Buruaga IS, Arotcarena M, Armitage PD, Gugliotta LM, Leiza JR, Asua JM (1996)
Chem Eng Sci 51:781
190. de Buruaga IS, Echevarria A,Armitage PD, de la Cal JC, Leiza JR,Asua JM (1997) AIChE
J 43:1069
191. Gugliotta LM, Leiza JR,Arotcarena M,Armitage PD,Asua JM (1995) Ind Eng Chem Res
34:3899
192. Nomura M, Ashizawa N (1998) US Patent 5 762 879, assigned to Todoroki Sangyo Co,
Fukui, Japan
193. Breitenbach VJW, Edelhauser H (1961) Macromol Chem 44–47:196
194. van der Hoff BME (1960) J Polym Sci 48:175
195. Al-Shahib WAG, Dunn AS (1980) Polymer 21:429
196. Barton J, Karpatyova A (1987) Makromol Chem 188:693
197. Il’menev PY, Litvinenko GI, Kaminskii VA, Gritskova IA (1988) Polym Sci USSR 30:826
198. Nomura M, Fujita K (1989) Makromol Chem Rapid Commun 10:581
199. Nomura M, Yamada A, Fujita S, Sugimoto A, Ikoma J, Fujita K (1991) J Polym Sci Pol
Chem 29:987
Emulsion Polymerization: Kinetic and Mechanistic Aspects 125

200. Nomura M, Ikoma J, Fujita K (1992) ACS Sym Ser 492:55

201. Nomura M, Fujita K (1992) DECHEMA Monogr 127:359
202. Nomura M, Ikoma J, Fujita K (1993) J Polym Sci Pol Chem 31:2103
203. Asua JM, Rodriguez VS, Sudol ED, El-Aasser MS (1989) J Polym Sci Pol Chem 27:3569
204. Alduncin JA, Forcada J, Barandiaran MJ, Asua JM (1991) J Polym Sci Pol Chem 29:1265
205. Vaskova V, Renoux D, Bernard M, Selb J, Candau F (1995) Polym Advan Technol 6:441
206. Mørk PC (1995) J Polym Sci Pol Chem 33:2305
207. Mørk PC, Ugelstad J, Aasen JO (1995) J Polym Sci Pol Chem 33:1759
208. Mørk PC, Makame Y (1997) J Polym Sci Pol Chem 35:2347
209. Suzuki K, Goto A, Takayama M, Muramatsu A, Nomura M (2000) Macromol Symp
155:99
210. Nomura M, Suzuki K (1997) Macromol Chem Phys 198:3025
211. Capek I (2001) Adv Colloid Interfac 91:295
212. Biggs S, Grieser F (1995) Macromolecules 28:4877
213. Ooi SK, Biggs S (2000) Ultrason Sonochem 7:125
214. Cooper C, Grieser F, Biggs S (1996) J Colloid Interf Sci 184:52
215. Bradley M, Grieser F (2002) J Colloid Interf Sci 251:78
216. Chou HCJ, Stoffer JO (1999) J Appl Polym Sci 72:797
217. Chou HCJ, Stoffer JO (1999) J Appl Polym Sci 72:827
218. Liao Y, Wang Q, Xia H, Xu X, Baxter SM, Slone RV, Wu S, Swift G, Westmoreland DG
(2001) J Polym Sci Pol Chem 39:3356
219. Xia H, Wang Q, Liao Y, Xu X, Baxter SM, Slone RV, Wu S, Swift G, Westmoreland DG
(2002) Ultrason Sonochem 9:151
220. Nomura M, Harada M, Eguchi W, Nagata S (1972) J Appl Polym Sci 16:835
221. Penlidis A, Macgegor JF, Hamielec AE (1988) J Appl Polym Sci 35:2023
222. Cunningham MF, Geramita K, Ma JW (2000) Polymer 41:5385
223. Bruyn HD, Gilbert RG, Hawkett BS (2000) Polymer 41:8633
224. Kokthoff IM, Harris WE (1947) J Polym Sci 2:49
225. Whang BCY, Lichti G, Gilbert RG, Napper DH (1980) J Polym Sci Pol Lett 18:711
226. Nomura M, Minamino Y, Fujita K, Harada M (1982) J Polym Sci Pol Chem 20:1261
227. Lichti G, Sangster DF (1982) J Chem Soc Farad T 1 78:2129
228. Maxwell IA, Morrison BR, Napper DH, Gilbert RG (1992) Makromol Chem 193:303
229. Weerts PA, van der Loos JLM, German AL (1991) Makromol Chem 192:2009
230. Huo BP, Campbell JD, Penlidis A, Macgregor JF, Hamielec AE (1987) J Appl Polym Sci
35:2009
231. Kemmere MF, Mayer MJJ, Meuldijk J, Drinkenburg AAH (1999) J Appl Polym Sci 71:2419
232. Barton J, Juranicova V (1991) Makromol Chem Rapid Commun 12:669
233. Echevarria A, Leiza JR, de la Cal JC, Asua JM (1998) AIChE J 44:1667
234. Salazar A, Gugliotta LM, Vega JR, Meira GR (1998) Ind Eng Chem Res 37:3582
235. Sayer C, Lim EL, Pinto JC, Arzamendi G, Asua LM (2000) J Polym Sci Pol Chem 38:1100
236. Sayer C, Lima EL, Pinto JC, Arzamendi G, Asua JM (2000) J Polym Sci Pol Chem 38:367
237. Plessis C, Arzamendi G, Leiza JR, Alberdi JM, Schoonbrood HAS, Charmot D, Asua JM
(2001) J Polym Sci Pol Chem 39:1106
238. Gugliotta LM, Salazar A, Vega JR, Meira GR (2001) Polymer 42:2719
239. Nomura M, Suzuki H, Tokunaga H, Fujita K (1994) J Appl Polym Sci 51:21
240. Okaya T, Suzuki A, Kikuchi K (2000) Macromol Symp 150:143
241. Bataille P, Dalpe J-F, Dubuc F, Lamoureux L (1990) J Appl Polym Sci 39:1815
242. Shunmukham SR, Hallenbeck VL, Guile RL (1951) J Polym Sci 6:691
243. Schoot CJ, Bakker J, Klaassens KH (1951) J Polym Sci 7:657
244. Evans CP, Hay PM, Marker L, Murray RW, Sweeting OJ (1961) J Appl Polym Sci 5:39
126 M. Nomura et al.

245. Omi S, Shiraishi Y, Sato H, Kubota H (1969) J Chem Eng Jpn 2:64
246. Weerts PA, van der Loos JLM, German AL (1991) Makromol Chem 192:1993
247. Arai K, Arai M, Iwasaki S, Saito S (1981) J Polym Sci Pol Chem 19:1203
248. Kim CU, Lee JM, Ihm SK (1999) J Appl Polym Sci 73:777
249. Ozdeger E, Sudol ED, El-Aasser MS, Klein A (1998) J Appl Polym Sci 69:2277
250. Brooks BW (1971) Brit Polym J 3:269
251. Zubitur M, Asua JM (2001) J Appl Polym Sci 80:841
252. Soares JBP, Hamielec AE (1997) In: Asua JM (ed) Polymeric dispersions (NATO ASI Ser
E:335). Kluwer Academic, London, p 289
253. Cunningham MF, Ma JW (2000) J Appl Polym Sci 78:217
254. Ma JW, Cunningham MF (2000) Macromol Symp 150:85
255. Zubitur M, Mendoza J, de la Cal JC, Asua JM (2000) Macromol Symp 150:13
256. Rimmer S, Tattersall P (1999) Polymer 40:5729
257. Rimmer S (2000) Macromol Symp 150:149
258. Leyrer RJ, Machtle W (2000) Macromol Chem Phys 201:1235
259. Lau W (1994) US Patent 5 521 266, assigned to Rohm and Haas Co, Philadelphia, PA, USA
260. Vanderhoff JW (1981) ACS Sym Ser 165:199
261. Lowry V, El-Aasser MS, Vanderhoff JW, Klein A (1984) J Appl Polym Sci 29:3925
262. Matejicek A, Ditl P, Pivonkova A, Kaska J, Formanek L (1988) J Appl Polym Sci
35:583
263. Tobita H, Takada Y, Nomura M (1994) Macromolecules 27:3804
264. Tobita H, Takada Y, Nomura M (1995) J Polym Sci Pol Phys 33:441
265. Tobita H (1995) Macromolecules 28:5128
266. Tobita H (1994) Polymer 35:3023
267. Tobita H (1994) Polymer 35:3032
268. Tobita H (1997) J Polym Sci Pol Phys 35:1515
269. Tobita H, Nomura M (1999) Colloid Surface A 153:119
270. Tobita H, Yamamoto K (1994) Macromolecules 27:3389
271. Tobita H, Uemura Y (1996) J Polym Sci Pol Phys 34:1403
272. Tobita H, Yoshihara Y (1996) J Polym Sci Pol Phys 34:1415
273. Tobita H (1995) Acta Polym 46:185
274. Min KW, Ray HW (1974) J Macromol Sci R M C 11:177
275. Lin CC, Chiu WY (1979) J Appl Polym Sci 23:2049
276. Lichti G, Gilbert RG, Napper DH (1980) J Polym Sci Pol Chem 18:1297
277. Lichti G, Gilbert RG, Napper DH (1982) In: Piirma I (ed) Emulsion polymerization.
Academic, New York, p 93
278. Katz S, Shinnar R, Saidel GM (1969) Adv Chem Ser 91:145
279. Giannetti E, Storti G, Morbidelli M (1988) J Polym Sci Pol Chem 26:1985
280. Storti G, Polotti G, Cociani M, Morbidelli M (1992) J Polym Sci Pol Chem 30:731
281. Ghielmi A, Storti G, Morbidelli M (1998) Macromolecules 31:7172
282. Benson SW, North AM (1962) Am Chem Soc 84:935
283. Ito K (1974) J Polym Sci Pol Chem 12:1991
284. Mahabadi HK (1985) Macromolecules 18:1319
285. Olaj OF, Zifferer G, Gleizner G (1986) Makromol Chem 187:977
286. Russell GT, Gilbert RG, Napper DH (1992) Macromolecules 25:2459
287. O’Shaughnessy B, Yu J (1994) Macromolecules 27:5067
288. Buback M, Egorov M, Kaminsky V (1999) Macromol Theor Simul 8:520
289. Russell GT (1994) Macromol Theor Simul 3:439
290. Clay PA, Gilbert RG (1995) Macromolecules 28:552
291. Wulkow M (1996) Macromol Theor Simul 5:393
Emulsion Polymerization: Kinetic and Mechanistic Aspects 127

292. Tobita H (1995) Macromolecules 28:5119

293. Olaj OF, Kauffmann HF, Breitenbach JB (1977) Makromol Chem 178:2707
294. Mayo FR (1943) J Am Chem Soc 65:2324
295. Mayo FR, Gregg RA, Matheson MS (1951) J Am Chem Soc 73:1691
296. Christie DI, Gilbert RG (1996) Macromol Chem Phys 197:403
297. Buback M, Gilbert RG, Russell GT, Hill DJT, Moad G, O’Driscoll KF, Shen J, Winnik MA
(1992) J Polym Sci Pol Chem 30:851
298. Tobita H, Shiozaki H (2001) Macromol Theor Simul 10:676
299. Morton M, Salatiello PP (1951) J Polym Sci 6:225
300. Britton D, Heatley F, Lovell PA (1998) Macromolecules 31:2828
301. Tobita H (1993) Polym React Eng 1:357
302. Tobita H, Saito S (1999) Macromol Theor Simul 8:513
303. Tobita H, Hamashima N (2000) J Polym Sci Pol Phys 38:2009
304. Tobita H, Hamashima N (2000) Macromol Theor Simul 9:453
305. Tobita H (2001) J Polym Sci Pol Phys 39:2960
306. Tobita H, Kawai H (2002) E-Polymers 048
307. Starkweather WH Jr, Han MC (1992) J Polym Sci Pol Chem 30:2709
308. Senrui S, Suwa T, Takehisa M (1974) J Polym Sci Pol Chem 12:93
309. Senrui S, Suwa T, Takehisa M (1974) J Polym Sci Pol Chem 12:105
310. Tobita H (2002) J Polym Sci Pol Chem 40:3426
311. Tobita H (2003) Bimodal molecular weight distribution formed in emulsion polymer-
ization with long-chain branching. Polym React Eng 11:855
312. Barabasi A-L, Albert R (1999) Science 286:509
313. Tobita H (2004) Scale-free power-law distribution of emulsion-polymerized nonlinear
polymers: Free-radical polymerization with chain transfer to polymer. Macromolecules
37:585
314. Friis N, Goosney D, Wright JD, Hamielec AE (1974) J Appl Polym Sci 18:1247
315. Friis N, Hamielec AE (1975) J Appl Polym Sci 19:97
316. Tobita H (1992) Macromolecules 25:2671
317. Tobita H, Kimura K, Fujita K, Nomura M (1993) Polymer 34:2569
318. Flory PJ (1953) Principles of polymer chemistry. Cornell University Press, Ithaca, NY
319. Tobita H, Hamielec AE (1989) In: Reichert K-H, Geiseler W (eds) Polymer reaction
engineering. VCH, Weinheim, Germany, p 43
320. Ding ZY, Ma S, Kriz D, Aklonis JJ, Salovey R (1992) J Polym Sci Pol Phys 30:189
321. Baker WO (1949) Ind Eng Chem 41:511
322. Obrecht W, Seitz U, Funke W (1976) ACS Sym Ser 24:92
323. Nakamura K, Imoto A, Aota H, Matsumoto A (1994) The 8th Polymeric Microsphere
Symposium, Fukui, Japan, p 37
324. Matsumoto A, Mori Y, Takahashi S, Aota H (1995) Netsukokasei-Jushi (J Thermoset
Plast Jpn) 16:131
325. Matsumoto A, Kodama K, Mori Y, Aota H (1998) Pure Appl Chem A35:1459
326. Tobita H, Kumagai M, Aoyagi N (2000) Polymer 41:481
327. Tobita H, Aoyagi N, Takamura S (2001) Polymer 421:7583
328. Hamielec AE, Tobita H (1992) In: Ulmann’s encyclopedia of industrial chemistry,
vol A21. VCH, Weinheim, Germany, p 305
329. Pohelein G (1997) In: Lovell PA, El-Aasser MS (eds) Emulsion polymerization and
emulsion polymers. Wiley, New York, p 277
330. Gershberg DB, Longfield JE (1961) Symp Polym Kinetics and Catalyst Systems, Preprint
10, 45th AIChE Meeting, New York
331. Nomura M (1981) ACS Sym Ser 165:121
128 Emulsion Polymerization: Kinetic and Mechanistic Aspects

332. Nomura M (1986) In: Reichert K-H, Geisler W (eds) Polymer eeaction engineering.
Hüthig & Wepf, Basel, p 41
333. Penlidis A, MacGregor JF, Hamielec AE (1989) Chem Eng Sci 44:273
334. Poehlein GW, Dougherty DJ (1977) Rubber Chem Technol 50:601
335. Omi S, Ueda T, Kubota H (1969) J Chem Eng Jpn 2:193
336. Degraff AW, Poehlein GW (1971) J Polym Sci A2 9:1955
337. Gerrens H, Kuchner K (1970) Brit Polym J 2:18
338. Gerrens H, Kuchner K, Ley G (1971) Chem Ing Tech 43:693
339. Gregor L, Gerrens H (1974) Macromol Chem 175:563
340. Nomura M, Sasaki S, Xue W, Fujita K (2002) J Appl Polym Sci 86:2748
341. Kiparissides C, MacGregor JF, Hamielec AE (1979) J Appl Polym Sci 23:401
342. Rawlings JB, Ray WH (1988) Polym Eng Sci 28:237
343. Rawlings JB, Ray WH (1988) Polym Eng Sci 28:257
344. Tauer K, Muller I (1995) DECHEMA Monogr 131:95
345. Greens RK, Gonzalez RA, Poehlein GW (1976) ACS Sym Ser 24:341
346. Feldon M, McCann RF, Laundrie RW (1953) India Rubber World 128:51
347. Geddes K (1983) Chem Ind 21:223
348. Geddes KR (1989) Brit Polym J 21:443
349. Rollin AL, Patterson I, Huneault R, Bataille P (1977) Can J Chem Eng 55:565
350. Iabbadene A, Bataille P (1994) J Appl Polym Sci 51:503
351. Lee D-Y, Kuo J-F, Wang J-H, Chen C-Y (1990) Polym Eng Sci 30:187
352. Lee D-Y, Wang J-H, Kuo J-F (1992) Polym Eng Sci 32:198
353. Abad C, de la Cal JC, Asua JM (1995) Polymer 36:4293
354. Abad C, de la Cal JC, Asua JM (1994) Chem Eng Sci 49:5025
355. Abad C, de la Cal JC, Asua JM (1995) J Appl Polym Sci 56:419
356. Abad C, de la Cal JC, Asua JM (1995) DECHEMA Monogr 131:87
357. Paquet DA Jr, Ray WH (1994) AIChE J 40:73
358. Mayer MJJ, Meuldijk J, Thoenes D (1996) Chem Eng Sci 51:3441
359. Scholetens CA, Meuleijk J, Drinkenburg AAH (2001) Chem Eng Sci 56:955
360. Imamura T, Saito K, Ishikura S, Nomura M (1993) Polym Int 30:203
361. Kataoka K, Ohmura N, Kouzu M, Okubo Y (1995) Chem Eng Sci 50:1409
362. Ohmura N, Kataoka K, Watanabe S, Okubo M (1998) Chem Eng Sci 53:2129
363. Schmidt W, Kossak S, Langenbuch J, Moritz H-U, Herrmann C, Kremeskötter J (1998)
DECHEMA Monogr 134:509
364. Wei X, Takahashi H, Sato S, Nomura M (2001) J Appl Polym Sci 80:1931
365. Xue W, Yoshikawa K, Oshima A, Nomura M (2002) J Appl Polym Sci 86:2755
366. Mead RN, Poehlein GW (1988) Ind Eng Chem Res 27:2283
367. Mead RN, Poehlein GW (1989) Ind Eng Chem Res 28:51
368. Shoaf GL, Poehlein GW (1989) Polym Plast Tech Eng 28:289
369. Poehlein GW (1995) Macromol Symp 92:179
370. Fang S-J, Xue W, Nomura M (2003) Polym React Eng 11:815

Received: February 2004

Adv Polym Sci (2005) 175: 129–255
DOI 10.1007/b100115
© Springer-Verlag Berlin Heidelberg 2005

Miniemulsion Polymerization
F. Joseph Schork ( )1 · Yingwu Luo 2 · Wilfred Smulders1 · James P. Russum1 ·
Alessandro Butté 3 · Kevin Fontenot 4
1
School of Chemical and Biomolecular Engineering, Georgia Insitute of Technology,
Atlanta Georgia30332–0100 USA
[email protected], [email protected],
[email protected]
2 The State Key Laboratory of Polymer Reaction Engineering, College of Material
and Chemical Engineering, Zhejiang University, Hangzhou, 310027 P. R. China
[email protected]
3 Prof. M. Morbidelli Group, ETH Zürich, Institute for Chemical- and Bioengineering,
ETH Hönggerberg/HCI F 135, 8093 Zürich, Switzerland
[email protected]
4
Eastman Chemical Company, P.O. Box 1972, Kingsport TN 37662, USA
[email protected]

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
1.1 Dispersed-Phase Polymerization . . . . . . . . . . . . . . . . . . . . . . . 133
1.2 The Concept of Miniemulsion Polymerization . . . . . . . . . . . . . . . . 135
1.3 Publication History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136

2 Miniemulsion Polymerization . . . . . . . . . . . . . . . . . . . . . . . . . 137

2.1 The Mechanism of Macroemulsion Polymerization . . . . . . . . . . . . . 137
2.1.1 Interval I – Particle Nucleation . . . . . . . . . . . . . . . . . . . . . . . . 138
2.1.1.1 Micellar Nucleation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
2.1.1.2 Homogeneous Nucleation . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
2.1.1.3 Droplet Nucleation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
2.1.1.4 Competition for Radicals . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
2.1.2 Interval II – Particle Growth . . . . . . . . . . . . . . . . . . . . . . . . . . 142
2.1.3 Interval III – Gel and Glass Effects . . . . . . . . . . . . . . . . . . . . . . . 143
2.2 The Mechanism of Miniemulsion Polymerization . . . . . . . . . . . . . . 144
2.3 Mathematical Modeling of Miniemulsion Polymerization . . . . . . . . . . 147

3 Properties of Miniemulsion Polymerization . . . . . . . . . . . . . . . . . 148

3.1 Shear Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
3.2 Choice of Surfactant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
3.3 Choice of Costabilizer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
3.3.1 Polymeric Costabilizers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
3.3.2 Monomeric Costabilizers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
3.3.3 Other Costabilizers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
3.3.4 Enhanced Nucleation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
3.4 Choice of Initiator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
3.5 Robust Nucleation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
3.6 Monomer Transport Effects . . . . . . . . . . . . . . . . . . . . . . . . . . 159
3.7 Droplet Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
3.7.1 Stability of Monomer Dispersions . . . . . . . . . . . . . . . . . . . . . . . 161
130 F. J. Schork et al.

3.7.2 Experimental Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170

3.8 Semibatch and Plug Flow Reactors . . . . . . . . . . . . . . . . . . . . . . 173
3.9 Continuous Stirred Tank Reactors . . . . . . . . . . . . . . . . . . . . . . . 174

4 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
4.1 Robust Nucleation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
4.1.1 Effect of Initiation and Inhibition . . . . . . . . . . . . . . . . . . . . . . . 178
4.1.1.1 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
4.1.1.2 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
4.1.2 Particle Size Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
4.1.2.1 Hexadecane as Costabilizer . . . . . . . . . . . . . . . . . . . . . . . . . . 184
4.1.2.2 Dodecyl Mercaptan as Costabilizer . . . . . . . . . . . . . . . . . . . . . . 186
4.1.2.3 Polymethyl Methacrylate as Costabilizer . . . . . . . . . . . . . . . . . . . 187
4.1.2.4 Influence of the Amount of the Surfactant . . . . . . . . . . . . . . . . . . 187
4.1.2.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
4.1.3 Shear Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
4.1.3.1 Relative Shear Stability of Miniemulsion and Macroemulsion Latexes . . . 190
4.1.3.2 Effect of Large Particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
4.1.3.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
4.2 Monomer Transport Effects . . . . . . . . . . . . . . . . . . . . . . . . . . 194
4.2.1 Polymerization of Highly Water-Insoluble Monomers . . . . . . . . . . . . 194
4.2.2 Copolymer Composition Distribution . . . . . . . . . . . . . . . . . . . . 195
4.2.2.1 Batch Copolymerization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
4.2.2.2 Semibatch Copolymerization . . . . . . . . . . . . . . . . . . . . . . . . . 200
4.2.2.3 Copolymerization in a Continuous Stirred Tank Reactor . . . . . . . . . . 203
4.2.3 Interfacial Polymerization . . . . . . . . . . . . . . . . . . . . . . . . . . . 204
4.3 Multiphase Particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
4.3.1 Hybrid Miniemulsion Polymerization . . . . . . . . . . . . . . . . . . . . . 208
4.3.1.1 Alkyds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
4.3.1.2 Polyester . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
4.3.1.3 Polyurethane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
4.3.1.4 Other Hybrids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
4.3.1.5 Artificial Miniemulsions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
4.3.2 Nanoencapsulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
4.4 Controlled Free Radical Polymerization . . . . . . . . . . . . . . . . . . . . 216
4.4.1 Nitroxide-Mediated Polymerization . . . . . . . . . . . . . . . . . . . . . . 216
4.4.1.1 Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
4.4.1.2 Effects Of Segregation And Heterogeneity . . . . . . . . . . . . . . . . . . 219
4.4.1.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
4.4.2 Atom Transfer Radical Polymerization . . . . . . . . . . . . . . . . . . . . 223
4.4.2.1 Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
4.4.2.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
4.4.3 Reversible Addition Fragmentation Polymerization . . . . . . . . . . . . . 228
4.4.3.1 Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
4.4.3.2 Effects of RAFT and Transfer Agents on Emulsion Polymerization Kinetics 230
4.4.3.3 Application of RAFT in Miniemulsion . . . . . . . . . . . . . . . . . . . . 234
4.4.4 Colloidal Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
4.5 Other Applications and Future Directions . . . . . . . . . . . . . . . . . . 242
4.5.1 Anionic/Cationic Polymerization in Miniemulsions . . . . . . . . . . . . . 242
4.5.2 Polycondensation in Miniemulsions . . . . . . . . . . . . . . . . . . . . . . 243
Miniemulsion Polymerization 131

4.5.3 Other Polymerizations in Miniemulsions . . . . . . . . . . . . . . . . . . . 244

4.5.4 Other Miniemulsion Applications . . . . . . . . . . . . . . . . . . . . . . . 245
4.5.5 Future Directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246

Abstract The subject of miniemulsion polymerization is reviewed. The approach taken is

one that combines a review of the technology with historical and tutorial aspects. Rather
than developing an absolutely exhaustive review, a tutorial approach has been taken, em-
phasizing the critical features and advantages of miniemulsion polymerization. In keeping
with this tutorial approach, a discussion of conventional emulsion polymerization is included
in order to be able to compare and contrast miniemulsion polymerization and conventional
emulsion polymerization later in the review.Areas where miniemulsion polymerization has
been adopted commercially, or where it is likely to be adopted are highlighted.

Keywords Miniemulsion · Polymerization · Emulsion · Free radical · Colloid

Abbreviations and Symbols

AMBN 2,2¢-azobis(2-methylbutyronitrile)
ATRP atom transfer radical polymerization
CA cetyl alcohol
CMC critical micelle concentration
CSA camphorsulfonic acid
CSTR continuous stirred tank reactor
CTA chain transfer agent
CTAB cetyltrimethylammonium bromide
DDM dodecyl mertcaptan
DLS dynamic light scattering
DOM dioctyl maleate
DPPH diphenylpicrylhydrazol
EHA 2-ethylhexyl acrylate
Fi copolymer composition, monomer i
fi monomer composition, monomer i
GPC gel permeation chromatography
HD hexadecane
Keq equilibrium constant
kp propagation rate constant
KPS potassium persulfate
LPO lauroly peroxide
MAETAC 2-(methacryloyoxy)ethyl]tri-methyl ammonium chloride
Maq molarity, aqueous phase
mij ratio of molar size, i to j
mM millimolar
MMA methyl methacrylate
[Mp] concentration of monomer in the particle
MWD molecular weight distribution
n– average number of radicals per particle
N number of particles per liter
NA Avogadro’s number
132 F. J. Schork et al.

nm nanometer
NMA n-methylol acrylamide
NMCRP nitroxide-mediated controlled radical polymerization
NMR nuclear magnetic resonance spectroscopy
PFR plug flow reactor
pMS paramethyl styrene
PS polystyrene
PSD particle size distribution
PST polystyrene
PVA polyvinyl alcohol
PVAc polyvinyl acetate
QSSA quasi-steady state approximation
r radius
ri reactivity ratio
R gas constant
RAFT reversible addition-fragmentation chain transfer
Rp rate of polymerization
RX organic halide
SE Smith Ewart
SPS sodium persulfate
T temperature
Tg glass transition temperature
–
Vi partial molar volume of i
VAc vinyl acetate
VD vinyl n-decanoate
VEH vinyl 2-ethylhexanoate
VEOVA vinyl neo-decanoate (vinyl versatate)
VH vinyl hexanoate
VS vinyl stearate
W Fuchs stability ratio
X halogen
cij Flory Huggins interaction parameter
–
DGi partial molar free energy or phase i
g interfacial tension
ji volume fraction of component i
m viscosity
mi chemical potential of component i
mm micron
%wt percent by weight

1
Introduction

Over the past 25 years, miniemulsion polymerization has grown from being
the subject of a single paper to being the focus of a great deal of academic
and industrial research. During that time, some products have been commer-
cialized based on this technology, and in the next few years a number of new
Miniemulsion Polymerization 133

commercializations of the technology are expected to take place. This text

attempts to trace the development of miniemulsion polymerization, the physics
and chemistry behind it, its unique features, and its potential for future ap-
plications. It is not an exhaustive bibliography of published work, but it does
cite significant and/or representative papers. Previous reviews of this field can
be found in [1–3].

1.1
Dispersed-Phase Polymerization

Free radical polymerization may be carried out in various media. Bulk poly-
merization is the simplest, but while the reactants (monomers) are most often
liquid, the product (polymer) is solid. This leads to problems when removing
the polymer from the reactor. In addition, since most free radical polymeriza-
tions are highly exothermic, the high viscosity of the monomer/polymer mix
inhibits the removal of the heat of reaction. Solution polymerization will
reduce, to some extent, the viscosity of the polymerizing mass, but it brings
with it the environmental and health issues of organic solvents. In addition, the
solvent reduces the monomer concentration, and hence the rate of polymer-
ization. Finally, recovery and recycling of the solvent can add substantially
to the cost of the process. Nevertheless, solution polymerization of vinylic
monomers is used in a number of commercial processes.
An alternative to solution polymerization is the whole realm of dispersed-
phase polymerization. In this class of processes, the liquid monomer is dispers-
ed in a second, continuous phase, usually water. As the monomer polymerizes,
the viscosity of the dispersion remains low, aiding the removal of the heat of
polymerization. If the dispersed phase is water, the high thermal conductivity
provides a very effective heat transfer medium. The high specific heat and large
latent heat of vaporization provide a large safety margin in the event of a run-
away polymerization. In addition, water is plentiful, nontoxic, environmentally
friendly, and inexpensive.
If an oil-soluble monomer is dispersed in a continuous aqueous phase with-
out the use of surfactants, suspension polymerization results. The viscosity of
the resulting suspension will remain essentially constant over the course of the
polymerization. Oil-soluble free radical initiators are used to effect polymer-
ization. The monomer is dispersed into beads by the action of an agitator. Since
little or no surfactant is used, no emulsification takes place, and, if the agita-
tion is stopped, the monomer will form a separate bulk phase, usually above
the aqueous phase. The monomer is polymerized by the initiator within
the droplets, forming polymer beads of approximately the same size as the
monomer droplets (0.1–10 mm diameter). The product can be readily sepa-
rated from the aqueous phase (via filtration or decantation) in the form of
macroscopic particles or beads, which can be easily packaged and/or trans-
ported. Heat transfer is facilitated by the presence of the continuous aqueous
phase. Blocking agents such as clays or talcs are used to prevent particle ag-
134 F. J. Schork et al.

glomeration. Small quantities of nonionic surfactants (such as polyvinyl

alcohol) may be used to impart particle stability and to disperse the blocking
agent. Viscosity enhancers such as carboxymethylcellulose may be used to
inhibit particle settling. The loci of polymerization are the monomer/polymer
beads. Due to the large sizes of the beads, such systems are suspensions rather
than emulsions or stable dispersions. The particles are kept suspended by
agitation throughout the course of the polymerization. The suspension poly-
merization process is described in detail by Trommsdorff and Schildknecht [4].
Kinetically, each bead acts as a small independent reactor; there is little
exchange of material between the beads. Since there is no solvent present at the
locus of polymerization, the kinetics are those of bulk polymerization, with the
molecular weight distribution (MWD) characteristics similar to those of bulk
or solution polymerizations. If water-soluble initiator is used in a suspension
polymerization, very little polymerization will occur, since few free radicals will
reach the locus of polymerization in the monomer beads.
If surfactant is added to a suspension polymerization system, a number of
phenomena may occur. If the surfactant is added in small amounts (below the
critical micelle concentration or CMC), the reduction in interfacial tension
between the organic and aqueous phases will result in smaller monomer
droplets, but it has hardly any other effect. If surfactant is added above the
CMC, and an oil-soluble initiator is used, the process is commonly termed a
microsuspension polymerization. Due to the reduced interfacial tension, the
droplet diameter (and hence bead diameter) is reduced to approximately
10–40 mm. Little polymerization takes place in the aqueous phase or in particles
generated from surfactant micelles because of the hydrophobic nature of the
initiator. However, some smaller particles initiated from surfactant micelles
may be found. The kinetics are still essentially those of a bulk free radical poly-
merization. Microsuspension polymerization is used to produce pressure-
sensitive adhesives for repositionable notes.
If a water-soluble initiator is used, both droplet nucleation (to form large
particles) and micellar nucleation (to form submicron particles) may occur.
The balance between these two mechanisms is a function of surfactant type
and amount and monomer water solubility. In general, small particles derived
from micellar nucleation will dominate, giving what is know as a conventional
emulsion polymerization system. The kinetics of conventional emulsion poly-
merization are no longer those of bulk free radical polymerization, since the
small sizes of the loci of polymerization introduce segregation effects, in which
bimolecular termination of the growing polymer chains is suppressed by
the small likelihood of two growing chains existing in the same particle. This
results in higher molecular weight at constant polymerization rate. The particle
diameter will range from 50 to 500 nm.
If the monomer droplet size in a conventional emulsion polymerization can be
reduced sufficiently (see below), the loci of polymerization become the monomer
droplets. This system is referred to as a miniemulsion polymerization and will be
discussed in detail below. The particle diameter will range from 50 to 500 nm.
Miniemulsion Polymerization 135

If the surfactant concentration in a macroemulsion is greatly increased,

or if the monomer concentration is greatly reduced, a microemulsion results.
Microemulsions are thermodynamically stable systems in which all of the
monomer resides within the micelles. At high surfactant concentration, the
micelles may form a bicontinuous network, rather than discrete micelles.
Polymerization (with water- or oil-soluble initiator) of the monomer within a
microemulsion is referred to as microemulsion polymerization. The particles
produced in this way are extremely small, ranging from 10 to 100 nm.

1.2
The Concept of Miniemulsion Polymerization

The mechanisms of conventional emulsion and miniemulsion polymerizations

are, in some ways, significantly different. A conventional unseeded (no small
particles added at the beginning) batch emulsion polymerization reaction can
be divided into three intervals. Particle nucleation occurs during Interval I and
is usually completed at low monomer conversion (2–10%) when most of the
monomer is located in relatively large (1–10 mm) droplets. Particle nucleation
takes place when radicals formed in the aqueous phase grow via propagation
and then enter into micelles or become large enough in the continuous phase to
precipitate and form primary particles which may undergo limited flocculation
until a stable particle population is obtained. Significant nucleation of particles
from monomer droplets is discounted because of the small total surface area of
the large droplets. Interval II involves polymerization within the monomer-
swollen polymer particles, with the monomer supplied by diffusion from the
droplets. Interval III begins when the droplets disappear – or at least reach a
polymer fraction similar to that of the particles – and continues to the end of the
reaction. Because the nucleation of particles can be irreproducible, commercial
emulsion polymerizations are often “seeded” with polymer particles of known
size and concentration, manufactured specifically for use as seed particles. In
this paper, for the purpose of clearly distinguishing between conventional emul-
sions and miniemulsions, the term macroemulsion will be used for the former.
In addition, a latex will be defined as a polymerized monomeric emulsion, while
the term emulsion will refer to an unpolymerized monomeric emulsion.
Miniemulsion polymerization involves the use of an effective surfactant/
costabilizer system to produce very small (0.01–0.5 mm) monomer droplets.
The droplet surface area in these systems is very large, and most of the sur-
factant is adsorbed at the droplet surface. Particle nucleation is primarily via
radical (primary or oligomeric) entry into monomer droplets, since little
surfactant is present in the form of micelles, or as free surfactant available to
stabilize particles formed in the continuous phase. Both oil- and water-soluble
initiators may be used; the important feature is that the reaction proceeds
by polymerization of the monomer in these small droplets, so there is no true
Interval II. The mechanisms of macro- and miniemulsion polymerization are
shown schematically in Fig. 1.
136 F. J. Schork et al.

Interval I Emulsified
Monomer
swollen monomer
micelle droplet
~ 100 Å ~ 10 μ
Emulsifier
Monomer
swollen
polymar
particle
~ 500 Å
Continuous aqueous phase

Interval II Emulsified
monomer
droplet
~ 10 μ
Monomer
swollen
polymar
particle
~ 500 Å

Continuous aqueous phase

Interval III

Monomer
swollen
polymar
particle
~ 1000 Å

Continuous aqueous phase

a

b
Fig. 1 Macroemulsion polymerization (a) versus miniemulsion polymerization (b)

1.3
Publication History

Miniemulsion polymerization began with a single paper [5]. Professor John

Ugelstad of Norway was visiting John Vanderhoff in the Department of
Chemistry at Lehigh University. Their discussions lead to speculation about the
possibility of nucleation and polymerization in very fine monomer droplets
during emulsion polymerization. Micellar nucleation is considered to be the
Miniemulsion Polymerization 137

Fig. 2 Growth in the number of publications on miniemulsion polymerization. (Courtesy

of Prof. M. S. El-Aasser)

dominant mechanism of particle nucleation in emulsion polymerization, but

might nucleation in droplets occur if the droplets were sufficiently small? The
task of investigating these ideas was given to a new postdoctoral fellow in
Vanderhoff ’s lab, Dr. (now Provost) Mohamed S. El-Aasser. From this point, the
field grew slowly. Figure 2 shows the total number of papers per year published
on miniemulsion polymerization. It may be seen that, after a slow start, the
number of contributions in the field has risen rapidly, as the scientific com-
munity made use of the basic research of the early years, and began to see the
commercial utility of the process.

2
Miniemulsion Polymerization

In order to adequately discuss miniemulsion polymerization, it will be neces-

sary to review the mechanism of macroemulsion polymerization.

2.1
The Mechanism of Macroemulsion Polymerization

Macroemulsion polymerization is a complex process. The literature con-

tains extensive reviews of emulsion polymerization theory [6–11]. Only a brief
review of the current state of the literature is given here. The theory of emul-
138 F. J. Schork et al.

sion polymerization revolves around one equation (neglecting the aqueous

phase),

Rp = kp[M]pNpn–/NA

where each of these terms is explained below.

The earliest qualitative theory of emulsion polymerization was developed
by Harkins [12] and was quickly quantified by Smith and Ewart (SE) [13]. Al-
though this theory only holds for the special case of water-insoluble monomers,
it is the typical starting point for most other theories. This theory is based on
the batch emulsion polymerization of styrene. It includes three intervals, as
depicted in the left half of Fig. 1. The first interval begins with the initiation
of the reaction and continues until all micelles become nucleated or are used
up as surface stabilizing agents.At this point particle formation ceases. During
Interval II, the particles grow at a constant rate in the presence of monomer
droplets. Once the monomer droplets disappear, Interval III begins. The
monomer concentration in the particle decreases and the reaction within the
particles becomes diffusion-limited throughout the remainder of the poly-
merization.
Micellar nucleation may not be the only, or even primary process of nuclea-
tion and growth. Other mechanisms are discussed later. To provide a com-
prehensive model for emulsion polymerization, the applicability of each
mechanism must be considered.
Equation 1 illustrates the concept of radical segregation, that is significant
in macroemulsion, miniemulsion and microemulsion polymerization. In a bulk
or solution free radical polymerization, the radical flux can be increased in
order to increase the rate of polymerization, but only at the cost of reducing
molecular weight. Equation 1 indicates that one might increase the rate of poly-
merization by increasing the surfactant level. This does not adversely impact
the molecular weight of the product. The ability to decouple reaction rate and
molecular weight come about because of the segregation of the free radicals in
segregated disperse-phase polymerization. Since each radical is confined to
a particle, bimolecular termination is suppressed, resulting in higher rates
of polymerization and molecular weight. Segregation occurs only when n– is
reasonably close to unity, or lower than unity.

2.1.1
Interval I – Particle Nucleation

Nucleation mechanisms are generally divided into three types: micellar, homo-
geneous, and droplet. Statistically, all three types can occur simultaneously in
every reaction. However it is the preponderance of one mechanism above the
others in a given system that causes authors to consider only one in their
studies. Numerous extensions have been made to the SE micellar nucleation
theory in an effort to furnish explanations for experimental results observed
Miniemulsion Polymerization 139

for monomers with slight water solubilities. Detailed reviews of these exten-
sions are readily available [10, 11, 14, 15]. Monomer droplet nucleation is nor-
mally neglected, except when considering mini- and microemulsions, due to its
insignificant contribution to the particle number and the size distribution.

2.1.1.1
Micellar Nucleation

All quantitative theories based on micellar nucleation can be developed from

balances of the number concentrations of particles, and of the concentrations
of aqueous radicals. Smith and Ewart solved these balances for two limiting
cases: (i) all free radials generated in the aqueous phase assumed to be absorb-
ed by surfactant micelles, and (ii) micelles and existing particles competing for
aqueous phase radicals. In both cases, the number of particles at the end of
Interval I in a batch macroemulsion polymerization is predicted to be propor-
tional to the aqueous phase radical flux to the power of 0.4, and to the initial
surfactant concentration to the power of 0.6. The Smith Ewart model predicts
particle numbers accurately for styrene and other water-insoluble monomers.
Deviations from the SE theory occur when there are substantial amounts of
radical desorption, aqueous phase termination, or when the calculation of
absorbance efficiency is in error.
Deviations with respect to order from the SE theory increase as the monomer
water solubility increases.

2.1.1.2
Homogeneous Nucleation

Although the SE micellar nucleation theory explains data for certain systems,
it fails for others. This has led some authors to propose a different mechanism
for nucleation. In the homogeneous nucleation theory, aqueous phase radicals
polymerize to form oligomers. These continue to grow until they reach a criti-
cal chain length, the size of a primary particle, and then precipitate. Through-
out the growth process, the oligomers may also flocculate or coagulate. This
theory is typically employed for relatively water-soluble monomers. Slight va-
riations of this theory have also been postulated.
Prior to 1952, little evidence for homogeneous nucleation existed [16, 17]. In
1952, Priest [18] studied the polymerization of vinyl acetate and presented a
qualitative theory for homogeneous nucleation. He concluded from experi-
mental work that aqueous phase nucleation is important in systems with
monomers that have a relatively high water solubility. Primary particle forma-
tion occurs throughout the course of the reaction. During later periods of
the reaction, large monomer-swollen polymer particles act as sinks for these
primary particles, encouraging coagulation.
In 1968, Roe [19] developed the SE limiting case equations for particle num-
ber from the homogeneous nucleation theory. He showed that the SE equation
140 F. J. Schork et al.

for particle nucleation is not unique to micellar nucleation, but results from
the SE assumptions. By assuming that (i) the nucleation stops upon depletion
of micelles, (ii) the volumetric growth rate is constant, and (iii) the radical
absorption is strictly a function of radical generation, he showed that the SE
dependency on radical flux and surfactant concentration could be generated
from homogeneous nucleation theory.
Fitch and Tsai [20, 21] developed a quantitative theory for homogeneous
nucleation. By using the collision theory for radical capture, Fitch [22] has shown
that the rate of radical capture is a function of radical production, particle
number, particle size, and diffusion distance. Primary particles may coagulate
with each other because of their small size and lower surface charge. As par-
ticles coagulate, the surface to volume ratio decreases, which causes an increase
in surface potential. When particles become sufficiently large, coagulation
ceases due to insufficient kinetic energy to overcome the biparticle surface
repulsion. Fitch and Tsai have provided experimental support for this theory by
polymerizing MMA with different initiators.
Ugelstad and Hansen [11] proposed that free radicals in the aqueous phase
propagate with dissolved monomer. Primary particles form by precipitation
when a critical chain length is reached. During growth from a monomer radical
to a primary particle, each oligomer can (i) terminate with other radicals, (ii)
precipitate if its length exceeds the critical chain length, or (iii) be captured by
particle.
A fundamental extension to the homogeneous nucleation theory was pro-
posed by Lichti et al. [23] and Feeney et al. [24]. Their theory is based on the posi-
tive skewness of the particle size distribution (PSD) as a function of volume
during Interval II. This implies that the rate of nucleation during Interval I
increases with time until it eventually drops off at the cessation of nucleation.
Lichti and Feeney claim that micellar nucleation or one step homogeneous
nucleation incorrectly predict either decreasing or constant nucleation rates.
This theory has been given the name coagulative nucleation. According
to Lichti and Feeney’s mechanism, precipitated “precursor particles” undergo
coagulation to form “true” or “mature” latex particles. A precursor particle is
unstable and said to be formed in either a micelle or the aqueous phase. Due
to their small size and hydrophilic nature, the precursors have low swelling
capacity and high radical desorption rates. Consequently, the propagation rate
is low and these precursors tend to coagulate with other precursors or mature
latex particles. These conclusions then rule out micellar nucleation as a possible
mechanism for precursor formation.
More recently, Maxwell et al. [25] suggested that the values to be used for
the critical chain length are much smaller than originally thought. They also
suggested that oligomeric radical capture is independent of particle size and
limited by the rate of propagation of the radical in the aqueous phase. However,
this theory does not consider variations in other parameters with particle size.
Miniemulsion Polymerization 141

2.1.1.3
Droplet Nucleation

Nucleation of monomer droplets has typically been neglected in emulsion

polymerization. The large diameter (1–10 mm) and small number (~1013 versus
1021 micelles) of droplets in macroemulsions usually makes their consideration
of no importance. Regardless of this, all droplets do get nucleated, because of
their large size. These droplets show up in TEM photographs as abnormally
large particles in very low concentrations. In 1973, Ugelstad et al. [5] showed
how submicron styrene monomer drops can be made stable enough to become
numerically significant in nucleation when a cosurfactant is used to enhance
the stability of the smaller droplets. Table 1 shows how the micelle number
varies with monomer droplet size at constant surfactant levels. Chamberlain
et al. [26] have presented experimental evidence that the efficiency of radical
capture by droplets is much lower than that for micelles or particles. This would
affect the results in Table 1. Ugelstad et al. [27, 28] have shown how nucleation
of monomer droplets can lead to latexes with large monodisperse particles.
However, if insufficient shear or cosurfactant is used, the potential for pro-
duction of bimodal PSD’s exists [29, 130]. This could be desirable in certain
instances.
The distinguishing feature of droplet nucleation as opposed to micellar or
homogeneous nucleation is the nature of the particle at “birth”. Droplets, which
are nucleated into particles, begin as nearly 100% monomer. Micellar or homo-
geneous nucleated particles start out with much lower monomer concen-
trations and eventually swell to around 60% (for MMA) in the presence of
monomer droplets. This fundamental difference may lead to large differences
between miniemulsion and macroemulsion polymerizations in radical desorp-
tion and/or intraparticle termination during Intervals I and II.

Table 1 Variation of number of micelles with droplet size for MMA droplets and SLS
surfactant in a basic emulsion recipe

Monomer droplet diameter (mm) 10.0 1.0 0.5 0.1

–13 –16 –17
Volume of monomer droplet (liter) 5.2¥10 5.2¥10 6.5¥10 5.2¥10–19
Number of monomer droplets (#/1) 7.8¥1011 7.8¥1014 6.3¥1015 7.8¥1017
Total area of droplets (m2/1) 247 2470 4930 24700
Surface area micelles (m2/1) 5246 3027 562 0
Number of micelles (#/1) 1.5¥1020 8.6¥1019 1.6¥1019 0.0
Total number of preparticles (#/1) 1.5¥1020 8.6¥1019 1.6¥1019 7.8¥1017
Droplet percent area 4.5 45 90 100
Droplet percent number 5.3¥10–7 9.2¥10–4 0.040 100

Base case parameters: temperature=50 °C; monomer/water ratio=0.4 g/g; surfactant CMC=
0.004 mol/L(aq); surfactant concentration=0.020 mol/L; surfactant surface area=0.57 nm2/
molecule; molecules per micelle=62; monomer watersaturation=0.137 mol/L(aq)
142 F. J. Schork et al.

2.1.1.4
Competition for Radicals

As pointed out above, particle nucleation includes all three mechanisms –

micellar, homogeneous, and droplet, since these mechanisms may compete and
coexist in the same system. Often one will dominate. Therefore, any general
model of emulsion polymerization should include all three mechanisms.
Hansen and Ugelstad [31] and Song [10] have presented probabilities for each
of these mechanisms in the presence of all three.
The competition for oligomeric radicals also includes particles that have
been created. In miniemulsion polymerizations, the nucleation of one droplet
results in the formation of one particle of equal surface area. Therefore, nucle-
ation therein has little effect on competition for radicals. This is not so with
macroemulsions, since both micellar and homogeneous nucleation result in
a large shift in the surface area from micelles to particles as the particles are
created and grow.

2.1.2
Interval II – Particle Growth

SE Interval II begins at the cessation of nucleation, or in light of the nucleation

theory just reviewed, when the particle number becomes relatively constant.
Most theories developed for this interval assume a constant particle number
and use the quasi-steady-state approximation (QSSA) for average number of
radicals per particle. The kinetics and mechanisms of Interval II have been
some of the most studied aspects of macroemulsion polymerization. SE Inter-
val II ends when the monomer droplets disappear and the monomer concen-
tration in the particles begins to decrease.
The rate of polymerization during Interval II is usually considered constant
for two reasons. The monomer concentration within the particle, as defined by
equilibrium thermodynamics, is approximately constant in the presence of
excess monomer. Mass transfer is assumed to be fast and particle size has little
effect on this concentration. Secondly, emulsion polymerization kinetics tend
to give a constant radical concentration within the particles during Interval II.
Therefore, the rate of polymerization given by Eq. 1 is approximately constant
until the end of Interval II, where [M]p, n–, and kp begin to change. These two
assumptions have been substantiated by experimental observations and are
considered reasonable. The challenge for Interval II is to determine the average
number of radicals per particle. Particle monomer concentration can be deter-
mined as a function of particle radius by an equilibrium relationship such as
the Morton [32] equation that considers both surface energy and mixing
energy. Propagation rate coefficients have been widely studied and are readily
available [33]. The particle number concentration is assumed constant.
Smith and Ewart developed three limiting cases for Interval II. Each of these
cases can be generated through a balance of Npi (the number of particles con-
Miniemulsion Polymerization 143

taining i radicals), where the number of particles is considered constant (no

nucleation). For Case 1, Smith and Ewart assume Np0
Np1
Np2
… . For this
case n– will be significantly less than 0.5. This case occurs when significant
monomeric radical desorption occurs, and is more common with monomers
of significant water solubility.
Case 2, assumes instantaneous termination of the existing radical with an
entering radical. In this case, each particle will contain either zero or one radical,
and n– becomes 0.5. Styrene generally follows Case 2 kinetics. Smith and Ewart
Case 3 assumes that both desorption from particles and aqueous phase termi-
nation may be neglected, and so n–
1.0. This occurs with large particles, and
in the limit results in bulk kinetics.
Coagulation of latex particles during Interval II is often neglected. If sur-
factant is available in great enough proportion, the particles will remain stable
throughout the reaction.

2.1.3
Interval III – Gel and Glass Effects

Interval III begins when all monomer droplets have vanished and/or the aque-
ous phase becomes unsaturated. Since each droplet in a macroemulsion actual-
ly absorbs radicals, they cannot disappear but rather shrink to a point where
they have no excess monomer. The monomer in the aqueous phase decreases
corresponding to the decrease in the particles. The conversion at which Inter-
val III begins varies for different monomers and systems, but is typically around
40 to 50%. However, it may not be as distinguishable in miniemulsions due to
early initiation of the gel effect.
As the monomer within the particles is consumed by polymerization, the
viscosity rises within the particles and the diffusion rate of the polymeric radi-
cals decreases. This causes a reduction in the rate of termination, which cor-
responds to a dramatic increase in the radical concentration. A higher radical
number within the particle results in an “auto acceleration” in the rate of poly-
merization. Common practice is to model this auto acceleration or gel effect by
decreasing the termination rate constant byseveral orders of magnitude as a
function of percent monomer in the particle.A free volume approach has been
used by Sundberg et al. [34]. Gilbert and coworkers [35] suggest a completely
empirical approach from precise experimental data. Empirical correlations
used in modeling the gel effect in bulk or solution polymerization have also
been modified for use in emulsion processes [36–38].
The problem with applying correlations derived from other systems to
emulsion polymerization is twofold. First, normal macroemulsion particles
are said to be created with 30 to 40% monomer in them and so the unbiased
(zero conversion) termination rate is unknown. Secondly, the diffusional
limitations in particles might be quite different from those observed in bulk or
suspension polymerizations. It is for these reasons that an empirical approach
is suggested.
144 F. J. Schork et al.

If the reaction temperature is below the polymer glass transition tempera-

ture and the amount of monomer in the particle decreases far enough, the glass
effect may become important. The polymerization rate virtually goes to zero
because the particle becomes so internally viscous, essentially glasslike, that
the diffusion of monomer to the radicals is limited. The glass transition point
varies for different polymers. This effect has also been studied by several
authors [34, 39, 40].

2.2
The Mechanism of Miniemulsion Polymerization

In the previous discussion of macroemulsion polymerization, all three forms

of particle nucleation were discussed. In macroemulsion polymerization,
micellar and homogenous nucleation dominate. This is because the large sizes
of the monomer droplets, and their consequent low interfacial area, makes
them ineffective in competing for water-borne free radicals. Droplet nucleation
undoubtedly takes place in macroemulsion polymerization, but it is generally
considered to be insignificant. If the monomer droplet size can be reduced to
below 0.5 µm, two phenomena will occur. First, the droplets will be able to com-
pete successfully for water-borne free radicals with any remaining micelles.
Second, the huge increase in interfacial area caused by the reduction in droplet
size will result in a huge increase in interfacial area. This new interface will
require a monolayer of surfactant to remain stable. The surfactant necessary to
support this large interfacial area will come from the break-up of surfactant
micelles. In a properly formulated miniemulsion, all micelles will be sacrificed
to support the droplet interfacial area. Therefore, not only do the small droplets
compete effectively for micelles, their presence causes the destruction of
the micelles, leaving droplet nucleation as the dominant particle nucleation
process.
Miniemulsions are produced by the combination of a high shear to break up
the emulsion into submicron monomer droplets, and a surfactant/costabilizer
system to retard monomer diffusion from the submicron monomer droplets.
Both are necessary to effect predominant droplet nucleation (nucleation in
which a preponderance of the particles originate from droplets rather than
from micelles or from homogeneous nucleation). High shear is provided by a
sonicator or a mechanical homogenizer. The surfactant is necessary to retard
droplet coalescence caused by Brownian motion, settling or Stokes law cream-
ing or settling. The costabilizer (also referred to in earlier works as a cosur-
factant) prevents Ostwald ripening [41]. When a liquid emulsion is subjected
to high shear, small droplets will result. There will still be a statistical distribu-
tion of droplet sizes. If the monomer is even slightly soluble in the continuous
aqueous phase (and most are, as evidenced by the fact that Interval II of macro-
emulsion polymerization takes place), the monomer will, over time, diffuse from
the smaller monomer droplets into the larger ones. This results in a lower
interfacial area (and interfacial energy), since the loss in interfacial area of the
Miniemulsion Polymerization 145

smaller droplets is larger than the gain in interfacial area of the larger ones. The
reduction in interfacial energy is the driving force for degradation of the small
droplets.
If Ostwald ripening is allowed to continue unchecked, creaming of the
monomer will occur as the droplet sizes become large enough for Stokes law
creaming to occur. This will occur in a matter of seconds to minutes. If the
system is initiated, bulk polymerization of the monomer layer will occur. If the
emulsion is stirred, an emulsion of large monomer droplets (of the order of
those of a macroemulsion) will result, and if the stirred emulsion is initiated,
macroemulsion polymerization will take place. A costabilizer functions to
prevent Ostwald ripening by retarding monomer diffusion from the smaller
droplets to the larger. Costabilizers should be highly insoluble in the aqueous
phase (so that they will not diffuse out of the droplets), and highly soluble in
the monomer droplets. Under these conditions, diffusion of the monomer out
of the smaller droplets results in an increase in the concentration of the co-
stabilizer in those particles (since, by definition costabilizers are too insoluble
in the aqueous phase to leave the droplet). The increase in free energy associ-
ated with the concentration of the costabilizer balances the decease from
reduced interfacial area caused by Ostwald ripening, and, at some point, ripen-
ing stops. Since all costabilizers are somewhat water-soluble, Ostwald ripening
will proceed, but on a timescale of months, which is unimportant since the
timescale of polymerization is minutes to hours. This phenomenon is shown
in Fig. 3 [42]. Here a macroemulsion and a miniemulsion of methyl methacry-
late (same recipe, but with no costabilizer, and no high shear for the macro-
emulsion) are shown after three hours without mixing. The macroemulsion is
completely creamed, since Stokes law creaming has taken place on the large
monomer droplets. The miniemulsion has not creamed, since Brownian motion
is sufficient to prevent creaming of the submicron monomer droplets. Similar
miniemulsions have remained stable for six months or more.

Fig. 3 Macroemulsion and miniemulsion after three hours

146 F. J. Schork et al.

In their original discovery of miniemulsion polymerization, Ugelstad and

co-workers [5] used either cetyl alcohol (CA: water solubility estimated at
6¥10–8 [43]) or hexadecane (HD: water solubility estimated at 1¥10–9 [43]) to
retard monomer diffusion from submicron monomer droplets. Both CA and
HD, referred to here as costabilizers, are volatile organic components and are
therefore not entirely desirable in the final product. Other researchers have
used polymers, chain transfer agents, and comonomers as stabilizers, as will be
discussed later.
Monomer droplet stability can be understood in terms of free energy. The
partial molar free energy of adding a second component to a droplet is made
up of two terms, the partial molar free energy of mixing and the interfacial
partial molar free energy. The partial molar free energy of mixing (the Flory
Huggins expression [44]) can be combined with the interfacial partial molar
free energy to give
– –
DGi 2
2Vig
= lnji = (1 + m )j
ij j + c j
ij j + (2)
RT RTr
6 7

Ugelstad et al. [45] have applied this equation to various monomers and sur-
factants. It is clear from this equation that the free energy increases as the
phase diameter decreases. The smaller the monomer droplet, the less stable it
is. Therefore, a driving force exists for the monomer to diffuse from a small
droplet to a larger one. Over time, non-monodisperse systems of droplets of
pure monomer will decrease in number as the smaller droplets swell the larger
ones and then disappear. Jannson [46] has shown that this occurs in unagitated
systems, and that the timescale for diffusional instability can be on the order
of seconds.
Prior to 1962, droplets below 1 mm were considered too unstable to partici-
pate in the nucleation process. In 1962, Higuchi and Misra [47] proposed that
the addition of a water insoluble compound to the monomer will enhance the
stability of small droplets by prohibiting diffusion. In 1973, Ugelstad et al. [48]
showed how submicron styrene droplets could be made stable enough to par-
ticipate in the nucleation processes by adding small amounts of cetyl alcohol.
Later, Ugelstad [48] used Eq. 2 to explain these experimental observations.
It can be shown [49] for two phases in equilibrium that the partial molar free
energies must be equal. In an emulsion (or miniemulsion) there are three
phases: monomer droplets, the aqueous phase and polymer particles. Since
monomer is soluble in all of these phases, the equilibrium condition requires
that the three phases have equal partial molar free energies. In the presence of
monomer droplets, emulsion polymer particles contain 30–80% monomer in
them. Therefore, they are said to be “swollen” with monomer. Ugelstad et al. [48]
and Azad and Fitch [50] have shown that addition of a third water-insoluble
component to a swollen polymer particle can increase the monomer to polymer
ratio. They have shown that an optimum chain size for the additive exists since
the solubility of the additive increases as the chain size decreases. They found
Miniemulsion Polymerization 147

that the optimum hydrocarbon stabilizer is hexadecane. Others have found that
if a fatty alcohol is used as the stabilizer, the minimum chain length required
is 12 carbon atoms [51].
Ugelstad et al. [52] have shown how this theory may be used to devise a
method to prepare large monodisperse particles of predetermined size. By
using the appropriate amount of cosurfactant, polymer particles can be swollen
with monomer to the desired size. Polymerization in conditions that prevent
additional nucleation results in large monodisperse polymer particles of size
1–100 µm. This method has been criticized by other groups as being in error
due to measurement selectivity.
If Ostwald ripening is retarded by using a costabilizer, predominant droplet
nucleation can be achieved. This is the basis of miniemulsion polymerization.
One of the first comprehesive studies of miniemulsion polymerization was
done on styrene by Choi et al. [53].

2.3
Mathematical Modeling of Miniemulsion Polymerization

Various mathematical treatments of specific mechanisms within the miniemul-

sion polymerization reaction abound. This section will be limited to those
papers that attempted to model the overall miniemulsion polymerization re-
action. Perhaps the earliest (1981) serious attempt to model this system was
that by Chamberlain, Napper and Gilbert [54]. Balances of the number of
droplets, number of polymer particles and monomer conversion were cons-
tructed for batch miniemulsion polymerization. Droplets and particles were
considered to be monodisperse. Comparison of the model with experimental
data led to the conclusion that free radical entry into monomer droplets is sub-
stantially less than for ordinary macroemulsion particles. Chen, Gothjelpsen
and Schork [55] published a model of approximately the same complexity
for continuous stirred tank miniemulsion polymerization with an oil-soluble
initiator. El-Aasser and coworkers [56–60] published a series of papers focus-
ing on the modeling of miniemulsion copolymerization, particularly in relation
to monomer transport. They described monomer transport in terms of a mass
transfer coefficient and a driving force derived from an equilibrium con-
centration calculated from equating the partial molar free energies. This same
group [61] modeled seeded miniemulsion polymerization (containing both
polymeric seed particles and miniemulsion droplets) with oil-soluble initiator.
Monomer transfer by collision of droplets with particles was found to be im-
portant. Fontenot and Schork [62, 63] published a very detailed model of batch
macro- and miniemulsion polymerization, indicating the differences between
the two mechanisms, and including both micellar and droplet nucleation me-
chanisms. Significant droplet coalescence was predicted. The model was in good
agreement with data. Samer and Schork [64] published a mathematical model
of continuous stirred tank (CSTR) and plug flow (PFR) miniemulsion polymer-
ization reactors. They were able to explain why the rate of polymerization
148 F. J. Schork et al.

for miniemulsion polymerization in a CSTR is substantially higher than for

macroemulsion polymerization in the same reactor.All of the models discussed
have been particle number models, containing no information about droplet
size distribution or particle size distribution. None have attempted to model the
formation of droplets during the miniemulsification stage.
Cunningham and coworkers [65–68] have completed detailed modeling of
nitroxide mediated radical polymerization in miniemulsion. They found that
issues of distribution of the control agent between the aqueous and organic
phases can be critical to maintaining livingness.

3
Properties of Miniemulsion Polymerization

After having described the mechanism of miniemulsion polymerization and

how it differs from macroemulsion polymerization in the previous section, this
section will focus on the various mechanisms and properties of miniemulsion
polymerization.

3.1
Shear Devices

Miniemulsions are produced by the combination of a high shear device to break

up the emulsion into submicron monomer droplets with a water-insoluble,
monomer-soluble component to retard monomer diffusion from the submi-
cron monomer droplets. Both steps are necessary to effect predominant droplet
nucleation. In the absence of a high-shear device, miniemulsion systems revert
to macroemulsion polymerizations, indicating that the presence of a costabil-
izer alone is not sufficient to cause predominant droplet nucleation. The for-
mation of submicron droplets is accomplished by placing a coarse emulsion
(of monomer in water) in a high shear field. In general, it is best to form a
coarse pre-emulsion before subjecting the system to high shear. This is because
most devices that impart high shear are poor mixers, so unless a coarse emul-
sion is created first, the monomer and water phases may not be in close pro-
ximity when they enter the high shear field. A coarse pre-emulsion may be
formed by vigorous stirring of the monomer, water, surfactant mix, as would be
done to create a macroemulsion. For reasons of practicality, the costabilizer
should be dissolved in the monomer before pre-emulsification.
For laboratory investigations of miniemulsions, a variety of high-shear
devices have been used, although sonication has been the most popular. Soni-
cation, however, may not be very practical for the large-scale production of
commercial miniemulsion polymers. An effective alternative to sonication is
also driven by the need to design an efficient miniemulsion polymerization
process. A continuous process places greater demand on the shear device in
terms of energy consumption and dissipation.
Miniemulsion Polymerization 149

Fig. 4 Sonicator with cooling jacket (from [69])

The mechanism for ultrasonic emulsification is primarily that of cavitation.

A typical sonicator for emulsification consists of a velocity transformer coupled
to a transducer, capable of oscillating in a longitudinal mode, where the velocity
transformer is immersed in the liquid. Figure 4 illustrates the basic parts of a
sonicator with a continuous flow attachment, like the one used in this work. In
this case, the flow cell is secured to the velocity transformer by a flange and a
Teflon O-ring. The intensity of cavitation depends on the power delivered to the
velocity transformer, which is relayed to the transducer from a variable trans-
former or some other control device not shown in Fig. 4.
The word homogenization is somewhat inconclusive and is typically defined
as used in context. Two processes are considered here; the first is a fine clearance
valve homogenizer, and the second is a rotor-stator-type mechanical homo-
genizer. Homogenization is similar to sonication and produces submicron
droplets by a combination of mechanical shearing and cavitation.
The fine clearance valve homogenizer has been in use for nearly 100 years
for the homogenization of milk and milk products. Raw milk is an emulsion of
fat globules dispersed in a continuous skim milk phase. Without homogeniza-
tion, the fat globules would rise to the top of the milk and form a cream layer.
Homogenization reduces the average diameter of these fat globules and sub-
sequently reduces their creaming rate, extending the shelf-life of the product.
The MicroFluidizer used by many miniemulsion investigators is an example of
this type of shear device.
The rotor-stator-type mechanical homogenizer generates submicron
droplets by forcing the emulsion through small openings in the stationary
150 F. J. Schork et al.

Fig. 5 Rotor-stator homogenizer (from [69])

stator at very high speeds, as illustrated in Fig. 5. The intensity of shearing

depends on the rotor speed, which can be set anywhere from 5–35 krpm for
most modern equipment. However, at higher speeds the shearing action gen-
erates a significant amount of heat, which may harm the sample being emulsi-
fied or the machine itself. This device has been used by Samer and Schork [69]
and others, and has been shown to be effective. However, in general, the
miniemulsion droplet size achievable with a rotor-stator device is larger than
that achievable with sonication or valve homogenizers.

3.2
Choice of Surfactant

The vast majority of miniemulsion polymerizations reported in the literature

have been stabilized with anionic surfactants, probably because of the wide-
spread application of anionic surfactants in macroemulsion polymerization,
and due to their compatibility with neutral or anionic (acid) monomers and
anionic initiators. However, Landfester and coworkers [70, 71] have used the
cationic surfactants cetyltrimethyl ammonium bromide (CTAB) and cetyltri-
methyl ammonium tartrate for the production of styrene miniemulsions. They
report that these surfactants produce similar particle sizes to anionic sur-
factants used at the same levels. Bradley and Grieser [72] report the use of
dodecyltrimethyl ammonium chloride for the miniemulsion polymerization of
MMA and BA.
Wang and Schork [73] miniemulsion polymerized vinyl acetate using the
nonionic surfactant polyvinyl alcohol (PVA). They found that stable miniemul-
sions could be made with PVA and HD, but when the HD was removed, the PVA
Miniemulsion Polymerization 151

alone was not capable of functioning as both surfactant and costabilizer. In

general, the droplet diameter was greater with PVA than would be expected
with an anionic surfactant. Chern and Chen [74, 75] used nonylphenol ethoxy-
late (40 ethylene oxides per molecule) with a monomeric costabilizer such as
docecyl methacrylate or stearyl methacrylate to form stable miniemulsions of
styrene.Wu and Schork [76] used polyoxyethylene-23 lauryl ether (BRIJ-35) as
the surfactant with HD as the costabilizer to form stable miniemulsions of vinyl
acetate. Landfester and coworkers [70, 71] also used polyethylene oxide for the
miniemulsion polymerization of styrene. Luo and Schork used nonylphenyl
ethoxylate (Triton X-405) and an interfacial initiation system to miniemulsion
polymerize BA and BA copolymerized with cationic monomer. Graillat and
Guyot [77] also used Triton X-405 to produce high solids vinyl acetate emulsions
via miniemulsion polymerization.
Guyot and coworkers [78] have produced stable miniemulsions of styrene
using the polymerizable surfactant, vinylbenzylsulfosuccinic acid sodium salt.

3.3
Choice of Costabilizer

Early work in miniemulsion polymerization [5, 48–50] used either cetyl alcohol
(CA) or hexadecane (HD) to retard Ostwald ripening in submicron monomer
droplets. Both CA and HD, referred to here as costabilizers, have the requisite
properties for a costabilizer: high monomer solubility, low water solubility and
low molecular weight. The need for these properties can be seen from Eq. 2.
High monomer solubility will give a large Flory Huggins interaction parame-
ter between the costabilizer and the monomer (cij). Low water solubility will
ensure a distribution coefficient for the costabilizer that very strongly favors the
monomer drops, giving a higher volume fraction of costabilizer in the droplet.
Low molecular weight will give a high ratio of costabilizer molecules to mo-
nomer molecules (mij) in the droplet.All of these factors will enhance swelling,
or retard monomer loss via Ostwald ripening.
With cetyl alcohol, there is the complication that the polarity of the mole-
cule may cause it to reside at the surface of the droplet, imparting additional
colloidal stability. Here, the surfactant and costabilizer form an ordered struc-
ture at the monomer-water interface, which acts as a barrier to coalescence and
mass transfer. Support for this theory lies in the method of preparation of the
emulsion as well as experimental interfacial tension measurements [79]. It is
well known that preparation of a stable emulsion with fatty alcohol costabiliz-
ers requires pre-emulsification of the surfactants within the aqueous phase
prior to monomer addition. By mixing the fatty alcohol costabilizer in the water
prior to monomer addition, it is believed that an ordered structure forms from
the two surfactants. Upon addition of the monomer (oil) phase, the monomer
diffuses through the aqueous phase to swell these ordered structures. For long
chain alkanes that are strictly oil-soluble, homogenization of the oil phase is
required to produce a stable emulsion.Although both costabilizers produce re-
152 F. J. Schork et al.

latively stable emulsions, Azad et al. [80] have shown that the alkanes will pro-
duce emulsions of higher stability. A 1:3 molecular ratio of surfactant to co-
stabilizer has been shown to provide optimal stability in emulsion systems
where the costabilizer is a fatty alcohol. Shah [81] postulated that this ratio is
due to an optimum alignment of surfactant and costabilizer molecules at the
interface in microemulsions. Hallworth and Carless [82] have proposed that
the stability of an emulsion containing long chain alkanes of fatty alcohols
comes from a film at the interface which makes collisions at the interface more
elastic.
Various researchers [83–88] have concluded from experimental data that
liquid crystals of surfactants exist at the interface. These observations have
been suggested through birefringence, interfacial tension, and viscosity mea-
surements. Lack et al. [87] studied the formation of liquid crystals with fatty
alcohol costabilizers as a function of concentration, chain length, and ratio of
emulsifiers using birefringence measurements. Tertiary phase diagrams were
presented which show two regions of normal micelle formation. There are also
three regions of homogeneous anisotropic mesophases. These three meso-
phases, are (i) hexagonal rodlike aggregates of mixed micelles, (ii) lamellar
double layers with overlapping tails, and (iii) lamellar double layers dispersed
in the aqueous phase. The concentration of surfactants used in miniemulsions
is found to fall within region (iii). These “liquid crystals” were shown to form
more easily at higher emulsifier concentrations, with shorter chain alcohols,
and with sonication.
Although evidence exists for liquid crystal formation with fatty alcohol
costabilizers, it does not for systems with long chain alkanes. Delgado et al. [89]
have presented evidence that the role of hexadecane costabilizer in miniemul-
sion polymerizations is one of diffusional control. Rodriguez [90] and Delgado
[91] have reported that no optional ratio of hexadecane to SLS exists in the
preparation of miniemulsions. This provides evidence for the lack of crystal
formation. Ugelstad et al. [45] have presented evidence that alkanes are more
likely to follow the diffusion mechanism.

3.3.1
Polymeric Costabilizers

The use of polymer as a costabilizer was first reported by Reimers et al. in 1995
[92]. Conventional thinking has been that effective costabilizers must be highly
water-insoluble, highly monomer-soluble, and of low molecular weight, as
required by Eq. 2. Polymer made from the same monomer from which the
miniemulsion is to be made will be highly water-insoluble, and most polymers
are quite soluble in their own monomers. The requirement that the costabilizer
must be of low molecular weight is based on reported swelling experiments and
theoretical swelling calculations [93]. Data from Schork and Reimers [94]
demonstrate that it is possible to create miniemulsion latexes with a poor co-
stabilizer (polymer). The inclusion of a small amount (~4%wt) of a monomer-
Miniemulsion Polymerization 153

soluble polymer can significantly reduce the diffusional degradation of an

emulsion. These emulsions are not thermodynamically stable, but they can be
kinetically stable. This means that the droplets resist diffusional degradation
long enough to allow nucleation to occur. The droplets are typically in the
miniemulsion range of 100 to 500 nm in diameter. The polymeric costabilizer
is thought to delay Ostwald ripening sufficiently to allow nucleation of the
monomer droplets by water-phase radicals (primary or oligomeric). Once
the droplets are nucleated, the polymer produced adds additional diffusional
stability. It should be noted that the monomeric miniemulsions formed are
not true miniemulsions in the sense that they are not stable over a period of
months. However, Ostwald ripening can be reduced to permit the polymeriza-
tion to be carried out. The latexes produced from polymer-stabilized emulsions
have all the characteristics of miniemulsion latexes, and derive from droplet
nucleation. The polymer has been shown to perform as well as hexadecane
in stabilizing the droplets for the short periods necessary to ensure nucleation.
It has the added advantages of being totally innocuous in the final product, very
soluble in the monomer, and very water insoluble.
Reimers and Schork [94, 95] report the use of PMMA to stabilize MMA
miniemulsions enough to effect predominant droplet nucleation. Emulsions
stabilized against diffusional degradation by incorporating a polymeric co-
stabilizer were produced and polymerized. The presence of large numbers of
small droplets shifted the nucleation mechanism from micellar or homo-
geneous nucleation, to droplet nucleation. Droplet diameters were in the mini-
emulsion range and reasonably narrowly distributed. On-line conductance
measurements were used to confirm predominant droplet nucleation. The
observed reaction rates were dependent on the amount of polymeric co-
stabilizer present. The latexes prepared with polymeric costabilizer had lower
polydispersities (1.006) than either latexes prepared from macroemulsions
(1.049) or from alkane-stabilized miniemulsions (1.037).
Wang and Schork [73] used PS, PMMA and PVAc as the costabilizers in
miniemulsion polymerizations of VAc with PVOH as the surfactant. They found
that, while PMMA and PS were effective kinetic costabilizers (at 2–4%wt on
total monomer) for this system, PVAc was not. While the polymeric costabi-
lizers did not give true miniemulsions, Ostwald ripening was retarded long
enough for predominant droplet nucleation to take place.
Aizpurua et al. [96] have studied the kinetics of vinyl acetate miniemulsions
stabilized with PS or PVAc. Guyot and coworkers [97] used PS as the costabi-
lizer for the miniemulsion encapsulation of pigment. Samer [67] has used
PMMA to stabilize MMA miniemulsions for continuous polymerization in a
CSTR.
Various papers on hybrid miniemulsion polymerization have used alkyd
[98, 99], polyester [100] or polyurethane [101] as both the costabilizer and a
component of the hybrid particle. Since most of these materials were added far
in excess of the levels normally used as costabilizers, it is not surprising that
they are effective.
154 F. J. Schork et al.

Polymeric materials are not costabilizers in the sense that costabilizers cause
superswelling. Rather, they slow the onset of Ostwald ripening and preserve the
number of monomer droplets, if not their size. However, this review will take
a functional, rather than thermodynamic definition of a costabilizer, and in-
clude a discussion of the use of polymers as agents to enhance droplet nuclea-
tion under the heading of costabilizers.

3.3.2
Monomeric Costabilizers

Reimers and Schork [102] first used highly water-insoluble comonomers as

costabilizers.Vinyl hexanoate, p-methyl styrene, vinyl 2-ethyl hexanoate, vinyl
decanoate, and vinyl stearate were copolymerized with MMA at 10%wt on the
total monomer. All formed stable miniemulsions with droplet diameters be-
tween 150 and 230 nm.All resulted in polymerization via predominant droplet
nucleation (miniemulsion). Chern and coworkers [43, 74, 75,103] have used
high molecular weight alkyl methacrylates at levels of 2–3%wt on the total
monomer as both costabilizers and comonomers in the miniemulsion poly-
merization of styrene. The advantage, of course, is that, after polymerization,
no low molecular weight costabilizer remains in the miniemulsion latex. Lauryl
methacrylate and stearyl methacrylate have been used, since these high me-
thacrylates have low water solubilities (10–8–10–9 g/g) and high solubilities in
styrene monomer. As might be expected, stearyl methacrylate is found to be
better at retarding Ostwald ripening than lauryl methacrylate, but neither is
found to be as effective as HD. Samer [104] found that 2-ethylhexyl acrylate
(2EHA) as a comonomer was not an effective costabilizer in a continuous
stirred tank (CSTR) copolymerization with MMA.

3.3.3
Other Costabilizers

The use of a chain transfer agent (CTA) as a costabilizer opens up new possibi-
lities for molecular weight control. Macroemulsion polymerizations which util-
ize higher molecular weight mercaptan chain transfer agents exhibit retarded
transport of the CTA from the monomer droplet into the growing polymer par-
ticles. This results in slower delivery of the mercaptan to the reaction sites over
the course of the polymerization. (In some commercial recipes this retarded
transport is used to “meter” the highly reactive CTA to the reaction site.) If the
mercaptan were at the site of polymerization, as in a miniemulsion, new degrees
of freedom in selecting chain transfer agents would exist. That is, the relative re-
activities of chain transfer versus propagation can be used to select the CTA, with-
out relying on retarded mass transfer. This may increase the efficiency of chain
transfer (since CTA will not be “trapped”in the shrinking monomer droplets near
the end of Interval II), or at least allow the chemist additional degrees of freedom
in tailoring the molecular structure by manipulating the reaction conditions.
Miniemulsion Polymerization 155

Mouran et al. [105] polymerized miniemulsions of methyl methacrylate with

sodium lauryl sulfate as the surfactant and dodecyl mercaptan (DDM) as the
costabilizer. The emulsions were of a droplet size range common to miniemul-
sions and exhibited long-term stability (of greater than three months). Results
indicate that DDM retards Ostwald ripening and allows the production of
stable miniemulsions.When these emulsions were initiated, particle formation
occurred predominantly via monomer droplet nucleation. The rate of polymer-
ization, monomer droplet size, polymer particle size, molecular weight of the
polymer, and the effect of initiator concentration on the number of particles all
varied systematically in ways that indicated predominant droplet nucleation.
For the MMA/DDM system, the value of the chain transfer constant (Cx) is
0.6–0.8, meaning that the chain transfer agent reacts slightly less rapidly than
the monomer. Hence, the DDM will be present throughout the course of the
reaction. In a system such as styrene/DDM, where Cx is 15–20, the rapid con-
sumption of the DDM might leave the particles subject to Ostwald ripening
before enough polymer is formed to stabilize the growing particles. In addition,
the rapid consumption of CTA early in the course of the polymerization might
give a clearly identifiable low molecular weight tail. Wang et al. [106] studied
the miniemulsion polymerization of STY with DDM as the costabilizer. In this
system, the chain transfer constant is at the other end of the kinetic spectrum.
The miniemulsion monomer droplets with dodecyl mercaptan as costabi-
lizer were very stable. Shelf lives ranged from 17 hours to three months. The
kinetics of miniemulsion polymerization were studied. Unlike other miniemul-
sion systems where the costabilizer does not act as a chain transfer agent, the
polymerization rate fell with costabilizer level because the chain transfer agent
enhances radical desorption from the particles. The polymerization rates in all
of the miniemulsions were lower than those of the corresponding macroemul-
sions. Polymerized particles were larger than in the corresponding macroemul-
sions, but molecular weights were lower. Results indicate that DDM can serve
as an effective costabilizer as well as a chain transfer agent, even when the chain
transfer constant is quite high. The fact that the molecular weights were lower
in the miniemulsion reactions indicates predominant droplet nucleation.
Reimers and Schork [107] used lauroyl peroxide (LPO) in the miniemulsion
polymerization of MMA as a costabilizer as well as an initiator. They showed that
lauroyl peroxide concentrations above 1 g/100 g of monomer are capable of sta-
bilizing droplets against Ostwald ripening. The stable droplets produced were in
the miniemulsion-size range and could then be nucleated. The ratio of the num-
ber of droplets to the number of particles was found to be close to unity. The over-
all rates of polymerization were high for the miniemulsions, as were the rates per
particle. Once again, it was shown that components other than conventional co-
stabilizers can stabilize small droplets against Ostwald ripening, causing droplet
nucleation. Asua et al. [108] used LPO (in addition to the traditional costabilizer
HD) to impart diffusional stability to styrene miniemulsions. They also evaluated
a number of oil-soluble initiators (LPO, BPO, AIBN) as costabilizers, and con-
cluded that only LPO was capable of acting as the sole costabilizer (without HD).
156 F. J. Schork et al.

3.3.4
Enhanced Nucleation

Miller et al. [109–111] report that the addition of a small amount (as small
as 0.05%wt) of polystyrene (PS) to the styrene phase of a miniemulsion poly-
merization of styrene causes an increase in both the rate of polymerization and
the number of final polymer particles. This is not just a polymeric costabilizer
effect, since these emulsions were also stabilized with what are known to be
effective levels of HD or CA, although the effect was more pronounced with CA.
With the addition of 1%wt styrene, the number of final particles was nearly the
same as the original number of droplets, indicating 100% droplet nucleation.
This was not the case for equivalent polymerizations without the PS. For poly-
merizations without the PS, the final particle number varied with the initial
initiator concentration to the power of 0.31.With 1%wt PS, the particle number
was independent of the initiator concentration, which is very clear evidence of
100% droplet nucleation. Miller hypothesized that miniemulsions prepared
from polystyrene in styrene solutions resemble the polymer particles formed
in normal (no polymer) miniemulsion polymerizations at early conversions.
This being the case, these polymer-containing droplets would be able to effec-
tively compete with growing polymer particles for free radicals, whereas their
counterparts that contain no polymer are not, and as a result a greater fraction
of the initial droplets become polymer particles. Based on this mechanism,
Miller speculated that the presence of the polymer increases the capture effi-
ciency of the droplets by modifying either their interior (by increasing the
interior viscosity, thereby increasing the probability of a radical propagating
rather than exiting) or the droplet/water interface (by disrupting the surfactant/
CA interfacial barrier to radical entry). Experimental results were reported
which support the latter explanation. (It should be noted that the effect was
most pronounced with CA, and that other investigators have reported near
100% droplet nucleation with HD and without added polymer.)
In a follow-on set of papers, Blythe et al. [112–115] studied the effect on the
polymerization kinetics of changing the properties of the polymer used to en-
hance nucleation. Miniemulsions were formed from CA with PS as the polymer
additive, and CA as the costabilizer. Varying the molecular weight of the PS
from 39,000 and 206,000 in systems containing 1% PS did not change the ki-
netics. Also, changing the end group of the polymer chain from a hydrophobic
end group to a hydrophilic end group had no effect on the kinetics in 1%
polymer systems. However, predissolving 1% PS in a miniemulsion always
results in a significant enhancement in the kinetics compared to similar sys-
tems that do not contain predissolved polymer. The authors conclude that this
enhancement in the kinetics is not due to either a change in the interior vis-
cosity of the droplets or a disruption of the condensed phase formed by cetyl
alcohol and sodium lauryl sulfate (since these effects would be altered by the
variations in the PS above). Instead, it was suggested that the enhancement can
primarily be attributed to a preservation of the droplet number due to the pre-
Miniemulsion Polymerization 157

sence of polymer in each of the miniemulsion droplets formed during homo-

genization. The authors rightly point out that the polymer is a poor costabilizer,
due to its high molecular weight, but that it does act to preserve the droplet due
to the thermodynamic balance between monomer and polymer. Therefore,
they conclude that the polymer is unable to preserve the size of the droplets
produced during ripening, only the number produced during homogenization.
They support this by using 1% PS with no other costabilizer, where they are able
to show that particle formation occurs via droplet nucleation. In other experi-
ments, they show that there is no enhancement unless the shear rate is high
enough to bring the droplet size down into the range where it is susceptible to
Ostwald ripening.
In another paper, Blythe et al. [116] studied enhanced droplet nucleation
when HD is used as the costabilizer. The enhancement in this case is much less.
The authors conclude that, since HD is a very effective costabilizer (much more
so than CA), the effect of the polymer in preserving the droplet number, if not
the droplet size distribution, is not pronounced. Therefore, this effect appears
to occur primarily in systems with CA (perhaps due to its polar nature, and so,
its probable interfacial activity). Multiple investigators have reported effective
(near 100%) droplet nucleation with HD and other costabilizers.
Blythe et al. argue (with justification) that polymer should not be termed a
costabilizer since it does not cause super-swelling; however, this review will take
a functional, rather than thermodynamic definition of a costabilizer, and treat
polymer-stabilized miniemulsions under the heading of costabilizers.

3.4
Choice of Initiator

Following the common practice in macroemulsion polymerization, most

miniemulsion polymerizations have been run using water-soluble initiators.
However, a number of researchers have looked at the possibility of using an oil-
soluble initiator instead.As discussed previously, Schork and Reimers [107] and
Asua et al. [108] have used LP as both the initiator and the costabilizer. In
addition, Asua et al. used other oil-soluble initiators in conjunction with HD
(as the costabilizer) to carry out miniemulsion polymerization of styrene.
Ghazaly et al. [117] used both water-soluble and oil-soluble initiators in the
copolymerization of n-butyl methacrylate (BA) with crosslinking monomers.
Variations in the particle morphologies were found between the water-soluble
and oil-soluble initiators, depending on the hydrophobicity of the crosslinking
monomer. It would seem that if the crosslinking monomer is quite hydro-
phobic, and therefore resides preferentially in the core of the particles (droplet),
then the oil-soluble initiator is more effective at carrying out crosslinking, since
the oil-soluble initiator will also reside preferentially in the core of the particle.
Luo and Schork [118] carried out emulsion and miniemulsion polymerization
using oil-soluble initiator in the presence of an aqueous phase free radical
scavenger. They concluded that, for miniemulsion particles up to 100 nm in
158 F. J. Schork et al.

diameter, even with an oil-soluble initiator, radicals originating in the aqueous

phase play an important role in initiating polymerization. This is attributed to
the fact that two radicals generated from the decomposition of an initiator
molecular within the particle may recombine before initiating polymerization.
Choi et al. [53] have successfully used both water-soluble and oil-soluble
initiators in the miniemulsion polymerization of styrene. Alducin and Asua
[119] have studied the MWD of polystyrene miniemulsion polymerized with
oil-soluble initiators. Rodriguez et al. [61] have developed a mathematical
model of seeded miniemulsion polymerization with oil-soluble initiator. Blythe
et al. [120] have successfully carried out miniemulsion polymerization of styrene
with AMBN (oil-soluble). Ghazaly et al. [117] have used AIBN for the mini-
emulsion copolymerization of a hydrophobic bifunctional macromer. The poly-
merization progressed much faster when KPS was used than when AIBN was
used. This may be due to the tendency of oil-soluble initiator radicals to recom-
bine before initiating polymerization, as discussed by Luo.
Oil-soluble initiators have commonly been used in hybrid miniemulsion
polymerization to improve monomer conversions. In most cases, the oil-soluble
initiator was used as a finishing initiator to increase final monomer conversion,
while a water-soluble initiator was used to carry out the majority of the poly-
merizations. These types of polymerizations will be discussed later.

3.5
Robust Nucleation

One of the problems with macroemulsion polymerization is the variability of

the particle number with initiation rate, monomer quality, inhibition levels,
and so on. This is a serious industrial problem, as shown by the fact that a great
many industrial macroemulsion polymerizations are carried out as seeded
polymerizations in which a known concentration of seed particles are added
to the emulsion, and the polymerization is run under conditions that suppress
nucleation of additional particles. The variance in particle number comes
about because there is a competition for surfactant between the growth of
existing particles (that need additional surfactant to stabilize their growing sur-
face area), and the nucleation of new particles.
If a miniemulsion could be run at 100% droplet nucleation (or near to this),
then a very robust nucleation system would result. The number of particles
could be determined by the number of initial monomer droplets, and this can
be controlled by adjusting surfactant, costabilizer and shear levels. In this case,
the number of particles would be independent of radical flux. In fact, the most
compelling evidence for droplet nucleation is experimental evidence that the
number of polymer particles is independent of the initiator level. (Once the
radical flux is high enough to nucleate all, or nearly all of the droplets, then
changes in radical flux caused by inconsistent initiator or unknown inhibitors
will not affect the final particle number.) We will discuss the results of such
robust nucleation later.
Miniemulsion Polymerization 159

3.6
Monomer Transport Effects

Macroemulsion polymerization relies on the transport of monomer from the

monomer droplets to the polymer particles. This transport is driven by the
equilibrium swelling of the polymer particles. This presumes rapid (relative to
the rate of polymerization) transport of monomer. For most monomers, this
is a good assumption. However, for monomers that are very water insoluble
(VEOVA [vinyl versatatex] or DOM [dioctyl maleate]), this may not be true.
In making this determination, the following assumptions can be made:
(i) The limiting resistance is the transport from the monomer droplets into
the aqueous phase.
(ii) Transport across the aqueous phase is by forced convection (stirring) and
it is not the rate-determining step.
(iii) Transport from the aqueous phase into the polymer particle may (or may
not) have an overall mass transfer coefficient equal to that for exit from the
monomer droplets, but the very large interfacial area of the particles (re-
lative to the monomer droplets) will ensure that this is not the limiting step.
(iv) Transport out of the monomer droplet can be modeled with an overall
mass transfer coefficient and a driving force based on the difference be-
tween the saturation concentration of monomer in the aqueous phase and
the concentration in the aqueous phase in equilibrium with the particle.
For monomers that are highly water insoluble, the driving force in (iv) will be
small, since the saturation value will be extremely small. Since accurate overall
mass transfer coefficients are hard to determine accurately, the likelihood of
transport limitation with highly water-insoluble monomers is an open ques-
tion. Some data (which will be discussed later) indicate the presence of trans-
port limitations in the copolymerization of highly water-insoluble monomers.
These potential transport limitations can be avoided by using miniemulsion
polymerization.
In the case of nanoencapsulations of solids, or the incorporation of high
molecular weight, highly water-insoluble additives (such as polymers, oligom-
ers, alkyds) into polymer particles, macroemulsion polymerization will not
work, since the high molecular weight material will remain in the monomer
droplet as the monomer is transported out. At the end of the reaction, the
additive will remain in the depleted monomer droplets, rather than in the
polymer particles. Clearly, these products can only be made via miniemulsion
polymerization.

3.7
Droplet Stability

Different techniques are available to carry out a free radical polymerization in

emulsion. In spite of the fact that their names (macro-, mini- and microemul-
160 F. J. Schork et al.

sion) seem to indicate a correlation between process and final particle size of
the polymer particle dispersion, this is usually not the case. On the contrary,
these techniques should rather be distinguished by either their nucleation
process or the stability of the initial oil (the monomer) dispersion. These two
aspects of the emulsion polymerization processes, nucleation and stability, are
strictly related and depend upon each other.
To clarify this point, let us first discuss the case of a conventional emulsion
recipe, or macroemulsion. In this case, the monomer is initially dispersed in
water under agitation in the presence of surfactants [121], resulting in a rather
coarse oil dispersion. Droplet size typically depends upon rate of stirring, and
the resulting droplet size is generally in the 1–10 mm range (the reason for the
name “macroemulsion”). This dispersion is unstable and, if the stirring is
stopped, the monomer phase separates very quickly. Because of the large sizes
of the droplets, the total surface area of the dispersion is small (the surface-to-
volume ratio for spherical droplets is proportional to the inverse of the droplet
diameter) and most of the surfactant is not used to stabilize the droplets; it
aggregates in water to form micelles. These micelles, and not the droplets, are
the main loci for polymerization, while the droplets just act as monomer reser-
voirs. In other words, the final particles in the macroemulsion do not correspond
to the initial droplets. The most important consequence is that, in order to have
a successful process, the monomer, as well as all other possible comonomers
and coreactants, must be water-soluble enough to diffuse from droplets to
particles.
For historical reasons [122], a stable dispersion of sub-micron droplets is
called a miniemulsion. Miniemulsions do not form spontaneously and require
high shear devices to form. The resulting dispersion is usually quite fine, with
the droplet size ranging from 50 to 500 nm [3], a large droplet surface area, with
all of the surfactant used to stabilize the droplets and with no more micelles
present in the system.As a result, the droplets become the predominant loci for
nucleation and polymerization. That is, in an ideal miniemulsion, there is a 1:1
correspondence between initial monomer droplets and final polymer particles.
(This 1:1 correspondence is not always attained, and remains a point of contro-
versy [1].) This was experimentally demonstrated for the first time by Ugelstad
and coworkers in 1973 [5]. The consequence of this virtual copying process
from droplets to particles is twofold: (i) final particle size is given by the initial
droplet dispersion, and both surfactant coverage and surface tension do not sig-
nificantly change during the process; (ii) any kind of hydrophobic component
can be conveniently included in the recipe, since we can be sure that it is going
to participate to the polymerization process.
A third type of emulsion process is the so-called microemulsion [123]. In
microemulsions, the polymerization starts in droplets as well. However, these
are thermodynamically stable and, in contrast to miniemulsions, they form
spontaneously by gentle stirring. They consist of large amounts of surfactants
or mixtures of them, and they possess an interfacial tension close to zero at the
water/oil interface, with droplet sizes usually ranging between 5 and 50 nm. In
Miniemulsion Polymerization 161

contrast to miniemulsions, the high amount of surfactant needed to prepare the

microemulsions leads to a complete coverage of the droplets surface. Given the
huge number of droplets initially present in the system, nucleation cannot take
place in all of the droplets, and a large number of empty micelles can still be
found in the final product.
Besides the similarities between mini- and microemulsions, especially with
respect to the nucleation process, microemulsions are characterized by very
peculiar thermodynamics; in this respect, there are many similarities between
a macro- and a miniemulsion. Accordingly, we should focus on the reason why
monomer dispersions are unstable in macroemulsions, while they are stable
in miniemulsions.

3.7.1
Stability of Monomer Dispersions

There are two ways by which a dispersion of monomer droplets can degrade:
(i) by droplet coalescence, and (ii) by diffusion degradation (often referred to
as Ostwald ripening).While the first mechanism of degradation can be avoided
by adding enough surfactant to the system, when two monomer droplets of dif-
ferent sizes, and stabilized by a surfactant, are put in water, they will start ex-
changing monomer without even making direct contact – through monomer
diffusion across the water (continuous) phase.
This process of molecular diffusion is governed by the difference in chem-
ical potentials of the monomer in the two droplets. Morton’s equation has been
successfully used to describe the swelling of polymer particles with monomer
[32]. According to this equation, the chemical potential of the monomer in a
droplet of radius rp(µ (d)
m ) in the presence of polymer is given by:

–


µ(d)
m 1 2
2gVm
= ln (jm) + 1 – 7 p j + c j
m,p p + (3)
RT mpm rpRT
6 9

where jm and jp represent the volume fraction of monomer and polymer, re-
spectively, in the particle (so jm+jp=1); mpm is the ratio of equivalent number
of molecular segments between monomer and polymer; cm,p is the interaction
parameter between monomer and polymer; g is the interfacial tension at the
–
water/oil interface; Vm is the molar volume of the monomer; R is the universal
gas constant; T the temperature. In the case of pure monomer droplets, the
partial molar free energy of mixing is zero, which is represented in Eq. 3 by the
first two terms, accounting for the entropy of mixing, and by the third term,
accounting for the enthalpy of mixing. Ugelstad and Hansen showed that this
expression could be satisfactorily extended to the case where any other species,
not necessarily a polymer, is involved [11].Accordingly, we will use Eq. 3 to con-
veniently describe the monomer chemical potential in a miniemulsion droplet.
According to Eq. 3, in the case of a macroemulsion of pure monomer
droplets, the monomer chemical potential is given by the last term with (jm=1),
162 F. J. Schork et al.

Fig. 6 Monomer chemical potential for a pure monomer droplet as a function of droplet
–
radius (g=25 Mn/M; Vm=1.1·10–4 m3 mol; T=298.15 K). Point A represents an unstable equi-
librium

accounting for the contribution of the interfacial energy. Therefore, given two
droplets with the same interfacial tension (we assume that surface tension is a
function of the degree of coverage only, and that this is the same for the two
droplets), their chemical potential is a function of the droplet radius only and
it is only the same if the two droplets have the same size. Therefore, when two
droplets of different radii rd,1 and rd,2 are put together, the difference in their
chemical potential is given by:



Dµ(d)
m,1–2 2gṼm 1 1 1 1
02 = 9 5 – 5 = y 5 – 5 (4)
RT RT rd,1 rd,2 rd,1 rd,2

In other words, if rd,2>rd,1, the difference in chemical potential is positive and

the monomer will diffuse from 1 to 2. This process is schematically represented
in Fig. 6. Point A represents the initial size of the monomer droplets. Equili-
brium exists only if all of the droplets have the same size. In the case where a
smaller droplet is created (symbolized by point B), monomer will flow from B
to A (positive flux to A, or J >0) as a result of the difference in chemical poten-
tial, making B smaller and smaller. The opposite happens for a bigger droplet
(point C), which will become bigger and bigger. In other terms, point A is an
unstable equilibrium. This process is usually referred to as monomer ripening.
The ripening process characteristic of oil dispersions is avoided in
miniemulsions by introducing a so-called costabilizer in the oil phase (a species
with no or very low water phase solubility). The mechanism that halts the
Miniemulsion Polymerization 163

Fig. 7 Schematic representation of the role of the hydrophobe in halting the ripening
process in a miniemulsion

ripening in a miniemulsion is sketched in Fig. 7. Let us suppose that the oil

phase is initially comprised of the monomer and a perfectly water insoluble
costabilizer. Let us also suppose that the system is initially made of two droplets
with different sizes but the same compositions (or monomer volume fractions).
Finally, let us suppose, for the sake of simplicity, that the mixture of monomer
and costabilizer behaves like an ideal mixture, so that cm,h=0. Under such hy-
potheses, the difference in chemical potential between the two droplets of
Fig. 7a is given by the following equation:


Dµ(d)
m,1–2 jm,1 1 1
= ln 7 + (1– mm,h)(jm,2 – jm,1) + y 5 – 5 (5)
RT
02 jm,2 rd,1 rd,2

where the indexes 1 and 2 refer to the first and the second droplet, respectively,
subscript h identifies the costabilizer, and y is defined as in Eq. 4. Knowing that
lnjm≈–(1–jm) for small volume fractions of the costabilizer, it is possible to
simplify the previous equation as follows:


Dµ(d)
m,1–2 1 1
= mm,h (jm,1 – jm,2) + y 5 – 5 (6)
RT
02 rd,1 rd,2

Since the two droplets have the same initial composition (jm,1=jm,2), it follows
from this equation that the difference in chemical potential is given by the dif-
ference in size only, and (as in a macroemulsion) monomer will flow from the
smaller to the bigger droplet. The main consequence of this process is that
the costabilizer is concentrated in the smaller droplet, while the big droplet
becomes more and more dilute. This creates a gradient in composition (see
Fig. 7b), which is represented by the first term on the right hand side of Eq. 6.
164 F. J. Schork et al.

Therefore, an opposite flow of monomer is generated, which eventually leads

to an equilibrium between the two terms of Eq. 6 (to a stable situation).
Just for completeness, these two contributions to the monomer equilibrium
in a miniemulsion are often expressed in terms of pressures in the literature.
The droplet composition gives rise to a so-called osmotic pressure:

Posm = RTcm (7)

while the presence of an interface generates a so-called Laplace pressure inside

the droplet, defined as follows:
3g
PLapl = 5 (8)
rd
Note that, when accounting for the differences in osmotic and Laplace pres-
sures between the two droplets, one obtains the same equation as Eq. 6, with
mm,h=8/3.
This same process is described in Fig. 8 in terms of the different contribu-
tions to the monomer chemical potential. In this figure, the initial droplet size
(rd) and monomer volume fraction (jm) have been fixed, and the correspond-
ing volume of costabilizer inside the droplet has been computed (Vh = 4/3p rd3
(1–jm). Assuming that the costabilizer cannot diffuse out of the particle, in
Fig. 8 it is possible to observe how chemical potential of the monomer changes
by increasing or decreasing the monomer volume fraction in the droplet. In
particular, two contributions to the global potential (entropy of mixing and sur-
face tension) have been reported. It is clear that if a second droplet is inserted
into the system, these two terms will act in opposite ways until the difference
in chemical potential between the droplets is compensated for.
Even though two droplets are always able to find an equilibrium when put to-
gether, because of the presence of the costabilizer, it is useful to check whether
this equilibrium is stable or not. Going back to the case of the macroemulsion
depicted in Fig. 6, point A is an unstable point because, if the system is perturbed
and a new droplet is formed, it will diverge from the equilibrium point. Clearly,
the necessarily condition to have a stable equilibrium is that the slope of the
chemical potential versus radius is positive at the point of equilibrium [122, 124].
For a macroemulsion, this condition leads to the following expression:


∂ µ(9)
m y
6 6 = – 32 (9)
∂rd RT rd

which is always a negative function. This is not true for a miniemulsion.

Referring to Fig. 8, we observe that the equilibrium point of the system is stable.
In fact, if a larger droplet is generated by perturbing the equilibrium, this has
a larger chemical potential and the monomer will flow back, bringing the
droplet back to the original equilibrium point. The opposite happens when pro-
ducing a smaller droplet.
Miniemulsion Polymerization 165

Fig. 8 Monomer chemical potential in a droplet comprising monomer and hydrophobe as

a function of the monomer volume fraction (top) and droplet radius (bottom). The global
potential (as given by Ugelstad’s equation) is given, as well as the entropic term due to mixing
and the Laplace term due to surface tension. Parameters: mm,h=1; cm,h=0; g=25 mN/m;
Ṽm=1.1 · 104 m3 mol; T=298.15 K; jm=0.96; rd=100 nm
166 F. J. Schork et al.

Let us check the conditions under which the derivative of the chemical
potential is positive in a miniemulsion. Under the hypothesis of small concen-
trations of costabilizer in the droplet (jmÆ1), and of ideal behavior of the
mixture (cm,h=0), the condition of local stability leads to the following expres-
sion:


∂ µ(9)
m 9mm,hVh y
6 6 = 04 –4 (10)
∂rd RT 4p r 4d rd2

and in turn to:


9mm,hVh
rd < r–d = 04 (11)
4py

The previous expression returns the value of the droplet radius corresponding
to the maximum of the monomer chemical potential curve (r–d), and says that,
for a given amount of costabilizer in the particle (Vh), the droplet radius must
be smaller than r–d in order to be locally stable. The same equation can be
expressed in terms of the critical costabilizer volume fraction, j–h, as follows:
3/2


y
jh > j–h = 9 (12)
3mrh

where rh=(3Vh/4p)1/3. Both these two critical quantities, r–d and j–h, are functions
of costabilizer concentration, costabilizer type and surface tension, as well as
the corresponding value of the monomer chemical potential, which is given by
the following simple expression:




µ(d)
m y 3/2
1 1/2
=2 4 (13)
RT rd =r–d 3 mm,h rh
6 01

Therefore, it is important to show how these quantities depend upon the

parameters involved, and what their effects are, in order to understand the
stability of the droplet.
This analysis is shown in Fig. 9, where the parameters y, mm,h, rd and jm
have been varied systematically. Let us start with the analysis of Fig. 9a. In this
figure, the volume of the droplet has been varied while the volume fraction of
the costabilizer is kept constant (the vertical dashed line represents the cor-
responding monomer volume fraction). This, in turn, changes the volume of
the costabilizer inside the droplet, and the monomer chemical potential curve
changes accordingly.As predicted by Eq. 12, when the droplet size is decreased,
the maximum of the chemical potential curve shifts to smaller monomer
volume fractions, and the height of the maximum increases. In other terms, as
droplet size decreases (or costabilizer volume fraction decreases) the system
approaches the unstable region. Figure 9b is conceptually identical to Fig. 9a,
Miniemulsion Polymerization 167

Fig. 9 Monomer chemical potential in a droplet comprising monomer and hydrophobe

as a function of the monomer volume fraction. (A) Effect of droplet size (rd=25, 37.5, 50,
75 and 150 nm); (B) effect of hydrophobe volume fraction (jh=0.005, 0.01, 0.02, 0.04 and 0.1)

but the amount of costabilizer in the droplets is varied while the droplet radius
is kept constant and the costabilizer volume fraction varies. Again, as the co-
stabilizer concentration decreases, the system moves from a region of complete
stability to a region of instability. By observing Fig. 9b, it is also possible to
distinguish between local and global equilibrium stability. The curve corre-
sponding to jh=0.04 clearly satisfies the requirements of local equilibrium sta-
bility. In fact, if the system is slightly perturbed, the equilibrium point is always
168 F. J. Schork et al.

Fig. 9c,d (C) effect of surface tension (g=5, 15, 25 and 35 mN/m); and (D) effect of hy-
drophobe to monomer segment ratio (1/mm,h=1, 1.29, 2, 5 and 10). Other parameters (oth-
erwise differently indicated): mm,h=1; cm,h=0; g=25 mN/m; Ṽm=1.1 · 104 m3 mol; T=298.15 K;
jh=0.04; rd=100 nm
Miniemulsion Polymerization 169

convergent. However, in the case of large perturbations, it is possible to create

droplets large enough to fall to the right of the maximum of the monomer
chemical potential curve; these droplets have chemical potentials lower than
that corresponding to equilibrium. Such a droplet will receive monomer from
the droplets at the equilibrium point and will grow indefinitely. On the other
hand, the same cannot happen for the equilibrium point corresponding to the
curve for jh=0.10. In fact, in this case, all of the droplets that are larger than
those at the equilibrium point also have larger monomer chemical potentials.
Therefore, these droplets will always shrink back to the equilibrium point.
In Fig. 9c, the effects of different surface tension values on the equilibrium
are examined. By decreasing the interfacial tension, the Laplace term becomes
less significant than the contribution given by the entropy of mixing, and there-
fore ripening is decreased and stability is enhanced. Theoretically, in a system
with zero surface tension at the oil/water interface, the total monomer chemi-
cal potential is given solely by the entropic terms, and it is always stable.
Finally, in Fig. 9d, the influence of the ratio of the equivalent number of
molecular segments in the costabilizer to that of the monomer, mm,h is shown.
This value can be thought of as the ratio between the molecular weights of the
two components, and for oligomers and polymers this can be replaced by the
average chain length. If we look closely at this figure, as well as Eqs. 12 and 13,
we can see that the maximum of the chemical potential curve becomes smaller
and shifts to larger droplet sizes as mm,h approaches unity, which facilitates the
formation of a stable equilibrium. When dealing with very bulky costabilizers
(such as a polymer), the maximum increases and shifts to low droplet sizes,
making the equilibrium unstable.
There are two issues we should remark upon at this point. First, even though
we have shown that the presence of a costabilizer in the system can lead to
complete thermodynamic stability, we have also shown that costabilizers with
smaller molecular weights are more effective. Therefore, even though these
species are hydrophobic, they always have a small but finite solubility in water.
As a result, the differences in costabilizer chemical potentials among the
droplets will lead to a very slow diffusion of the costabilizer and eventually to
the destabilization of the system. Second, Eqs. 11 and 12 do not contradict those
reported by (for example) Sood and Awasthi, that point to a minimum droplet
diameter below which there is no stability [124]. In our analysis, we set the
equilibrium and the corresponding volume of costabilizer, and then we per-
turbed the system by changing the monomer concentration, and observed what
happened. In their work, they computed the corresponding equilibrium point
as a function of droplet radius for a given costabilizer concentration, and they
checked the conditions at which this point corresponds to a stable equilibrium.
As in our analysis, they conclude that decreasing droplet radius sufficiently
eventually pushes the system into the region of instability.
170 F. J. Schork et al.

3.7.2
Experimental Validation

Our understanding of miniemulsion stability is limited by the practical dif-

ficulties encountered when attempting to measure and characterize a distribu-
tion of droplets. In fact, most of the well-known, established techniques used
in the literature to characterize distributions of polymer particles in water are
quite invasive and generally rely upon sample dilution (as in dynamic and static
laser light scattering), and/or shear (as in capillary hydrodynamic fractiona-
tion), both of which are very likely to alter or destroy the sensitive equilibrium
upon which a miniemulsion is based. Good results have been obtained by
indirect techniques that do not need dilution, such as soap titration [125],
SANS measurements[126] or turbidity and surface tension measurements
[127]. Nevertheless, a substantial amount of experimental evidence has been
collected, that has enabled us to establish the effects of different amounts of
surfactant and costabilizer, or different costabilizer structures, on stability.
This work has been summarized very effectively by Landfester and co-
workers [127]. They showed that many costabilizers act as osmotic agents,
blocking monomer ripening. All of these costabilizers have relatively low
molecular weights and, according to the analysis above, the polymers can
barely be used to prevent ripening, even though the resulting miniemulsion can
be stable for enough time to run the polymerization.
However, the most significant result of their work was that they clearly
demonstrated the role of surfactant in the formation of miniemulsions. They
showed that it is possible to effectively control droplet size by tuning the
surfactant concentration. The more surfactant, the smaller the droplets one can
obtain. They also showed that, apart from the case where extremely large
amounts of surfactant are used, the surfactant coverage of the droplets is always
incomplete. In particular, larger droplets exhibit very low coverage and, there-
fore, large interfacial tensions at the oil/water interface. On the other hand,
small droplets need large surfactant coverage in order to be stable.
In the same work, it is also supposed that colloidal stability, rather than
monomer ripening, plays an effective role in determining the final droplet size.
Such a conclusion was supported by two different experimental results. First,
it was noticed that droplet size increases right after the emulsification process
stops, and a stable situation is typically achieved after just a few hours. How-
ever, if surfactant is added immediately after, this growth in size does not occur.
Second, it is shown that there is a clear correlation between final droplet size
and amount of oil phase used in the recipe. In particular, when the oil fraction
in the system increases, droplet size also increases.
These results can be effectively explained by supposing that colloidal
stability plays a major role in determining miniemulsion stability. In fact, it is
clear that addition of surfactant stops the droplet growth, which is explained
by the enhanced colloidal stability. Moreover, in more concentrated systems,
where the rate of droplet coalescence is larger, one obtains larger droplets, as
Miniemulsion Polymerization 171

expected. However, some droplet flocculation or the formation of a limited

layer of monomer at the water/air interface is commonly observed in many
miniemulsion recipes. Moreover, there are well documented cases in the litera-
ture where either a bimodal distribution has been obtained at the end of the
polymerization [124], or instability problems are evident at the beginning of
the polymerization [128].
When speculating about the colloidal stability of a monomer droplet disper-
sion in water, one could use the Deryaguin-Landau-Verwey-Overbeek theory,
also known as DLVO theory, to analyze the stability of the system. This has been
done in Fig. 10a, where we show the effect of different surface potentials upon
the rate of coagulation, b, defined in the case of two droplets of the same size as:

8 kbT
b=8 (14)
3µW

where m represents the water viscosity and W is Fuch’s stability ratio (see [129]
for a further explanation of how this value has been computed). In Fig. 10b, the
quantity 1/bNd is reported, which expresses the characteristic time for coales-
cence. We can see that even small surface potentials are enough to give great
colloidal stability. These values correspond to rather low surfactant coverage
(around 5%) [124, 129], which is well below the typical coverage measured
experimentally for miniemulsions [127]. It is therefore apparent that colloidal
stability is a very strong function of droplet size, and it is quite difficult to
explain the formation of bimodalities or monomer separation by colloidal
stability only.
On the other hand, previous analyses of the monomer chemical potential in
miniemulsions may justify some of the previous results. In Fig. 8, it is shown
that miniemulsion formulations are often very close to instability, with positive
monomer chemical potentials (which corresponds to conditions of monomer
oversaturation, or superswelling). Experimental proof of this was reported by
Landfester et al. [127], who said that the Laplace pressure is much larger than
the osmotic pressure at equilibrium. We must also consider that, in reality, we
are never dealing with perfectly monodispersed distributions of droplets, as the
previous analysis supposed. Therefore, it is realistic to suppose that in the pre-
sence of a droplet size distribution, a fraction of the droplets lie in the unstable
region and can lead to the formation of either pronounced bimodalities,
monomer phase separation, or droplet flocculation [124]. Moreover, from the
analysis of Fig. 9a to d, we observe that critical stability is almost reached when
small droplets are formed. Therefore, it is not surprising to observe that smaller
droplets require larger surfactant coverage to be stable, since lower surface
tension helps the droplets to move away from the instability region. The same
influence of surface tension on the chemical potential of the monomer could
also explain why, on the addition of surfactant, the sizes of the droplets do not
change significantly after emulsification. In fact, by adding surfactant, one
decreases the interfacial tension and stabilizes the miniemulsion ripening.
172 F. J. Schork et al.

Fig. 10 Coagulation rate constant (b) and characteristic time for coagulation (1/bNd) as a
function of droplet size for various droplet surface potential values (z=15, 25 and 40 Mv),
as computed by DLVO theory. In order to compute Nd, a system with 20% oil fraction was
supposed
Miniemulsion Polymerization 173

It is certain that we do not know what the leading effect in determining

droplet stability and droplet distribution in miniemulsion is at this point; both
colloidal and ripening effects probably play a role. Future work is therefore
needed to clarify these problems.

3.8
Semibatch and Plug Flow Reactors

Semibatch (also known as semicontinuous) reactors are used commercially for

two primary reasons: to limit the rate of polymerization by effecting some level
of monomer starvation, or to correct for copolymer composition drift in
copolymerization. Since the use of semibatch reactors for miniemulsion co-
polymerization involves important concepts (relative rates of mass transfer and
reactivity ratios), a discussion of semibatch copolymerization reactions will
be deferred until the section on copolymerization. In the area of semibatch
homopolymerization, Tang et al. [116] studied seeded semibatch polymeriza-
tion of BA. They found that when the monomer was added neat, a small
number of new particles were formed. However, when the semibatch feed was
a miniemulsion, a large number of new particles were formed, presumably by
droplet nucleation. Monomer droplet nucleation decreased with increasing
seed concentration, presumably because of monomer transport to the existing
particles. Leiza, Sudol and El-Aasser [130] used semibatch miniemulsion poly-
merization (starting from a seed latex) to prepare high solids (>60%) lattices
of BA. It is well-known that semibatch polymerization can be an effective
method for making high solids latex. Part of the advantage of semibatch when
making high solids is the broad PSD brought on by nucleation of particles over
most of the reaction time. Miniemulsion can be effective in this regard, since
a miniemulsion feed is likely to produce additional polymer particles, as shown
by Tang et al. [131].
Sajjadi and Jahanzad [132] have used semibatching to study the effects
of monomer-starved and monomer-flooded conditions on the seeded poly-
merization of styrene. Seed particles were grown via macroemulsion poly-
merization and added to the initial charge of the reactor. Feeds of styrene
miniemulsion (using HD as the costabilizer) were then added, either batch-wise
or in a semibatch mode. Under starved conditions, the miniemulsion droplets
were depleted of their monomer by transport into the growing particles. (Even
though HD prevents the loss of monomer to droplets of a larger size, it cannot
prevent monomer depletion to polymer particles with a low degree of
monomer saturation, as would be found in a starved reactor; this same effect
will be seen later in experiments where fresh miniemulsion droplets are intro-
duced into a CSTR at high monomer conversion.) When the miniemulsion was
added batch-wise (flooded conditions) the final particle number was greater,
since a greater portion of the miniemulsion droplets were nucleated before
being depleted of monomer. When polymer was predissolved in the monomer
prior to miniemulsion formation, the final number of particles was indepen-
174 F. J. Schork et al.

dent of the method of addition, and was equal to the seed particles plus the
number of miniemulsion droplets added. Therefore, with predissolved polymer,
the droplet number (but not the droplet size distribution as inferred from the
final PSD) was preserved. This result reinforces the idea of polymer as an agent
for preserving droplet number.
Macro- and miniemulsion polymerization in a PFR/CSTR train was modeled
by Samer and Schork [64]. Since particle nucleation and growth are coupled for
macroemulsion polymerization in a CSTR, the number of particles formed in
a CSTR only is a fraction of the number of particles generated in a batch re-
actor. For this reason, their results showed that a PFR upstream of a CSTR has
a dramatic effect on the number of particles and the rate of polymerization in
the CSTR. In fact, the CSTR was found to produce only 20% of the number of
particles generated in a PFR/CSTR train with the same total residence time as
the CSTR alone. By contrast, since miniemulsions are dominated by droplet
nucleation, the use of a PFR “prereactor” had a negligible effect on the rate of
polymerization in the CSTR. The number of particles generated in the CSTR
was 100% of the number of particles generated in a PFR/CSTR train with the
same total residence time as the CSTR alone.
Durant [133] has used a PFR to achieve a miniemulsion with a solids con-
tent in the industrially relevant range. Ouzineb and McKenna [134] have used
a PFR to obtain miniemulsion latexes with high solids contents.

3.9
Continuous Stirred Tank Reactors

The first work on continuous miniemulsion polymerization was by Chen,Goth-

jelpsen and Schork [55]. Theirs was a very simple model of CSTR miniemulsion
polymerization using an oil-soluble initiator. The first experimental work was
that of Barnette [35, 135]. This work showed two significant facts:

(i) Miniemulsion polymerization in a CSTR is not subject to the sustained

or decaying oscillations very often found in CSTR macroemulsion poly-
merization. The sustained oscillations in macroemulsion take place at low
surfactant concentration due to the competition for micellar surfactant
between existing particles requiring additional surfactant to stabilize the
increased interfacial area produced by particle growth, and nucleation of
new particles from micelles. Since, in miniemulsion polymerization, the
nucleation occurs from monomer droplets, and micelles do not exist, no
such competition exists, and the monomer conversion climbs monotoni-
cally on start-up to its steady-state value.
(ii) In miniemulsion polymerization, the steady-state monomer conversion is
approximately twice that found in macroemulsion polymerization (after
the oscillations have died away). This means that a single CSTR will yield
approximately twice as much polymer as the same reactor carrying out a
macroemulsion polymerization.
Miniemulsion Polymerization 175

Observation (i) above can be understood in terms of droplet nucleation and the
lack of competition between nucleation and growth.A mechanistic understand-
ing of observation (ii) above was provided by Samer and Schork [64]. Nomura
and Harada [136] quantified the differences in particle nucleation behavior for
macroemulsion polymerization between a CSTR and a batch reactor. They
started with the rate of particle formation in a CSTR and included an expression
for the rate of particle nucleation based on Smith Ewart theory. In macroemul-
sion, a surfactant balance is used to constrain the micelle concentration, given
the surfactant concentration and surface area of existing particles. Therefore,
they found a relation between the number of polymer particles and the resi-
dence time (reactor volume divided by volumetric flowrate). They compared this
relation to a similar equation for particle formation in a batch reactor, and con-
cluded that a CSTR will produce no more than 57% of the number of particles
produced in a batch reactor. This is due mainly to the fact that particle forma-
tion and growth occur simultaneously in a CSTR, as suggested earlier.
An approach similar to that taken by Nomura and Harada was used by
Samer to quantify the effects of droplet nucleation on emulsion polymerization
kinetics in a CSTR. In their simplified analysis, it was assumed that radical cap-
ture by particles and droplets is proportional to the ratio of particle and droplet
diameters. This assumption is reliable at low to moderate residence times, when
polymer particles still closely resemble monomer droplets with respect to com-
position and surface characteristics. For predominant droplet nucleation, the
maximum particle generation is limited by the concentration of monomer
droplets in the feed. In Fig. 11 the steady state particle generation is given as
a function of the residence time and temperature. Nucleation efficiency is
defined as the number of particles divided by the number of droplets in the

Fig. 11 Model predictions for the number of particles in CSTR miniemulsion polymeriza-
tion expressed as the number of particles divided by the number of droplets in the feed
(from [64])
176 F. J. Schork et al.

feed and is shown to increase with increasing temperature. It can be seen in this
figure that the nucleation efficiency approaches unity for residence times
greater than 30 minutes. Therefore, whereas the particle number is limited by
surfactant in macroemulsion emulsion polymerization in a CSTR, in mini-
emulsion polymerization, the particle number is limited by monomer droplet
concentration, but to a much smaller extent. Therefore, since nucleation effi-
ciencies in miniemulsion polymerization approach 100%, rather than the 57%
predicted for macroemulsion polymerization, the steady-state conversion in
a miniemulsion (proportional to particle number) found by Barnette [36] is
approximately twice that in a macroemulsion.
Samer and Schork [69, 137] confirmed Barnette’s findings on the lack of
oscillations in the CSTR miniemulsion polymerization of MMA.Aizpurua and
Barandiaran [138] confirmed the lack of oscillations in CSTR miniemulsion
polymerization over a wide range of surfactant and initiator concentrations
for VAc. They also observed near-identical MWD for macroemulsion and mini-
emulsion polymerizations under the same conditions. This is understandable,
since MWD should be determined by particle size (and number of radicals per
particle) rather than by particle nucleation mechanism. Aizpurua et al. [139]
successfully used polymeric costabilizers in the CSTR miniemulsion polymer-
ization of VAc at high solids levels. Neither sonication alone, nor the presence
of costabilizer alone, was able to eliminate the oscillations found in macro-
emulsion polymerization. However, sonication and costabilizer together were
capable of eliminating oscillations, indicating droplet nucleation. The results
cited above are particularly significant, since, as fresh miniemulsion droplets
are introduced into a CSTR, they must compete to retain their monomer with
existing particles that may contain more than 50% polymer (50% conversion).
Therefore, for instance, miniemulsion droplets stabilized with 5% polymer
would need to compete with existing particles containing 50% polymer. The
fact that the particle number approaches the droplet number in the feed sug-
gests that the droplet number (if not droplet size distribution) is conserved.
Samer also demonstrates the existence of multiple steady states in isothermal
miniemulsion polymerization in a CSTR. This is not surprising, since multipli-
city is a function of gel or Trommsdorf effect, and not of nucleation mechanism.
Miniemulsion copolymerization in a CSTR involves some very interesting
features. However, in the interest of clarity, these systems will be discussed
along with results for batch copolymerization.

4
Applications

It should be apparent by now that miniemulsion polymerization systems have

some properties that ought to be exploitable when making polymer colloidal
products with unique or improved properties. This section will discuss some
of these documented and potential applications.
Miniemulsion Polymerization 177

Overall, there are two areas in which miniemulsion polymerization differs

significantly from macroemulsion polymerization. First, miniemulsions exhibit
a significant robustness of nucleation. Macroemulsion polymerization relies
heavily on micellar nucleation. Micellar nucleation is notoriously nonrobust.
It has been called “near-chaotic”, because final particle number depends on the
operating variables (temperature, initiator concentration and addition method,
mixing, and so on) and on the quality of the reagents (including initiator purity
and level of inhibitor in the monomer). As stated above, this is due to the com-
petition for surfactant between the nucleation of new particles from micelles
and the adsorption of surfactant onto the surface of growing polymer particles.
In fact, the sensitivity to inhibitor levels has actually been used to manipulate
particle size: addition of an oil-soluble initiator is known to result in the
nucleation of more particles, and hence smaller particles. The introduction of
a water-phase inhibitor is known to extend Interval I, allowing particle growth
to occur simultaneously with particle nucleation, producing fewer and hence
smaller particles. Finally, the use of seeded systems to control particle number
and size in commercial macroemulsion polymerization highlights the poor
robustness of particle nucleation in macroemulsion polymerization. This
should be contrasted with miniemulsion polymerization where, if droplet
nucleation efficiency can be driven close to unity, the number of particles is
determined by the levels of surfactant and costabilizer, and so should be con-
trollable independent of the initiation system.While the number of particles is
often preserved during miniemulsion polymerization, we have seen before that
the PSD may not be. We have also seen the effect of robustness on CSTR poly-
merization, in that CSTR miniemulsion polymerization exhibits a substantially
higher rate of particle nucleation and a lack of oscillations. This section will
explore the ramifications of robust nucleation in miniemulsions.
The second property of miniemulsions that makes them unique when com-
pared to macroemulsions is the virtual lack of monomer transport. Recall that,
during Interval II of a macroemulsion polymerization, monomer must diffuse
from the monomer droplets, across the aqueous phase, and into the growing
polymer particles. Most macroemulsions are presumed to be reaction-limited.
In other words, the rate of monomer diffusion is rapid in comparison with the
rate of propagation, so that polymerization is reaction-limited. This is certainly
true for monomers with water solubilities as low as that of styrene, but for
monomers of much lower water solubility, this assumption may be questioned.
Also, if any sort of non-monomeric materials are to be incorporated into
the polymer particles, these are not likely to be transported easily across the
aqueous phase. Examples of such systems would be preformed polymers or
oligomers in hybrid miniemulsion polymerization, or solid particles in the case
of nanoencapsulation. Certainly high molecular weight prepolymers or solid
particles will not traverse the aqueous phase, and so, if introduced into a
macroemulsion system, they will remain outside the loci of polymerization. In
addition, in living radical polymerization, the molecular weight control agents
are often only sparingly soluble in water. If the control agent is not at the locus
178 F. J. Schork et al.

of polymerization, the polymerization will proceed by uncontrolled free radical

polymerization, and any advantages of livingness are lost. This section will
explore areas where miniemulsion polymerization may be superior to
macroemulsion polymerization because of its virtual lack of interphase mass
transfer.

4.1
Robust Nucleation

4.1.1
Effect of Initiation and Inhibition

Reimers [95] used polymeric costabilizer to carry out miniemulsion polymer-

ization of MMA. Droplet nucleation was found to be the dominant nucleation
mechanism in the polymerization.As a result, the nucleation was more robust,
and the polymerizations were less sensitive to variations in the recipe or con-
taminant levels. This was evident in the rates of polymerization and in the par-
ticle numbers. The miniemulsion polymerizations were subjected to changes
in initiator concentration, water-phase retarder, and oil-phase inhibitor, and
were shown to be significantly more robust.
Batch miniemulsion polymerization of MMA using PMMA as the costabi-
lizer was carried out with SLS as the surfactant and KPS as the initiator. Solids
content was kept at ~30%. A low surfactant level was used with the miniemul-
sions to ensure droplet nucleation. The initiator concentration of the polymer-
stabilized miniemulsion polymerizations was varied from 0.0005 to 0.02 Maq ,
based on the total water content. An aqueous phase retarder, (sodium nitrite)
or an oil-phase inhibitor (diphenylpicrylhydrazol [DPPH]), was added to
both the miniemulsions and the macroemulsions prior to initiation. Particle
numbers and rates of polymerization for both systems were determined.

4.1.1.1
Results

Results from the polymer-costabilized miniemulsion polymerizations are

shown in Table 2. Droplet sizes were found to vary between 115.1 and 121.0 nm.
These are in accord with measurements made by Fontenot [140] for MMA
miniemulsions stabilized with hexadecane. The sizes of the particles in the final
products were close to the sizes of the droplets, ranging from 102.6 to 108.1 nm,
with polydispersities ranging from 1.011 to 1.027. The ratio of the number of
particles to the number of droplets (Np/Nd) was found to be between 0.95 and
1.08. Therefore, the majority of the droplets were nucleated to form polymer
particles. Droplet nucleation led to polymerization rates comparable to those
for the corresponding macroemulsions. For equal concentrations of initiator,
0.01 Maq, the rates are 0.199 and 0.233 gmol/min Laq for the mini- and the macro-
emulsion polymerizations, respectively.
Miniemulsion Polymerization 179

Table 2 Results from polymer-stabilized miniemulsion polymerizations (from [95])

Dd (nm) Dp (nm) PDIp Np¥10–17 Np/Nd Rp (mol/

(L–1) min Laq)

[I]= 0.0005 118.1 105.1 1.013 4.452 0.99 0.096

0.001 117.5 104.3 1.015 4.508 1.00 0.102
0.002 116.8 105.1 1.018 4.428 0.98 0.151
0.005 120.2 103.1 1.017 4.328 1.04 0.263
0.01 117.4 105.1 1.016 4.548 1.01 0.199
0.02 – – – – – 0.176

[NaNO2]= 0.0 117.4 105.1 1.016 4.548 1.01 0.199

0.0001 115.1 103.3 1.027 4.432 0.99 0.253
0.0005 117.3 104.1 1.016 4.780 1.06 0.203
0.001 118.7 102.7 1.011 4.840 1.08 0.201
0.002 117.1 102.9 1.017 4.640 1.03 0.180
0.005 118.2 118.4 1.014 3.304 0.47 0.016

[DPPH]= 0.0 117.4 105.1 1.016 4.548 1.01 0.199

0.00005 118.2 103.2 1.013 4.704 1.05 0.155
0.0001 118.4 103.3 1.012 4.720 1.05 0.149
0.0005 117.8 102.6 1.014 4.732 1.05 0.146
0.001 117.3 102.1 1.017 4.752 1.06 0.090

RPM= 100 119.2 107.0 1.012 4.328 0.95 0.086

200 120.2 103.1 1.017 4.096 1.04 0.111
300 120.8 108.1 1.012 4.096 0.97 0.138
400 120.8 107.9 1.015 4.072 0.98 0.144
500 121.0 107.4 1.012 3.996 0.95 0.143

The effects of the (water-soluble) initiator concentration on the polymer-

ization of polymer-stabilized miniemulsion are shown in Table 2. An increase
in the initiator concentration does not change the number of particles, but does
increase the rate of polymerization. This is due to an increase in the number of
radicals per particle. However, the number of radicals per particle ranged from
just 0.5 to 0.8, indicating that the kinetics (after nucleation) are still essentially
Smith Ewart Case II. The number of particles was found to be proportional to
the initiator concentration raised to the power of 0.002±0.001. Macroemulsion
polymerizations, in contrast, show a dependence of 0.2 and 0.4 for methyl
methacrylate and styrene, respectively [141]. The fact that the exponent ap-
proaches zero indicates that all or nearly all of the droplets are being nucleated.
A water-phase retarder (sodium nitrite) was added to both the mini- and
macroemulsion polymerizations. The rate of polymerization was reduced with
increasing level of retarder, as would be expected. However, the number of
particles increased with increasing retarder concentration. This result would
only be expected with an oil-soluble retarder. The reason for this anomaly is
180 F. J. Schork et al.

Fig. 12 The effect of a water-phase retarder on the number of particles in macroemulsion

polymerization (from [95])

unknown. If we plot the log of the nitrite concentration against the log of the
number of particles, we find a linear relationship between them, as depicted by
Fig. 12. The slope of this line is 0.153±0.009. This value is close to the value of
0.2 reported for the initiator dependence, perhaps implying that the function
of the water-phase retarder is simply to reduce the effective radical flux to the
particles. The polymer-stabilized miniemulsions are far less sensitive to the
presence of the retarder than are the macroemulsions. The retarder has little ef-
fect on the particle number. Particle numbers remained fairly constant up to a
concentration of 5 mMaq. Up to this amount, the dependence of the retarder
concentration on the number of particles was calculated to be 0.020±0.007
(Fig. 13). This is significantly less than the value found for macroemulsions.
(It is presumed that the highest level of retarder prevents a large fraction of the
droplets from ever being nucleated.) Monomer conversions exhibit prolonged
nucleation periods, but the rates are not significantly affected.Again the nitrite
is acting as a retarder, since no induction period is observed.
Macroemulsion polymerizations carried out in the presence of an oil-phase
inhibitor (DPPH) resulted in an increase in the number of particles. Presum-
ably initiator radicals that enter droplets are terminated by the inhibitor,
resulting in dead particles. These particles do not grow, and hence do not con-
sume surfactant to stabilize their increasing surface area, until they absorb
another radical. The surfactant not adsorbed by dead particles is available to
Miniemulsion Polymerization 181

Fig. 13 The effect of a water-phase retarder on the number of particles in a polymer-stabil-

ized miniemulsion polymerization (from [95])

Fig. 14 The effect of an oil-phase inhibitor on the number of particles in macroemulsion

polymerization (from [95])
182 F. J. Schork et al.

Fig. 15 The effect of an oil-phase inhibitor on the number of particles in miniemulsion

polymerization (from [95])

stabilize new particles, thereby increasing the total number of particles. Since
the nucleation period is lengthened, the polydispersity increases. Figure 14
shows that the dependence of the inhibitor concentration on the number of
particles is 0.176±0.010. Conversion time curves indicate that an induction
period results from the presence of the inhibitor. Since polymer-stabilized
miniemulsion polymerization occurs via droplet nucleation, it should be less
sensitive to oil-phase inhibition. Initiator radicals will enter the droplet one
after the other until all of the inhibitor is used up, and the monomer polymer-
izes. This does not affect the number of droplets or particles.As seen in Fig. 15,
the number of particles is proportional to the DPPH concentration raised to the
power of 0.0031±0.0001. Therefore, the number of particles is essentially
independent of the presence of inhibitor.

4.1.1.2
Summary

Shifting the site of nucleation to the droplets greatly enhances the robustness
of the nucleation process to recipe variations, inhibition levels, and changes in
operating procedure (initiation rate and/or agitation rate).As a result of droplet
nucleation, polymer-stabilized miniemulsion polymerizations are far less sen-
Miniemulsion Polymerization 183

Table 3 Summary of dependence of particle number on impurities and operational varia-

tions ([95])

X a (macroemulsion) a (miniemulsion)

[I]a 0.23 0.002±0.001

[NaNO2] 0.153±0.009 0.020±0.007
[DPPH]b 0.176±0.010 0.0031±0.0001
RPM – –0.026±0.001

a [I] and [NaNO2] in gmol/Laq .

b [DPPH] in gmol/Lmon .

sitive to these variations in operation. The dependence of the particle number

on the concentration of initiator, water-phase retarder, oil-phase inhibitor,
and agitation are shown in Table 3. The exponents for the variation of particle
number with each of these variations were 0.002, 0.02, 0.0031, and –0.026,
respectively. The corresponding values for the macroemulsions were one to two
orders of magnitude larger. Therefore, nucleation in polymer-stabilized
miniemulsion polymerizations was found to be more robust than in macro-
emulsion polymerizations.
An enhanced robustness can benefit a process in a number of ways. Since the
polymer-stabilized miniemulsions are less susceptible to disturbances, their
polymerization is less likely to be affected by operator error, fluctuations in feed
stream concentrations and residual contaminants in the reaction vessel. Many
monomers contain species that can act as inhibitors or retarders as a result of
monomer production, storage, or processing. These contaminants also cause
batch-to-batch variability in particle number in macroemulsions. Therefore,
miniemulsion polymerization may be an alternative to seeded polymerization
as a way of maintaining robust control of particle number.

4.1.2
Particle Size Distribution

There has been a belief that, due to the fact that the original miniemulsion
droplets are formed by a shear process, the droplet size distribution will be
broad, and so the resulting PSD will have a large polydispersity (as measured
by the polydispersity index, defined as the mass average over the number av-
erage particle radius). In a recent note [142] Landfester et al. discuss particle
size polydispersity in miniemulsions and attempt to dispel the idea that
miniemulsions necessarily have broader PSD than the equivalent macroemul-
sions. Rather, they argue that the PSD of a miniemulsion can be either broader
or narrower than its macroemulsion counterpart, and that, in most cases, the
miniemulsion will have a polydispersity equal to, or only very slightly greater
than, the equivalent macroemulsion.
184 F. J. Schork et al.

4.1.2.1
Hexadecane as Costabilizer

Fontenot and Schork [140] studied the miniemulsion polymerization of methyl

methacrylate using hexadecane as the costabilizer. A portion of their results
are shown in Table 4. Polydispersities are listed for macroemulsion, and mini-
emulsions subjected to varying durations of sonication, at two levels of initia-
tor. In this and all cases following, the macroemulsions and miniemulsions
were made from the same recipe, but with the costabilizer left out of the
macroemulsion. The miniemulsions and macroemulsions were polymerized by
the same procedure except that the sonication was eliminated for the macro-
emulsions. It may be seen that at both initiator levels, the macroemulsion is
slightly more narrow than some of the miniemulsions, but broader than others.
An estimate of the standard deviation of the polydispersity measurement is
given as ±0.01–0.02, and may be applied to all of the polydispersity data report-
ed. With this standard deviation estimate, it may be seen that the differences
in polydispersity between the macro- and miniemulsions are not likely to be
significant.
Landfester et al. [143] studied the miniemulsion polymerization of styrene
using hexadecane as the costabilizer. When styrene miniemulsions were sub-
jected to varying sonication times (see Table 5), very similar trends are seen as
for the MMA miniemulsions. The particle size and the polydispersity of
miniemulsion droplets rapidly polymerized after sonication either do not
depend on the amount of the costabilizer, or are very weak functions of the
amount of costabilizer (see Table 6). It was found that doubling the amount of
costabilizer does not decrease the radius nor have any effect on the polydis-

Table 4 Polydispersity index as a function of initiator concentration and sonication time,

with hexadecane as costabilizer and MMA as monomer ([140])

Sonication time Polydispersity index

[I] 0.005 mol/L(aq)

Macroemulsion 1.05
2 min 1.08
4 min 1.06
6 min 1.04
12 min 1.05

[I] 0.01 mol/L(aq)

Macroemulsion 1.05
2 min 1.07
4 min 1.06
6 min 1.05
8 min 1.07
12 min 1.04
Miniemulsion Polymerization 185

Table 5 Polydispersity index as a function of sonication time, with hexadecane as costabil-

izer, styrene as monomer, and [I]=0.3 mol/L (aq) (from [143])

Sonication time Diameter di (nm) Polydispersity index

Macroemulsion – 1.04
0.5 min 135 1.01
1 min 112 1.03
2 min 96 1.00
5 min 87 1.03
10 min 84 1.02
20 min 83 1.01

Table 6 Hexadecane as costabilizer, with styrene monomer (from [143])

Hexadecane level (gm) Particle diameter di (nm) Polydispersity index

Macroemulsion 98 1.04
0.33 109 1.03
0.66 108 1.01
1.66 108 1.01
3.33 102 1.04
5 100 1.03
6.66 99 1.05
8.33 95 1.01

persity. It was also found that the droplet size is initially a function of the
amount of mechanical agitation. The droplets are rapidly reduced in size
throughout sonication in order to approach a pseudo-steady state [143]. Once
this state is reached, the size of the droplet does not change. Higher sonication
time causes a slight reduction in polydispersity.
After halting sonication, a rather rapid equilibration process must occur.
Since the droplet number after sonication is fixed, this process does not in-
fluence the average size, but the droplet size distribution usually undergoes
very rapid change. It was found that steady-state miniemulsification results in
a system “with critical stability”; in other words the droplet size is the product
of a rate equation of fission by ultrasound and fusion by collisions, and the
droplets are as small as possible for the timescales involved. The equality of
droplet pressures makes such systems insensitive against net mass exchange by
diffusion processes (after the very fast equilibrium process at the beginning),
but the net positive character of the pressure makes them sensitive to all changes
in the droplet size. Steady-state homogenized miniemulsions, which are criti-
cally stabilized, undergo droplet growth on a timescale of hundreds of hours,
presumably by collisions or by costabilizer exchange. As can be seen from
Table 7, during this growth, the polydispersity does not change significantly.
186 F. J. Schork et al.

Table 7 Influence of time delay between the ultrasonication and the polymerization
(from [143])

Time delay between start of Particle diameter di (nm) Polydispersity index

polymerization and sonication (h)

0 82 1.01
1 87 1.05
6 108 1.03
48 152 1.03
96 164 1.04

Therefore, we may conclude that there is indeed no significant difference in

polydispersity between the miniemulsion and the equivalent macroemulsion.

4.1.2.2
Dodecyl Mercaptan as Costabilizer

Mouron et al. [105] and Wang et al. [106] have used dodecyl mercaptan (DDM)
as the costabilizer in styrene and MMA miniemulsion polymerizations, re-
spectively. Some of the results are shown in Table 8 and Table 9. For styrene
(Table 8), the macroemulsion is compared with miniemulsions containing
varying levels of DDM (costabilizer). In this case, the macroemulsion has a
broader particle size distribution than all but one of the miniemulsions. For
MMA (Table 9), miniemulsions and the equivalent macroemulsions have been
compared at varying initiator concentrations. In this case, the macroemulsions

Table 8 Dodecyl mercaptan as costabilizer with styrene monomer (from [106])

DDM level (gm) Polydispersity index

Macroemulsion 1.02
1 1.01
2 1.01
3 1.02
4 1.04

Table 9 Dodecyl mercaptan as costabilizer with methyl methacrylate monomer (from [105])

Initiator (mol/L(aq)) Macroemulsion PDI Miniemulsion PDI

0.005 1.02 1.02

0.01 1.01 1.02
0.02 1.01 1.02
Miniemulsion Polymerization 187

all have narrower particle size distributions, although the difference is hardly
significant.

4.1.2.3
Polymethyl Methacrylate as Costabilizer

Reimers and Schork [144] have used polymethyl methacrylate as the co-
stabilizer for methyl methacrylate miniemulsion polymerization. A portion of
the results are shown in Table 10. In this case, the miniemulsion has a narrower
particle size distribution than the equivalent macroemulsion.

4.1.2.4
Influence of the Amount of the Surfactant

Colloidal stability is usually controlled by the type and amount of surfactant

employed. In miniemulsions, the fusion-fission rate equilibrium during soni-
cation, and therefore the size of the droplets directly after primary equilibra-
tion, depends on the amount of surfactant. For styrene miniemulsions that use
SLS as surfactant, droplet sizes between 180 nm down to 32 nm can be obtain-
ed. The polydispersity slightly increases with decreasing size, but is still quite
low (see Table 11). Using similar molar amounts of the simple cationic surfac-

Table 10 Polymethyl methacrylate as costabilizer with methylmethacrylate monomer (from

[144])

Polydispersity index

Macroemulsion 1.02
Miniemulsion 1.01

Table 11 Polydispersity index as a function of SDS concentration (from [105])

SDS concentration Particle diameter di (nm) Polydispersity index

(% compared to monomer)

0.3 180 1.03

0.5 134 1.07
1.0 108 1.02
1.5 94 1.02
2.1 89 1.08
3.5 82 1.08
4.9 82 1.03
6.8 65 1.03
10.3 55 1.05
17.0 46 1.06
25.2 42 1.07
188 F. J. Schork et al.

Table 12 Polydispersity index as a function of CTAB concentration (from [70])

CTAB concentration Particle diameter di (nm) Polydispersity index

(% compared to monomer)

0.4 347 1.01

0.7 159 1.05
1.2 125 1.05
2.4 102 1.04
3.6 86 1.01
10.0 59 1.09
16.7 59 1.13

tant cetyltrimethylammonium bromide (CTAB) or the anionic surfactant SLS

results in similar particle sizes, showing that the particle size is essentially con-
trolled by the limit of the surfactant coverage of the latex particles [71, 145].
Again, the polydispersity increases with decreasing size (see Table 12), but is
only slightly higher than in the SDS miniemulsions.

4.1.2.5
Summary

Based on the data above, it would appear that it is possible, via miniemulsion
polymerization, to make a polymer latex with a particle size distribution that
approaches that made by macroemulsion polymerization. In some cases, the
miniemulsion product may be even narrower than the macroemulsion. There
are two significant mechanisms leading to this narrowness. First, the monomer
droplet size distribution is to some extent determined by the thermodynamics
of swelling, and not solely by the droplet size distribution induced by the soni-
cator or homogenizer. For this to be true, the process should include a ripening
time between sonication and polymerization. During this ripening time, the
droplets will come to swelling equilibrium. Studies show that the ripening time
is of the order of seconds to minutes, and is naturally included in the prepara-
tion of batch polymerizations. Second, the narrowness of the particle size dis-
tributions depends on the ability to nucleate nearly all of the droplets over a
short period of time. If droplet nucleation takes place over a longer period of
time, some particles will have polymerized for a longer time, and some droplets
will lose monomer by mass transfer to growing particles before the droplets
begin to polymerize. Using hexadecane or polymer as a costabilizer will facili-
tate one hundred percent droplet nucleation, while the use of cetyl alcohol does
not. Miller et al. [109] have shown that a small amount of polymer dissolved in
the monomer droplets enhances droplet nucleation.Also, the initiator flux must
be high enough to nucleate all of the droplets within a short time interval.
In summary, the miniemulsion route to polymer latexes should not be dis-
missed solely due to a requirement for narrow particle size distribution, par-
Miniemulsion Polymerization 189

ticularly when the unique properties of the miniemulsion process may be of

particular advantage.

4.1.3
Shear Stability

The shear stabilities of mini- and macroemulsion latexes were compared and
quantitatively evaluated with respect to their particle size distributions by
Rodrigues and Schork [146]. Although miniemulsion latexes exhibit many of
the properties of macroemulsion latexes, there may be subtle differences in
particle size distribution and surface characteristics due to differences in their
polymerization mechanisms. To study the effects of these differences on the
shear stabilities of the miniemulsions, a quantitative approach was developed
where changes in the average diameter and total number of particles have been
related to the particle size distribution before and after shearing.
Two pairs of MMA mini- and macroemulsion latexes were polymerized for
this study. HD was used as the costabilizer for the miniemulsions, and the
polymerizations were carried out at 60 °C. Efforts were made to make the main-
and macroemulsion in each pair as similar as possible. The two pairs were:
Pair I
Macroemulsion (Sample A)
0.02 mol. SLS/Laq; 0.0115 mol. KPS/Laq; 2 g DDM;
PSD range: 141–188 nm (diameter)
Miniemulsion (Sample E)
0.02 mol. SLS/Laq; 0.0115 mol. KPS/Laq; 2 g DDM;
PSD range: 96–123 nm (diameter)
Pair II
Macroemulsion (Sample H)
0.01 mol. SLS/Laq; 0.0115 mol. KPS/Laq; 4 g DDM
PSD range: 167–241 nm (diameter)
Miniemulsion (Sample D)
0.01 mol. SLS/Laq; 0.0115 mol. KPS/Laq; 4 g DDM
PSD range: 145–209 nm (diameter)
The samples were sheared using a rotational viscometer with a coaxial cylinder
system, based on the Searle-type, where the inner cylinder (connected to a
sensor system) rotates while the outer cylinder remains stationary. The outer
cylinder surrounding the inner one was jacketed, allowing good temperature
control, and the annular gap was of constant width. The sensor system used was
the NV type, with a rotor with a recommended viscosity range of 2¥103 mPa,
a maximum recommended shear stress of 178 Pa, and a maximum recom-
mended shear strain rate of 2700 s–1; this rotor could work with volumes from
10–50 ml. Flow was laminar.
190 F. J. Schork et al.

The optimum shear rate for each pair was arrived at by trial and error, such
that the shearing produced aggregation, but not massive coagulation. That is,
the shear rate was not increased beyond the point at which the particle dia-
meters stopped increasing (and maybe even started decreasing again). All
the tests were conducted at 25 °C; each miniemulsion latex within a pair was
sheared for the same time interval as its macroemulsion latex counterpart. Only
changes in particle sizes were followed during these series of experiments, with
the particle size distribution (analyzed by dynamic light scattering, DLS) re-
corded before and after shearing for each of the shear experiments.
The ratio of total number of particles initially present to the total number
of particles after shearing (Ni/Nf) was computed and used both to compare the
two types of latexes, and to determine the extent of aggregation and the nature
of the aggregates formed.

4.1.3.1
Relative Shear Stability of Miniemulsion and Macroemulsion Latexes

The particle size range and average particle diameter before and after shear-
ing, and the shear rate and time of shear used for each of the sample pairs is
shown in Table 13. For both pairs, the shifts in the particle size range and in the
average diameter are substantially greater for the macroemulsion latex than the
corresponding miniemulsion latex. In all cases, the particle size distribution
broadened after shearing. The percentage change in average diameter is greater
for the macroemulsions as well.
The ratio of the initial total number of particles to the final total number
of particles after shearing (N0/Nf) is shown in Table 14. It is clear that the
macroemulsion latexes showed greater shear instability. The ratio No/Nf gives
an indication of the extent of aggregation in each of the latexes; a No/Nf ratio
of two would indicate that average aggregation up to doublet formation has
taken place, and so on. It can be seen from Table 14 that aggregation has taken
place in all the latexes essentially up to doublet formation, with slightly higher

Table 13 Particle size ranges and average particle diameters before and after shearing, and
the shear rates and times of shear for samples A, E, H, and D (from [146])

Sample Shear Time of Initial PSD Average PSD range Average % Change
rate shear range diameter after diameter in average
(s–1) (s) (diameter) before shearing after diameter
(nm) shearing (nm) shearing (nm)
(nm) (nm)

A (macro) 200 1260 141–188 155 181–244 204 31.61

E (mini) 200 1260 96–123 108 106–143 121 12.04
H (macro) 200 1260 167–241 195 218–319 264 35.38
D (mini) 200 1260 145–209 168 170–223 195 16.07
Miniemulsion Polymerization 191

Table 14 Ratio of the initial total number of particles to the final number of particles after
shearing for samples A, E, H, and D (from [146])

Sample % Solids Ni/Nf

A (macro) 27.2 2.28

E (mini) 25.9 1.41
H (macro) 20.9 2.87
D (mini) 24.8 2.43

aggregate formation in the macroemulsion latexes than in the miniemulsion

latexes, as evidenced by their correspondingly higher No/Nf values.

4.1.3.2
Effect of Large Particles

The effects of a few externally-added large particles on the shear stabilities of

miniemulsion- and macroemulsion latexes were also investigated. This was
done to determine if the greater shear instability of the macroemulsion latex
was due to the presence of a few large particles, possibly formed by droplet
nucleation during the synthesis of the latex itself. To test this, portions of a
larger particle size macroemulsion (Sample C) were added to the two macro-
emulsion latexes (Samples A and H), and portions of a larger miniemulsion
(Sample G) were added to the two miniemulsion latexes (Samples E and D)
used in the first part of this analysis. The large particles were as follows:
Macroemulsion Seed Latex (Sample C)
PSD range: 252–298 nm (diameter)
Average diameter: 276 nm
Miniemulsion Seed Latex (Sample G)
PSD range: 320–380 nm (diameter)
Average diameter: 344 nm
The percentages by weight of the larger particle size latexes used were 2, 5, 10
and 25% of total sample weight. The latexes were then sheared, and the shift in
the particle size distribution range and average particle diameter, the percent
change in this average diameter after shearing, and the No/Nf values were de-
termined for both sets of samples. These results are shown in Tables 15 and 16.
Table 15 shows that, for all four samples analyzed, the percent change in
average diameter increases with the fraction of larger particles, reaches a maxi-
mum, and then decreases. In all cases, the maximum shear aggregation occurs
at approximately 5–10% large particles. This supports (but by no means proves)
the hypothesis that macroemulsion latexes are more susceptible to shear co-
agulation than miniemulsion latexes due to the presence of a small fraction of
large particles originating from droplet nucleation. It is important to remember
192 F. J. Schork et al.

Table 15 Particle size ranges and average particle diameters before and after shearing,
and the shear rates and times of shear for samples A, E, H, and D with large particles added
(from [146])

Sample Shear Time of Initial PSD Average PSD range Average % Change
rate shear range diameter after diameter in average
(s–1) (s) (diameter) before shearing after diameter
(nm) shearing (nm) shearing (nm)
(nm) (nm)

A (original- 200 1260 141–188 155 181–244 204 31.61

macro)
A (2% big C) 200 1260 143–310 170 194–334 228 34.12
A (5% big C) 200 1260 144–317 176 208–358 249 41.48
A (10% big C) 200 1260 143–312 182 202–344 246 35.16
A (25% big C) 200 1260 143–308 199 196–333 250 25.63

E (original- 200 1260 96–123 108 106–143 121 12.04

mini)
E (2% big G) 200 1260 102–372 118 114–380 153 29.66
E (5% big G) 200 1260 103–379 124 124–389 168 35.48
E (10% big G) 200 1260 106–384 131 130–397 177 35.11
E (25% big G) 200 1260 105–377 140 123–394 186 32.86

H (original- 250 1740 167–241 195 233–348 277 42.05

macro)
H (2% big C) 250 1740 169–323 199 236–352 290 45.73
H (5% big C) 250 1740 171–338 207 264–388 324 56.52
H (10% big C) 250 1740 170–333 212 253–369 305 43.87
H (25% big C) 250 1740 170–312 222 242–355 297 33.78

D (original- 250 1740 145–209 168 197–281 226.00 34.52

mini)
D (2% big G) 250 1740 149–370 179 224–377 247.00 37.99
D (5% big G) 250 1740 151–374 193 233–381 266 37.82
D (10% big G) 250 1740 154–373 208 242–399 284 36.54
D (25% big G) 250 1740 155–375 226 237–389 305 34.96

that in all four latexes used here, the size differential between these latexes and
the two latexes used as large particles (Samples C and G) is not very large.
Therefore, the particles that qualify as large particles by definition lie at the
upper end of the particle size distribution of samples C and G; in other words,
only a small fraction of the externally-added larger latex actually functions as
large particles in terms of influencing shear stability.
The results shown in Table 16 support the observations drawn from Table 15.
For all of the latexes, there is an increase in No/Nf value as the percentage
of externally added large particles increases, up to approximately 5–10%wt
of large particles. Any further addition of large particles decreases the No/Nf
value.
Miniemulsion Polymerization 193

Table 16 Ratio of the initial total number of particles to the final number of particles after
shearing for samples A, E, H, and D with large particles added (from [146])

Sample % Solids Ni/Nf

A (original-macro) 27.2 2.28

A (2% big C) 27.1 2.41
A (5% big C) 26.9 2.83
A (10% big C) 26.7 2.47
A (25% big C) 26.2 1.98
E (original-mini) 25.9 1.41
E (2% big G) 25.7 2.18
E (5% big G) 25.6 2.49
E (10% big G) 25.3 2.47
E (25% big G) 25.2 2.34
H (original-macro) 20.9 2.87
H (2% big C) 21.1 3.09
H (5% big C) 21.2 3.83
H (10% big C) 21.2 2.98
H (25% big C) 21.3 2.39
D (original-mini) 24.8 2.43
D (2% big G) 24.9 2.63
D (5% big G) 25.3 2.62
D (10% big G) 25.5 2.55
D (25% big G) 25.6 2.46

4.1.3.3
Summary

Under controlled shearing conditions, miniemulsions were shown to be more

shear stable than similar conventional or macroemulsions. This may be due to
macroemulsion shear instability resulting from the presence of a small number
of large particles (derived from droplet polymerization) that act as seeds for
aggregation. Intentional seeding of mini- and macroemulsions with larger
particles induced increased shear instability, supporting this hypothesis. How-
ever, it may be that miniemulsion polymerization avoids the burying of some
of the initiator end groups. For KPS initiation, these end groups, if on the
surface of the particle, will add substantially to the colloidal stability of the
particles. Due to the nature of macroemulsion polymerization (particle growth
at the expense of monomer droplets), end groups tend to become buried in the
particle where they contribute nothing to colloidal stability. Since there is less
particle growth (none in the ideal case) for miniemulsion polymerization, most
of the initiator end groups should remain on particle surfaces.
While the mechanism is an open question, there clearly seems to be a dif-
ference in shear stability between miniemulsions and macroemulsions.
194 F. J. Schork et al.

4.2
Monomer Transport Effects

One of the most unique properties of miniemulsion polymerization is the lack

of monomer transport. Recall from Fig. 1 that with macroemulsion polymer-
ization, the monomer must diffuse from the monomer droplets, across the
aqueous phase, and into the growing polymer particles. In contrast, in an ideal
miniemulsion (nucleation of 100% of the droplets), there is no monomer trans-
port, since the monomer is polymerized within the nucleated droplets. This
lack of monomer transport leads to some of the most interesting properties of
miniemulsions. For most monomers, macroemulsion polymerization is con-
sidered to be reaction, rather than diffusion limited. However, for extremely
water insoluble monomers, this might not be the case. In this instance, poly-
merization in a miniemulsion might be substantially faster than polymeriza-
tion in an equivalent macroemulsion. For copolymerization in a macroemul-
sion, where one of the comonomers is highly water insoluble, the comonomer
composition at the locus of polymerization might be quite different from the
overall comonomer composition, resulting in copolymer compositions other
than those predicted by the reactivity ratios.

4.2.1
Polymerization of Highly Water-Insoluble Monomers

Balic [147] has made a complete study of the macroemulsion polymerization

of vinyl neo-decanoate (vinyl versatate or VEOVA). This monomer is highly
water-insoluble (4¥10–5 mol/L at 25 °C). Balic reports low rates of polymeriza-
tion and long inhibition periods in macroemulsions. He asserts that this is not
due to monomer transport limitations, and provides calculations to support this.
He attributes the low rates to impurities in the monomer, although he could not
remove these. It could be that the extremely low solubility of the monomer in the
aqueous phase retards the formation of oligomeric radicals of sufficient length
(hydrophobicity) to enter the polymer particles. Under these conditions of very
slow aqueous phase polymerization, the oligomers might be particularly sus-
ceptible to low levels of aqueous phase inhibitor. With the resultant low radical
flux into the particles, the rate of polymerization would be low. It could also be
that the rate of monomer diffusion is not sufficient to allow the polymerization
to be reaction limited, since Balic’s arguments do not necessarily rule out this
possibility. It would be interesting to see the same polymerizations run in
miniemulsion, since this would rule out the monomer transport effect.
Kitzmiller et al. [148] have found the rate of copolymerization for VAc and
vinyl 2-ethylhexanoate to be much slower in macroemulsion than in mini-
emulsion. They attribute this to monomer transport effects for the less water-
soluble monomer.
Miniemulsion Polymerization 195

4.2.2
Copolymer Composition Distribution

While the rate of monomer transport in macroemulsions may or may not

limit the rate of polymerization, it is quite possible that unequal rates of dif-
fusions for comonomers may make the comonomer composition at the locus
different (richer in the more water-soluble monomer) from the overall com-
position.
Copolymerization refers to the process by which two monomers (M1 and
M2) are simultaneously polymerized. Mayo and Lewis [149] developed the
following equation to describe copolymerization kinetics

dM1 M1(r1M1 + M2)

7 = 004 (15)
dM2 M2(M1 + r2M2)

where Mi is the molar concentration of monomer i at the site of propagation.

The reactivity ratios, r1 and r2, are the homopropagation rate constants divided
by the cross propagation rate constants, such that

kp11 kp22
r1 = 6 r2 = 6 (16)
kp12 kp21

where r1 and r2 are determined experimentally, typically from bulk or solution

polymerization experiments. Different types of copolymerization behavior are
observed, depending on the values of r1 and r2. Equation 15 is rewritten in a
more useful form, where the comonomer compositions ( f1 and f2) are related
to the instantaneous copolymer compositions (F1 and F2):

r1 f1
6 +1
dM1 F1 f2
7 = 4 = 02 (17)
dM2 F2 r2 f2
6 +1
f1

Equation 17 is known as the copolymerization or Mayo Lewis equation.

Schuller [150] and Guillot [98] both observed that the copolymer composi-
tions obtained from emulsion polymerization reactions did not agree with the
Mayo Lewis equation, where the reactivity ratios were obtained from homo-
geneous polymerization experiments. They concluded that this is due to the fact
that the copolymerization equation can be used only for the exact monomer
concentrations at the site of polymerization. Therefore, Schuller defined new
reactivity ratios, r¢1 and r¢2, to account for the fact that the monomer concen-
trations in a latex particle are dependent on the monomer partition coefficients
(K1 and K2) and the monomer-to-water ratio (y):
196 F. J. Schork et al.

1 1
1+ 7 1+ 7
K2y K1y
r¢1 = r1 03 r¢2 = r2 03 (18)
1 1
1+ 7 1+ 7
K1y K2y

The Mayo Lewis equation, using reactivity ratios computed from Eq. 18, will
give very different results from the homogenous Mayo Lewis equation for mini-
or macroemulsion polymerization when one of the comonomers is sub-
stantially water-soluble. Guillot [151] observed this behavior experimentally for
the common comonomer pairs of styrene/acrylonitrile and butyl acrylate/vinyl
acetate. Both acrylonitrile and vinyl acetate are relatively water-soluble (8.5 and
2.5%wt, respectively) whereas styrene and butyl acrylate are relatively water-
insoluble (0.1 and 0.14%wt, respectively). However, in spite of the fact that
styrene and butyl acrylate are relatively water-insoluble, monomer transport
across the aqueous phase is normally fast enough to maintain equilibrium
swelling in the growing polymer particle, and so we can use the monomer parti-
tion coefficient.
Schuller’s equation is appropriate when one of the monomers has significant
water solubility and the other does not. In the case where one monomer is water
insoluble, and the other is extremely water insoluble, Schuller’s equation does
not hold. For the comonomer pair of MMA/VS (vinyl stearate), K1K2, so
Schuller’s equation predicts that, for macroemulsion polymerization, the poly-
mer formed early in a batch reaction will be richer in vinyl acetate (VA) than
in the homogenous case. The opposite was found by Reimers [102]. It has long
been accepted that monomer transport is not a rate-limiting step in the con-
ventional emulsion polymerization of relatively water-insoluble monomers,
such as MMA, styrene, and butyl acrylate (BA). However, the water solubility
of VS is as much as three orders of magnitude smaller than these typical emul-
sion polymerization monomers. In this case, VS cannot readily cross the
aqueous phase to saturate the growing polymer particles. (The major transport
resistance is actually from the monomer droplets into the aqueous phase.)
Therefore, Schuller’s partition coefficient model cannot be used here, since
it assumes that the particles are saturated with both monomers. The homo-
geneous polymerization model is not useful either since the monomer con-
centration in the particles is not identical to the bulk monomer concentration.
Therefore, a pseudo-partition coefficient k2 was proposed by Samer [137] for
extremely water-insoluble comonomers in order to interpret the experimental
data. k2 is not a partition coefficient and does not have any physical signi-
ficance.As defined by Samer, it is simply an adjustable parameter that replaces
K2 in Eq. 18. If K1 is set to infinity and k2 is set to unity, Eq. 17 and Eq. 18 can be
used to correlate the copolymer composition with the total monomer conver-
sion data in the MMA/VS (and other) systems.
Extremely water insoluble comonomers are only selectively used in emulsion
polymerization because of concerns about monomer transport limitations.
Miniemulsion Polymerization 197

Typically, copolymer composition can be manually adjusted by slowly feeding

the more reactive monomer in throughout the reaction; but this may not be
helpful when trying to overcome monomer transport limitations. Therefore,
Reimers and Schork [102] performed identical copolymerization experiments
in miniemulsions, where monomer transport is less significant, in order to
determine what effect this would have on the evolution of the copolymer
composition. Data on the MMA/VS (and other) copolymerizations indicate
that the Schuller equation (and not the Samer adaptation) fits the copolymer
composition data. This points to the effect of extremely low monomer water
solubility on copolymer composition in macroemulsion polymerization, and
the relative insensitivity of miniemulsion polymerization to this effect.
We will now look at the type of reactor system (batch, semibatch or CSTR)
from the point of view of miniemulsion copolymerization.

4.2.2.1
Batch Copolymerization

Reimers [102] carried out batch copolymerizations in both macro- and equiv-
alent miniemulsions. MMA was used as the main monomer. The MMA was
copolymerized in macroemulsion- or miniemulsion with p-methylstyrene
(pMS), vinyl hexanoate (VH), vinyl 2-ethylhexanoate (VEH), vinyl n-decanoate
(VD) or vinyl stearate (VS). The comonomers were copolymerized at 10%wt
comonomer, 90%wt MMA. SLS was used as the surfactant and KPS as the
initiator. The comonomers (all highly water insoluble) were used as the co-
stabilizer. Miniemulsions were sonicated, while equivalent macroemulsions were
only subjected to vigorous mixing. Polymerizations were carried out at 60 °C.
Copolymer composition was obtained by integrating characteristic peaks in
the 1H NMR spectra. The integrated peaks correspond to signals produced by
the methyl ester protons (three) of MMA, 3.56 ppm, the aromatic protons (four)
of pMS, 6.7 ppm, and the a-proton of the vinyl esters at 4.8 to 5.6 ppm. The
relative ratios of the peaks allowed the mole fraction of each comonomer in the
copolymer to be assessed. Samples were taken throughout the reaction at
different times to monitor the incorporation of comonomer as a function of the
extent of total monomer conversion. These results were then compared with the
results for average comonomer fraction given by an integrated copolymer
(Mayo Lewis) equation [149], using r1 and r2 from bulk polymerization. The
mole fractions used were based on the total moles of monomer in the droplets
and were corrected for the water solubility.
Both the mini- and macroemulsion copolymerizations of pMS/MMA tend
to follow bulk polymerization kinetics, as described by the integrated copoly-
mer equation. MMA is only slightly more soluble in the aqueous phase, and
the reactivity ratios would tend to produce an alternating copolymer. The mini-
emulsion polymerization showed a slight tendency to form copolymer that
is richer in the more water-insoluble monomer. The macroemulsion formed
a copolymer that is slightly richer in the methyl methacrylate than the co-
198 F. J. Schork et al.

polymer equation would predict. This may be explained by different nucleation

mechanisms operating in each polymerization. Droplets would have a higher
concentration of the less water-soluble comonomer throughout the reaction.
Therefore their nucleation should yield a copolymer with a higher pMS fraction.
In contrast, micellar nucleation should lead to a copolymer with a higher methyl
methacrylate content, because it can cross the aqueous phase more readily.
As the water solubility of the comonomer decreases, the difference in in-
corporation of the hydrophobic monomer between the mini- and macroemul-
sion polymerization becomes more pronounced. This was seen in the copoly-
merization of VH/MMA. The fraction of the hexanoate in the copolymer
formed in the miniemulsion polymerization was substantially higher than
that found with the macroemulsion. This incorporation closely follows the
copolymer equation. The VEH/MMA miniemulsion copolymerization also
followed the copolymer equation. Differences between the mini- and macro-
emulsion polymerization are not as pronounced in this system. For the
VD/MMA and VS/MMA systems there were large differences between the two
copolymerizations. In addition, none of the mini- or macroemulsion copoly-
merizations of vinyl decanoate or vinyl stearate are predicted by the copolymer
equation. The miniemulsion copolymerizations fall above the prediction curve
(more hydrophobic monomer incorporation than predicted), and the macro-
emulsions fall below. In these cases, both micellar and droplet nucleation took
place in the miniemulsion polymerizations, and the presence of micelles tended
to enrich the concentration of the hydrophobic monomer in the droplets, since
the micelles would likely be richer in the more water-soluble MMA.
Samer [104] carried out similar copolymerizations with similar results.
An example of his data is given in Fig. 16. Here 2-ethylhexyl acrylate (EHA) was
copolymerized with MMA in batch. The miniemulsion polymerizations (two
are shown) follow the copolymer equation, while the macroemulsion poly-
merization gives EHA incorporation that is lower than predicted by the co-
polymer equation, presumably due to the low concentration of EHA at the
locus of polymerization. The dotted line in Fig. 16 is for a model derived by
Samer that accurately predicts the copolymer composition. Samer derived this
model by adapting the work of Schuller [149]. Schuller modified the reactivity
ratios for the macroemulsion polymerization of water-soluble monomers to
take into account that the comonomer concentration at the locus of polymer-
ization is different from the comonomer composition in the reactor due to the
water solubilities of the monomers. Samer used the same approach to account
for the fact that the comonomer concentration at the locus of polymerization
might be different from that of the reactor due to transport limitations of
water insoluble comonomers.
Delgado et al. published a series of papers [56, 58, 91, 153–157] on the mini-
emulsion copolymerization of vinyl acetate and butyl acrylate.A very compre-
hensive mathematical model of the polymerization system was developed.
Equilibrium swelling was accounted for, since the model did not presume com-
plete droplet nucleation, and so monomer transport from unnucleated mini-
Miniemulsion Polymerization 199

Fig. 16 Copolymer composition during the batch miniemulsion copolymerization of

methyl methacrylate with 2-ethylhexyl acrylate (from [104])

emulsion droplets to polymer particles occurred. Monomer transport limita-

tions were minor, but with these two monomers, that is not surprising. Delgado
did report [91] a higher BA content (up to 70% conversion) for miniemulsion-
polymerized copolymers relative to those polymerized in an equivalent
macroemulsion. Since BA is significantly less water-soluble than VAc, this
result is in agreement with Reimers. Rodriguez et al. [60, 61] developed a com-
prehensive model of the miniemulsion copolymerization of styrene and
MMA. This system presumes significantly less than 100% nucleation of
monomer droplets, and so there was significant monomer transport from the
monomer droplets to the polymer particles. However, since neither of these
monomers are extremely water insoluble, no effects on copolymer composi-
tion were observed. Ghazaly et al. [117] studied the miniemulsion copoly-
merization of n-butyl methacrylate with crosslinking macromonomers (two
double bonds per molecule). Since the comonomers were macromers, pre-
sumably, it was important to use a miniemulsion to avoid common transport
issues. Wu and Schork [152] studied the batch and semibatch miniemulsion
copolymerization of VAc with BA. They found the copolymer composition
for batch miniemulsion polymerization to be in good agreement with the
copolymer equation. However, there was significant deviation for batch macro-
emulsion polymerization, especially at low conversion. Delgado [157] attri-
buted this effect to the small amount of water that exists in the monomer-
swollen polymer particles. Because of the high water solubility of VAc, the
particles then contain a higher level of VAc, and so BA incorporation is sup-
pressed.
200 F. J. Schork et al.

4.2.2.2
Semibatch Copolymerization

Controlling copolymer composition has long been of prime interest in polymer

reaction engineering. Because of possible differences in reactivity ratios, the
copolymer composition distribution may be broad, and only the overall
(average) copolymer composition at full conversion need be at the ratio of
monomer feeds. As noted above, Mayo and Lewis [149] studied the kinetics of
copolymerization and developed an equation to describe the relationship be-
tween the molar concentrations at the site of propagation and reactivity ratios
of monomers for homogeneous copolymerization, such as bulk or solution
polymerization. Semibatch polymerization (both solution and macroemulsion)
has long been used to produce copolymers of desired copolymer composition
distribution. This may be done on one of two ways. First, in a binary polymer-
ization the more reactive monomer may be fed in a semibatch manner. This will
result in a high concentration of the less reactive monomer at the locus of poly-
merization, and the formation of polymer that is higher in the less reactive
monomer than would be the case for batch polymerization. Alternatively, the
reaction may be run monomer-starved. In this case, the monomers are fed
into the system in the desired ratio, but at a slow rate. The rate of polymeriza-
tion is then controlled by the rate of monomer feed. Since polymerization
occurs under monomer-starved conditions, the copolymer composition will be
that of the comonomer feed. Obviously, monomer-starved conditions will
result in low rates of polymerization. If either policy for semibatch monomer
feeding is to be used, issues of relative transport of the monomers must be
considered. Also, the decision must be made as to whether to feed in neat
monomer, a macroemulsion of monomer droplets, or a miniemulsion of
monomer droplets. In general, neat monomer feed will result in the least
nucleation, while miniemulsion feed will result in the most. Depending on the
goals of the polymerization, nucleation of new particles during the semibatch
feeding may be desirable or undesirable.
In 1993, Unzue and Asua [158] studied the semibatch miniemulsion ter-
polymerization of BA, MMA and VAc. The aim was to produce a miniemulsion
of 65% solids. It is well known that semibatch polymerization can be an effec-
tive method of making high solids latex. Part of the advantage of semibatch in
making high solids is the broad PSD brought on by nucleation of particles over
most of the reaction time. Miniemulsion can be effective in this regard, since
a miniemulsion feed is likely to produce additional polymer particles, as shown
by Tang et al. [131].
Wu and Schork [152] compared batch and semibatch and mini- and macro-
emulsion polymerization for three monomer systems, VAc/BA, VAc/dioctyl
maleate (DOM) and VAc/n-methylol acrylamide (NMA), with large differences
in reactivity ratios and water solubilities. HD was used as the costabilizer.
(It should be noted that DOM could function as a costabilizer itself, but for the
sake of consistency, HD was added to the DOM polymerizations.) KPS and the
Miniemulsion Polymerization 201

emulsifier SLS was used as the initiator. The semibatch miniemulsion poly-
merization involved two stages, a miniemulsion batch stage and semibatch
stage. The miniemulsion batch stage was performed as above. When the
monomer conversion was estimated to be about 80%, the feeding stage was
initiated by pumping monomer emulsion (miniemulsion or macroemulsion,
which was continuously formed while feeding) and initiator solution at set
flow rates simultaneously into the reactor. The composition of the monomer
feed was identical to the composition of the monomer in the batch stage.
Copolymer composition was determined by NMR, MWD by GPC and PSD by
dynamic light scattering.
Miniemulsion and macroemulsion copolymerizations of VAc and
comonomers with extremely different physical and kinetic properties (BA,
NMA and DOM), were investigated in batch and semibatch systems. The results
for polymer particle size and number, monomer conversion, composition and
molecular weights of copolymers indicated that there was an obvious diver-
gence between macroemulsion and miniemulsion copolymerization. In all
cases, the particle size was smaller and the particle number was higher in
macroemulsion copolymerization than in miniemulsion copolymerization. For
the systems VAc/BA and VAc/DOM, the particle number increased with in-
creasing conversion throughout the reaction for both batch macroemulsion
and miniemulsion runs. This was taken to indicate that the nucleation of new
particles takes place via homogeneous nucleation throughout these reactions.
For the batch runs, the rate of polymerization of the macroemulsion polymer-
ization runs was faster than that of the miniemulsion.
An investigation of the copolymer composition demonstrated the impor-
tant effect of monomer transport on the copolymerization. The droplets in
the macroemulsion act as monomer reservoirs. In this system, the effect of
monomer transport will be predominant when an extremely water-insoluble
comonomer, such as DOM, is used. In contrast with the macroemulsion system,
the miniemulsion system tends to follow the integrated Mayo Lewis equation
more closely, indicating less influence from mass transfer.
Likewise, for the semibatch operation, the influence of monomer was seen
in the differences between macro- and miniemulsion feeds. For extremely wa-
ter-insoluble monomers, the miniemulsion-feed mode lessens the departure of
the copolymer composition from the feed composition during semi-starved
semibatch polymerization. However, this is accomplished by simultaneously
broadening the PSD. Results from the GPC analysis indicated that the polymers
with lower molecular weight and broader distribution were formed in the
semibatch process, in contrast to the batch run.
Wu [159] also investigated the miniemulsion and macroemulsion copoly-
merization of VAc and vinyl versitate (VEOVA). At room temperature, the
water solubility of VAc is 2.58%wt, and vinyl versitate (also known as neo-
decanoate, one of the isomers in VEOVA-10) is 7.5¥10–4%wt. The extreme dif-
ference in water solubility between the two comonomers may impact on co-
polymer composition and the properties of the final polymer, due to the mass
202 F. J. Schork et al.

transfer of monomer.In this work, mini- and macroemulsion polymerizations

of VAc/vinyl versatate were designed to investigate the effects of monomer
transport and feeding strategies (for semibatch runs) on the reaction rate,
particle size distribution, molecular weight distribution, copolymer composi-
tion, and glass transition temperature (Tg) of the resultant polymer. Polymer-
izations were run at 55 °C. HD was used as the costabilizer, and SLS as the
surfactant. The initiator was KPS.VEOVA was added either as an emulsion with
the vinyl acetate, or as a neat liquid stream. In the polymerization of VAc
miniemulsion (or macroemulsion) plus neat vinyl versatate, the VAc mini-
emulsion (or macroemulsion) was pre-formed. The neat vinyl versatate was
then injected into the polymerization system at the same time as the injection
of initiator solution. For the semibatch processes, 20%wt of the polymer solids
was in the form of seed and the remaining 80% was fresh monomer emulsion.
The seed latex was prepared as a miniemulsion polymerization. The mini-
emulsion was made (sonicated) in-line, immediately prior to feeding into the
reactor. The first shot of initiator solution was introduced when the feed
of monomer emulsion started.A subsequent shot followed the removal of each
latex sample for analysis.
In the semibatch runs of VAc miniemulsion (or macroemulsion) plus neat
VEOVA, with simultaneous feeding of VAc miniemulsion (or macroemulsion),
the neat vinyl versatate was injected into the polymerization system two (for
the feedrate of 0.6 ml/min) or three (for 0.3 ml/min) times during each samp-
ling interval. Copolymer composition was determined by NMR, MWD by GPC,
Tg by DSC, and PSD by dynamic light scattering.
The effect of mass transfer of vinyl versatate on the mini/macroemulsion
polymerization of VAc/VEOVA in batch and semibatch systems was explored.
For the batch experiments, the addition of neat VEOVA formed poor disper-
sions of VEOVA, which resulted in smaller particles, lower polymerization rates
and different polymer composition tracks compared to normal mini/macro-
emulsion polymerization of VAc/VEOVA. The well-dispersed VEOVA seemed
to help the monomer-swollen particle to gain more radicals in the nucleation
period.
In the semibatch experiments, the particle size distributions of the final
latexes were affected by the residual surfactant in the seed latex, which tended
to facilitate homogeneous nucleation during the entire feed period. The
monomer feedrate determined the polymerization rate and had little effect on
copolymer composition. The polymer compositions for the runs with different
monomer feeding modes tended to be identical at very low feedrate.
For all runs, thermal analysis of the resulting polymers showed that only one
glass transition temperature could be found. This corresponded to the Tg of the
VAc/VEOVA copolymer. Lower glass transition temperatures were found for the
semibatch runs, perhaps due to slightly improved VEOVA incorporation.
In summary, the semibatch feeding of neat monomer or a macroemulsion
of monomer to a miniemulsion does not differ substantially from the equi-
valent semibatch feeding into a macroemulsion. The semibatch feeding of a
Miniemulsion Polymerization 203

miniemulsion tends to cause an increase in the particle number (due to partial

nucleation of the monomer droplets in the feed) and copolymer compositions
that more closely follow the Mayo Lewis Equation (due to the ability of the
miniemulsion droplets to, at least partially, retain their monomer, rather than
being depleted of monomer to feed the existing higher conversion polymer
particles).

4.2.2.3
Copolymerization in a Continuous Stirred Tank Reactor

Samer [137] studied miniemulsion copolymerization in a single CSTR. Two

separate feed streams, miniemulsion (or macroemulsion for comparative
studies) and initiator were fed at constant rates into the reactor. SLS was used
as the surfactant, HD as the costabilizer, and KPS was the initiator. In the
miniemulsion configuration (costabilizer included in recipe), the emulsion
stream was continuous. Constant volume was provided by an overflow outlet.
Salt tracer experiments were used to validate the ideal mixing model assumed
for a CSTR. Total monomer conversion was measured via in-line densitometry,
and copolymer composition via offline NMR.
Continuous macroemulsion copolymerization of MMA with 0.033–0.05mol%
2-ethylhexyl acrylate (EHA): On the basis of the batch copolymerization
experiments for this system, one would expect the average compositions to
follow the relation F2 Mini>F2 Macro. (The subscript 2 refers to the EHA.) However,
this behavior was not observed in a CSTR.Although the steady-state conversion
was significantly greater (as expected) for miniemulsions than for macroemul-
sions, the copolymer composition is nearly identical for both reactions. In this
case, it does not appear that droplet nucleation leads to an increase in the
amount of the extremely water-insoluble comonomer incorporated into the
copolymer, as observed in batch reactors. The experimental copolymer com-
positions were compared with the predicted copolymer compositions calcu-
lated from the Mayo Lewis equation where the reactivity ratios were obtained
from homogeneous copolymerization experiments and where the pseudo-
partition coefficient k2 was adjusted to fit the data. Surprisingly, both the macro-
emulsion and miniemulsion data showed good agreement with Samer’s modi-
fication of Schuller’s modified reactivity ratio model. This included both mini-
and macroemulsion copolymer compositions at different initial comonomer
compositions. Since the water solubility cannot change, it was suggested that
monomer transport, or at least the relative difference in monomer transport
between MMA and EHA, changes between batch and continuous miniemulsion
copolymerization reactions.
The difference in copolymer composition between miniemulsion and
macroemulsion copolymerization in a batch reactor was not observed in a
CSTR. In this case, the copolymer composition for the extremely water insoluble
comonomer in a miniemulsion recipe decreases from a batch reactor to a
CSTR. This difference can be attributed to the fact that monomer transport is
204 F. J. Schork et al.

enhanced in the steady-state CSTR where fresh monomer droplets are in

contact with “monomer-starved” particles. The comonomers cannot cross the
aqueous phase at similar rates because of their water-solubility differences,
which favors incorporation of the more water-soluble comonomer. All of the
droplets are nucleated at roughly the same time in a batch reactor, so little
monomer is available to quench monomer-starved particles.
However, this does not preclude miniemulsion copolymerization in a CSTR
for extremely water-insoluble comonomers. In spite of the fact that the copoly-
mer composition in the continuous miniemulsion is less than that predicted
using the homogeneous copolymerization reactivity ratios, the miniemulsion
copolymer might be more uniform than the macroemulsion copolymer, where
the possibility of significant droplet nucleation could lead to two separate homo-
polymers or, at the very best, copolymers of various composition. Therefore, it
is very important to use CSTR data to scale up a continuous miniemulsion
copolymerization product to take into account the different particle growth
kinetics for batch and continuous reactors.

4.2.3
Interfacial Polymerization

Interest in the design and controlled fabrication of composite nanoparticles con-

sisting of hydrophobic polymer cores coated with hydrophilic polymer shells con-
tinues to increase. These particles have potential technological applications in di-
agnostic testing, bioseparations, controlled release of drugs, gene therapy,
catalysis, and water-borne coatings and adhesives [160–170]. Emulsion copoly-
merization of hydrophobic and hydrophilic monomers seems to be a straight-
forward approach to fabricating such nanoparticles. However, this kind of
emulsion copolymerization presents a big challenge, because the hydrophilic
monomer resides almost exclusively in the aqueous phase while the hydrophobic
monomer resides almost exclusively in the organic phase. AIndeed, most of the
studies made so far have been limited to a very low hydrophilic monomer level.
In all cases, the incorporation of water-soluble monomer was very limited, re-
gardless of the initial amount of water-soluble monomer loaded into the system
[171]. It was proposed that interfacial graft-polymerization of hydrophilic
monomer onto hydrophobic polymer should be possible in an emulsion process
by selecting an appropriate initiator system [172]. In the work of Luo et al. [170],
the idea of interfacial polymerization was used to develop a novel repulpable pres-
sure-sensitive adhesive. In order to maximize the incorporation of cationic
monomer (hydrophilic monomer), an interfacial redox initiator system was used.
In this initiator system, oil-soluble cumene hydroperoxide (CHP) was used as the
oxidizer while hydrophilic tetraethylenepentamine was employed as the reducer.
It was hoped that by using CHP/TEPA, the hydrophobic CHP would meet the hy-
drophilic TEPA at the particle-water interface, where hydrophobic and hydrophilic
monomer are both present. In the meanwhile, non-ionic surfactant (Triton X-405)
was used to facilitate the adsorption of cationiccomonomer at the interface.
Miniemulsion Polymerization 205

In order to gain evidence for interfacial initiation, the redox initiator system
was compared with a water-soluble initiator (VA-044) in terms of the emulsion
polymerization behavior of butyl acrylate (BA)/[2-(methacryloyoxy)ethyl]tri-
methyl ammonium chloride (MAETAC). It was found that for the water-solu-
ble initiator system, only homopoly(MAETAC) was formed and BA did not
polymerize at all. In the case of VA-044, it was suggested that it may be difficult
for polymeric free radicals in the aqueous phase to penetrate the viscous
surfactant layer to initiate the polymerization of the BA monomer. On the other
hand, it has also been found that BA could be rapidly polymerized under the
same conditions if VA-044 is replaced with CHP/TEPA, indicating that radicals
are formed in the interface, where they do not need to penetrate through vis-
cous surfactant layer.
The polymerization kinetics of BA/MAETAC macroemulsion and mini-
emulsion copolymerization was investigated with the interfacial redox initiator
system. It was found that adding MAETAC had a complex effect on the poly-
merization kinetics of BA, as shown in Figs. 17 and 18 [170].
In comparison with the homopolymerization of BA, adding 5%wt MAETAC
greatly increases the BA polymerization rate for both macroemulsion and
miniemulsion polymerization. However, when the MAETAC level is increased
further, the polymerization rate begins to decrease. At higher MAETAC levels,
the effect of MAETAC is different for macroemulsion and miniemulsion
polymerizations. For macroemulsion polymerization at high MAETAC level,
there seems to be an induction period before polymerization begins.At higher
MAETAC levels, the length of the induction period increases. For miniemulsion
polymerization, there is no induction period; instead, the BA polymerization
rate decreases to zero at less than full conversion. BA homopolymerization with
CHP/TEPA levels off at ~30% conversion. At higher MAETAC levels, the BA
conversions levels off at substantially higher conversion. The leveling-off has

Fig. 17 BA monomer conversion for macroemulsion copolymerization with varying levels

of MAETAC (from [170])
206 F. J. Schork et al.

Fig. 18 BA monomer conversion for miniemulsion copolymerization with varying levels

of MAETAC (from [170])

been ascribed to the depletion of TEPA. Comparing Fig. 17 with Fig. 18, it is
clear that the influence of MAETAC levels on macroemulsion polymerization
is much greater than on miniemulsion polymerization, especially in the early
stage of polymerization. With increasing MAETAC levels, the polymerization
rate of BA, especially in the early stage of polymerization, rapidly drops off
in the macroemulsion polymerization. It is suggested that adding MAETAC
would interfere with the nucleation of macroemulsion polymerization. At low
MAETAC levels, introducing MAETAC leads to homogeneous nucleation, so
the polymerization rate increases. However, at high MAETAC level, adding
MAETAC did not result in homogeneous nucleation but suppressed micellar
nucleation. It has been suggested that polymer formed in the aqueous phase
is too hydrophilic to lead to homogeneous nucleation at high MAETAC. The
suppression of micelle nucleation can be illustrated by Fig. 19. The hydrophobic
molecules such as BA and CHP will be present in the micelles at the beginning
of the polymerization. When TEPA was added to the reactor, many of the free
radicals will be formed by the following mechanism in the surface of a micelle,
where CHP and TEPA meet each other [173].
Because of their different hydrophilicities, the two free radicals formed at the
same time can separate from each other quickly which can eliminate the cage
effect. In a micelle, the local BA concentration may be quite high. Once a micelle
is initiated, a number of BA molecules may be added quickly. As a result, some
short BA blocks would be incorporated into a poly(MAETAC) chain to form
something like multi-block copoly(MAETAC-BA), as shown in Fig. 19 [170].
Surfactant should stabilize the BA blocks so that the block copolymer remains
in the aqueous phase.
The BA blocks in the copolymer, which is surrounded by surfactant, swell
with BA monomer, which polymerizes there until particles form. Therefore,
micellar nucleation is diminished or even eliminated. As a result, micellar
nucleation is retarded, and Np decreases. It is clear that the nucleation is very
Miniemulsion Polymerization 207

Fig. 19 Schematic of the formation of multi-block poly(MAETAC-BA) (from [170])

sensitive to the level of hydrophilic monomer for macroemulsion polymeriza-

tion. As a result, the polymerization rate and particle size change dramatically
with changes of hydrophilic monomer level. For miniemulsion polymerization,
the monomer is dispersed into droplets of 50–500 nm prior to polymerization
and no micelles exist. Therefore, TEPA comes into contact with CHP at the
droplet-water interface. MAETAC and BA copolymerize at the interface. The
number of BA molecules in a monomer droplet is far larger than that in a
micelle, so the resultant copolymer has a much higher composition of BA and
is anchored at the interface. Therefore, droplet nucleation during miniemulsion
polymerization is hardly affected, although it was found that when the hydro-
philic monomer was low, homogeneous nucleation could occur in the mini-
emulsion polymerization, as shown in Fig. 20 [170]. As a result, there is little
initial dependence of the polymerization rate on the hydrophilic monomer level

Fig. 20 Particle number development for miniemulsion copolymerization with varying

levels of MAETAC (from [170])
208 F. J. Schork et al.

during miniemulsion polymerization, in sharp contrast to macroemulsion poly-

merization, as shown in Figs. 17 and 18. Clearly, nucleation during miniemulsion
polymerization is highly robust to the level of hydrophilic comonomer.

4.3
Multiphase Particles

4.3.1
Hybrid Miniemulsion Polymerization

It would often be desirable to create submicron particles containing two or

more polymers. These could be in the form of a blend or in the form of a graft
copolymer. In a simple blend, the two polymers may or may not be compatible.
If they are compatible, the particle will be homogenous. If the polymers are not
compatible, then microphase separation is likely. However, if the phase sepa-
ration occurs in submicron particles, the phase domains will be small, and
decent dispersion of the two polymers will occur. Homogenous, grafted, or
phase-separated morphologies might conceivably be of practical value.
One method of creating such polymer blends or grafts in submicron partic-
les is through hybrid miniemulsion polymerization. In this technology, a pre-
formed polymer (or oligomer) is dissolved into a monomer (or monomer
solution). A miniemulsion is then created from the monomer-polymer solu-
tion, and this miniemulsion is polymerized via standard techniques. Care must
be taken in creating the miniemulsion, because the polymer solution will likely
have a high viscosity (higher than for a simple monomer miniemulsion) and
so the droplet break-up by the shear device may be more difficult. On the other
hand, the preformed polymer may act as the costabilizer, eliminating the need
for HD or other costabilizers. For some chemistries, grafting will take place
during the polymerization of the monomer, and for some polymers phase
separation will occur. In any case, the product is a submicron dispersion of one
polymer in another, and may well have practical value. Since most of the
oligomers or prepolymers are polymeric and probably highly water-insoluble,
macroemulsion polymerization will not result in a graft copolymer or an
intimate blend. Since the prepolymer is not transported from the monomer
droplets to the polymer particles, the prepolymer does not reach the locus of
polymerization. Macroscopic phase separation and colloidal instability often
result.
Details of the chemistry and process (and the product) are very specific to
the choice of prepolymer and monomer; for this reason, each system will be
discussed separately here.
Miniemulsion Polymerization 209

4.3.1.1
Alkyds

Water-based coatings have become more widely used over the past few decades
because they are environmentally friendly, offer easy clean-up, and their pro-
perties and application performance characteristics have improved. Solvent-
based systems such as alkyd resins and polyurethanes have remained impor-
tant for some applications because of superior properties, such as gloss and
hardness. This is due to the curing mechanism of oil-based coatings in which
the oils react with atmospheric oxygen to form very hard crosslinked materials.
This mechanism is generally lacking in water-based coatings, which tend to
be soft and pliable, due to the fact that the coatings are made soft to allow film
formation, and since there is no curing chemistry available, remain soft on
drying. Several researchers have focused on the use of the hybrid miniemulsion
polymerization of acrylic monomers in the presence of alkyd and polyurethane
resins to develop alternative coatings which have the advantages of water-based
systems (like low VOC) with the drying (air cure) properties of solvent-based
systems.Alkyd/acrylate coatings are targeted as replacements for solvent-based
architectural coatings, and oil-modified polyurethane (OMPU)/acrylate coat-
ings may provide a low VOC alternative to solvent-based clear coats. Since
U.S. architectural coating sales in 1995 amounted to 625 million gallons [174],
a conservative estimate of the VOC reduction if all of these coatings were water-
based is approximately 500 million pounds of solvent that would not be re-
leased into the air.
Nabuurs and German [175] developed an alkyd-acrylic hybrid system via
emulsion polymerization. They were able to produce a stable product using
MMA as the acrylic.When the alkyds were functionalized by sulfonation, there
was no evidence of heterogeneity in the particles.When unfunctionalized alkyd
was used, the MMA appeared as microdomains within the particles. Grafting
of acrylic to alkyd was low. The presence of the alkyd led to low rates of poly-
merization and limited conversion that were both attributed to retardation
through radical delocalization following radical transfer to the unsaturated
groups in the fatty acids of the alkyd. Although this work was not reported to
be miniemulsion polymerization, the emulsions were subjected to high shear
prior to polymerization. No costabilizer was added, but the alkyd presumably
functioned as such. Since the alkyd-acrylic droplet size was probably quite
small, droplet nucleation was probably the dominant nucleation mechanism.
This would explain the good colloidal stability of the resulting system, and the
fact that both alkyd and acrylic domains were found in the same particles.
Wang et al. [98] carried out macroemulsion and miniemulsion polymeriza-
tion of acrylic monomers in the presence of alkyd resins. Miniemulsion and
macroemulsion polymers were produced using a commercial medium soya-
linseed alkyd and a mix of acrylic monomers consisting of 50% BA, 49% MMA,
and 1% acrylic acid (AA). PMMA polymer with a weight average molecular
weight of 100,000 was used as the costabilizer. Alkyd levels were 5, 30, 60 or
210 F. J. Schork et al.

100% based on total acrylic monomer. SLS was used as the surfactant, and
sodium persulfate (SPS) was used as the initiator. The miniemulsions were pre-
pared by dispersing the desired amount of monomer-PMMA-alkyd solution in
the aqueous SLS solution by mixing with stirring at room temperature. The re-
sulting emulsion was sheared further by sonication. Polymerization was carried
out at 60–80 °C. The reaction was followed by gravimetric conversion analysis.
Macroemulsion polymerizations were carried out in the same manner except
that no sonication process was used and the PMMA costabilizer was not em-
ployed. Droplet and particles sizes were determined by dynamic light scatter-
ing, and double bond content of the alkyd by NMR. Grafting was determined
by extraction, and degree of crosslinking by exhaustive extraction.
The monomer miniemulsions with PMMA as costabilizer were prepared
with different amounts of alkyd resin. The PMMA costabilizer was effective in
the preparation of stable miniemulsions, especially in conjunction with the
alkyd. The size of monomer droplets was below 300 nm. After five days, the
unpolymerized macroemulsions with alkyd separated into three phases,
monomer on the top, clear water in the middle, and alkyd resin on the bottom.
The miniemulsion without alkyd showed two phases, monomer and water.
All miniemulsions with alkyd resin appear to remain uniform. Very stable
miniemulsions were obtained when the alkyd content was higher than 30%.
The shelf life of macroemulsions was only 2–8 minutes.
The polymerization rate in the presence of alkyd was slower than that with-
out alkyd. Doubling the initiator and emulsifier concentration increased the
reaction rate, but not to the level achieved with the miniemulsion polymeriza-
tion without alkyd. This retardation (as reported also by Nabuurs) increased
with increasing alkyd level. The latexes obtained from the miniemulsion poly-
merization of the alkyd-acrylate mixtures were uniform emulsions, and no
coagulation occurred during polymerization. Macroemulsion polymerization
with alkyd resulted in colloidal instability, probably due the inability of the
alkyd to reach the locus of polymerization.
NMR analysis indicated that approximately 30% of the alkyd double bonds
had been consumed in the polymerization process. This is important in that it
indicates that a substantial fraction of the double bonds remain for oxidative
crosslinking while a coating made from this material is dried. Selective ex-
traction indicated that approximately 60% of the acrylate was grafted to alkyd.
Exhaustive extraction indicated less than 5% crosslinked material. The poly-
merized latex formed good films with acceptable hardness.
On the whole, the miniemulsion polymerization process proved to be effec-
tive for incorporating an alkyd resin into acrylic coated copolymers. The reac-
tion produced stable, small particle size latexes that contain graft copolymer of
the acrylic and alkyd components. Attempts at macroemulsion hybrid poly-
merization were unsuccessful.
Van Hamersveld et al. [176, 177] carried out hybrid miniemulsion poly-
merization of MMA, using HD as the costabilizer. In an attempt to encourage
grafting, oxidized triglycerides (such as sunflower oil) were used as initiators.
Miniemulsion Polymerization 211

This produced homogenous particles (presumably due to higher levels of graft-

ing), whereas the use of conventional initiators resulted in particle inhomo-
geneity due to phase separation of the alkyd and acrylic.
Wu et al. [99, 174] further investigated the hybrid miniemulsion polymer-
ization of alkyd-acrylic systems, using the same monomer mix as Wang [98].
Retardation by the alkyd was reported; high polymerization temperature and
mixed (oil and water-soluble) initiators were used to improve monomer con-
version. The polymers obtained had a very wide molecular weight distribution,
with polydispersities of more than 19 and a number average molecular weight
slightly larger than alkyd. This indicates that a fraction of the alkyd remains in
the ungrafted form. On the other hand, extraction results indicate that a large
fraction of polyacrylate chains contain at least some grafted alkyd. Approx-
imately 20% of the double bonds in the alkyd are consumed in grafting reac-
tions. Two glass transition temperatures were observed, indicating the presence
of at least two forms of polymer. The two glass transition temperatures cor-
respond to those of poly(acrylate-graft-alkyd) and polyacrylate respectively.
The proportion of the two kinds of polymers in the samples was determined
by extraction, and this indicated that poly(acrylate-graft-alkyd) is the pre-
dominant form.
Tsavalas et al. [178] studied the limiting conversion phenomenon brought on
by alkyd retardation. He concluded that retardive chain transfer to the alkyd
double bonds was not adequate, especially for systems that graft through
addition through double bonds (acrylates, but not methacrylates) as well as
through chain transfer via hydrogen abstraction at alkyd double bonds (acryl-
ates and methacrylates). Without transfer, no radicals of low activity would be
created, and so the dramatic reduction in polymerization rate would have to be
attributed to another cause. He concluded that there are two mechanisms for
limiting conversion, one kinetic, and one physical. MMA, which has a high Tg,
and grafts primarily through chain transfer, was found to produce a plateau in
the kinetic profile of monomer conversion when the monomer glass transition
temperature was near the reaction temperature. However, simple calculations
suggest that transfer alone could not produce such a dramatic change in
kinetics. The physical mechanism is thought to play a significant role, partic-
ularly since MMA/alkyd particles display core shell morphology, with the
possibility of residual monomer trapped in the alkyd-rich core.
Butyl acrylate has a low Tg and no steric hindrance to prevent direct addi-
tion to alkyd double bonds. In this case, the limiting conversion is not absolute,
but more of a very significant reduction in the rate of polymerization at high
conversion. Again, simple calculations suggest that retardive chain transfer
alone could not produce such a dramatic change in kinetics. Instead, the
physical mechanism was found to be significant as well. PBA and alkyd both
have glass transitions well below the reaction temperature, so a barrier to entry
is never formed in that type of system. However, a viscous environment forms
after appreciable conversion that slows the mobility of both monomer and
initiator. BA monomer is thought to be dissolved in small alkyd domains dis-
212 F. J. Schork et al.

tributed throughout a continuous BA particle phase. These islands eventually

act as reservoirs diffusing monomer to the polymerization of BA in the con-
tinuous particle phase. The viscosity and diffusion rate are then what retard the
rate in this type of hybrid system. This is evidenced by the fact that although
there is a point of dramatic rate change, afterwards the new rate continues
to complete conversion. That point of rate change likely corresponds to a
morphology transformation to those alkyd island domains. Butyl methacrylate
exhibits a reduced rate of polymerization like that of BA, rather than a true
limiting conversion like that of MMA. The grafting kinetics of BMA are similar
to those of MMA, but its limiting conversion behavior is similar to that of
BA. Therefore it was concluded that the physical mechanism (based on Tg) is
dominant over the kinetic mechanism (retardive chain transfer) for all three
monomers.
Tsavalas [179] also studied the grafting of alkyd-acrylic hybrid systems. He
attributed differences in levels of grafting for different monomers to differences
in grafting mechanism. Grafting was observed between methacrylates and
typical alkyds, but steric hindrance at the methacrylate reactive center directs
addition to an alkyd double bond. This method was shown to give optimal
grafting efficiency. Instead, methacrylates tend towards allylic hydrogen ab-
straction, a process that creates a relatively stable and unreactive radical on the
resin along with terminating the abstracting methacrylate chain. These effects
degrade both the grafting efficiency and the rate of polymerization. Acrylate
monomers were found to produce high levels of grafting. Direct addition to
resin double bonds is facilitated and virtually complete grafting of the com-
ponent is observed. This was attributed to the lack of steric hindrance of the
acrylate reactive center. The double bond content of the resin was shown to be
important to grafting. Double bonds are needed, even in systems where
abstraction is the dominant route of attack, since hydrogens allylic to them are
good leaving groups. The double bond density correlates directly with the
concentration of possible grafting sites and was shown to lead to higher levels
of grafting. Tsavalas showed that the choice of monomer(s) is the most im-
portant variable in determining the level of grafting. Chain transfer dominates
the interaction of methacrylate with resin, and so there is less opportunity for
grafting. Conversely, the interaction of acrylate with resin is dominated by
direct addition to a resin double bond, a highly efficient mode of grafting. In a
third paper, Tsavalas [180] confirmed the differences in particle morphologies
between acrylates and the methacrylates described above.
Shoaf and Stockl [181] optimized the formulations for hybrid miniemulsion
polymerization. By adjusting the Tg of the polymer phase (acrylate-styrene),
they were able to create latex that gave good film formation with little coalesc-
ing aid, and hence, very low VOC. By adding a latent oxidative functional
monomer, they were able to get very hard film. Latexes containing up to appro-
ximately 50% alkyd (based on total solids) were produced. The coatings exhi-
bited high gloss, which is sometimes unattainable with water-based systems.
No information on the heterogeneity of the particles was provided.
Miniemulsion Polymerization 213

4.3.1.2
Polyester

Tsavalas et al. [100] carried out hybrid miniemulsion polymerization with a

three component acrylic system of methyl methacrylate, butyl acrylate, and
acrylic acid in the presence of a Bayer Roskydal TPLS2190 unsaturated polyes-
ter resin. Latexes were obtained in which the polyester resin was grafted to the
acrylic polymer, forming a water-based crosslinkable coating. Both emulsions
and latexes were shelf stable for over six months, shear stable, and resistant
to at least one freeze/thaw cycle. Resin to monomer ratios as high as 1:1
(wt:wt) and total emulsion solids as high as 45% were studied. The sizes of the
monomer droplets and latex particles were similar, suggesting predominant
droplet nucleation.A high level of crosslinking (>70%) during polymerization
was observed in this particular hybrid system in contrast to those involving
alkyd, as reported above. Homogeneous and hard films were achieved with
exceptional adhesion. Electron microscopy showed the hybrid particle mor-
phology to have internal domains of polyester resin in an acrylic matrix.
Kinetic studies showed that as resin content increased in comparison to mono-
mer content, the polymerization rate decreased, suggesting retardive chain
transfer as found with the alkyds.

4.3.1.3
Polyurethane

Oil-modified polyurethanes (OMPU) are, in terms of volume produced and

sold, the most important polyurethane coatings, with superior properties such
as gloss, chemical resistance and film formation. Most urethane coatings are
solvent-based, and solvent-based coatings are less than desirable due to the
environmental impact of their high VOC. To meet the increasing concern for
health, safety and the environment, there has been a strong preference in recent
years for water-borne coatings. Dong et al. [101] carried out hybrid mini-
emulsion polymerization with acrylic monomers (methyl methacrylate, butyl
acrylate and acrylic acid) in the presence of oil-modified polyurethane resin.
The OMPU served as the costabilizer. Latexes with different ratios of resin
to acrylic monomer were synthesized. The monomer emulsions prepared for
hybrid miniemulsion polymerization showed excellent shelf-life stability (more
than five months) and the polymerization was run free of coagulation. Solvent
extraction indicated that the grafting efficiency of polyacrylates was greater
than 29% for all of the samples produced. The 13C solution NMR spectrum
showed that a substantial fraction of the original carbon double bonds (>61%)
in oil-modified polyurethane remained after polymerization for film curing.
Films obtained from the latexes presented good adhesion properties and
fair hardness properties.
Li et al. [182] used hybrid miniemulsion polymerization to prepare urethane/
BMA latexes with particle sizes of about 50 nm. Hexadecane was used as the
214 F. J. Schork et al.

costabilizer. In this case, the presence of the prepolymer (polyurethane, MW

<10,000) resulted in an increase, rather than a decrease, in the rate of polymer-
ization due to the fact that the presence of the polyurethane prepolymer resulted
in smaller initial droplet sizes. (Presumably there was no retardive chain trans-
fer, since there was no significant unsaturation in the polyurethane.) Free iso-
cyanate groups remaining on the polyurethane reacted with the aqueous phase,
causing an increase in particle size over several days due to flocculation.
Wang et al. [183] carried out hybrid miniemulsion polymerization of acryl-
ates in the presence of polyurethane. The polyurethane was used as the cos-
tabilizer, and SLS as the surfactant. When MMA was used as the monomer,
some homogenous nucleation was observed. This is in agreement with Tsavalas
[179] who reported evidence of homogenous nucleation in the hybrid mini-
emulsion of MMA in the presence of alkyd.
Barrere and Landfester [184] prepared a hybrid miniemulsion in which
isophorone diisocyanate was condensation polymerized with dodecanediol to
form polyurethane at the same time that the polystyrene or polyBA was free
radical polymerized. Unlike previous work, the polyurethane was not prepared
in organic solvent in advance. Therefore, in this one-pot synthesis, polyaddition
and free radical polymerization both take place in the same particle. HD was
used as the costabilizer. After miniemulsification, the polycondensation was
allowed to take place, and then a free radical initiator was added to polymerize
the styrenic or acrylic monomer. Molecular weight distributions were bimodal;
the PU had a substantially lower molecular weight than the polyacrylate.
Neither intra- nor interparticle phase separation could be detected by TEM; the
particles appeared to be homogeneous. No measurements of grafting were
made, but since there was no unsaturation in the PU, none was expected.

4.3.1.4
Other Hybrids

El-Aasser and coworkers [185, 186] have carried out miniemulsion polymeriza-
tion of styrene monomer in the presence of Kraton D1102 thermoplastic
elastomer, to form hybrid composite latexes approximately 100–150 nm in size.
A costabilizer other than the rubber was used. The miniemulsification was
carried out via homogenization, and resulted in a very broad droplet size dis-
tribution. This resulted in the formation of inhomogeneous hybrid composite
particles due to monomer diffusion during polymerization. When an oil-
soluble initiator was used, an induction period was observed, resulting from the
presence of radical scavengers such as antioxidants and UV stabilizers within
the Kraton-styrene droplets. This also lead to inhomogeneous particles, but was
eliminated with a water-soluble initiator. TEM showed domains of polystyrene
within the rubber particles. Some evidence of homogeneous (or micellar)
nucleation was found.
Kawahara et al. [187] prepared acrylic/epoxy composite latexes via hybrid
miniemulsion polymerization. Landfester et al. [188] have incorporated PMMA
Miniemulsion Polymerization 215

macromer as a phase capatibilizing agent in the core shell polymerization of

styrene and MMA. Roberts et al. [189, 190] copolymerized vinyltriethoxysilane
with various acrylates. Phase separation resulted. Hydrolysis of the triethoxy-
silane and subsquent condensation resulted in a crosslinked siloxane phase,
separate from the polyacrylate.

4.3.1.5
Artificial Miniemulsions

El-Aasser and coworkers [191–193] have used the miniemulsification process to

create water dispersions of preformed polymers. In this technique, a polymer is
dissolved in a volatile organic solvent. The polymer-solvent solution is then
miniemulsified and the solvent is evaporated off. This leaves a submicron water
dispersion of the polymer. This technology has some distinct advantages. First,
it can be applied to any polymer that is insoluble in water, but soluble in an
organic solvent. Second, the polymerization need not be via free radical poly-
merization, since it is accomplished previous to miniemulsification. For this rea-
son, polymerizations that are not water-tolerant may be used to form the poly-
mer. This technique has been applied to a number of commercial applications.

4.3.2
Nanoencapsulation

Nanoencapsulation, or the encapsulation of solids within submicron particles,

has been the subject of considerable work in recent years. Most have attempted
nanoencapsulation via emulsion and miniemulsion polymerization techni-
ques. Lee and coworkers [194, 195] have carried out emulsion polymerization
in the presence of layered silicate. The term emulsion polymerization is un-
fortunate, since the polymerization takes place between the layers of silicate
rather than in latex particles, forming an intercalated silica nanocomposite.
Garcés et al. [196] and many others have reported marked improvement in
mechanical properties if the inorganic reinforcing material (such as silica) is
dispersed on the nano-scale. None of these applications are truly nanoen-
capsulation, since the polymer resides in the interstices of large inorganic
particles, rather than having the inorganic fully encapsulated within a polymer
particle.
Van Herk and German [197] have surveyed the true nano (micro) encapsu-
lation of inorganic and organic pigments and fillers via emulsion polymeriza-
tion. In this technique, small inorganic particles are used as nuclei for the for-
mation of polymer particles.A number of materials have been encapsulated in
this way, including clays, limestone, alumina, silica, carbon black, and magnetic
materials. The practical challenges associated with these encapsulations include
stabilizing the inorganic colloids, and controlling particle nucleation.
A more direct and reproducible route to nanoencapsulation is that of
miniemulsion polymerization. If an organic or inorganic solid is contained
216 F. J. Schork et al.

within the monomer droplets, nanoencapsulation can be accomplished; unlike

in macroemulsion polymerization, no transport of solids from the monomer
droplets to the locus of polymerization is required. Erdem et al. [198–200] have
published a series of papers on the nanoencapsulation of titanium dioxide
(TiO2) particles in styrene. The challenges in this work involved getting the
hydrophilic TiO2 particles to reside in the hydrophobic environment of the
interior of the miniemulsion monomer droplet, rather than in the hydrophilic
environment of the continuous aqueous phase. This was accomplished through
the use of specific surfactant systems. Landfester et al. [3, 201–204] studied
the nanoencapsulation of solid materials via miniemulsion polymerization.
They successfully encapsulated a wide variety of materials. For hydrophobic
solids (like carbon black) that will easily transport into the monomer drop-
lets, the technique is straightforward. For hydrophilic solids (like TiO2) that
prefer the aqueous phase, the authors used a combination of oil-in-water and
water-in-oil surfactants to move the solids into the miniemulsion monomer
droplets.
While the literature has demonstrated the viability of nanoencapsulation via
miniemulsion technology, a great many issues remain. These include the use of
surfactant systems to stabilize the solid colloidal particles and bring them
into the monomer droplets, uniformity of encapsulation (a uniform [small]
number of solid particles per polymer particle for a minimum number of
polymer particles, not including the encapsulated solid), and complete cover-
age of the solids by the resultant polymer coating.

4.4
Controlled Free Radical Polymerization

4.4.1
Nitroxide-Mediated Polymerization

Nitroxide-Mediated Controlled Radical Polymerization (NMCRP) was first dis-

covered by Solomon et al., who patented their discovery in 1985 [205]. This
opened up new pathways in the field of free-radical polymerization. Polymer
architectures, which were the domain of the anionic polymer chemist, became
accessible to the free-radical polymer chemist. However, it was not until the
work of Georges et al. [206] was published in 1993, that the world of polymer
chemistry became aware of the possibilities of this new class of free-radical
polymerization. This was the beginning of what is today one of the leading
topics in free-radical polymer chemistry: Controlled or “Living” Free Radical
Polymerization. This initiated the search for new Controlled or “Living” Free
Radical Polymerization techniques, and soon afterwards other methods (which
will be discussed later) were developed.
Miniemulsion Polymerization 217

4.4.1.1
Mechanism

In processes based on reversible termination, like NMCRP and ATRP (Sect.

4.4.2), a species is added which minimizes bimolecular termination by rever-
sible coupling. In NMCRP this species is a nitroxide. The mechanism of
nitroxide-mediated CRP is based on the reversible activation of dormant
polymer chains (Pn-T) as shown in Scheme 1. This additional reaction step in
the free-radical polymerization provides the living character and controls the
molecular weight distribution.

Scheme 1 Reversible activation of dormant polymer chains

When a dormant species or alkoxyamine dissociates homolytically, a carbon-

centered radical and a stable nitroxide radical are formed (Scheme 2). This is
a reversible process and the reversible reaction is very fast – close to diffusion-
controlled rates. With increasing temperature, the dissociation rate will in-
crease, which will increase the concentration of the polymeric radicals (Pn•).
These will have a chance to add to monomer before being trapped again, which
allows growth of the polymer chains. The nitroxide is an ideal candidate for this
process since it only reacts with carbon-centered radicals, is stable and does
not dimerize, and in general couples nonspecifically with all types of carbon-
centered radicals (at close to diffusion-controlled rates).

Scheme 2 Dissociation of a typical alkoxyamine into a carbon-centered radical (ethylben-

zene radical) and a nitroxide (TEMPO)

Polymerization can be started using an alkoxyamine as initiator such that,

ideally, no reactions other than the reversible activation of dormant species and
the addition of monomer to carbon-centered radicals take place. The alkoxy-
amine consists of a small radical species, capable of reacting with monomer,
trapped by a nitroxide. Upon decomposition of the alkoxyamine in the presence
of monomer, polymeric dormant species will form and grow in chain length
over time. Otherwise, polymerization can be started using a conventional free-
radical initiator and a nitroxide. The alkoxyamine will then be formed in situ
when an initiator molecule decomposes, and, after adding a monomer unit or
two, is trapped by a nitroxide.
218 F. J. Schork et al.

Since the nitroxide and the carbon-centered radical diffuse away from each
other, termination by combination or disproportionation of two carbon-cen-
tered radicals cannot be excluded. This will lead to the formation of “dead”
polymer chains and an excess of free nitroxide. The build-up of free nitroxide
is referred to as the Persistent Radical Effect [207] and slows down the poly-
merization, since it will favor trapping (radical-radical coupling) over propa-
gation. Besides termination, other side reactions play an important role in
nitroxide-mediated CRP. One of the important side reactions is the decompo-
sition of dormant chains [208], yielding polymer chains with an unsaturated
end-group and a hydroxyamine, TH (Scheme 3, reaction 6).Another side reac-
tion is thermal self-initiation [209], which is observed in styrene polymeriza-
tions at high temperatures. Here two styrene monomers can form a dimer,
which, after reaction with another styrene monomer, results in the formation
of two radicals (Scheme 3, reaction 7). This additional radical flux can compen-
sate for the loss of radicals due to irreversible termination and allows the poly-

Scheme 3 Mechanism of nitroxide-mediated CRP. R-T represents an alkoxyamine, T· rep-

resents a nitroxide
Miniemulsion Polymerization 219

merization to proceed successfully, providing that the number of initiating

radicals is small compared to the number of nitroxide-trapped polymer chains
[210]. Systems that do not show thermal self-initiation can also be controlled
by using an additional initiator, which will provide the additional radical flux
[210]. In addition, the dimer formed (Scheme 3, reaction 7) can react with a
nitroxide molecule to provide the dimer radical and a hydroxyamine. The most
important reactions in nitroxide-mediated CRP are shown in Scheme 3. An
excellent overview of the kinetics and mechanism, supported by simulations,
is given by Fukuda [211].

4.4.1.2
Effects Of Segregation And Heterogeneity

Similar to conventional free radical polymerization, heterogeneity and segre-

gation effects make the kinetics of NMCRP more complex when applied in
miniemulsions. This issue has also been discussed in a review article by Qiu
et al. [212], which covers controlled free-radical polymerization in heteroge-
neous media up to 2001 (and so also covers miniemulsion polymerization).
Butté et al. [213, 214] and Charleux [215] both discussed compartmentalization
effects. Butté et al. came to the conclusion that the effects of segregation in
NMCRP miniemulsions are very small, while Charleux predicted an increased
polymerization rate for small (50–100 nm) particles. Both groups come to
different conclusions, because Butté et al. did not account for the possibility
that a nitroxide molecule exits a particle, while Charleux did. In summary,
the segregation effect in NMCRP miniemulsions is not large unless the par-
ticles are small or a lot of the nitroxide partitions to the aqueous phase. This
is nicely illustrated by work of Pan et al. [216], who performed NMCRP mini-
emulsions with varying surfactant concentrations. Although the surfactant
concentration had a large effect on particle size, it hardly affected the poly-
merization rate.
Since both NMCRP and ATRP (Sect. 4.4.2) are based on reversible termina-
tion, the effects observed for these will be similar and are further discussed
later.

4.4.1.3
Results

A review article by Qiu et al. [212] and references herein [217–226] covers
NMCRP in miniemulsions up to 2001. Cunningham wrote a related review
in 2002, also covering controlled radical polymerization in dispersed phase sys-
tems [227]. Here, the main results reported in the Qiu review will be sum-
marized, and new developments in the field since then will be reviewed.
For several reasons, miniemulsion polymerization is the preferred technique
for NMCRP in aqueous dispersed systems. Since NMCRP in general requires
high reaction temperatures (above 100 °C) for the thermal polymerization of
220 F. J. Schork et al.

styrene, the presence of a monomer phase is not desired. This would lead to
colloidal instability. Furthermore, the presence of a large monomer reservoir
makes the nitroxide partition out of the loci of polymerization, which will
reduce control of the polymerization. Another reason for using miniemulsion
polymerization is the fact that pre-formed alkoxyamines or nitroxide-termi-
nated oligomers are often used. Generally, these are too water-insoluble to be
transported through the aqueous phase, which excludes the use of conventional
or seeded emulsion polymerization. Finally, the poorly understood and
complex particle nucleation step is avoided in miniemulsion polymerization,
which allows the use of oil-soluble initiators and makes results easier to
understand.
The first results for NMCRP in miniemulsion were reported by Propdan
et al. [218, 221] and Macleod et al. [219]. Both groups performed miniemulsion
polymerizations of styrene at high temperatures (above the boiling point of
water) using high pressure reactors. Propdan et al. used oil-soluble benzoyl
peroxide (BPO) free-radical initiator and TEMPO, 1 (see Scheme 4), in a Dow-
fax 8390 surfactant system at 125 °C, while Macleod et al. used water-soluble
potassium persulfate (KPS) initiator and sodium dodecylbenzene sulfonate
(SDBS) surfactant at 135 °C. The Propdan system resulted in stable latexes and
90% conversion was reached in 12 hours. The polydispersity was 1.15–1.60 and
molecular weights were up to 40 kg/mol. The Macleod system, if a proper
nitroxide/initiator ratio was chosen, also gave stable latexes with good control,
with polydispersities in the range of 1.1–1.2. Conversion reached 87% in six
hours, although a later publication [225] showed that this relatively fast poly-
merization also resulted in a large proportion of dead chains.

Scheme 4 Nitroxides used for NMCRP in miniemulsion

Later on, these same research groups started using pre-formed TEMPO-
terminated polystyrene as a one component initiator system instead of a bi-
component nitroxide/initiator system. Pan et al. [224] prepared TEMPO-termin-
ated polystyrene in bulk and isolated this to use it as the initiator in their
miniemulsions. This led to slower polymerization rates, molecular weights
lower than predicted and relatively broad molecular weight distributions.
Keoshkerian et al. [225], on the other hand, reported very high conversions in
six hours (99.6%) and narrow polydispersities (1.15) by preparing TEMPO-
terminated polystyrene in bulk up to a conversion of about 5% and applying
Miniemulsion Polymerization 221

the mixture directly in miniemulsion without purification.As a proof of living-

ness, they were able to extend the chains with styrene in bulk and to produce
block copolymers with butyl acrylate (BA) by directly adding the BA to the
miniemulsion, yielding a block copolymer with a polydispersity of 1.18.
A disadvantage of TEMPO mediated systems is that a high temperature
(above the boiling point of water) is required and so conventional emulsion
polymerization reactors cannot be used. Another disadvantage of TEMPO is
the limited monomer choice. The use of the so-called SG1 nitroxide, 3 (see
Scheme 4), partially overcomes these problems, since it has been reported to
work at 90 °C and to work with both styrene and BA [217, 220, 223, 226, 228, 229,
230].When AIBN, an oil-soluble initiator, was used, poor results were obtained
in styrene miniemulsion polymerizations [217, 220]. Conversion was low and
the polydisperity was around 1.6. On the other hand, the use of a water-soluble
initiator was more successful [223, 226, 228]. Conversion reached 90% within
eight hours.Another important observation was that the pH was a very impor-
tant parameter in the SG1-mediated polymerizations. This was assigned to
side-reaction of the SG1 nitroxide. The best results were obtained when the pH
was close to 7. Also, the monomer/water ratio appeared to have an important
effect on the controllability of the polymerizations. Increasing the monomer/
water ratio led to better-controlled (lower polydispersity) reactions. This was
assigned to the fact than at a higher monomer/water ratio, less nitroxide
partitions to the aqueous phase.
As already mentioned, the SG1 nitroxide is also capable of controlling poly-
merizations other than styrene [226, 229, 230]. Farcet et al. showed that it also
worked with BA, although it required higher reaction temperatures (above
100 °C) because of the lower activation rate constant of SG1-terminated poly-
butyl acrylate compared to polystyrene. Polydispersities as low as 1.19 were
obtained for BA miniemulsion homopolymerizations, while Mn increased
linearly with conversion. Addition of styrene, after the majority of BA had
reacted, resulted in the formation of block copolymers with a narrow poly-
dispersity of 1.27. Additional evidence for the livingness of the polybutyl
acrylate chains was given by thorough analyses of the materials formed via 1H
NMR, 13C NMR, SEC and MALDI-TOF [230]. These analyses showed that the
majority of chains consisted of polybutyl acrylate with one initiator-derived
and one SG1 chain-end. Besides block copolymers, gradient copolymers of
styrene and BA were also synthesized [230] using miniemulsion copolymer-
ization of styrene and BA. Due to the composition drift and the livingness of
the chains, this gives gradient block copolymers that contain relatively more
styrene in the beginning of the chain and relatively more BA closer to the
chain end.
Keoshkerian et al. used another nitroxide, 4 (see Scheme 4), to perform
miniemulsion polymerizations with acrylates [231]. First the nitroxide was
reacted with styrene and BPO in bulk to form nitroxide-terminated oligomers.
These oligomers were used in a miniemulsion polymerization of BA at 135 °C.
86% conversion was reached after three hours, and at this point the polymer
222 F. J. Schork et al.

had a Mn of 12 kg/mol with a polydispersity of 1.27. When the same procedure

was followed with TEMPO as the nitroxide, conversion was less than 8% and
did not proceed any further. The addition of a small amount of ascorbic acid,
which destroys free nitroxide, led to conversions close to 65% after three hours,
although the polydispersity of 1.62 was broader than with 4.
Tortosa et al. [228] tried to synthesize styrene/BA block copolymers using
both TEMPO and OH-TEMPO, 2 (see Scheme 4). OH-TEMPO was used be-
cause of the aqueous phase partitioning of this nitroxide, which would reduce
the nitroxide concentration in the particles and so result in higher polymer-
ization rates. It was indeed found that the conversion in the OH-TEMPO-
mediated polymerizations was much higher. However, it was also found that the
TEMPO-mediated polymerizations showed a greater living character and that
the OH-TEMPO-mediated polymerizations also gave pBA homopolymer. In
a later publication, Cunningham et al. [232] studied the effects of camphorsul-
fonic acid (CSA), a nitroxide destroyer known to accelerate bulk polymeriza-
tions, on styrene miniemulsion polymerizations with TEMPO and OH-
TEMPO. It was found that the CSA effectively accelerates the polymerization
rate, especially at high nitroxide/initiator ratios, although the effects were not
as large as seen in bulk experiments. This was ascribed to the aqueous phase
partitioning of the CSA, which reduces the CSA concentration in the particles.
Unlike the large differences in polymerization rate seen in the BA polymer-
izations with TEMPO and OH-TEMPO [228], the experiments with styrene
showed about equal polymerization rates for both the TEMPO and OH-
TEMPO-mediated systems. Also, the increase in rate caused by the addition of
CSA was equal in both systems, despite the large difference in water-solubili-
ties between TEMPO and OH-TEMPO.
In another publication, Cunningham et al. [233] studied the effects of the
KPS concentration and the TEMPO/KPS ratio on conversion, molecular weight
and particle size in styrene miniemulsion polymerizations. It was found that
most characteristics were similar to those for bulk polymerizations, although
some unique features for heterogeneous systems were identified. A much
higher initiator efficiency (approaching 100% at TEMPO/KPS=4) compared to
conventional emulsion and miniemulsion polymerization was also observed,
which emphasized the role of aqueous phase TEMPO in deactivating aqueous
phase radicals. These results inspired the authors to model these systems in or-
der to gain an even better understanding and to identify the operating condi-
tions for optimal process performance [234, 235]. Ma et al. were the first to
model the interfacial mass transfer of TEMPO and they found that phase equi-
librium is achieved before TEMPO has an opportunity to react with active poly-
mer radicals, and this is fast enough to maintain phase equilibrium through-
out the polymerization [234]. In a second publication [235], they modeled the
whole system, and by varying the KPS and TEMPO concentration they were
able to find operating conditions at which the polydispersity was minimized
and the degree of polymer livingness was maximized. In addition, it was found
that the polymerization rate and the degree of livingness could be further im-
Miniemulsion Polymerization 223

Table 17 Different NMCRP miniemulsion systems, with associated references

Nitroxide T Monomers Initiators Surfactants Reference

1 125 Sty BPO Dowfax 8390 [219, 222]

1 125 Sty pSty-1 Dowfax 8390 [225]
1 135 Sty KPS SDBS [220, 223, 234]
1 135 Sty, BA pSty-1 SDBS [226]
3 90 Sty AIBN SDS [218, 221]
3 90 Sty KPS/SPS SDS [218, 221, 224]
3 112–120 BA, Sty 3-alkoxyamine SDS/Forafac [227, 229, 230,
Dowfax 8390 231]
1, 2 135 Sty, BA BPO, KPS SDBS [229, 233]
1,4 135 BA pSty-1 pSty-4 SDBS [232]
1 125 Sty pSty-1 Dowfax 8390 [217]

proved by increasing the volume fraction of water, although from an industrial

point of view this might not be desirable.
The final paper that will be discussed in this section is a kinetic investiga-
tion by Pan et al. [216]. They created a series of styrene miniemulsions, using
TEMPO-terminated polystyrene, in which they varied the Dowfax 8390 sur-
factant concentration from 1.25 to 25 mM. Although this had a large effect on
the particle size, the effect on molecular weight and polymerization rate was
small, while for conventional miniemulsion polymerization the effect of the
particle size on polymerization rate is generally very significant. This was ex-
plained by the low average number of radicals per particle as a result of the cou-
pling between TEMPO and active radicals, which overwhelms the compart-
mentalization effect in this case.
All systems discussed are summarized in Table 17.

4.4.2
Atom Transfer Radical Polymerization

4.4.2.1
Mechanism

Atom Transfer Radical Polymerization (ATRP) was first reported in 1995 by

Matyjaszewski et al. [236–238] who investigated its potential for copper com-
plexes, and Sawamoto et al. [239–241] who utilized ruthenium complexes.ATRP
belongs to a class of living polymerizations known as reversible termination,
that includes nitroxide-mediated radical polymerizations (NMRP). They are so
named because the growth of the chain is controlled by a reversible termination
event where the chain-end is exchanged between an active and dormant species.
The lifetime of the active species is very short, such that only a few monomer
units are added during each active cycle, giving the reaction its living character.
224 F. J. Schork et al.

To induce this reversible termination, ATRP employs a transition metal

complex with sufficient redox potential to deactivate propagating radicals.
A halide atom, typically Cl or Br, is transferred reversibly (hence the name
“atom transfer”) to the metal complex. In the process the metal alternates
between a lower and higher oxidation state. A general mechanism is shown in
Scheme 5.

Scheme 5 General ATRP mechanism

The metal undergoes a one-electron oxidation with the simultaneous ab-

straction of the halogen, generating radicals via a reversible redox process. The
success of the process depends upon the fact that the equilibrium is shifted
heavily in the direction of the dormant species.While complexes of copper and
ruthenium have been most widely studied, complexes of nickel, palladium and
iron can also be used [242–244]. Molybdenum, rhenium and rhodium ATRP
have also been reported [245–247]. ATRP reactions are very well behaved and
can easily produce polymers of controlled molecular weight and narrow poly-
dispersity. Most classes of monomers have been successfully polymerized via
ATRP. These include styrenes, (meth)acrylates, (meth)acrylamides, dienes,
acrylonitrile and other monomers containing radical stabilizing substituents
[248, 249]. Ring-opening polymerizations are also possible with ATRP [250,
251]. The initiators used are typically alkyl halides, but any compound with a
weak halogen-heteroatom bond will suffice. The halogen end-group has the
advantage of offering the ability to add functionality to the polymer.
In general, reaction rates in ATRP are slower than conventional free radical
polymerizations. The unique nature of the ATRP equilibrium, which is shifted
strongly towards the dormant species, effectively lowering the active, propa-
gating radical concentration as compared to the conventional analog, is the
source of the lower rates. This can be overcome to a certain degree by adjust-
ing the metal/ligand ratio or through the use of additives [252–254]. Another
factor contributing to the rate is that each bimolecular termination event
releases two metal complexes in the higher oxidation state. As such, a shift in
equilibrium in order to increase the propagation rate results in increased
termination and an increase in the concentration of the complex in the higher
oxidation state. The system tends to self-regulate and maintain the rate of poly-
merization. ATRP with transition metal complexes is extremely sensitive to
oxygen owing to its reliance on the redox reaction between the halide and the
metal complex. Reactions must be conducted in an inert environment or the
metal will oxidize, effectively killing the polymerization. Because the quantity
of metal complex required is relatively large and much of it will remain in the
polymer, its residue must be removed for both environmental and economic
reasons.
Miniemulsion Polymerization 225

Although ATRP behaves differently from conventional free radical polymer-

ization, the fundamental reactions involved are very similar and include initia-
tion, propagation, transfer and termination (see Scheme 6). Since chain termi-
nation does not occur in a truly living polymerization, the “living” character of
the chains in ATRP derives from the fact that chain propagation is first order
with respect to radical concentration and irreversible bi-molecular termination
is second order. As such, the concentration of the radicals is kept very low, the
rate of bi-molecular termination is greatly reduced, and typically less than 10%
of all of the chains will terminate. Unlike conventional free radical polymer-
ization, where the rate is dictated by a steady state between the initiation
and termination rates, the rate and concentration of propagating radicals in
ATRP is controlled entirely by the equilibrium between activation and de-
activation [255].

Scheme 6 Fundamental ATRP reactions

The initiators used are typically alkyl halides with similar structures to that
of the monomer being employed. Of the halogens, chlorine and bromine have
been shown to produce the best overall results with respect to molecular weight
control [256]. However, iodine works well in certain cases, for example, rhe-
nium-mediated styrene polymerizations [246] and copper-mediated acrylate
polymerizations [257]. The C–F bond with fluorine is too strong to cleave homo-
lytically, ruling out its use as an effective initiator. In general, any alkyl halide
with activating substituents on the a-carbon (aryl, carbonyl or allyl groups) can
act as an ATRP initiator. In addition, polyhalogenated compounds (like CCl4
or CHCI3) and compounds with a weak R–X bond (like N–X, S–X) can also
be used.
There are several requirements that are generally recognized as essential to
an effective ATRP catalyst [256, 258, 259, 260]. The metal center should be able
to assume at least two oxidation states, separated by one electron, like Cu(I) and
Cu(II). It should also be attractive to halogens, it should possess an expandable
226 F. J. Schork et al.

coordination sphere such that when oxidized it can contain the halogen, and it
should have a low affinity for alkyl radicals and the hydrogen atoms on alkyl
groups.
Conventional free radical polymerizations in miniemulsions benefit kineti-
cally from the effects of radical segregation. In solution, any radical could ter-
minate with another theoretically. However, when the radicals are segregated
into isolated reaction loci (such as miniemulsion droplets or particles) termi-
nation is no longer possible. Because the total concentration of radicals is
distributed throughout the particles, the probability that any two radicals will
terminate bi-molecularly is greatly reduced. The ideal situation is one in which
a lone radical in a particle can terminate only with a radical that enters the
particle. This is known as the “zero-one” limit [121], and in this case the rate of
bimolecular termination is controlled by radical entry alone. In the absence of
other effects, this lowering of the incidence of bi-molecular termination events
tends to increase the overall rate of polymerization while simultaneously nar-
rowing the molecular weight distribution. However, because of the mechanism
involved, the same benefit is not seen with ATRP in miniemulsion. With ATRP,
the dominating rate-controlling factor is the equilibrium between the dormant
and active species. Since the equilibrium heavily favors the dormant species, the
lifetime of the active species is extremely short. As such, the concentrations of
these active radicals are always minute [261]. Therefore, the probability that a
water phase radical will enter a particle containing another radical and termi-
nate is exceedingly low. Any kinetic benefit that might otherwise be gained
from segregation is overshadowed by this low radical concentration. It should
also be noted that, because of the ATRP mechanism, increases in reaction rate,
whatever their origin, will come at the expense of the deactivator species. The
resulting decrease in the concentration of deactivator tends to produce broader
polydispersities.

4.4.2.2
Results

There are few reported instances of ATRP in miniemulsion in the current

literature. Matyjaszewski [262] and co-workers employed both forward (direct)
and reverse-ATRP of n-butyl methacrylate in miniemulsion stabilized with a
non-ionic surfactant, polyoxyethylene(20) oleyl ether (Brij 98), and using 4,4¢-
di(5-nonyl)-4,4¢-bipyridine (dNbpy) with either CuBr or CuBr2. They looked
at the effects of using both oil and water-soluble initiators (AIBN and 2,2¢-
azobis(2-methylpropionamidine) dihydrochloride, or V-50) with the reverse
process. With AIBN, the final polydispersity was relatively narrow (~1.4) but
increased with conversion. Also, the number average molecular weight began
higher than predicted and exhibited some curvature. This was attributed to
slow decomposition of the AIBN, causing a slow and less quantitative forma-
tion of chains. Polydispersities were slightly lower with the V-50 and the
number average weight progressed linearly with conversion, although the over-
Miniemulsion Polymerization 227

all rate was much slower than with AIBN. Additionally, the progression of Mn,
though linear, was much higher than found theoretically. In this case, the ini-
tiator decomposed faster but the radicals formed tended to terminate in the
aqueous phase, contributing both to the slower overall rate and the higher than
predicted actual molecular weights. The authors performed direct ATRP with
an oil-soluble initiator, ethyl 2-bromoisobutyrate (EBiB). At a lower tempera-
ture than the reverse process, 70 °C vs 90 °C, the polymerization rate of the
direct process was significantly faster. The polydispersity remained relatively
flat throughout the experiment, with a final value of approximately 1.3. The
molecular weight evolution, though roughly linear, was higher than predicted.
Additionally, a semilog plot of the conversion data revealed some curvature,
indicating a larger than expected number of chain termination events. It was
postulated that the cause lay in the partitioning of the Cu(II) species into the
aqueous phase. Recent studies with ATRP in aqueous dispersions lend credence
to this argument [263]. They observed that the polymerization rate was insen-
sitive to the size and number of particles and controlled entirely by the atom
transfer equilibrium. The researchers also studied the effect of removing the
costabilizer, hexadecane, from the recipe in order to determine if the hydro-
phobic dNbpy ligand alone would act as a sufficient droplet stabilizer. However,
it was noted that the droplet size increased dramatically in the absence of hexa-
decane, indicating that the osmotic pressure would be insufficient to prevent
Ostwald ripening.
Li and Matyjaszewski [264], building on the earlier work mentioned here
[262], conducted reverse ATRP in miniemulsions using n-butyl methacrylate
(BMA) with a more active catalyst system and a faster initiator. The solids
content was roughly double that of the previous effort, jumping from approx-
imately 13% to over 20%. More importantly, the surfactant (Brij 98) concen-
tration was reduced from 13.5%wt to 2.3%wt based on monomer, decreasing
the likelihood of micellar nucleation. Because of their high activities in bulk
and solution ATRP, complexes of hexasubstituted tris(2-aminoethyl)amine
(TREN) with Cu/Br2 were utilized as the metal activator/deactivator. The water-
soluble initiator used was 2,2¢-azobis[2-(2-imidazolin-2-yl)propane] dihydro-
chloride (VA-044). The VA-044 was chosen for its fast decomposition rate,
which facilitates a well-controlled ATRP and contributes to colloidal stability.
The authors looked at the effect of the ligand and noted that there was no
induction period with the TREN ligand as compared to dNbpy. While linear
progressions of Mn with conversion were seen, the authors found that the mole-
cular weight was not quantitative in initiator concentration. Instead Mn tended
to be much higher than that calculated based on the initiator concentration and
more closely followed a trajectory calculated from the initial concentration of
deactivator, Cu(II).Assuming 100% initiator efficiency, the authors explain the
deviation in terms of an excess radical concentration over the concentration of
Cu(II), leading to the generation and subsequent termination of some of the
oligomeric chains. Since a 1:1 molar ratio of [Cu(II)]0/[I]0 was used in most
of the experiments, in theory there would be two initiator radicals for every
228 F. J. Schork et al.

molecule of Cu(II) deactivator very early on in the polymerization. Many of

the radicals would indeed quickly terminate until their concentration reached
low enough levels that bi-molecular termination was insignificant compared
to deactivation and they no longer competed for Cu(II). When a 2:1 ratio of
[Cu(II)]0/[I]0 was employed, marginally better control of Mn was observed and
the progression still tended to follow that calculated based upon [Cu(II)]0, lead-
ing the authors to conclude that higher levels of Cu(II) would be needed for
polymerizations that provide better control of Mn and relatively low polydis-
persities. The effect of surfactant concentration was studied using several non-
ionic surfactants in addition to Brij 98. The principal finding was that at higher
surfactant concentrations (>13 wt% based on monomer), bi-modal molecular
weight distributions and bi-modal particle size distributions of the latex
were observed, indicating micellar and droplet nucleation. By increasing the
initiator concentration and lowering the deactivator concentration, the authors
also demonstrated that the kinetics of the polymerization are controlled
primarily by the atom transfer equilibrium.

4.4.3
Reversible Addition Fragmentation Polymerization

4.4.3.1
Mechanism

The third (and also the most recently developed) controlled free radical tech-
nique discussed in this review is RAFT. In 1998 Rizzardo et al. published a novel
“controlled” free-radical polymerization technique, which they designated the
RAFT process [265–267] because the mechanism involves Reversible Addition-
Fragmentation chain Transfer. This technique allowed the production of poly-
mer with a narrow molecular weight distribution. In fact, this concept was not
entirely new, and stemmed from the same researchers’ previously published
work to produce block copolymers using methacrylate macromonomers as
reversible addition-fragmentation chain transfer agents in 1995 [268]. However,
these macromonomers were not very effective RAFT agents. The breakthrough
came with the discovery of a more reactive double bond species, S=C(Z)SR.
During styrene polymerization, the propagating radicals were very reactive
to the dithioesters and to a much lesser extent to the xanthates [269]. A brief
description of the RAFT process is given below, and a schematic representation
is given in Scheme 7.
A conventional free-radical initiator is added (contrary to some other con-
trolled free-radical polymerization techniques) that generates radicals, which
can add either to the monomer or the S=C moiety of the RAFT agent (step 1). In
most cases the addition of small carbon-centered radicals to the RAFT agent is
rapid and is not rate determining. Therefore, step (1) involves polymeric radical
addition to 1 to form an intermediate radical species 2 that will fragment back
to the original polymeric radical species or fragment to a dormant species 3
Miniemulsion Polymerization 229

Scheme 7 Schematic representation of the proposed RAFT mechanism. It should be noted

that in these equilibria any radical can react with any dormant species or RAFT agent.
(1) Addition of a propagating polymeric radical to the initial RAFT agent 1, forming the
intermediate radical 2. The intermediate radical can either fragment into the two species it
was formed from or into a dormant polymeric RAFT agent 3 and a small radical R•. (2) The
small radical initiates polymerization, forming a polymeric radical, rather than reacting with
3 (reforming 1). Therefore R should be a good leaving group and should have the ability
to be added to the monomer. (3) Equilibrium between propagation polymeric radicals and
dormant polymeric RAFT agents. (4) Intermediate radical termination

and a small radical, R•. R• can then further propagate to form a polymeric rad-
ical (step 2 in Scheme 7), rather than adding to 3. The dormant polymeric RAFT
agent acts in a similar way to a RAFT agent, so growing polymeric radicals can
also add to the dithiocarbonyl double bond of the polymeric RAFT agent,
thereby forming an intermediate radical 4. This intermediate has an equal
probability of fragmenting back into its starting species or into a dormant poly-
meric RAFT agent and a polymeric radical, in which the dithiocarbonate
moiety has been exchanged between the active and dormant polymer chains
of the starting species. This equal probability of fragmenting to either side of
the equilibrium is a result of the symmetry of 4. There might be a difference in
the chain length of both sides, but this will not have an effect, unless one of the
two sides is extremely short. This mechanism of addition of radicals to the
dithiocarbonyl double bond and fragmentation of the intermediate was shown
by Moad et al. [270], who observed the intermediate radical directly by ESR.
Overall, polymer chains with a dithiocarbonate end-group are formed. If
addition to the dithiocarbonyl double bond is fast compared to propagation,
230 F. J. Schork et al.

and termination is suppressed by keeping the radical concentration low, all

of the chains will grow in a sequential process, leading to a low polydispersity.
The number of chains is determined by the amount of RAFT agent and initiator
that has been consumed. Assuming termination by combination, the number
of dead chains will be equal to the amount of initiator that is consumed. The
number of chains with a dithiocarbonate end-group, the dormant chains, is
equal to the amount of consumed RAFT agent. One should therefore keep the
initiator to RAFT agent ratio low in order to obtain a high percentage of dor-
mant chains. This criterion is especially important in the preparation of block
copolymers [271–273].
In fact, the RAFT process resembles the degenerative transfer (DT) process
[274]. In a polymerization in which an alkyl iodide is used as the degenerative
transfer agent, the iodine atom is exchanged between a polymeric radical and
a dormant chain, similar to the dithiocarbonate exchange in RAFT. However,
in the case of degenerative transfer there is a direct equilibrium between the
dormant and growing chains, without formation of an intermediate radical.
If reactions 1 to 3 in Scheme 7 are considered, there is no reason to assume
that addition of a RAFT agent to a conventional free radical polymerization will
have an effect on the polymerization rate, since the equilibrium concentration
of propagating radicals will not be affected. However, it has been found that
considerable retardation does take place in RAFT polymerization [275–281].
The intermediate radical was postulated to be the reason for the significant
retardation of the polymerization rate. Two explanations for retardation have
been put forward:
(i) slow fragmentation of the intermediate radical [277–280]
(ii) termination of the intermediate radical (reaction 4 in Scheme 7) [275, 276,
281]
The key variable in both explanations is the fragmentation rate constant. There
is a difference of six orders of magnitude between the fragmentation rate con-
stants obtained via explanations (i) and (ii). The question of which explanation
is correct has been hotly debated for over three years, but is still unresolved.

4.4.3.2
Effects of RAFT and Transfer Agents on Emulsion Polymerization Kinetics

As far back as 1948, Smith and Ewart [13] included the effects of radical desorp-
tion in emulsion polymerization kinetics, and in 1965 Romatowski et al. [282–
284] showed that radicals resulting from chain transfer to monomer indeed
escape from the particles.
Nomura et al. [285] and Lichti et al. [286] studied the effects of transfer
agents on the kinetics of ab initio and seeded emulsion polymerization of
styrene, respectively. Nomura et al. found that the polymerization rate per
particle decreased with increasing amounts of carbon tetrachloride, carbon
tetrabromide and primary mercaptans, and that the effects were stronger when
Miniemulsion Polymerization 231

the transfer constant or the water-solubility was higher. Lichti et al. observed
the same in seeded experiments with carbon tetrachloride and carbon tetra-
bromide as chain transfer agents. They were the first to actually measure the
exit rate coefficient, using g-radiolysis relaxation data. They found an increas-
ing exit rate coefficient with increasing amounts of transfer agents and a higher
exit rate coefficient when the transfer constant was higher. They also found that
the entry rate coefficient increased with increasing amounts of carbon tetra-
bromide. Due to the counterbalancing effects of an increased exit rate and an
increased entry rate the polymerization rate passed through a minimum.
Maxwell et al. [287] extended their own model for entry by taking the effect
of transfer agent into account and used this model to explain the increase
observed in emulsion polymerizations with monomers with a high critical
chain length z and thiols of intermediate chain length. They also used this
model to show that longer chain thiols are too water-insoluble to have an effect
and that short chain thiols might suffer aqueous phase termination and
increase the exit rate, and so they can reduce the polymerization rate instead
of increasing it. No effect was expected for styrene from Maxwell’s model,
which was confirmed by the work of Asua et al. [288]. They found that n-
dodecyl mercaptan had no effect on the polymerization rate.
The work of Monteiro et al. [289] showed that when RAFT agents are applied
in emulsion, the rate of polymerization is significantly retarded. This effect
is stronger when a RAFT agent with a more water-soluble leaving group is used.
Exit from the particles after fragmentation was proposed to be the main
reason for the observed retardation. Because of the high reactivity of the RAFT
agents used, it is expected that all of the RAFT agent is consumed after a few
percent conversion, and so it should no longer should have an effect. However,
it was observed that the rate of polymerization decreased with increasing RAFT
in interval II. Monteiro et al. claimed that this was due to transport limitation
of RAFT from the monomer droplets to the particles, meaning that there is a
constant flux of RAFT agent to the particles, even if all of the RAFT agent has
been consumed in the particles. Therefore, not all of the chains start to grow
simultaneously, resulting in broad polydispersities. However, retardation was
also observed in Interval III, and this could not be ascribed to exit and trans-
port limitations.When this work was published, intermediate radical termina-
tion [275, 276] had not yet been put forward by Monteiro et al. as a source of
retardation. However, since the system might not be under zero-one conditions
in Interval III due to the increased particle size, intermediate radical termina-
tion might explain these results.
Another observation Monteiro et al. made was that a red layer was observed
during Interval II, consisting of low molecular weight dormant chains, swollen
with monomer.At the crossover to Interval III, this red layer coalesced, forming
red coagulant. The same red layer was also observed by De Brouwer et al. [290]
in miniemulsions stabilized with ionic surfactants. When polymer was used
as the so-called cosurfactant, this polymer was not present in the red layer,
indicating that this layer was not due to droplet coalescence.Also, the use of an
232 F. J. Schork et al.

oil-soluble initiator did not reduce the formation of the red layer. Using a
higher radical flux (to enhance droplet nucleation) did not have an effect.
Indeed, the formation of the red layer was correlated to the polymerization rate,
which indicates that the product formed during the polymerization plays a
crucial role in the destabilization. However, when nonionic surfactants were
used, destabilization did not occur, and controlled miniemulsion polymeriza-
tions could be performed without destabilization. Later, Luo et al. suggested
that the instabilities are the result of the large number of oligomers formed in
the early stages of RAFT miniemulsions, causing a superswelling state [128],
which could be prevented by increasing the amount of cosurfactant.
Moad et al. [291] showed that the type of RAFT agent is important. Using a
very reactive RAFT agent (with a transfer constant of about 6000), similar to
that used in the work of De Brouwer and Monteiro, resulted in a broad poly-
dispersity in ab initio styrene polymerizations with ionic surfactant, which was
ascribed to the fact that the RAFT agent was not uniformly dispersed in the
polymerization medium. The use of less reactive RAFT agents (with transfer
constants of 10–30) did not result in destabilization and the final polymer had
a polydispersity close to 1.4.
Prescott et al. [292] used acetone to transport a water insoluble RAFT agent
to the seed particles. The polymerization was initiated in Interval III after
removing the acetone. No destabilization was observed, which according to our
previous discussion might indicate that the transfer constant of the RAFT agent
used was not extremely high. However, it was high enough to result in a linear
increase in molecular weight, and polydispersities between 1.2 and 1.4. Al-
though the RAFT agent is consumed at the beginning of the reaction (the mole-
cular weight follows the theoretical linear increase), a reduction in rate is ob-
served throughout the reaction. In these experiments, a small seed was used
and the amount of monomer was such that the particle size does not increase
much, which means that the system is likely to be under zero-one conditions
throughout the polymerization and so intermediate radical termination can-
not explain the retardation observed.
Monteiro et al. [293] also studied the effect of xanthates (RAFT agents with
low transfer constants) with styrene, in ab initio styrene polymerizations.
Again rate retardation was observed throughout the polymerization. This is not
surprising, since the low transfer constants of these RAFT agents mean that
they are present during the whole polymerization, which results in an increased
exit rate throughout the reaction.
This was later confirmed by Smulders et al. [294], who experimentally deter-
mined the exit rate in similar systems using g-relaxation experiments. The exit
rate was found to increase linearly with the RAFT concentration, although the
decrease in rate could not be ascribed to the increase in exit rate alone.
Summarizing, we know that RAFT can be applied in emulsion, although the
mechanism for this is not yet fully understood. Highly reactive RAFT agents
can lead to destabilization, although the use of nonionic surfactants seems to
prevent this destabilization.Rate retardation is observed in all cases. This can
Miniemulsion Polymerization 233

be partly ascribed to the increased exit rate, although the retardation is still
observed even when all of the RAFT agent has been consumed. In that case,
intermediate radical termination might explain the reduction in rate. However,
even when the system is under zero-one conditions and all RAFT is consumed,
retardation still occurs. This cannot be ascribed to intermediate radical termi-
nation or to exit. In fact, an explanation for this might be quite simple, as shown
by Smulders [295]. Retardation with RAFT in zero-one systems in which all
of the RAFT has been consumed cannot be ascribed to increased exit rate any-
more, since the leaving groups of the dormant polymer chains cannot exit.
Intermediate radical termination is also not a dominant mechanism, since each
particle contains only one radical. However, the fact that each particle contains
only one radical explains why retardation is observed in these systems. This
one radical is either present as a “normal” radical, R•, capable of propagating
and so consuming monomer, or as a “intermediate” radical, I•.While the radical
is in the intermediate state it does not consume monomer, which in turn leads
to retardation. Since the system is under zero-one conditions, the system does
not reach steady state at the microscopic level (inside a particle), because a
particle contains either no radical, one “normal” radical, or one “intermediate”
radical. The lifetimes of R• and I• are given by:

1 1
t R• = 0005 t I • = 91 (19)
kadd [dormant chains] 2k–add

Using the rate parameters for dithiobenzoate RAFT polymerization of styrene

at 70 °C, as reported by Monteiro et al. [275] (kadd=4¥106 dm3mol–1s–1, k–add=
1¥105 s–1), and a dormant chain concentration of 0.06 M, this results in a life-
time of 4.2¥10–6 s for a “normal” radical, and a lifetime of 5.0¥10–6 s for an
intermediate radical. This means that the fraction of time that a radical is pre-
sent as a propagating radical in this system is 4.2¥10–6/(4.2¥10–6+5.0¥ 10–6)=
0.46. This also means that the polymerization rate in this example would
only be 46% of the polymerization rate without RAFT. The mechanism pro-
posed by Monteiro et al., which included intermediate radical termination,
was supported by Fukuda et al. [281, 296]. However, the latter authors proposed
a value of kdd on the order of 104 s–1. Following the same pathway, this value
means that the polymerization rate with RAFT is only 7.7% of the rate without
RAFT in a zero-one system. Since no good experimental data is currently avail-
able for zero-one emulsion systems with dithiobenzoate RAFT agents, at the
moment no definitive statement can be made about which value of the frag-
mentation rate constant best describes a zero-one system. However, these
results indicate that zero-one experiments can be a useful tool for determining
this rate parameter, and they provide data useful as we attempt to pinpoint
the correct value for the fragmentation rate constant (see previous section).
Davis et al. fitted conversion-time data for a styrene polymerization with RAFT
at 60 °C using kadd=5.4¥105 dm3 mol–1 s–1 and k–add=3.3¥10–2 s–1. For a zero-
one system with 1 mol% RAFT, these values would lead to tR• =3.1¥10–5 s and
234 F. J. Schork et al.

tI• =15.2 s. This means that zero-one polymerization should not proceed,
because the radicals are present as intermediate radicals for more than 99.99%
of time. Since we have experimental evidence that RAFT systems do proceed
under zero-one conditions [292], these rate parameters seem highly unlikely,
although it should be noted that Prescott et al. used a RAFT agent with a less
stable intermediate. Zero-one experiments with dithiobenzoate RAFT agents
might be the key to closing the “six-orders-of-magnitude-gap” for the frag-
mentation rate constant.

4.4.3.3
Application of RAFT in Miniemulsion

As with nitroxide-mediated polymerizations and ATRP, and with RAFT, mini-

emulsion systems are often preferred over conventional emulsion systems,
although not as exclusively as with NMCRP and ATRP. In this section we will
discuss some applications of RAFT in miniemulsions.
In 2000, Moad et al. reported the synthesis of controlled polystyrene using
RAFT in miniemulsion [291]. Using phenyl ethyl dithiobenzoate in a SDS/cetyl
alcohol stabilized system at 70 °C, 25% conversion was obtained in four hours,
while a control experiment without RAFT reached 82% conversion in one
hour. Molecular weight increased with conversion and the polydispersity went
down to 1.18. No problems with stability were reported. On the other hand, De
Brouwer et al. [290] and Tsavalas et al. [296] were unable to obtain stable latexes
using dithiobenzoate RAFT agent in either anionic- or cationic-stabilized mini-
emulsions. They reported the formation of a red organic layer on top of the
miniemulsion as soon as the polymerization started. This layer consisted of low
molecular weight polymer and monomer. Luo et al. later ascribed the observed
phenomena to a superswelling state, caused by the large number of oligomers
formed at the beginning of the polymerization [128]. However, when nonionic
surfactants like Igepal 890 and Brij 98 were used by Brouwer et al. [290], they
could perform stable RAFT miniemulsion polymerizations. Miniemulsion poly-
merizations of EHMA, STY, MMA, BMA, and MA all resulted in stable latexes
with polydispersities below 1.4, and sometimes as low as 1.1, at very high con-
versions. When the miniemulsions formed were used as seed latexes in either
batch or semi-batch polymerization with a second monomer, block copolymers
with a low polydispersity and a high level of block purity were obtained.
Butté et al. [261] were able to perform miniemulsion polymerizations
stabilized with SDS and hexadecane using dithiobenzoate and “pyrrole” RAFT
agents. In most cases oligomerized RAFT agents were used, but “monomeric”
RAFT agents were also applied successfully. Although they used basically the
same systems, Butté et al. did not observe the red layer formation reported
earlier by De Brouwer and Tsavalas. Linear molecular weight growth and rela-
tively narrow polydispersities were reported, although they were broader than
for bulk polymerizations. This was ascribed to the presence of dead chains
in the oligomers and differences in miniemulsion droplet sizes, leading to dif-
Miniemulsion Polymerization 235

ferences in monomer-to-RAFT ratios. Smaller particles have a larger sur-

face/volume ratio and are therefore preferentially entered by z-mers, leading to
monomer consumption in these particles which is replaced by monomer from
the larger particles.
Butté found that the polymerization rate decreased with RAFT. This was
supposed to be the result of an increasing exit rate. However, even when oligo-
meric RAFT agents were used, which should not lead to an increased exit rate,
a decrease in rate was observed, which was ascribed to the presence of “mono-
meric” RAFT agent in the oligomer mixture. Finally, Butté reported the
synthesis of block copolymers in miniemulsion by adding styrene to a fully
polymerized MMA miniemulsion and by adding BA to a 63% polymerized
styrene miniemulsion. In both cases it was shown that block copolymer was
formed, although polydispersities were relatively high.
In order to prevent the formation of the red layer observed by De Brouwer
and Tsavalas, Vosloo et al. performed SDS-stabilized miniemulsion polymer-
izations of styrene using pre-formed dithiobenzoate-end-capped styrene oli-
gomers, formed in bulk [297]. Two types of cosurfactant (hexadecane and cetyl
alcohol) and two oligomers with different molecular weights were used. Red
layer formation was not observed in any of the miniemulsion polymerizations,
and the results were better – lower polydispersities, and molecular weights that
were closer to the theoretical values – when hexadecane and the lower mole-
cular weight oligomers were used.
Lansalot et al. studied the influence of the structure of the RAFT agent on
styrene miniemulsion polymerization [298]. The use of 1-phenylethyl phenyl-
dithioacetate (PEPDTA) was compared to cumyl dithiobenzoate (CDB) and
1-phenylethyl dithiobenzoate (PEDB). It was shown that PEPDTA did not show
retardation in bulk experiments, while CDB and PEDB show a large decrease
in rate with increasing RAFT concentration. This was ascribed to the less stable
PEPDTA macroRAFT radical. When the same RAFT agents were used in
styrene miniemulsion polymerizations, stabilized by SDS/hexadecane, again
the PEPDTA showed much higher polymerization rates than CDB and PEDB.
However, in contrast to the bulk experiments with PEPDTA, a decrease in rate
with an increase in RAFT was observed in the miniemulsion. This was ascribed
to the exiting of radicals formed after addition and the fragmentation of the
initial RAFT agent. This was confirmed by miniemulsion polymerization
experiments performed using oligomerized PEPDTA, where the leaving radical
cannot exit to the aqueous phase. In that case, using the same concentration as
in the experiment with “monomeric” PEPDTA, the polymerization rate drama-
tically increased to almost the same polymerization rate as without RAFT.

4.4.4
Colloidal Stability

After being frustrated by poor colloidal stability (phase separation or coagula-

tion) in controlled macroemulsion polymerization (NMP, ATRP, RAFT), re-
236 F. J. Schork et al.

searchers turned to living miniemulsion polymerization [299–304]. It was ex-

pected that the colloidal stability should be improved in living miniemulsion
polymerization on the premise that molecular weight controlling agents (RAFT
agent, nitroxide, and ATRP catalyst) do not need to be transferred from the
monomer reservoir to the polymerization loci. However, this strategy only gave
limited success. Problems included loss of colloidal stability, large particle
size, broad particle size distributions, and irreproducible particle sizes were
observed [221, 222, 226, 296, 305–307]. It is evident the stabilization of colloids
during living polymerization is more difficult than during regular miniemulsion
polymerization. The stability of latex seems to be sensitive to the recipe.
Georges and co-workers [222, 307] reported that, for the styrene miniemulsion
polymerization of TEMPO-mediated living polymerization, when the surfac-
tant concentration was reduced from 1.4%wt to 0.7%wt but the HD was kept
constant at 3 wt%, the system could be made stable. In an effort to commer-
cialize nitroxide-mediated miniemulsion polymerization, Georges [226] pro-
posed a modified miniemulsion SFRP process in which TEMPO-terminated
polystyrene oligomers were used to initiate the polymerization with 10 wt%
HD and 6.7 wt% sodium dodecylbenzenesulfonate (surfactant), yielding
polymers that have a high degree of livingness and stable latex. Charleux [223]
found that about 5 wt% hexadecane is needed for ST miniemulsion SFRP to get
stable latex, but with n-butyl acrylate, the proportion of HD must be decreased
to less than 1 wt%. For butyl acrylate SFRP in miniemulsion, it has been re-
ported that the particle size is not reproducible [226]. The El-Aasser group
reported successful miniemulsion polymerization that employed nitroxides
with controlled molecular growth and good miniemulsion stability [221]. It is
interesting to note that the level of costabilizer was more than 5%wt, far more
than the typical 2%wt. Even so, El-Aasser’s particle size is much larger than the
non-living counterpart, and the particle size distribution is much broader. In
the next paper of the El-Aasser group [308], TEMPO-terminated oligomers of
polystyrene were prepared via bulk polymerization of styrene, and they were
used as initiator in miniemulsion polymerization. The stable latexes were ob-
tained with 3 wt% HD and smaller particles. The Matyjaszewski group [309]
studied ATRP-controlled free radical miniemulsion polymerization and found
that using either anionic (SLS) or cationic (dodecyltrimethylamomonium
bromide) surfactant led to instability.It was argued that the catalyst may inter-
fere with SLS, while no reason was given for the instability of dodecyltri-
methylamomonium bromide. The final particle size was more than 1 mm in the
normal ATRP polymerization of butyl methacrylate, and more than 250 nm in
reverse ATRP. It has been reported that the use of nonionic polymeric surfac-
tant improved the stability. Also, the longer the PEO segment in the surfactant,
the better the stability. In the case of Tween 20, a portion of the monomer phase
separated, forming small pools in the reaction mixture, but these would dis-
sipate as the polymerization progressed, and there was no coagulation of
polymer at higher conversion. In another ATRP paper by Matyjaszewski et al.
[262], dNbpy, nonionic Brij 98, hexadecane, and the water-soluble azo com-
Miniemulsion Polymerization 237

pound 2,2¢-azobis (2-methylpropinonamide) dihydrocholoride (VA-50), were

used as ligand, surfactant, costabilizer, and initiator, respectively. The resulting
latexes showed improved stability. It is worth noting that a large amount of sur-
factant (5.0%wt, 13.5%wt based on monomer) and HD (1/10 [v/v monomer])
was used but particle sizes are large too (around 300 nm). In the most recent
paper by the same group [264], it was reported that the amount of Brij 98 and
HD can be reduced to 2.3%wt and 3.6%wt, respectively, by replacing VA-50 by
the more reactive VA-44, and replacing dNbpy by hydrophobic hexa-substituted
TREN with high catalytic activity, which gives polymerization with good
colloidal stability and particle sizes of 200–250 nm. De Brouwer et al. [296, 305]
reported stability problems in controlled free radical miniemulsion polymer-
ization using RAFT. It was reported that when an ionic surfactant (either
cationic or anionic) was used, a monomer bulk phase constituting up to 35%
of the total organic material in the system could be observed at low monomer
conversion. The phenomenon can be seen with various different monomers,
different initiator systems, and different costabilizers. Sanderson [310] also
reported these instability issues at less than 5 wt% SDS. De Brouwer et al. [305]
found that when nonionic surfactant is used, the polymerization is well
controlled in terms of molecular weight, and the colloidal stability is good.
However, both Butté et al. [261] and Lansalot et al. [298] did not report any in-
stability issues in RAFT miniemulsion polymerization. In Butté’s work, 3.3%wt
HD and 1.67%wt SDS was employed. The miniemulsion were prepared by a
three-step approach: after mechanical preemulsification for 10 minutes and
sonication for 20 minutes, the mixture was passed into a microfluidizer ten
times. No colloidal characteristics were reported. In Lansalot’s work, 2%wt
hexadecane, 1%wt PST, and 0.01 mol/L water SLS were used. The miniemulsion
was prepared by ultrasonication for a period of 7 minutes at 35% amplitude,
30–35 W power, in a Branson 450 sonicator. In Lansalot’s work, although stable
latex was obtained, the particle size was much bigger than the corresponding
non-living system (except when oligomeric RAFT agent was used), and particle
size distribution was broad. This seems to indicate that miniemulsification
procedure plays some role in obtaining stable latex. To avoid colloidal in-
stability, Sanderson [310, 311] successfully developed two approaches based on
the predictions of Luo’s theory [128]:
a. Form RAFT-endcapped oligomers in bulk, dissolve the RAFT agent into the
monomer, then disperse the monomer into water under shear with surfactant
and costabilizer, and recommence polymerization.
b. Use high surfactant concentrations (10%wt with respect to monomer), and
high costabilizer (n-HD) concentration (>4%wt with respect to monomer).
From the above summary of experimental investigations into living miniemul-
sion polymerization, we can see that controlled miniemulsion polymerization
(SFRP,ATRP, and RAFT) is less colloidal-stable during polymerization than its
non-living counterpart. Colloidal instability leads to phase separation in the
worst cases. In improved cases, the latex that results from the controlled mini-
238 F. J. Schork et al.

emulsion polymerization has much bigger particle size and much broader
particle size distribution than that obtained using regular miniemulsion poly-
merization. One can also see that colloidal instability is very sensitive to the
polymerization recipe. Monomer, surfactant, dormant agent, and levels of
surfactant and costabilizer all have large influences on the stability. High levels
of surfactant and costabilizer, and use of a nonionic surfactant and an oligo-
meric control agent all proved to aid stability, although larger particle sizes and
broader particle size distributions were still seen. On the other hand, the litera-
ture indicates that instability is a general problem in controlled free radical
miniemulsion polymerization, regardless of the living control mechanism,
monomer, surfactant system (except for polymeric surfactant), or initiator sys-
tem. It seems reasonable to assume that the instability is caused by the “living”
nature of the systems. In a controlled free radical polymerization system, the
kinetics of polymer chain formation is totally different from that for classical
free radical polymerization. In classical free radical polymerization, a polymer
chain is fully polymerized in about 1 s. At the very beginning of polymeriza-
tion, a few polymer chains of high molecular weight are formed. During poly-
merization, monomer is consumed to form more and more large polymer
chains. However, in controlled free radical polymerization, a large number of
oligomers are formed at the beginning of polymerization. During polymer-
ization, the oligomers gradually grow into large polymers. The feature common
to all of the controlled free radical systems is the presence of large concentra-
tions of oligomers early in the polymerization. In miniemulsion polymeriza-
tion, polymerization occurs in the particles (around 100 nm in size). Ugelstad
et al. [40] showed that oligomers are very efficient swelling agents, and
hence the existence of oligomers may dramatically modify the state of the
miniemulsion. A theoretical model has been developed by Luo et al. [128] to
simulate the swelling of oligomers formed in the controlled miniemulsion poly-
merization.
The chemical potential of monomer droplets is determined by [128]
–

 
1 2 c +
2V1
µd = RT (ln jd1 + 1 – 5 jd2 + j d2 12 RT (20)
m2 g rd
6

Before the start of polymerization, the monomer chemical potentials in all

droplets can be assumed to equal because the monomer has at least limited
solubility in the aqueous phase. However, once a monomer droplet is initiated,
a part of the monomer will polymerize into oligomers, and the droplet is con-
verted into a particle. The particle chemical potential is described by [312]

1
m2 
1

µd = RT ln jd1 + 1 – 5 jd2 + 1 – 5 jp3 + j 2p2 c12 + j p3
m3
2
c13
–

 
c13 2V1
+ jp2 jp3 c12 + c13 – 5 + 6 RT (21)
m2 g rd
Miniemulsion Polymerization 239

Fig. 21 Variation in the droplet chemical potential (top) and particle chemical potential
(bottom) during particle swelling (from [170])

The monomer chemical potential in the particles is lower than that in the
droplets, so that the monomer in the droplets will diffuse across the aqueous
phase and into the particles, leading to changes in the monomer chemical po-
tentials of the the particles and droplets. The change in the monomer chemical
potential is illustrated in Fig. 21, where Y is defined as the swelling capacity: the
ratio of the weight of a swollen particle to its weight before it is swollen.
As shown in Fig. 21, the particles swell with monomer diffusion (increase
in Y). During swelling, the monomer chemical potential in the particles first
rapidly increases and then decreases gradually down to zero with more and
more monomer swelling. On the other hand, the droplets shrink and the co-
stabilizer is concentrated since it cannot (by definition) diffuse out with the
monomer. The monomer chemical potential decreases monotonically. During
the process of monomer diffusion, if the monomer chemical potential in the
droplets is equal to that of the particles, equilibrium is established and monomer
diffusion ceases. In Fig. 21, the formation of high MW polymer is shown for
contrast. Three intersections of the droplet and particle chemical potential curves
can be seen. Two of these occur at low Y and the third at a much higher Y. As
is often the case with three equilibrium points, the middle point is unstable.
When the system arrives at the first intersection during swelling (lowest Y),
monomer transfer stops and the system reaches an equilibrium state that is
called the normal swelling state. In this case, the other two equilibrium points
will never be reached. However, in the case of controlled polymerization, the
formation of oligomers rather than high molecular weight polymer leads to a
lower mixing free energy so that the monomer in the particles has a lower
chemical potential. If the effect is large enough so that the chemical potential
of the droplets remains higher than that of the particles at the peak of the
particle chemical potential curve, the system will move to the right-most equi-
librium point. This will be denoted as the super-swelling state. In this case, a
large amount of monomer will transfer from the droplets to the particles.
240 F. J. Schork et al.

Super-swelling may be the cause of the stability problems encountered in

controlled miniemulsion polymerization. As the super-swelling equilibrium
point is approached, a large amount of monomer would transfer from a large
number of droplets to a small number of particles, which would cause the
monomer droplets to shrink and the particles to swell. This would broaden the
particle size distribution, or even destroy the miniemulsion. In the worst case,
the super-swelling would lead to a very large size difference between the
droplets and particles. The particles may be swollen to around 1 mm, a critical
size where the system becomes shear-sensitive and buoyant forces dominate.
Because shear is low in a miniemulsion polymerization reactor, the particles
would rapidly approach a breaking-coagulating dynamically balanced particle
size, as in suspension polymerization. In this case, particle size could be more
than 10 mm, or they may even form a bulk phase, depending on the shear field.
Alternatively, it is possible to destroy the miniemulsion using so-called hetero-
coagulation (small particles/droplets coagulating onto large ones) when the
size difference becomes large. The hypothesis that super-swelling causes the
instability is also supported by two other papers on controlled free radical
miniemulsion polymerization that used a degenerative transfer agent (C6F13I)
to control molecular weight. It was reported that the final latex morphology was
well controlled. C6F13I is a relatively inefficient transfer agent (Ctr =1.4 at 70 °C),
so the degree of polymerization of the product was rather high at the beginning
of polymerization. Super-swelling is less likely to occur in such a case.
Based on Luo’s simulations [128], it has been found that the super-swelling
state is rather sensitive to recipe variations. Simply increasing the costabilizer
level and/or using a nonionic polymeric surfactant would probably eliminate
super-swelling, and hence, the instability. More recently, Sanderson [310]
reported that two strategies could successfully form stable latex from RAFT
miniemulsion polymerization:
a. Replace common RAFT by RAFT-endcapped oligomers
b. Use high levels of SDS and HD
The fact that these two quite different approaches are both successful can
largely be reconciled using Luo’s theory. Replacing common RAFT by RAFT-
endcapped oligomers can avoid the super-swelling because the dangerous stage
of oligomer formation is avoided. Using oligomers of a molecular weight con-
trolling agent has also proved very helpful with SFRP in miniemulsion [226,
232]. High levels of SDS would lead to low interfacial tension, which helps to
suppress the super-swelling, and high HD levels would also suppress the super-
swelling. The super-swelling could also explain the instability or broad particle
size distribution in most cases, as discussed in the literature [128].
Super-swelling was postulated as a cause of instability in controlled free
radical miniemulsion polymerization. Asua [312] thought that the interfacial
tension used in the simulations was too high and the droplet size was too small.
However, it is turned out that the question about droplet size is due to a mis-
understanding. It is well accepted that a well-performed miniemulsion has
Miniemulsion Polymerization 241

Fig. 22 Plot of monomer conversion versus the number of input free radicals in a droplet
(dotted: D=50 nm; solid: D=100 nm) (from [170])

increased interfacial tension.Actually, the interfacial tension used in the simu-

lations was cited from Landfester’s work on miniemulsions [313]. Another
controversy is about the view of the nucleation process in controlled free radical
miniemulsion polymerization. In Luo’s work [312], it was found that super-
swelling happens only in the presence of a small fraction of particles with
rather high conversion (10%); the majority of droplets have zero conversion.
It is difficult to understand what would lead to such scenarios. However, Luo’s
simulations [179] showed that the nucleation process of RAFT miniemulsion
polymerization could be very different to that of regular miniemulsion poly-
merization. As shown in Fig. 22, the simulations suggest that, by introducing a
highly reactive RAFT agent, a large number of free radicals (Nc) need to be cap-
tured by a droplet before rapid polymerization in the droplet can take place,
which is totally different from the situation in regular miniemulsion polymer-
ization. More interestingly, it was found that droplet size had a significant
influence on Nc. In Fig. 23, droplets larger than 150 nm (about 8.24% of all
droplets) have been nucleated. Interestingly, monomer conversion for droplets
less than 130 nm in size (unnucleated) is less than 2.5%, while conversion for
droplets larger than 150 nm in size (nucleated particles) is much higher,
depending the radical flux. The nucleation process for ATRP or SFRP in mini-
emulsion has not been reported yet, though it is very important. However,
Charleux [215] has theoretically studied the segregation effect of the emulsion
for SFRP. The results showed that segregation is only effective for particles far
smaller than 130 nm (for example 50 nm). The polymerization rate in
miniemulsion, in most cases, is similar to that in homogeneous polymerization
because the droplet size miniemulsion polymerization usually is around
100 nm. However, in reality, droplet size is polydispersed, so we cannot exclude
the existence of very small droplets, where the segregation is effective. In such
a case, it is possible that a minor fraction of the very small particles has a much
higher polymerization rate than the majority of particles with larger particle
242 F. J. Schork et al.

Fig. 23 The monomer conversion in particles with different initial droplet sizes at various
average free radical fluxes (average of ten runs, in s–1). Filled squares: 0.02; filled triangles:
0.05; filled circles: 0.1 (from [107])

sizes where segregation is not effective. The same argument is suitable for ATRP
in miniemulsion.Additionally, it is well-known that it takes some time for SFRP
and ATRP to set up the propagation/dormant equilibrium. It is likely that the
time required to build up the equilibrium is droplet-size dependent. In such
cases, the scenarios above might also occur. In fact, it has been reported that the
reverse ATRP in miniemulsion has a higher colloidal stability than the direct
ATRP in miniemulsion [264].
The above argument suggests that the initial droplet size distribution may
play an important role in super-swelling or colloidal instability. This indicates
that one should monitor the emulsification procedure for controlled free
radical emulsion polymerization closely.

4.5
Other Applications and Future Directions

Recently, many new reactions have been carried out in miniemulsions. Most of
these are polymerizations, but a number of nonpolymerization reactions have
been proposed. This section will survey these applications, and close with some
speculation on the future of miniemulsions.

4.5.1
Anionic/Cationic Polymerization in Miniemulsions

Maitre et al. [314] carried out anionic polymerization of phenyl glycidyl ether
(PGE) in miniemulsion using didodecyldimethylamonium hydroxide as an
inisurf (combination initiator and surfactant). Long chain alcohols were used
as the costabilizer and stable miniemulsions were created by sonication.
Monomer conversion was low, as was the degree of polymerization, which only
Miniemulsion Polymerization 243

reached eight. The degree of polymerization was dependent on the initiator

concentration, the type of alcohol, and its concentration. Comparison with bulk
polymerization suggests that polymerization takes place near the droplet sur-
face, since no high molecular weight material (expected from bulk polymer-
ization in the droplet core) is formed. The authors postulate that initiation and
propagation take place near the droplet-water interface, initiated by the inisurf.
Termination with water takes place after a few propagations. However, the
oligomers formed would be surface active and are thought to adsorb at the
interface and increase the solubility of PGE in the locus of interfacial polymer-
ization, enhancing subsequent propagation reactions. One might conjecture
that with a hydrophobic initiator, polymerization in the droplet core might be
encouraged, resulting in higher degrees of polymerization.
The same researchers [315, 316] reported the anionic ring opening poly-
merization of 1,3,5-tris(trifluoropropylmethyl)cyclotrisiloxane in miniemul-
sion using didodecyldimethylammonium bromide as the surfactant and sodium
hydroxide as the initiator. Molecular weights were 2000–30,000. A two state
mechanism was put forward, consisting of anionic kinetically-controlled ring
opening polymerization continuing to complete conversion, followed by con-
densation and backbiting reactions. The delay between the two stages was long
enough to allow high polymer yield.
These same researchers [317] reported the anionic polymerization of n-butyl
cyanoacrylate in macroemulsion and miniemulsion. Dodecylbenzenesulfonic
acid (DBSA) was used as the surfactant. The DBSA slows the rate of interfacial
anionic polymerization through reversible termination, preventing an un-
desirably high degree of polymerization. Polymerization in macroemulsion
resulted in a much higher degree of polymerization, perhaps due to droplet
polymerization where the interface is less significant.
This same research group also reported [318] the cationic polymerization of
p-methoxystyrene in miniemulsion. DBSA was used as both a protonic initiator
and surfactant. A monomer conversion of 100% was achieved in eight hours
at 60 °C. Molecular weights were low (approximately 1,000) and solids of up to
40% could be achieved with good colloidal stability. Polymerization takes place
at the interface, initiated by the proton, and terminated by water. Molecular
weight increased with conversion, suggesting either reversible termination or
decreasing termination.
While all of these results are far from providing commercial products, they
highlight the possibilities for alternative polymerization chemistries in mini-
emulsions.

4.5.2
Polycondensation in Miniemulsions

Barrère and Landfester [319] reported the synthesis of polyester in mini-

emulsions. Hydrophobic polyesters were synthesized in miniemulsion. DBSA
was used as a catsurf (catalyst and surfactant) and HD was used as the co-
244 F. J. Schork et al.

stabilizer. The pH was kept below 4 in order to prevent deprotonation of the

acid monomers. Molecular weights of approximately 1000–2000 were found,
with yields generally 70–80%. Both molecular weights and yields varied with
monomer choice. Most interestingly, the presence of the particle-water inter-
face did not change the polymerization-depolymerization equilibrium; the
yield was the same in 100 nm particles as in very large droplets. However, the
water concentration within the particle was found to be critical. Highly hydro-
phobic monomers reduced the monomer concentration in the particles, push-
ing the equilibriums toward esterification, and increasing yield.Alcohols bear-
ing electron-donating groups were found to displace the equilibrium toward
ester formation. They also report the formation of polyester-polystyrene hybrid
particles using a one-pot procedure. In a separate paper [319] the same authors
report one-pot polymerizations of polyurethane/acrylic hybrids. The proce-
dures for both the polyester/polystyrene and the polyurethane/polyacrylate
polymerizations were similar. First, the entire system was miniemulsified; then
the polycondensation took place. After the polycondensation, a free radical
initiator was added and the addition monomers were polymerized. Two mole-
cular weight peaks were found for the polyurethane/polyacrylate system, a low
one for the polyurethane, and a much higher MW peak for the polyacrylate.
Hydroxybutyl acrylate, when added, was found to be a crosslinking agent, since
it can undergo polycondensation and polyaddition.

4.5.3
Other Polymerizations in Miniemulsions

Extremely hydrophobic monomers do not polymerize well via macroemulsion

polymerization due to their very low rates of monomer transport across the
aqueous phase. Obviously, these monomers can be polymerized much more
effectively in a miniemulsion system. One example of this is provided by
Landfester et al. [320]. In this paper, fluoroalkyl acrylates are polymerized in a
miniemulsion with low levels of a protonated surfactant. When fluorinated
monomers were copolymerized with standard hydrophobic and hydrophilic
monomers, either core-shell structures or statistical copolymers were formed.
A similar situation occurs with vinyl chloride (VC) for a very different
reason. Vinyl chloride is very soluble in water, but polyvinyl chloride (PVC) is
not soluble in its own monomer. VC does swell PVC, and for that reason, there
is a driving force for VC transport across the aqueous phase in macroemulsion
polymerization. This transport is aided by the fact that VS is very soluble in
water. However, this is one macroemulsion system that might greatly benefit
from the miniemulsion synthesis route.
Willert and Landfester [321] have polymerized amphiphilic copolymers
from miniemulsion systems. The hydrophobic monomer was miniemulsified,
while the hydrophilic monomer resides in the continuous phase. Polymeriza-
tion was found to take place in the droplet phase, at the interface, or in the
continuous phase; the quality of the product depended strongly on the primary
Miniemulsion Polymerization 245

locus of polymerization. Inverse miniemulsions (oil as the continuous phase,

with water-soluble monomer dissolved in water droplets) was also used.
Marie et al. [322] have reported the synthesis of polyaniline particles via
inverse and direct miniemulsion. Inverse miniemulsions of anilinium hydro-
chloride were oxidized by hydrogen peroxide, resulting in highly crystalline
polyaniline. Oxidation of aniline miniemulsions in water also leads to highly
crystalline polyaniline. The same research group reports the use of chitosan as
a surfactant for miniemulsions [338], resulting in latex particles with functional
biopolymer surfaces for grafting in biological applications. Taden et al. [324]
report the enzymatic polymerization of lactone to form biodegradable nano-
particles.
Claverie et al. [325] have polymerized norbornene via ROMP using a con-
ventional emulsion polymerization route. In this case the catalyst was water-
soluble. Particle nucleation was found to be primarily via homogenous nuclea-
tion, and each particle in the final latex was made up of an agglomeration of
smaller particles. This is probably due to the fact that, unlike in free radical
polymerization with water-soluble initiators, the catalyst never entered the
polymer particle. Homogeneous nucleation can lead to a less controllable
process than droplet nucleation (miniemulsion polymerization). This system
would not work for less strained monomers, and so, in order to use a more
active (and strongly hydrophobic) catalyst, Claverie employed a modified mini-
emulsion process. The hydrophobic catalyst was dissolved in toluene, and sub-
sequently, a miniemulsion was created. Monomer was added to swell the toluene
droplets. Reaction rates and monomer conversion were low, presumably be-
cause of the proximity of the catalyst to the aqueous phase due to the small
droplet size.

4.5.4
Other Miniemulsion Applications

The miniemulsification technique can also be applied to nonpolymerization

systems. The number of nonpolymerization applications of miniemulsions is
small, but seems to be growing. Landfester [326] has reviewed the generation
of nanoparticles in miniemulsions. Revelino et al. [327] has studied crystal-
lization from direct and inverse miniemulsions. It was found that, since each
droplet must be nucleated separately, the undercooling necessary to effect
crystallization increases dramatically. The crystallization rate was higher in
miniemulsions and proportional to droplet size. Taden et al. [328] studied the
crystallization of polyethylene oxide from miniemulsions. Crystallization only
occurred at large supercooling. Drying of the crystallized dispersion resulted
in a highly ordered arrangement of polyethylene oxide platelets. Montenegro
et al. [329] studied crystallization of alkanes from miniemulsions.Wegner et al.
[330] have used polymeric nanospheres produced via miniemulsion polymer-
ization to control nucleation and growth of inorganic crystals from aqueous
media.
246 F. J. Schork et al.

Vaihinger et al. [331] have molecularly imprinted polymer nanospheres as

synthetic affinity receptors via miniemulsion techniques. Landfester et al. [332]
have created semiconducting polymer nanospheres in aqueous dispersions
via the synthetic miniemulsion technique. That is, conducting polymers were
dissolved in solvent and then miniemulsified. Films produced by spin coating
retained the nanosphere character until they were annealed above the glass
transition temperature. In similar work, Piok et al. [333–335] formed organic
light-emitting devices fabricated from semiconducting nanospheres created
by the miniemulsification process. Willert et al. [336] created inorganic and
metallic nanoparticles by the miniemulsification of molten salts and metals. Zu
Putlitz et al. [337] created “armored latexes” and hollow inorganic shells made
of clay sheets by templating cationic miniemulsions and latexes.

4.5.5
Future Directions

It would seem that miniemulsions have finally moved from being a laboratory
curiosity to being a viable commercial process and a useful synthetic technique
for producing interesting materials with nano-scale structure. It would appear
that polymer-polymer hybrids and polymer-inorganic hybrids achieved via
miniemulsion polymerization will result in new classes of water-borne mate-
rials. Other, traditionally solvent-based, polymerization chemistries may soon
be carried out routinely via the miniemulsion route due to improvements in
polymerization catalysts. The use of miniemulsions in ROMP has been cited
above. Metallocene polymerization of ethylenic monomers has been carried
out in macroemulsion. A short review and discussion of this work is given in
[338]. If these water-tolerant polymerization chemistries are successful, it
cannot be long before they are ported into miniemulsions. Controlled radical
polymerization, particularly using RAFT chemistry, is a natural application for
miniemulsion technology. Perhaps most importantly, miniemulsion techniques
will be used in a variety of nonpolymerization technologies to produce nano-
scale, highly structured materials.

References
1. Asua JM (2002) Prog Polym Sci 27:1283
2. Guyot A (2001) Curr Trends Polym Sci 6:47
3. Antonietti M, Landfester K (2002) Prog Polym Sci 27:689
4. Trommsdorf E, Schlidknecht E (1956) Suspension polymerization in polymer
processes. Interscience, New York
5. Ugelstad J, El-Aasser MS, Vanderhoff JW (1973) J Polym Sci Pol Lett 11:503
6. Gardon JL (1970) Brit Polym J 2:1
7. Poehlein GW (1982) In: Piirma I (ed) Emulsion polymerization. Academic, New York,
p 357
8. Poehlein GW (1985) ACS Sym Ser 285:131–150
Miniemulsion Polymerization 247

9. Poehlein GW, Dougherty DJ (1976) Rubber chemistry and technology 50:601

10. Song Z (1988) PhD Thesis, Georgia Institute of Technology, Atlanta, GA
11. Ugelstad J, Hansen FK (1976) Rubber Chem Technol 49:536
12. Harkins WD (1947) J Am Chem Soc 69:1428
13. Smith WV, Ewart RH (1948) J Chem Physics 16:592
14. Alexander AE, Napper DH (1971) Prog Polym Sci 3:145
15. Blackley DC (1982) Emulsion polymerization. Academic, New York
16. Baxendale JH, Bywater S, Evans MG (1946) T Faraday Soc 42:675
17. Baxendale JH, Evans MG, Kilham JK (1946) T Faraday Soc 42:668
18. Priest WJ (1952) Phys Chem J 56:1077
19. Roe CP (1968) Ind Eng Chem 60:20
20. Fitch RM, Tsai CH (1971) In: Fitch RM (ed) Polymer colloids. Plenum, New York,
p 73
21. Fitch RM, Tsai CH (1971) In: Fitch RM (ed) Polymer colloids. Plenum, New York,
p 103
22. Fitch RM (1973) Brit Polym J 5:467
23. Lichti G, Gilbert RG, Napper DH (1983) J Polym Sci Pol Chem 21:269
24. Feeney PJ, Napper DH, Gilbert RG (1984) Macromolecules 17:2520
25. Maxwell IA, Morrison BR, Napper DH, Gilbert RG (1991) Macromolecules 24:1629
26. Chamberlain BJ, Napper DH, Gilbert RG (1982) J Chem Soc Farad T I 78:591
27. Ugelstad J, Kaggerud KH, Hansen FK, Berge A (1979) Makromol Chem 180:737
28. Ugelstad J, Mfutakamba HR, Mork PC, Ellingsen, Berge A, Schmidt R, Holm L, Jorgedel
A, Hansen FK, Nustad K (1985) J Polym Sci Pol Sym 72:225
29. Hansen FK, Ofstad EB, Ugelstad J (1976) Theory and practice of emulsion technology.
Academic, New York
30. Ugelstad J, Flagstad H, Hansen FK, Ellingsen T (1973) J Polym Sci 42:473
31. Hansen FK, Ugelstad J (1979) J Polym Sci 17:3047
32. Morton M, Kaizermann S, Altier MW (1954) J Colloid Interf Sci 9:300
33. Beuermann S, Buback M (2002) Prog Polym Sci 27:191
34. Sundberg DC, Hsieh JY, Soh SK, Baldus RF (1982) ACS Sym Ser 165:327
35. Buback M, Garcia-Rubio LH, Gilbert RG, Napper DH, Gulliot J, Hamielec AE, Hill D,
O’Driscoll KF, Olaj OF, Shen J, Solomon D, Moad G, Stickler M, Tirrell M, Winnick MA
(1988) J Polym Sci Pol Phys 26:293
36. Barnette DT, Schork FJ (1987) Chem Eng Prog 83:25
37. Fris N, Hamielec AE (1973) J Polym Sci Pol Chem 11 3321
38. Rawlings JB, Ray WH (1988) Polym Eng Sci 28:257
39. Ballard MJ, Gilbert RG, Napper DH, Pomery PJ, O’Sullivan PW, O’Donnell JH (1986)
Macromolecules 19:1303
40. Soh SK, Sundberg DC (1982) J Polym Sci Pol Chem 20:1331
41. Ostwald WZ (1901) Z F Phys Chem 37:495
42. Fontenot K, Schork FJ (1993) Ind Eng Res 32:373
43. Chern CS, Chen TJ (1998) Colloids Surface A 138:65
44. Flory PJ (1953) Principles of polymer chemistry. Cornell Univ Press, Ithaca, NY
45. Ugelstad J, Mork PC, Kaggerud KH, Ellingsen T, Berge A (1980) Adv Colloid Interfac
13:101
46. Jansson LH, Wellons MC, Poehlein GW (1983) J Polym Sci Pol Lett 21:937
47. Higuchi WI, Misra J (1962) J Pharm Sci 51:459
48. Ugelstad J (1978) Makromol Chem 179:815
49. Fontenot K (1991) PhD Thesis, Georgia Institute of Technology, Atlanta, GA
50. Azad ARM, Fitch RM (1980) In: Polymer colloids II. Plenum, New York, p 95
248 F. J. Schork et al.

51. El-Aasser MS, Lack CD, Choi YT, Min TI, Vanderhoff JW, Fowkes FM (1984) Colloids
Surface 12:79
52. Ugelstad J, Kaggerud KH, Fitch RM (1980) In: Fitch RM (ed) Polymer colloids II.
Plenum, New York, p 83
53. Choi YT, El-Aasser MS, Sudol ED, Vanderhoff JW (1985) J Polym Sci Pol Chem 23:
2973
54. Chamberlain BJ, Napper DH, Gilbert RG (1982) J Chem Soc Farad T 1 78:591
55. Chen CM, Gothjelpsen L, Schork FJ (1986) Polym Proc Eng 4:1
56. Delgado J, El-Aasser MS, Silebi CA, Vanderhoff JW (1986) Polym Mat Sci Eng 54:444
57. Rodriguez VS, Delgado J, Silebi CA, El-Aasser MS (1988) Polym Mat Sci Eng 58:761
58. Delgado J, El-Aasser MS, Vanderhoff JW (1986) J Polym Sci Pol Chem 24:861
59. Asua JM, Rodriguez VS, Silebi CA, El-Aasser MS (1990) Makromol Chem M Symp
35–36:59
60. Rodriguez VS, Delgado J, Silebi CA, El-Aasser MS (1989) Ind Eng Chem Res 28:65
61. Rodriguez VS, Asua JM, El-Aasser MS, Silebi CA (1991) J Polym Sci Pol Physics 29:483
62. Fontenot K, Schork FJ (1992) Polym Reaction Engineering 1:75
63. Fontenot K, Schork FJ (1993) Polym React Eng 1:289
64. Samer CJ, Schork FJ (1997) Polym React Eng 5:85
65. Ma JW, Cunningham MF, McAuley KB, Keoshkerian B, Georges MK (2003) Macromol
Theor Simul 12:72
66. Ma JW, Cunningham MF, McAuley KB, Keoshkerian B, Georges M (2003) Chem Eng Sci
58:1177
67. Ma JW, Smith JA, McAuley KB, Cunningham MF, Keoshkerian B, Georges MK (2003)
Chem Eng Sci 58:1163
68. Ma JW, Cunningham MF, McAuley KB, Keoshkerian B, Georges MK (2002) Macromol
Theor Simul 11:953
69. Samer CJ, Schork FJ (1999) Ind Eng Res 38:1801
70. Landfester K, Bechthold N, Tiarks F, Antonietti M (1999) Macromolecules 32:2679
71. Bechthold N, Tiarks F, Willert M, Landfester K, Antonietti M (2000) Macromol Symp
151:549
72. Bradley M, Grieser F (2002) J Colloid Interf Sci 251:1
73. Wang S, Schork FJ (1994) J Appl Polym Sci 54:2157
74. Chern CS, Chen TJ (1997) Colloid Polym Sci 275:1060
75. Chern CS, Liou YC (1999) Polymer 40:3763
76. Wu XQ, Schork FJ (2001) J Appl Polym Sci 81:1691
77. Graillat C, Guyot A (2003) Macromolecules 36:6371
78. Boisson F, Uzulina I, Guyot A (2001) Macromol Rapid Comm 22:1135
79. Lack CD, El-Aasser MS, Vanderhoff JW, Fowkes FM (1985) ACS Sym Ser 272:345
80. Azad ARM, Ugelstad J, Hansen FK (1976) ACS Sym Ser 24:1
81. Hawkett BS, Napper DH, Gilbert RG (1980) J Chem Soc Farad T I 76:1323
82. Hallworth GW, Carless JE (1974) Theory and practice of emulsion technology. Acade-
mic, New York, p 305
83. Choi YT, El-Asser MS, Sudol ED, Vanderhoff JW (1985) J Appl Polym Sci 23:2973
84. Choi YT (1986) PhD Thesis, Lehigh University, Bethlehem, PA
85. Friberg SE, Neogi P (1986) Disp Sci Tech J 7:50
86. Kislalioglu S, Friberg S (1976) Theory and practice of emulsion technology. Academic,
New York, p 257
87. Lack CD, El-Aasser MS, Silebi CA, Vanderhoff JW, Fowkes FM (1987) Langmuir 3:1155
88. Shah DO, Schechter RS (1977) Improved oil recovery by surfactant and polymer flood-
ing. Academic, New York
Miniemulsion Polymerization 249

89. El-Aasser MS, Lack CD, Choi YT, Min TI, Vanderhoff JW, Fowkes FM (1984) Colloids
Surface 12:79
90. Rodriguez VS (1988) PhD Thesis, Lehigh University, Bethlehem, PA
91. Delgado J, El-Aasser MS, Vanderhoff JW (1986) J Polym Sci Pol Chem 24:861
92. Reimers JL, Skelland AHP, Schork FJ (1995) Polym React Eng 3:235
93. Ugelstad J (1980) Adv Colloid Interfac 13:101
94. Reimers JL, Schork FJ (1996) J Appl Polym Sci 60:251
95. Reimers JL, Schork FJ (1996) J Appl Polym Sci 59:1833
96. Aizpurua I, Amalvy JI, Barandiaran MJ (2000) Colloids Surface A 166:59
97. Lelu S, Novat C, Graillat C, Guyot A, Bourgeat-Lami E (2003) Polym Int 52:542
98. Wang ST, Schork FJ, Poehlein GW, Gooch JW (1996) J Appl Polym Sci 60 2069
99. Wu XQ, Schork FJ, Gooch JW (1999) J Polym Sci Pol Chem 37:4159
100. Tsavalas JG, Gooch JW, Schork FJ (2000) J Appl Polym Sci 75:916
101. Dong H, Gooch JW, Schork FJ (2000) J Appl Polym Sci 76:105
102. Reimers JL, Schork FJ (1996) Polym React Eng 4:135
103. Chern C-S, Sheu J-C (2000) J Polym Sci Pol Chem 38:3188
104. Samer CJ, Schork FJ (1999) Ind Eng Res 38:1792
105. Mouron D, Reimers J, Schork FJ (1996) J Polym Sci Pol Chem 34:1073
106. Wang S, Poehlein GW, Schork FJ (1997) J Polym Sci Pol Chem 35:595
107. Reimers JL, Schork FJ (1997) Ind Eng Resh 36:1085
108. Asua J, Alduncin J, Forcada J (1994) Macromolecules 27:2256
109. Miller CM, Blythe PJ, Sudol ED, Silebi CA, El-Aasser MS (1994) J Polym Sci Pol Chem
32:2365
110. Miller CM, Sudol ED, Silebi CA, El-Aasser MS (1995) Macromolecules 28:2754
111. Miller CM, Sudol ED, Silebi CA, El-Aasser MS (1995) Macromolecules 28:2765
112. Miller CM, Sudol ED, Silebi CA, El-Aasser MS (1995) Macromolecules 28:2772
113. Blythe PJ, Morrison BR, Mathauer KA, Sudol ED, El-Aasser MS (1999) Macromolecules
32:6944
114. Blythe PJ, Klein A, Sudol ED, El-Aasser MS (1999) Macromolecules 32:6952
115. Blythe PJ, Klein A, Sudol ED, El-Aasser MS (1999) Macromolecules 32:4225
116. Blythe PJ, Morrison BR, Mathauer KA, Sudol ED, El-Aasser MS (2000) Langmuir
16:898
117. Ghazaly HM, Daniels ES, Dimonie VL, Klein A, El-Aasser MS (2001) J Appl Polym Sci
81:1721
118. Luo Y, Schork FJ (2002) J Polym Sci Pol Chem 40:3200
119. Alduncin JA, Asua JM (1994) Polymer 35:3758
120. Blythe PJ, Klein A, Phillips JA, Sudol ED, El-Aasser MS (1999) J Polym Sci Pol Chem
37:4449
121. Gilbert RG (1995) Emulsion polymerization: A mechanistic approach.Academic, London
122. Kabalnov AS, Pertzov AV, Shchukin ED (1987) J Colloid Interf Sci 118:590
123. Candau F, Pabon M, Anquetil J-Y (1999) Colloid Surface A 153:47
124. Sood A, Awasthi KJ (2003) Appl Polym Sci 88:3058
125. Erdem B, Sully Y, Sudol ED, Dimonie VL, El-Aasser MS (2000) Langmuir 16:4890
126. Landfester K, Bechthold N, Föster S,Antonietti M (1999) Macromol Rapid Comm 20:81
127. Landfester K, Bechthold N, Tiarks F, Antonietti M (2000) Macromolecules 32:5222
128. Luo Y, Tsavalas JG, Schork FJ (2001) Macromolecules 34:5501
129. Melis S, Kemmere M, Meuldijk J, Storti G, Morbidelli M (2000) Chem Eng Sci 55:3101
130. Leiza JR, Sudol ED, El-Aasser MS (1997) J Appl Polym Sci 64:1797
131. Tang PL, Sudol ED, Adams M, El-Aasser MS, Asua JM (1991) J Appl Polym Sci 42:
2019
250 F. J. Schork et al.

132. Sajjadi S, Jahanzad F (2003) Euro Polym J 39:785

133. Durant YG (1999) Book of Abstracts 217th ACS National Meeting (PMSE-326),
Anaheim, CA, 21–25 March 1999, American Chemical Society, Washington, DC
134. Ouzineb K, Graillat C, McKenna TF (2001) DECHEMA Monographien (7th Int
Workshop on Polymer Reaction Eng), Hamburg, Germany, 8–10 October 2001, 137:
293
135. Barnette DT, Schork FJ (1989) Chem Eng Commun 80:113
136. Nomura M, Harada M (1981) ACS Sym Ser 165:121
137. Samer CJ, Schork FJ (1999) Ind Eng Res 38:1792
138. Aizpurua I, Barandiaran M (1999) Polymer 40:4105
139. Aizpurua I, Amalvy JI, de la Cal JC, Barandiaran MJ (2000) Polymer 42:1417
140. Fontenot K, Schork FJ (1993) J Appl Polym Sci 49:633
141. Odian G (1988) Principles of polymerization, 3rd edn. Wiley-Interscience, New York
142. Landfester K, Schork FJ, Kusuma VA (2003) CR Acad Sci 6(11–12):1337–1342
143. Landfester K, Bechthold N, Tiarks F, Antonietti M (1999) Macromolecules 32:5222
144. Reimers JL, Schork FJ (1996) J Appl Polym Sci 59:1833
145. Landfester K, Bechthold N, Tiarks F, Antonietti M (1999) Macromolecules 32:2679
146. Rodrigues JC, Schork FJ (1997) J Appl Polym Sci 66:317
147. Balic R (2000) PhD Thesis, University of Sydney, Sydney, Australia
148. Kitzmiller EL, Miller CM, Sudol ED, El-Aasser MS (1995) Macromol Symp 92:157
149. Mayo VF, Lewis FM (1944) J Am Chem Soc 66:1594
150. Schuller H (1986) In: Reichert K, Geisler W (eds) Polymer reaction engineering. Hüthig
and Wepf, Heidelberg, Germany, p 137
151. Guillot J (1986) In: Reichert K, Geisler W (eds) Polymer reaction engineering. Hüthig
and Wepf, Heidelberg, Germany, p 147
152. Wu XQ, Schork FJ (2000) Ind Eng Chem Res 39:2855
153. Delgado J, El-Aasser MS, Silebi CA, Vanderhoff JW (1987) Polym Mat Sci Eng 57:976
154. Delgado J, El-Aasser MS, Silebi CA, Vanderhoff JW, Guillot J (1987) In: El-Aasser MS,
Fitch RM (eds) NATO ASI Ser E: Appl Sci 138:79
155. Delgado J, El-Aasser MS, Silebi CA, Vanderhoff JW (1988) Makromol Chem M Symp
20/21:545
156. Delgado J, El-Aaser MS, Silebi CA, Vanderhoff JW (1989) J Polym Sci Pol Chem
27:193
157. Delgado J, El-Aasser MS, Silebi CA,Vanderhoff JW (1990) J Polym Sci Pol Chem 28:777
158. Unzue MJ, Asua JM (1993) J Appl Polym Sci 49:81
159. Wu XQ, Hong XM, Schork FJ (2002) J Appl Polym Sci 85:2219
160. Caruso F, Caruso RA, Mohwald H (1998) Science 282:1111
161. Caruso F (2001) Adv Mater 13:11
162. Numo-Donlunas S, Rhoton AI, Corona-Galvan S, Puig JE, Kaler JE (1993) Polym Bull
30:207
163. Antonova LF, Leplyanin GV, Zayev YY, Rafikov SR (1978) Polym Sci USSR 20:778
164. Turner SR, Weiss RA, Lundberg RD (1985) J Polym Sci Pol Chem 37:535
165. Kim JH, Chainey M, El Aasser MS, Vanderhoff JW (1992) J Polym Sci Pol Chem 30:535
166. Emelie B, Pichot D, Guillot J (1988) Makromol Chem 189:1879
167. Shoaf GL, Poehlein GW (1992) Polym React Eng 9:1
168. Charmot D, D’Allest JF, Dobler F (1996) Polymer 37:5237
169. Ganachaud F, Sauzedde F, Elaissari A, Pichot C (1997) J Appl Polym Sci 65:2315
170. Luo YW, Schork FJ (2001) J Polym Sci Pol Chem 39:2696
171. Kim JH, Chainey M, El Aasser MS, Vanderhoff JW (1990) J Polym Sci Pol Chem
28:3188
Miniemulsion Polymerization 251

172. Gilbert RG, Anstey JF, Subramaniam N, Monteiro MJ (1999) ACS Polym Prepr Div
Polym Chem 40:102
173. Blackley DC (1975) Emulsion polymerization. Wiley, New York, p 244
174. Dong H, Gooch JW, Poehlein GW, Wang ST, Wu X, Schork FJ (2001) ACS Sym Ser
766:8–17
175. Nabuurs T, Baijards RA, German AL (1996) Prog Org Coat 27:163
176. Van Hamersveld EMS, van Es GS, Cuperus FP (1999) Colloids Surface A 153:285
177. Van Hamersveld EMS,Van Es GS, German AL, Cuperus FP,Weissenborn P, Hellgren AC
(1999) Prog Org Coat 35:235
178. Tsavalas JG, Luo Y, Hudda L, Schork FJ (2003) Polym React Eng 11:277
179. Tsavalas JG, Luo Y, Schork FJ (2003) J Appl Polym Sci 87:1825
180. Tsavalas JG, Schork FJ, Landfester K (2003) J Coating Technol (in press)
181. Shoaf GL, Stockl RR (2003) Polym React Eng 11:319
182. Li, Daniels ES, Dimonie VL, Sudol ED, El-Aasser MS (2001) Polym Mat Sci Eng 85:258
183. Wang C, Chu F, Graillat C, Guyot A (2003) Polym React Eng 11:541
184. Barrere M, Landfester K (2003) Macromolecules 36:5119
185. El-Aasser MS, Li M, Jeong P, Daniels ES, Dimonie VL, Sudol ED (2001) DECHEMA
Monographien (7th Int Workshop on Polymer Reaction Engineering), Hamburg,
Germany, 8–10 October 2001, 137:1
186. Jeong P, Dimonie VL, Daniels ES, El-Aasser MS (2002) ACS Sym Ser 801:357
187. Kawahara H, Goto T, Okamoto Y, Kage H, Ogura H, Matsuno Y (2002) Kagaku Kogaku
Ronbun 28:175
188. Landfester K, Dimonie VL, El-Aasser MS (1998) DECHEMA Monographien 134 (6th
Int Workshop on Polymer Reaction Engineering 1998), Berlin, 5–7 October 1998,
134:469
189. Roberts JE, Marcu I, Dimonie V, Daniels E, El-Aasser MS (2003) ACS Polym Prepr Div
Polym Chem 44:277
190. Marcu I, Daniels ES, Dimonie VL, Hagiopol C, Roberts JE, El-Aasser MS (2003) Macro-
molecules 36:328
191. El-Aasser MS, Vanderhoff JW, Poehlein GW (1977) Coating Plastic Prepr 37:92
192. El-Aasser MS, Hoffman JD, Manson JA,Vanderhoff JW (1980) Org Coating Plast Chem
43:136
193. Vanderhoff JW, El-Aasser MS, Hoffman JD (1978) US Patent 4070323
194. Noh MH, Jang LW, Lee DC (1999) J Appl Polym Sci 74:179
195. Noh MH, Lee DC (1999) J Appl Polym Sci 74:2811
196. Garcés JM, Moll DJ, Bicerano J, Fibinger R, McLeod DG (2000) Adv Mater 12:1835
197. Van Herk AM, German AL (1999) Microsph Microcaps Lipos 1:457
198. Erdem B, Sudol ED, Dimonie VL, El-Aasser MS (2000) J Polym Sci Pol Chem 38:4419
199. Erdem B, Sudol ED, Dimonie VL, El-Aasser MS (2000) J Polym Sci Pol Chem 38:4431
200. Erdem B, Sudol ED, Dimonie VL, El-Aasser MS (2000) J Polym Sci Pol Chem 38:4441
201. Tiarks F, Landfester K, Antonietti M (2001) Langmuir 17:5775
202. Antonietti M, Landfester K (2002) Chem Ing Tech 74:543
203. Landfester K, Montenegro R, Scherf U, Guntner R, Asawapirom U, Patil S, Neher D,
Kietzke T (2002) Adv Mater 14:651
204. Ramirez LP, Landfester K (2003) Macromol Chem Phys 204:22
205. Solomon DH, Rizzardo E, Cacioli P (1985) Eur Pat Appl EP 135280
206. Georges MK, Moffat KA, Veregin RPN, Kazmaier PM, Hamer GK (1993) Polym Mater
Sci Eng 69:305
207. Fischer H (1997) Macromolecules 30:5666
208. Gridnev AA (1997) Macromolecules 30:7651
252 F. J. Schork et al.

209. Moad G, Solomon DH (1995) The chemistry of free radical polymerization, 1st edn.
Elsevier, Amsterdam
210. Fukuda T, Terauchi T, Goto A, Ohno K, Tsujii Y, Miyamoto T, Kobatake S, Yamada B
(1996) Macromolecules 29:6393
211. Fukuda T (2002) Handbook of radical polymerization. Wiley, New York, Ch 9
212. Qiu J, Charleux B, Matyjaszewski K (2001) Prog Polym Sci 26:2083
213. Butte A, Storti G, Morbidelli M (1998) In: Reicher KH, Moritz HU (eds) 6th Int Work-
shop on Polymer Reaction Engineering. DECHEMA Monographien, Berlin, 5–7 October
1998, 134:497
214. Butte A, Storti G, Morbidelli M (2000) Macromolecules 33:3485
215. Charleux B (2000) Macromolecules 33:5358
216. Pan G, Sudol ED, Dimonie VL, El-Aasser MS (2002) Macromolecules 35:6915
217. Lansalot M, Farcet C, Charleux B, Vairon J-P, Pirri R, Tordo P (2000) ACS Sym Ser
768:138
218. Prodpan T, Dimonie VL, Sudol ED, El-Aasser M (1999) Polym Mater Sci Eng 80:534
219. MacLeod PJ, Keoshkerian B, Odell P, Georges MK (1999) Polym Mater Sci Eng 80:539
220. Lansalot M, Charleux B, Vairon J-P, Pirri R, Tordo P (1999) ACS Polym Prepr 40:317
221. Prodpan T, Dimonie VL, Sudol ED, El-Aasser M (2000) Macromol Symp 155:1
222. MacLeod PJ, Barber R, Odell P, Keoshkerian B, Georges MK (2000) Macromol Symp
155:31
223. Farcet C, Lansalot M, Charleux B, Pirri R, Vairon JP (2000) Macromolecules 33:8559
224. Pan G, Sudol ED, Dimonie VL, El-Aasser MS (2001) Macromolecules 34:481
225. Keoshkerian B, MacLeod PJ, Georges MK (2001) Macromolecules 34:3594
226. Farcet C, Charleux B, Pirri R (2001) Macromolecules 34:3823
227. Cunningham MF (2002) Prog Polym Sci 27:1039
228. Tortosa K, Smith JA, Cunningham MF (2001) Macromol Rapid Comm 22:957
229. Farcet C, Belleney J, Charleux B, Pirri R (2002) Macromolecules 35:4912
230. Farcet C, Charleux B, Pirri R (2002) Macromol Symp 182:249
231. Keoshkerian B, Szkurham AR, Georges MK (2001) Macromolecules 34:6531
232. Cunningham MF, Tortosa K, Lin M, Keoshkerian B, Georges MK (2002) J Polym Sci Pol
Chem 40:2828
233. Cunningham MF, Xie M, McAuley KB, Keoshkerian B, Georges MK (2002) Macromole-
cules 35:59
234. Ma JW, Cunningham MF, McAuley KB, Keoshkerian B, Georges MK (2002) Macromol
Theor Simul 11:953
235. Ma JW, Cunningham MF, McAuley KB, Keoshkerian B, Georges MK (2003) Macromol
Theor Simul 12:72
236. Matyjaszewski K, Wang J (1995) J Am Chem Soc 117:5614
237. Matyjaszewski K, Wang J (1995) Macromolecules 28:7901
238. Matyjaszewski K, Xia J, Patten T, Abernathy T (1996) Science 272:866
239. Sawamoto M, Kamigaito M, Higashimura T, Kato M (1995) Macromolecules 28:1721
240. Sawamoto M, Kamigaito M, Ando T, Kato M (1996) Macromolecules 29:1070
241. Sawamoto M, Kamigaito M, Kotani Y, Kato M (1996) Macromolecules 29:6979
242. Uegaki H, Kotani Y, Kamigato M, Sawamoto M (1998) Macromolecules 31:6756
243. Moineau G, Minet M, Dubois P, Teyssié P, Sennenger T, Jérôme R (1999) Macromolecules
32:27
244. Lecomte P, Drapier I, Dubois P, Teyssié P, Jérôme R (1997) Macromolecules 30:7631
245. Brandts JAM, van de Geijn P, van Faassen EE, Boersma J, van Koten G (1999) J Organo-
met Chem 584:246
246. Kotani Y, Kamigaito M, Sawamoto M (1999) Macromolecules 32:2420
Miniemulsion Polymerization 253

247. Percec V, Barboiu B, Neumann A (1996) Macromolecules 29:3665

248. Patten TE, Matyjazsweski K (1998) Adv Mater 10:901
249. Matyjaszewski K (1999) Chem Eur J 5:3095
250. Pan CY, Lou XD (2000) Macromol Chem Phys 201:1115
251. Chen XP, Padias AB, Hall HK (2001) Macromolecules 34:3514
252. Matyjaszewski K, Wei M, Xia J, Gaynor S (1998) Macromol Chem Phys 199:2289
253. Hamasaki S, Kamigaito M, Sawamoto M (2002) Macromolecules 35:2934
254. Hamasaki S, Sawauchi C, Kamigaito M, Sawamoto M (2002) J Polym Sci Pol Chem
40:617
255. Fischer H (1999) J Polym Sci Pol Chem 37:1885
256. Sawamoto M, Kamigaito M (2000) Macromol Symp 161:11
257. Davis K, O’Malley J, Paik H, Matyjaszewski K (1997) ACS Polym Prepr 38:687
258. Matyjaszweski K, Xia J (2001) Chem Rev 101:2921
259. Takhashi H, Ando T, Kamigaito M, Sawamoto M (1999) Macromolecules 32:3820
260. Matyjaszewski K (1996) Macromol Symp 111:47
261. Butté A, Storti G, Morbidelli M (2001) Macromolecules 34:5885
262. Matyjaszewski K, Qiu J, Tsarevsky NV, Charleux B (2000) J Polym Sci Pol Chem 38:
4724
263. Zhang B, Zhang ZB, Xan XL, Hu CP, Ying SK (2002) Chinese J Polym Sci 20:445
264. Li M, Matyjaszewski K (2003) Macromolecules 36:6028
265. Chiefari J, Chong YK, Ercole F, Krstina J, Jeffery J, Le TPT, Mayadunne RTA, Meijs GF,
Moad CL, Moad G, Rizzardo E, Thang SH (1998) Macromolecules 31:5559
266. Le TP, Moad G, Rizzardo E, Thang SH (1998) PCT Int Appl, p 88
267. Rizzardo E, Thang SH, Moad G (1999) PCT Int Appl, p 40
268. Krstina J, Moad G, Rizzardo E, Winzor CL, Berge CT, Fryd M (1995) Macromolecules
28:5381
269. Corpart P, Charmot D, Biadatti T, Zard S, Michelet D (1998) PCT Int Appl, p 70
270. Hawthorne DG, Moad G, Rizzardo E, Thang SH (1999) Macromolecules 32:5457
271. Chong YK, Le TPT, Moad G, Rizzardo E, Thang SH (1999) Macromolecules 32:2071
272. De Brouwer H, Schellekens MAJ, Klumperman B, Monteiro MJ, German A (2000) J Polym
Sci Pol Chem 38:3596
273. Monteiro MJ, Sjoberg M, Van der Vlist J, Gottgens CM (2000) J Polym Sci Pol Chem
38:4206
274. Matyjaszewski K, Gaynor S, Wang JS (1995) Macromolecules 28:2093
275. Monteiro MJ, De Brouwer H (2001) Macromolecules 34:349
276. Kwak Y, Goto A, Tsujii Y, Murata Y, Komatsu K, Fukuda T (2002) Macromolecules
35:3026
277. Barner-Kowollik C, Quinn JF, Morsley DR, Davis TP (2001) J Polym Sci Pol Chem
39:1353
278. Barner-Kowollik C, Quinn JF, Nguyen TLU, Heuts JPA, Davis TP (2001) Macromolecules
34:7849
279. Barner-Kowollik C, Vana P, Quinn JF, Davis TPJ (2002) J Polym Sci Pol Chem 40:1058
280. Barner-Kowollik C, Coote ML, Davis TP, Radom L,Vana P (2003) J Polym Sci Pol Chem
41:2828
281. Wang AR, Zhu S, Kwak Y, Goto A, Fukuda T, Monteiro MJ (2003) J Polym Sci Pol Chem
41:2833
282. Jovanovic S, Romatowski J, Schulz GV (1965) Makromol Chem 85:187
283. Schulz GV, Romatowski J (1965) Makromol Chem 85:195
284. Romatowski J, Schulz GV (1965) Makromol Chem 85:227
285. Nomura M, Minamino Y, Fujita K, Harada M (1982) J Polym Sci Pol Chem 20:1261
254 F. J. Schork et al.

286. Lichti G, Sangster DF, Whang BCY, Napper DH, Gilbert RG (1982) J Chem Soc Farad T
1 78:2129
287. Maxwell IA, Morrison BR, Napper DH, Gilbert RG (1992) Makromol Chem 193:303
288. Mendoza J, De La Cal JC, Asua JM (2000) J Polym Sci Pol Chem 38:4490
289. Monteiro MJ, Hodgson M, De Brouwer HJ (2000) Polym Sci Pol Chem 38:3864
290. de Brouwer H, Tsavalas JG, Schork FJ, Monteiro MJ (2000) Macromolecules 33:9239
291. Moad G, Chiefari J, Chong YK, Krstina J, Mayadunne RTA, Postma A, Rizzardo E, Thang
SH (2000) Polym Int 49:993
292. Prescott SW, Ballard MJ, Rizzardo E, Gilbert RG (2002) Macromolecules 35:5417
293. Monteiro MJ, de Barbeyrac J (2001) Macromolecules 34:4416
294. Smulders WW, Gilbert RG, Monteiro MJ (2003) Macromolecules 36:4309
295. Smulders WW (2002) PhD Thesis, Technische Universiteit Eindhoven, The Nether-
lands
296. Tsavalas JG, Schork FJ, de Brouwer H, Monteiro MJ (2001) Macromolecules 34:3938
297. Vosloo JJ, De Wet-Roos D, Tonge MP, Sanderson RD (2002) Macromolecules 35:4894
298. Lansalot M, Davis TP, Heuts JPA (2002) Macromolecules 35:7582
299. Uzulina I, Kanagasbapathy S, Claverie J (2000) Macromol Symp 150:33
300. Monteiro MJ, Hodgson M, de Brouwer H (2000) J Polym Sci Pol Chem 38:3864
301. Bon SAF, Bosveld M, Klumperman B, German AL (1997) Macromolecules 30:324
302. Marestin C, Noel C, Claverie J (1998) Macromolecules 31:4041
303. Lansalot M, Farcet C, Charleux B, Vairon JP, Pirri R, Tordo O (2000) ACS Sym Ser
768:138
304. Jousset S, Qiu J, Matyjaszewski K, Granel C (2001) Macromolecules 34:6641
305. De Brouwer H, Tsavalas JG, Schork FJ, Monteiro MJ (2000) Macromolecules 33:9239
306. Prodpan T, Dimonie VL, Sudol ED, El-Aasser M (1999) Polym Mater Sci Eng 80:534
307. MacLeod PJ, Keoshkerian B, Odell P, Georges MK (1999) Polym Mater Sci Eng 80:539
308. Pan GF, Sudol ED, Dimonie VL, El-Aasser MS (2002) Macromolecules 35:6915
309. Matyjaszewski K, Qui J, Shipp DA, Gaynor SG (2000) Macromol Symp 155:15
310. Tonge MP, McLeary JB, Vosloo JJ, Sanderson RD (2003) Macromol Symp 193:289
311. Vosloo JJ, Roos DW, Tonge MP, Sanderson RD (2002) Macromolecules 35:4894
312. Ausa JM (2002) Prog Polym Sci 27:1283
313. Landfester K, Bechthold N, Tiarks F, Antonietti M (1999) Macromolecules 32:5222
314. Maitre C, Ganachaud F, Ferreira O, Lutz JF, Paintoux Y, Hemery P (2000) Macromole-
cules 33:7730
315. Barrere M, Maitre C, Dourges MA, Hemery P (2001) Macromolecules 34:7276
316. Barrere M, Ganachaud F, Bendejacq D, Dourges MA, Maitre C, Hemery P (2001) Polymer
42:7239
317. Limouzin C, Caviggia A, Ganachaud F, Hemery P (2003) Macromolecules 36:667
318. Cauvin S, Sadoun A, Santos R, Belleney J, Ganachaud F, Hemery P (2002) Macromole-
cules 35:7919
319. Barrere M, Landfester K (2003) Polymer 44:2833
320. Landfester K, Rothe R, Antonietti M (2002) Macromolecules 35:1658
321. Willert M, Landfester K (2002) Macromol Chem Phys 203:825
322. Marie E, Rothe R, Antonietti M, Landfester K (2003) Macromolecules 36:3967
323. Marie E, Landfester K, Antonietti M (2002) Biomacromolecules 3:475
324. Taden A, Antonietti M, Landfester K (2003) Macromol Rapid Comm 24:512
325. Claverie JP, Viala S, Maurel V, Novat C (2001) Macromolecules 34:382
326. Landfester K (2001) Adv Mater 13:765
327. Montenegro R, Antonietti M, Mastai Y, Landfester K (2003) J Phys Chem B 107:5088
328. Taden A, Landfester K (2003) Macromolecules 36:4037
Miniemulsion Polymerization 255

329. Montenegro R, Landfester K (2003) Langmuir 19:5996

330. Wegner G, Baum P, Muller M, Norwig J, Landfester K (2001) Macromol Symp 175:
349
331. Vaihinger D, Landfester K, Krauter I, Brunner H, Tovar GEM (2002) Macromol Chem
Phys 203:1965
332. Landfester K, Montenegro R, Scherf U, Guntner R, Asawapirom U, Patil S, Neher D,
Kietzke T (2002) Adv Mater 14:651
333. Piok T,Wenzl FP, Gamerith S, Gadermaier C, Patil S, Montenegro R, Kietzke T, Neher D,
Scherf U, Landfester K, List EJW (2003) Mater Res Soc Symp Proc 738:245
334. Piok T, Romaner L, Gadermaier C, Wenzl FP, Patil S, Montenegro R, Landfester K,
Lanzani G, Cerullo G, Scherf U, List EJW (2003) Synthetic Met 139:609
335. Piok T, Gamerith S, Gadermaier C, Plank H, Wenzl FP, Patil S, Montenegro R, Kietzke T,
Neher D, Scherf U, Landfester K, List EJW (2003) Adv Mater 15:800
336. Willert M, Rothe R, Landfester K, Antonietti M (2001) Chem Mater 13:4681
337. zu Putlitz B, Landfester K, Fischer H, Antonietti M (2001) Adv Mater 13:500
338. Manders B, Sciandrone L, Hauck G, Kristen M (2001) Angew Chem Int Edit 40:4006

Received: June 2004

Adv Polym Sci (2005) 175: 257–298
DOI 10.1007/b100117
© Springer-Verlag Berlin Heidelberg 2005

Microemulsion Polymerizations and Reactions

Pei Yong Chow ( ) 1 · Leong Ming Gan 2
1
Institute of Bioengineering and Nanotechnology, 31 Biopolis Way, The Nanos,
#04–01, 138669 Singapore
[email protected]
2
Institute of Materials Research & Engineering, 3 Research Link, 117602 Singapore
[email protected]

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259

2 Polymerizations in Globular and Bicontinuous Microemulsions

for Producing Microlatexes . . . . . . . . . . . . . . . . . . . . . . . . . . . 260
2.1 Inverse Microemulsion Polymerization . . . . . . . . . . . . . . . . . . . . . 261
2.2 Bicontinuous Microemulsion Polymerization . . . . . . . . . . . . . . . . . . 262
2.3 Polymerization in Oil-in-Water Microemulsions . . . . . . . . . . . . . . . . 263
2.4 Microemulsion Polymerization for Microlatexes with High Polymer-
to-Surfactant Weight Ratios . . . . . . . . . . . . . . . . . . . . . . . . . . . 266

3 Bicontinuous-Microemulsion Polymerization for Nanostructured

Solid-Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
3.1 Nanostructured Polymers Produced by Bicontinuous-Microemulsion
Polymerizations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
3.1.1 Ion-Conductive Membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
3.1.2 Proton-Exchange Membranes . . . . . . . . . . . . . . . . . . . . . . . . . . 274
3.2 Polymer Nanocomposites Produced by Bicontinuous-Microemulsion
Polymerizations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274
3.2.1 Ruthenium (II) Complexes in Polymerized Bicontinuous-Microemulsions . . 274
3.2.2 Aligned Nanocomposites of Ferrite-Polymer from Bicontinuous-
Microemulsion Polymerization . . . . . . . . . . . . . . . . . . . . . . . . . 276
3.3 Synthesis of Nanocomposites Via In-Situ Microemulsion Polymerization . . 277

4 Microemulsion Reactions for Processing Inorganic Nanomaterials . . . . . . 278

4.1 Synthesis of Inorganic Nanoparticles in Inverse Microemulsion . . . . . . . . 280
4.2 Materials Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
4.2.1 Doped and Un-Doped CdS/ZnS Nanoparticles . . . . . . . . . . . . . . . . . 283
4.2.2 Magnetic Ferrites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
4.2.3 Silica and Silica-Supported Ru-Cu Oxides . . . . . . . . . . . . . . . . . . . . 288
4.2.4 Perovskites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
4.2.5 Zirconia, Lead Zirconate and Lead Zirconate Titanate . . . . . . . . . . . . . 289
4.2.6 Hydroxyapetite (HA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290
4.2.7 PtRu/C Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
4.2.8 Polymer-Coated Inorganic Nanoparticles . . . . . . . . . . . . . . . . . . . . 293

5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294
258 P. Y. Chow · L. M. Gan

Abstract This review describes how the unique nanostructures of water-in-oil (W/O), oil-
in-water (O/W) and bicontinuous microemulsions have been used for the syntheses of some
organic and inorganic nanomaterials. Polymer nanoparticles of diameter approximately
10–50 nm can easily be obtained, not only from the polymerization of monomers in all three
types of microemulsions, but also from a Winsor I-like system. A Winsor I-like system with
a semi-continuous process can be used to produce microlatexes with high weight ratios of
polymer to surfactant (up to 25). On the other hand, to form inorganic nanoparticles, it is
best to carry out the appropriate chemical reactions in W/O- and bicontinuous micro-
emulsions.
Recent developments in the cross-polymerization of the organic components used in
bicontinuous microemulsions ensure the successful formation of transparent nanostruc-
tured materials. Current research into using polymerizable bicontinuous microemulsions
as a one-pot process for producing functional membranes and inorganic/polymer nano-
composites is highlighted with examples.

Keywords Microemulsion polymerization · Microemulsion reaction · Water-in-Oil (W/O)

microemulsion · Oil-in-Water (O/W) microemulsion · Bicontinuous microemulsion ·
Functional membranes and inorganic/polymer nanocomposites

Abbreviations
AN Acrylonitrile
AA Acrylic acid
AM Acrylamide
APTAC 2-acrylamido-2-propane trimethylammonium chloride
AUTMAB (Acryloyloxy)-undecyltrimethylammonium bromide
AUDMAA (Acryloyloxy)-undecyldimethylammonioacetate
AIBN Azobisisobutyronitrile
AOT 1,4-Bis(2-ethylhexyl)sulfosuccinate sodium salt
APS Ammonium persulfate
BL g-butyrolactone
BUA Butylacrylate
C1-PEO-C11-MA-40 w-methoxypoly(ethylene oxide)40-undecyl-a-methacrylate
CTAB Cetyltrimethylammonium bromide
DMC Dimethyl carbonate
DTAB Dodecyltrimethylammonium bromide
EGDMA ethyleneglycol dimethacrylate
EC Ethylene carbonate
HC Hydrocarbon
HA Hydroxyapetite
HEMA 2-hydroxylethylmethacrylate
MADQUAT 2-Methacryloyloxyethyltrimethyl ammonium chloride
MMA Methyl methacrylate
NaA Sodium acrylate
NaAMPS Sodium-2-acrylamido-2–2-methylpropane sulfonate
NaSS Sodium styrenesulfonate
NIPAM N-isopropylacrylamide
NP Poly(oxyethylated alkylphenyl ether)
OTAC Octyldecyltrimethylammonium chloride
PC Propylene carbonate
PMMA Poly(methyl methacrylate)
PNIPAM Poly(N-isopropylacrylamide)
Microemulsion Polymerizations and Reactions 259

PS Poly(styrene)
SDS sodium dodecyl sulfate
SEAAU Sodium 11-(N-ethylacrylamido)-undecanoate
St Styrene
TEM Transmission electron microscope
THFM Tetrahydrofurfuryl methacrylate
TTAB Tetradecyltrimethylammonium bromide
VA Vinyl acetate
VBSLi 4-Vinylbenzene sulfonate, lithium salt

1
Introduction
Microemulsions are transparent liquid systems consisting of at least ternary mix-
tures of oil, water and surfactant. Sometimes a cosurfactant is needed for the
formation of a thermodynamically-stable microemulsion. A transparent micro-
emulsion is, in fact, heterogeneous (nanostructured) on a molecular scale. Micro-
emulsion domains fluctuate in size and shape and undergo spontaneous coales-
cence and break-up [1]. They can exhibit water continuous and bicontinuous
structures, with typical equilibrium domain sizes ranging from about 10 to 100 nm
[2]. The transparent microemulsions can be in the form of nano-globules of
oil-swollen micelles dispersed in the water continuous phase as oil-in-water (O/W)
microemulsions, or water-swollen micellar globules dispersed in oil as water-in-
oil (W/O) microemulsions. In between the regions of O/W and W/O microemul-
sions, there may also exist a composition region of bicontinuous (sponge-like
structure) microemulsions [3], whose oil and water domains are randomly dis-
persed in two phases, as can be seen from the sketched phase diagram of Fig. 1.

Fig. 1 Illustration of some phase equilibria encountered in multicomponent systems

260 P. Y. Chow · L. M. Gan

In addition to single phase microemulsions, several phase equilibria known

as Winsor systems [4] are also shown at low surfactant concentrations.A Winsor I
(WI) system consists of an O/W microemulsion that is in equilibrium with an
oil phase, while a Winsor II (WII) system is a W/O microemulsion in equili-
brium with an aqueous phase.A WIII system has a middle phase (bicontinuous)
microemulsion that coexists with both oil and aqueous phases.
When a single chain anionic surfactant (such as sodium dodecyl sulfate,
SDS) is used, it generally requires a cosurfactant for the formation of a micro-
emulsion. A cosurfactant may not be needed to form a microemulsion if non-
ionic surfactant(s), certain types of cationic surfactants, or double-chain sur-
factants such as sodium 1,4-bis(2-ethylhexyl)sulfosuccinate (Aerosol OT or
simply AOT) are used.
Due to the enormous inner surfaces of the nanostructures of W/O and O/W
globular microemulsions, they can provide novel reaction sites for some inor-
ganic/organic reactions and polymerizations. Nano-sized particles of inorganic
materials are generally prepared in W/O microemulsion, and microlatexes
(d<50 nm) of polymers from both O/W- and W/O-microemulsion polymer-
izations. The recently developed approach of using polymerizable surfactants
in bicontinuous microemulsions also enables the syntheses of transparent
nanoporous polymeric materials and nanocomposites. Hence, microemulsions
have been used as novel chemical nanoreactors for producing nanophase
materials such as polymer latexes [5, 6], inorganic particles [7], and industrial-
ly useful materials [8]. This review relates the development of microemulsion
polymerizations and reactions to the formation of nanostructured materials.

2
Polymerizations in Globular and Bicontinuous Microemulsions
for Producing Microlatexes

Over the past two decades, free radical polymerization studies have mainly
been carried out in globular microemulsions (both O/W microemulsions and
W/O microemulsions, also known as inverse microemulsions). Each milliliter
of the globular microemulsion usually contains 1015–1017 W/O- or O/W nano-
sized (5–10 nm in diameter) droplets (globules). The enormous number of
nano-globules are potential loci for fast polymerization, producing microlatex
particles much less than 50 nm in diameter. Though small polymer particles,
molecular weights exceeding one million can easily be obtained from these
polymerization systems. Most of the microemulsion polymerization studies [5]
have dealt with hydrophobic monomers, such as styrene (St) or methyl meth-
acrylate (MMA), within oil cores of O/W microemulsions [9] and with the poly-
merization of water-soluble monomers, such as acrylamide (AM), within
aqueous cores of inverse microemulsions [10, 11]. For both O/W and inverse
microemulsion systems, the amount of monomer was usually restricted to less
than 10 wt% with respect to the total weight of microemulsion. Moreover,
Microemulsion Polymerizations and Reactions 261

higher amounts of surfactant (5–15 wt%) were normally needed for the stabil-
ity of the polymerization. For those microemulsions requiring a cosurfactant,
the compatibility between the cosurfactant and the polymers formed becomes
an issue. For instance, if styrene is polymerized within an O/W microemulsion
that contains an alcohol cosurfactant, this alcohol will not dissolve polystyrene.
However, this is not the case for the polymerization of acrylamide (AM) in
alcohol-free inverse microemulsion.

2.1
Inverse Microemulsion Polymerization

The polymerization of a water-soluble monomer such as AM, acrylic acid (AA),

sodium acrylate (NaA), or 2-hydroxyethylmethacrylate (HEMA), can be carried
out easily in inverse microemulsion or/and bicontinuous microemulsion. These
water-soluble monomers also act as cosurfactants, increasing the flexibility and
the fluidity of the interfaces, which enhances the solubilization of the monomer.
A cosurfactant effect during the polymerization of vinyl acetate in anionic
microemulsions has also been reported [12].
Candau and co-workers were the first to address the issue of particle nu-
cleation for the polymerization of AM [13, 14] in an inverse microemulsion
stabilized by AOT. They found that the particle size of the final microlatex
(d~20–40 nm) was much larger than that of the initial monomer-swollen
droplets (d~5–10 nm). Moreover, each latex particle formed contained only one
polymer chain on average. It is believed that nucleation of the polymer particle
occurs for only a small fraction of the final nucleated droplets. The non-
nucleated droplets also serve as monomer for the growing particles either by
diffusion through the continuous phase and/or by collisions between droplets.
But the enormous number of non-nucleated droplets means that some of the
primary free radicals continuously generated in the system will still be cap-
tured by non-nucleated droplets. This means that polymer particle nucleation
is a continuous process [14]. Consequently, each latex particle receives only one
free radical, resulting in the formation of only one polymer chain. This is in
contrast to the large number of polymer chains formed in each latex particle
in conventional emulsion polymerization, which needs a much smaller amount
of surfactant compared to microemulsion polymerization.
The polymerization of acrylamide (AM) and the copolymerization of acryl-
amide-sodium acrylate in inverse microemulsions have been studied exten-
sively by Candau [10, 11, 13–15], Barton [16, 17], and Capek [18–20]. One of
the major uses for these inverse microlatexes is in enhanced oil recovery pro-
cesses [21]. Water-soluble polymers for high molecular weights are also used
as flocculants in water treatments, as thickeners in paints, and retention aids in
papermaking.
262 P. Y. Chow · L. M. Gan

2.2
Bicontinuous Microemulsion Polymerization

The polymerization of AM in inverse microemulsions originally required a

high weight ratio of surfactant/monomer. This high ratio was subsequently
drastically reduced by polymerizing water-soluble monomers in nonionic
bicontinuous microemulsions [10, 22]. Besides AM, other water-soluble mo-
nomers investigated were sodium acrylate [23], sodium 2-acrylamido-2–2-
methylpropane sulfonate [24] (NaAMPS, CH2=CHCONHC(CH3)2CH2SO3Na)
and 2-methacryloyloxyethyltrimethylammonium chloride [25] (MADQUAT,
CH2=C(CH3)COOCH2CH2H(CH3)3Cl). This type of polymerization has been
extended to syntheses of copolymers that possess both positively and nega-
tively charged moieties along the macromolecular backbone (polyampholytes).
High charge density polyampholytes can be formed from the microemulsion
copolymerization [25] of NaAMPS and MADQUAT. Polymerization of water-
soluble monomers in bicontinuous microemulsions produced transparent
microlatexes [26] with particle sizes (50–100 nm) that remain unchanged for
years.
Microemulsion formulation was further optimized by Candau [27] using
cohesive energy ratio (CER) and hydrophile-lipophile balance (HLB) concepts.
The best formulation was obtained for high HLB values (8–11) using a blend
of nonionic surfactants. This favors microemulsions with bicontinuous struc-
tures in the presence of AM and NaAMPS or AM and NaA. The process yielded
clear, stable microlatexes of moderate particle size (d<100 nm) for high mole-
cular weight copolymers up to 25 wt% solids. High-solid content (~30 wt%) of
water-soluble poly(vinyl acetate) latexes [28] could also be obtained by the
multi-stage addition of monomer to a latex pre-produced by the polymeriza-
tion of 3 wt% vinyl acetate in three-component microemulsions stabilized by
low concentration of AOT (<1 wt%). These microemulsion-made latexes con-
tained latex particles two to three times smaller than those obtained by emul-
sion polymerization.
Syntheses of thermosensitive polyampholytes by polymerization in bicon-
tinuous microemulsions have also been studied by Candau’s group [29, 30]
recently. Poly(N-isopropylacrylamide) (PNIPAM) exhibits a well-defined lower
critical solution temperature (LCST) in water around 32 °C, and it is the most
extensively studied temperature-responsive polymer. The microemulsion
contained three-monomer-mixtures, namely, N-isopropylacrylamide (NIPM,
neutral), NaAMPS (anionic), and 2-acrylamido-2-propanetrimethylammo-
nium chloride (APTAC, cationic). Bicontinuous microemulsions were produced
in an HLB domain ranging between around 10.6 and 11.3 using a blend of non-
ionic surfactants. After polymerization, it formed thermosensitive polymers,
namely NIPAM-based polyelectrolytes and polyampholytes. The final products
consisted of stable and clear microlatexes of small particle size containing up
to 20 wt% high molecular weight copolymers. In order to overcome the high
surfactant-to-monomer ratio used, they performed the polymerization by a
Microemulsion Polymerizations and Reactions 263

semi-continuous process that reduced the surfactant content required to sta-

bilize the final microlatexes by more than 40%.

2.3
Polymerization in Oil-in-Water Microemulsions

The main difficulty encountered by most researchers working on cosurfactant-

O/W microemulsion polymerization over the past decade [31–36] is the in-
stability of microlatex produced from systems that use lower weight ratios (<1)
of surfactant to monomer for a higher content of polymer (>8 wt%). The in-
stability of the resulting microlatexes may be due to the incompatibility be-
tween the higher content of polymer formed and the cosurfactant [37]. The
polymerization of a hydrophobic monomer in a ternary O/W microemulsion
without a cosurfactant was first reported in 1989 by Ferrick et al [38]. This
spurred a new interest in systematic polymerization studies for cationic O/W
microemulsions [39–42]. The cationic surfactants can be dodecyltrimethyl-
ammonium bromide (DTAB), tetradecyltrimethylammonium bromide (TTAB),
or cetyltrimethylammonium bromide (CTAB). Nonionic surfactants were also
used in ternary O/W microemulsions for the polymerization of styrene and
methylmethacrylate (MMA) [43] and anionic AOT microemulsion for the poly-
merization of tetrahydrofurfuryl methacrylate (THFM) [44]. Anionic SDS
ternary O/W microemulsions have also been used to study the polymerization
of butylacrylate (BUA) [45, 46], alkyl acrylates [47], and the copolymerization
of BUA with acrylonitrile [48]. All of the microemulsion polymerizations pro-
duced stable microlatexes with particle sizes ranging from 20 to 60 nm in dia-
meter. However, these ternary O/W microemulsions usually required a high
surfactant concentration (7–15 wt%) to solubilize a relatively low monomer
content of less than 10 wt%. These conventional O/W microemulsion polymer-
ization recipes call for the use of at least an equal amount of surfactant and
monomer, which makes the polymerization and surfactant removal processes
prohibitively expensive. Nevertheless, there have been a few studies in which
the formulation has been modified from these polymerization conditions in
order to dramatically reduce the use of surfactant, as we will discuss in the next
section.
It is generally accepted that an O/W microemulsion polymerization proceeds
via a continuous particle nucleation mechanism, as in the case of inverse micro-
emulsion polymerization [14]. The particle nucleation of polymer is generally
postulated to occur in microemulsion droplets (micellar nucleation mecha-
nism) for less water-soluble monomers, such as styrene and BUA. Homoge-
neous nucleation for this type of monomer may be insignificant because of an
extremely large number of microemulsion droplets (~1016–1017 m/L) that can
effectively capture the free radicals generated. The primary radicals generated
in the aqueous phase probably react first with some dissolved monomers in the
aqueous phase to form oligomeric radicals of higher hydrophobicity. These
oligomeric radicals can then be captured more favorably by microemulsion
264 P. Y. Chow · L. M. Gan

droplets. Therefore, the newly formed latex particles continue to grow in the
microlatex through the constant supply of monomer from nucleated droplets
until the chain in the particle is terminated by chain transfer to monomer [42,
49, 50]. However, Kaler’s group recently consistently obtained very high mole-
cular weights of ~15¥106 Da for styrene O/W microemulsion polymerizations
[51]. Such molecular weights are considerably higher than those typically
obtained via free-radical polymerizations. In the free-radical polymerization
of styrene, the chain transfer to monomer is limited and hence can only produce
molecular weights of up to ~2¥10–6 Da, as suggested by the most recent mea-
surements of chain-transfer constants by Kukulj et al [52]. Therefore, the
diffusion-limited exit of monomer radicals to the aqueous phase coupled with
chain transfer to polymer are probable reasons for the enhanced molecular
weight of polystyrene [51].
In contrast to emulsion polymerization, the reaction kinetics of micro-
emulsion polymerization is characterized by two polymerization rate intervals;
the interval of constant rate characteristic of emulsion polymerization is miss-
ing [42, 49, 53], as shown in Fig. 2. Polymer particles are generated continuously
during the reaction by both micellar and homogeneous mechanisms. As the
solubility of the monomer in the continuous domain increases, homogeneous

Fig. 2 Experimental and model rate versus conversion profiles for the polymerization of
hexylmethacrylate in a microemulsion stabilized by the surfactant DTAB. The two curves are
for initiator concentrations of 0.045 wt% (top) and 0.015 wt% (bottom) relative to the
amount of monomer in the microemulsion. The solid lines are predictions from the Mor-
gan model [56]
Microemulsion Polymerizations and Reactions 265

nucleation becomes more important [54]. In O/W microemulsion polymer-

ization, monomer partitioning between polymer particles and uninitiated
micelles via diffusion through the aqueous phase determines the concentration
of monomer at the polymerization loci. This partitioning plays an important
role in determining polymer particle formation and growth [55–57]. The rate
of polymerization initially increases to a maximum around 20–25% of total
monomer conversion (Interval I) and then decreases on further polymerization
(Interval II). This is generally interpreted as indicating that the particle nuclea-
tion occurs mainly in Interval I, and polymer growth occurs in Interval II.
The decreased polymerization rate in Interval II is due to the progressively
depletion of monomer in the latex particles.
The growth pattern for PMMA particles was found to be different from that
for styrene polymerization in ternary cationic microemulsions [58, 59] con-
taining either TTAB, TTAC, CTAB or OTAC (octyldecyltrimethylammonium
chloride). For the MMA microemulsions, the average hydrodynamic particle
radius (Rh) increased continuously from about 20 to 50 nm during the poly-
merization. The improved interactions between the cosurfactant MMA and the
cationic surfactant at interfaces may restrict the swelling of MMA-swollen
PMMA particles. Hence a gradual increase in the sizes of the PMMA particles
during polymerization was observed, as expected [59]. However, this was not
the case for the weaker interaction between styrene and a cationic surfactant
in the corresponding microemulsion. The maximum swelling of PS particles by
styrene monomer, and hence the largest Rh, was attained during the early stage
of polymerization (4–7% polymer conversion) [58]. As the styrene polymer-
ization continued, Rh generally decreased due to the diminishing styrene con-
centration in the swollen PS particles. For some polymerized MMA-micro emul-
sions, particle sizes might increase significantly towards the final stage of
polymerization. The agglomeration of some PMMA latex particles stabilized by
TTAB was evidenced by the prolonged heating (long term storage) of micro-
latexes [60] at 60 °C, as shown in Fig. 3. Cationic surfactants with larger carbon
chain lengths, such as OTAC or CTAB, form thicker interfacial layers which
prevent latex flocculation through steric stabilization. It should be noted that the
stability of PS microlatex was not affected by the alkyl chain length of the cationic
surfactant due to stronger surfactant adsorption on the surface of PS particles
than on PMMA particles [60]. Indeed, it is known that surfactant adsorption
on the latex/water surface decreases with increasing polymer polarity [61].
Core-shell nanoparticles can also be fabricated using microemulsions. This
was performed using a two-stage microemulsion polymerization beginning
with a polystyrene seed [62]. Butyl acrylate was then added in a second step
to yield a core-shell PS/PBA morphology. The small microlatex led to better
mechanical properties than those of similar products produced by emulsion
polymerization. Hollow polystyrene particles have also been produced by micro-
emulsion polymerization of MMA in the core with crosslinking of styrene on
the shell. After the synthesis of core-shell particles with crosslinked PS shells,
the PMMA core was dissolved with methylene chloride [63]. The direct cross-
266 P. Y. Chow · L. M. Gan

Fig. 3 Changes in PMMA particle size during long term storage at 60 °C for microlatexes
stabilized by different surfactants: (filled triangles) TTAB; (filled squares) TTAC; (filled
circles) CTAB; (empty triangles) OTAC

linking polymerization of styrene at the interfaces of isoctane microdroplets

of microemulsion also produced hollow nanocapsules [64].

2.4
Microemulsion Polymerization for Microlatexes
with High Polymer-to-Surfactant Weight Ratios

A major drawback of conventional microemulsion polymerization is the high

surfactant-to-monomer ratio usually needed to form the initial microemulsion.
Surfactant can be used more efficiently in semi-continuous or fed polymer-
ization processes. Several polymerization cycles can be run in a short period
of time by stepwise addition of new monomer. After each cycle of monomer
addition, most of the surfactant is still available to stabilize the growing hydro-
phobic polymer particles, or to forms microemulsion again when a polar
monomer is used. For instance, in the polymerization of vinyl acetate (VA) by
a semi-continuous microemulsion process [21], latexes with a high polymer
content of about 30 wt% were obtained at relatively low AOT concentrations of
about 1 wt%. Moreover, their particle sizes and molecular weights were much
smaller than those obtained by conventional emulsion polymerization.
A new system associated with an O/W microemulsion for styrene poly-
merization of up to 15 wt% solids at about 1 wt% DTAB was first reported [65]
in 1994. This new two-phase system is rather similar to a Winsor I system, as
depicted in Fig. 1; an organic phase containing small portions of water and
surfactant in equilibrium with an O/W microemulsion in the bottom phase.
Microemulsion Polymerizations and Reactions 267

The new system consists of a pure styrene upper phase placed on top of a
ternary O/W microemulsion. Strictly speaking, this new system is not identical
to the Winsor I system. Hence the new polymerization system has been referred
as a “Winsor I-like” system, which can be prepared by simply topping up a
ternary O/W microemulsion with hydrophobic monomer without disturbing
the microemulsion phase. Styrene polymerization took place only in the initial
ternary microemulsion containing 0.5–1.0 wt% styrene, 1.0 wt% DTAB and a
redox initiator of ammonium persulfate/tetramethylethylenediamine (APS/
TMEDA). The upper styrene phase acted only as a monomer reservoir to conti-
nuously supply monomer to the polymerization loci in the lower micro-
emulsion phase through diffusion without stirring. PS particles initially formed
in the microemulsion phase became the seed for the further growth of PS
particles to a more uniform size, up to about 80 nm in diameter. Therefore,
microlatexes of about 15 wt% PS could be obtained using only 1 wt% DTAB.
It should also be noted that the Winsor I-like system with a small amount of
monomer in the upper phase becomes a milky emulsion on stirring. But this
milky emulsion will transform into a transparent or bluish microlatex after
polymerization.
The polymerization rate of the Winsor I-like system was very slow due to the
low monomer diffusion rate through the limited interface to the polymeriza-
tion loci for the unstirred system. In order to increase the interfacial areas for
the infusion of monomer to the polymerization loci, hollow fiber monomer
feeds were later used to polymerize not only styrene, but also MMA and butyl-
acrylate [66]. Such a set-up using continuous monomer feeding via a hollow-
fiber is shown in Fig. 4. About 100 polypropylene (PP) hollow fibers (pore size
70 nm) were bundled together. One of the openings of the hollow-fiber bundle
was used for monomer feeding and the other end was sealed with epoxy resin.
The initial microemulsion usually consisted of 0.5 wt% styrene or BA, or 2 wt%
MMA, 1 wt% CTAB or 1.5 wt% SDS together with 0.2 wt% 1-pentanol, 4 mM
equimolar redox initiator (APS/TMEDA), and water to make it up to 100 wt%.
After polymerization of the microemulsion for about 15 min at room temper-
ature with about 97–98% conversion, the additional monomer was then contin-
uously introduced to the polymerization system via the infusion of monomer
from hollow fibers connected to monomer reservoir. The rate of monomer
infusion into the microemulsion could be regulated by the external nitrogen
pressure applied to the monomer reservoir. The slow bubbling by nitrogen gas
into the polymerizing microemulsion was to ensure homogeneous mixing.
When the infusion rate of monomer was optimal, latex particles would grow to
more uniform sizes at the expense of forming secondary particles via homo-
geneous nucleation. This method was able to produce almost uniform micro-
latexes of various sizes (15–65 nm) at high polymer/surfactant weight ratios up
to about 15 with high molecular weights (106 g/mol) within 2–3 h.
Though the continuous addition of monomer via hollow fibers into a
Winsor I-like system is rather novel, it is not as convenient as the drop-wise
addition of monomer under a semi-continuous process. Ming et al [67, 68]
268 P. Y. Chow · L. M. Gan

Fig. 4 Schematic presentation of the set-up (not to scale) for the polymerization of mo-
nomer in a microemulsion via hollow-fiber feeding of additional monomer

slightly modified the Winsor I-like method by directly adding monomer very
slowly to pre-polymerized ternary microemulsions for producing microlatexes
of PS, PMMA, poly(butyl methacrylate), or poly(methacrylate) (PMA). With
1 wt% DTAB, 24 wt% PMMA microlatexes of 33–46 nm in size were obtained.
They showed that this semi-continuous process was very effective for produc-
ing small microlatexes (15 nm) containing up to 30 wt% PMA at a high PMS to
SDS weight ratio of 25:1. The semi-continuous process works at the monomer-
starved condition to ensure that no empty micelles exist during polymerization.
The polymerization of styrene in Winsor I-like systems by semi-continuous
feeding of monomer stabilized by either DTAB, TTAB or CTAB has been sys-
tematically investigated by Gan and coworkers [69a]. Rather monodisperse
polystyrene microlatexes of less than 50 nm with molecular weights of over one
million were obtained at a polymer/surfactant weight ratio of 14:1. The Win-
sor I-like (micro)emulsion polymerization of styrene stabilized by non-ionic
surfactant and initiated by oil-soluble initiators has also been reported very re-
cently [69b]. The sizes of the large monomer-swollen particles decreased with
conversion and they merged with growing particles at about 40–50% conversion.
High PMMA content (30–40%) microlatexes [70] stabilized with low con-
centrations of anionic SDS were also prepared by microemulsion polymeriza-
Microemulsion Polymerizations and Reactions 269

tion via semi-continuous feeding of MMA. The synthesis of monodisperse

poly(dimethylsiloxane) microlatexes (<80 nm) [71] stabilized by several sur-
factants were also obtained at polymer/surfactant weight ratios of up to 3:1.
Nanosized polystyrene (PS) microlatexes stabilized by a mixture of cationic/
cationic, anionic/anionic, or anionic/cationic surfactants of various types [72]
with PS-to-surfactant weight ratios up to 10:1 have been synthesized by a semi-
continuous microemulsion polymerization process. For cationic or anionic sys-
tems, spherical latex particles ranging from about 22 to 53 nm were produced
almost linearly, independent of the weight ratio of the mixed surfactants
of similar charges. High molecular weight (Mw) PS ranging from 1.1¥106 to
1.9¥106 g/mol could easily be obtained from all three systems. The present poly-
merization method allows one to synthesize nanoparticles of PS or other
polymers with high polymer/surfactant weight ratios at some particle sizes that
cannot be achieved with a single type of surfactant.
High polymer/surfactant weight ratios (up to about 15:1) of polystyrene
microlatexes [73] have been produced in microemulsions stabilized by poly-
merizable nonionic surfactant by the semi-continuous process. The copoly-
merization of styrene with the surfactant ensures the long-term stability of the
latexes. Nanosized PS microlatexes with polymer content (≤25 wt%) were also
obtained from an emulsifier-free process [74] by the polymerization of styrene
with ionic monomer (sodium styrenesulfonate, NaSS), nonionic comonomer
(2-hydroxyethylmethacryalte, HEMA), or both. The surfaces of the latex particles
were significantly enriched in NaSS and HEMA, providing better stabilization.
Nanoparticles of PS (Mw=1.0¥106–3.0¥106 mol–1) microlatexes (10–30 nm)
have also been successfully prepared from their respective commercial PS for
the first time [75]. The dilute PS solutions (cyclohexane, toluene/methanol or
cyclohexane/toluene) were induced to form polymer particles at their respec-
tive theta temperatures. The cationic CTAB was used to stabilize th micro-
latexes. The characteristics of these as-formed PS latex particles were quite
similar to those obtained from the microemulsion polymerization of styrene
as reported in literature. These microlatexes could also be grown to about
50 nm by seeding the polymerization of styrene with a monodisperse size dis-
tribution of Dw/Dn=1.08. This new physical method for preparing polymer
nano-sized latexes from commercial polymers may have some potential appli-
cations, and therefore warrants further study.

3
Bicontinuous-Microemulsion Polymerization for Nanostructured
Solid-Materials

Numerous attempts to prepare nanostructured materials by polymerization of

suitable monomers (like MMA and styrene) in water-in-oil [76, 77] or oil-in-
water [39, 51, 78–81] and bicontinuous microemulsion [27, 82–84] have been
made. Polymeric materials were traditionally stabilized by non-polymerizable
270 P. Y. Chow · L. M. Gan

surfactants, such as sodium dodecyl sulfate (SDS), and their polymerized micro-
emulsions were usually found to be opaque and phase-separated. By incorpo-
rating polymerizable acrylic acid as co-surfactant in a microemulsion, stable
transparent polymeric materials can be obtained but only at low solid contents
(<15 wt%) [85]. However, the polymerized microemulsions did not reveal any
microstructures when viewed by scanning electron microscope (SEM). But
when SDS was replaced by a polymerizable surfactant, transparent nanoporous
polymeric materials were obtained.

3.1
Nanostructured Polymers Produced by Bicontinuous-Microemulsion
Polymerizations

There is increasing interest in the study of the formation of transparent nano-

porous polymers (film or sheet) by bicontinuous-microemulsion polymeriza-
tion [86–91a]. The real success of forming transparent solid polymers pos-
sessing various nanostructures arises from the polymerization of bicontinuous
microemulsions containing a polymerizable surfactant with other monomers.
The polymerizable surfactants are also known as “surfers” [91b], and they can
be anionic sodium 11-(N-ethylacrylamindo)-undecanoate (SEAAU), cationic
(acryloyloxy)-undecyltrimethylammonium bromide (AUTMAB), zwitterionic
(acryloyloxy)-undecyldimethylammonioacetate (AUDMAA) or nonionic w-
methoxypoly(ethylene oxide)40-undecyl-a-methacrylate (C1-PEO-C11-MA-40),
as listed in Table 1. The polymerizable surfactants should not contain allylic
hydrogens, to avoid active allylic chain transfer reactions.
The development of polymerizable microemulsions consisting of only three
basic components (except a water component) for producing transparent solid
polymers with nanostructure is a recent achievement [87]. For example, Fig. 5
shows the SEM micrograph of the fractured polymer prepared by the UV-
initiated polymerization of a bicontinuous microemulsion consisting of 35 wt%
water, 35 wt% AUDMAA and 30 wt% MMA. This micrograph reveals random-
ly distributed bicontinuous nanostructures of water channels and polymer
domains. The widths of the bicontinuous nanostructures were about 40–60 nm.
The sizes of the nanostructures can be readily reduced by adding 2-hydro-

Table 1 Some polymerizable surfactants that have been synthesized

Surfactant Structure CMC Reference

(mol l–1,
25 °C)

SEAUU CH2=CHCON(C2H5)(CH2)10COO–Na+ 4.2¥10–5 [92]

AUTMAB CH2=CHCOO(CH2)11-N+(CH3)3Br– 1.4¥10–2 [88]
AUDMAA CH2=CHCOO(CH2)11-N+(CH3)2CH2COO– 9.3¥10–3 [87]
C1-PEO-C11-MA-40 CH2=CHCOO(CH2)11-(OCH2CH2)40OCH3 6.6¥10–5 [93]
Microemulsion Polymerizations and Reactions 271

Fig. 5 SEM micrograph of a fractured polymer prepared by polymerization of a bicontinous

microemulsion consisting of 35 wt% H2O, 35 wt% AUDMAA and 30 wt% of MMA. The black
and gray areas refer to water and polymer domains respectively

Fig. 6 SEM micrograph of a fractured polymer prepared by polymerization of a bicontinous

microemulsion consisting of 30 wt% H2O, 28 wt% AUTMAB, 38 wt% of MMA and 4 wt%
HEMA. The black and gray areas refer to water and polymer domains respectively
272 P. Y. Chow · L. M. Gan

Fig. 7 SEM micrograph of a fractured polymer prepared by polymerization of a bicontinous

microemulsion consisting of 40 wt% aqueous solution containing 3% NaCl, 33 wt%
AUTMAB and 27 wt% of MMA. The black and gray areas refer to water and polymer do-
mains respectively

xyethylmethacrylate (HEMA) or increased by adding 1–3 wt% NaCl solution,

as shown in Figs. 6 and 7 [91a] respectively. In all of the bicontinuous micro-
emulsions investigated, about 1 wt% of cross-linker ethyleneglycol dimeth-
acrylate (EGDMA) was usually used to enhance the mechanical strengths of the
polymers.
This type of cross-polymerization of all of the organic components (like
MMA, HEMA and a polymerizable surfactant) in a bicontinuous micro-
emulsion is an important area of recent development in microemulsion poly-
merization, which can be used to produce nanostructures of transparent poly-
mer solids. The polymerization can be readily initiated using either redox or
photo-initiators. The gel formation usually occurred within 20 minutes. The
use of this novel type of microemulsion polymerization for preparing trans-
parent inorganic-polymer nanocomposites in the form of films or sheets is
emerging and exciting. However, very little published information about
this type of nanocomposite is available, as will be described in the following
sections.

3.1.1
Ion-Conductive Membranes

The development of transparent polymer electrolyte membrane from the bi-

continuous-microemulsion polymerization of 4-vinylbenzene sulfonic acid
lithium salt (VBSLi), acrylonitrile and a polymerizable non-ionic surfactant, w-
methoxypoly(ethylene oxide)40-undecyl-a-methacrylate (C1-PEO-C11-MA-40)
was reported in 1999 [94, 95]. The ionic conductivities of the polymer electro-
Microemulsion Polymerizations and Reactions 273

Fig. 8 SEM micrograph of the polymerized microemulsion solid, that contains the poly-
merizable non-ionic surfactant C1-PEO-C11-MA-40 [96], after ethanol extraction

lyte membranes were rather high, in the range of (2–3)¥10–3 S/cm. This rela-
tively high conductivity is attributed to the inclusion of the polymerizable ionic
salt VBSLi and the existence of the interconnected aqueous channels in the
polymerized bicontinuous network, which facilitate ionic conduction for the
membrane electrolyte [94]. However, some unusual ionic conduction pheno-
mena were also observed for these ion-containing membranes – large ions
exhibited higher mobilities than smaller ones [95]. A possible explanation
comes from the larger hydration shells of the lighter cations. This is further
supported by the sharp drop in conductivity when the system is cooled below
the freezing point of water.
The morphology of the microemulsion-polymerized solid after ethanol
extraction, as revealed by SEM, is shown in Fig. 8. The extracted samples show
globular microstructures and voids (pores). These pores might be derived from
the interconnected water-filled voids generated from numerous coalescences of
growing particles during polymerization.
The water in the membranes [96] could easily be replaced by polar organic
solvents, such as BL, PC-EC, or EC-DMC, by immersion. The conductivities of
the membranes decreased sharply (by about two orders of magnitude) with the
use of organic solvents.After soaking these membranes in electrolyte solutions
of 1 M LiSO3CF3/PC-EC, 1 M LiBF4/BL, or 1 M LiClO4/EC-DMC, their conduc-
tivities were restored to 10–3 S/cm [96]. This demonstrates that the water con-
tent in these microporous membranes can be freely exchanged for organic
solvents or electrolyte solutions, indicating that they may be further developed
as composite polymeric electrolytes for lithium/lithium ion rechargeable
batteries.
274 P. Y. Chow · L. M. Gan

3.1.2
Proton-Exchange Membranes

The bicontinuous-microemulsion polymerization technique has also been used

to develop novel proton exchange membranes (PEM) for fuel cell evaluation
[97]. A series of hydrocarbon-based membranes were prepared based on the
formulation shown in Fig. 5, with additional ionic vinyl monomers such as VB-
SLi or bis-3-sulfopropyl-itaconic acid ester. After polymerization, the mem-
branes were treated with dilute H2SO4 (0.5 M) to convert them to PEM mem-
branes. The good performance of these PEM membranes in a single fuel cell is
illustrated in Fig. 9.

Fig. 9 Membrane performance of a single fuel cell (area: 5 cm2, Pt loading on anode/cathode:
0.4 mg/cm2) using commercial membranes 1 and 2, and microemulsion-synthesized mem-
branes HC-1 and HC-2

3.2
Polymer Nanocomposites Produced by Bicontinuous-Microemulsion
Polymerizations

3.2.1
Ruthenium (II) Complexes in Polymerized Bicontinuous-Microemulsions

Bicontinuous microemulsions consisting of cationic surfactant AUTMAB

(30–40 wt%), MMA (30–40 wt%), and 20–40 wt% of an aqueous 50 mM solu-
tion of water-soluble metal complexes such as Ru(dip)3Cl2 (dip=4,7-diphenyl-
1,10-phenanthroline) have been investigated [98]. After polymerization, the
microemulsion transformed into a transparent polymer film which showed a
remarkable enhancement in luminescence intensity. The emission lifetime also
Microemulsion Polymerizations and Reactions 275

Fig. 10 The effect of potassium ferrocyanide on the luminescence spectra of [Ru(dip)3]2+

in polymerized cationic microemulsion: (1) 0 ppm, (2) 1 ppm, (3) 3 ppm, (4) 5 ppm, (5)
7 ppm and (6) 9 ppm. The contact time between the potassium ferrocyanide solution and
the polymer film was 10 min. The insert is the Stern-Volmer plot for the quenching of
[Ru(dip)3]2+ by potassium ferrocyanide

increased substantially from 0.98 ms (aqueous), to 2.0 ms (fluid microemulsion),

to 5.8 ms for the polymerized microemulsion. The substantial increases in emis-
sion intensity and lifetime of [Ru(dip)3]2+ are ascribed to rigidochromism re-
sulting from the transformation of the dynamic nanostructure of the fluid
microemulsion into its solidified state after polymerization. In addition, the
fluorescence of [Ru(dip)3]2+ could largely be quenched within 5–10 min by im-
mersing the polymer film in a dilute (several ppm) aqueous solution of po-
tassium ferrocyanide, as demonstrated in Fig. 10. This was possible because the
polymer film possessed nanoporous open structures which enabled water and
[Fe(CN)6]4– ions to freely and rapidly pass through the nanochannels of the
polymer network. The use of the polymerized bicontinuous microemulsion as
a matrix allowed [Ru(dip)3]2+ complexes to be immobilized, resulting in its
high sensitivity and fast response to the quencher from the external phase. This
novel approach can be further developed for sensors and nanocomposites with
specific functionality.
276 P. Y. Chow · L. M. Gan

3.2.2
Aligned Nanocomposites of Ferrite-Polymer from Bicontinuous-Microemulsion
Polymerization

It has been reported [99] that flame spraying of microemulsions containing

nanoparticles can lead to the deposition of thin films. This is an alternative
method to laser-assisted ablation and other thermal methods to evaporate
metallic substrates and condense the vapors onto thin film.A novel strategy has
recently been employed in the assembly of magnetic nanoparticles in polymer-
nanoparticle composites using an external magnetic field [100]. A new type of
hybrid colloidal assembly, combining a bicontinuous microemulsion matrix
with a suspension of colloidal magnetic particles (ferrofluid), was investigated
in the study. One distinct feature of this system is that the polymer matrix con-
sists of numerous randomly interconnected channels that allow solid magnetic
nanoparticles to be incorporated after the structure of the bicontinuous micro-
emulsion matrix has been locked into place by polymerization.
The precursor, magnetic nickel ferrite nanoparticles, was prepared from
mixing two W/O microemulsions, one of which contained Ni-Fe nitrate solu-
tion and the other containing ammonium hydroxide. The nickel-iron hydrox-
ide particles formed were recovered by centrifugion followed by calcining at
500 °C for 5 h, giving the complete conversion of the hydroxide particles into
magnetic nickel ferrite (NiFe2O4). The ferrofluid was then prepared by dis-
persing the ferrite particles in the aqueous solution stabilized by PEO-macro-
monomer surfactant (C1-PEO-C11-MA-40). The ferrite-microemulsion, consist-
ing of about 33 w% water, 1–2 wt% of NiFe2O4, 35 wt% of C1-PEO-C11-MA-40,
21 wt% acrylonitrile (AN) and 9 wt% MMA was polymerized by redox initia-
tion in a glass cell at room temperature. The transparent solid sample was
microtoned to thin film, for TEM observation as shown in Fig. 11.
In the absence of a magnetic field during the polymerization, the nanopar-
ticles of NiFe2O4 were randomly distributed in the polymerized microemulsion
sample, as shown in Fig. 11a. However, under a magnetic field of a particular
field strength, NiFe2O4 particles in the aqueous channels of the microemulsion
could be aligned with the field without causing much agglomeration of the
magnetic particles. As the polymerization of microemulsion advanced to the
solidified state, the magnetic particles were locked within the water channels
of the polymerized matrix along the direction of the magnetic field, as revealed
by Fig. 11b. By further refining the polymerization method, magnetic particles
may be patterned in polymer substrate by careful control of the external mag-
netic field. Many other magnetic/polymer nanocomposites can be prepared in
a similar way.
This bicontinuous-microemulsion polymerization method can also be used
to synthesize polymer nanocomposites containing SiO2 [101], TiO2, ZnO and
many other semiconductors. The advantage of this method is that the nano-
particles of inorganic materials can be dispersed in the polymer matrix fairly
uniformly. The only requirement is that nanomaterials should be first stabilized
Microemulsion Polymerizations and Reactions 277

b
Fig. 11 TEM micrographs of polymerized ferrite-microemulsion nanocomposites contain-
ing 2 wt% NiFe2O4 . a No magnetic field during polymerization; b under the influence of
H=50 Oe field during polymerization

in an aqueous phase of microemulsions without causing phase separation dur-

ing polymerization. Hybrid inorganic polymer materials can also be prepared
from ternary microemulsions [102].

3.3
Synthesis of Nanocomposites Via In-Situ Microemulsion Polymerization

Some non-oxide nanoparticles such as PbS and CdS can be used to prepare
polymer-inorganic nanocomposites by a double-microemulsion process [103].
In this case, two precursor microemulsions must be prepared separately first
and then mixed together for polymerization. Using CdS-polymer nanocom-
278 P. Y. Chow · L. M. Gan

Fig. 12 TEM micrograph of Au-polymer nanocomposite prepared via in situ microemul-

sion polymerization

posite as an example, common compositions of two precursor microemulsions

were 35 wt% MMA and 35 wt% AUDMAA. One of the microemulsions con-
tained 30 wt% of 10 mM CdBr2 aqueous solution, while the other contained
30 wt% of 10 mM (NH4)2S aqueous solution. Two precursor microemulsions
were then mixed for polymerization at 55 °C using AIBN initiator.
A dispersion of nanoparticles of Au or other metals in a polymer matrix may
also be obtained by a one-pot process of microemulsion polymerization. For
instance, the UV-polymerization of a microemulsion of 35 wt% MMA, 35 wt%
AUDMAA and 30 wt% of 0.1 M HAuCl4 aqueous solution would produce a Au-
polymer nanocomposite, as shown in Fig. 12 [104]. This TEM micrograph
shows a microtoned thin film of the sample. It is clearly apparent that Au par-
ticles of about 10–15 nm are well dispersed in the polymer matrix.

4
Microemulsion Reactions for Processing Inorganic Nanomaterials

The existence of microdomains or droplets with large interfacial area per unit
volume in inverse and bicontinuous microemulsions opens up possibilities for
controlling the inorganic reaction rates, the pathways, the stereochemistry, and
the morphology of the products. By employing microemulsions, monodisperse
particles of a few nanometers in diameter can be prepared for the study of
quantum size effects. Many new and unusual physical and chemical properties
also arise as particles attain nano-sized dimensions [105–107]. There is increas-
ing recognition that aqueous synthesis offers growth control capabilities that
can be conveniently exploited to prepare these desirable fine particles [107,
108]. The dependence of the size and polydispersity of the particles obtained
on the size and concentration of the microemulsion droplets, the concentration
Microemulsion Polymerizations and Reactions 279

of the dispersed reactants, the exchange rate between the droplets, and the
reactions and interactions between surfactants and metal ions or particles has
already been partly addressed [109–114].
Compared to conventional solid-state reaction methods, solution-based syn-
thesis results in higher levels of chemical homogeneity. In solution systems,
mixing of the starting materials is achieved at the molecular level, and this is
especially important when multi-component oxides are being prepared. As a
solution-based materials synthesis technique, the microemulsion-mediated
method offers a unique ability to affect particle synthesis and particle stabil-
ization in one step. A wide variety of nanosize metal oxides – binary oxides
from alumina to zirconia, as well as complex metal oxides – can be prepared by
exploiting the ability of microemulsions to solubilize, compartmentalize, and
concentrate reactants and products. Most of the focus in the past has been on
the successful synthesis of specific metal oxide compounds as nanoparticles.
Water-in-oil microemulsions (inverse microemulsions) have been the most
widely used for the preparation of spherical nanoscale particles. Reagents dis-
solved in the water domains of an inverse microemulsion can react via inter-
micellar communication through dynamic collision processes. Nucleation and
growth occur as the process proceeds.
Many potential applications of microemulsion-derived materials appear
in the literature. The list includes catalysts, nanoporous membranes, nanocom-
posites and precursor powders for functional ceramics [115–139]. However,
before microemulsion-derived materials become more than laboratory curios-
ities, two key issues must be addressed: product recovery and yield. Materials
generated in microemulsion media can be used in two ways: as prepared in the
fluid phase, or as recovered solids. A common practice found in the literature
involves phase destabilization of microemulsions via the addition of polar
solvents such as ethanol and acetone to precipitate the products. This approach
works well as a laboratory technique that permits particles to be recovered
for later characterization. However, it is unlikely to be an economically viable
approach for practical processing, as we eventually need to separate the added
solvents in order to reformulate the original microemulsion. The susceptibility
of microemulsion to destabilization by electrolytes severely limits the highest
concentrations that can be used for precipitation reactions. This has an adverse
impact on prospects for large-scale microemulsion processing. The feasibility
of microemulsion-derived synthesis depends on the availability of surfactant/
oil/water formulations that give stable microemulsions during reaction, but not
during precipitation reactions. Therefore, investigations that emphasize the
effects of reactants and products on the stability domains of microemulsions
will play an important role. Hopefully, such investigations will lead to guide-
lines for formulating microemulsion compositions that are compatible with
large-scale material synthesis.
280 P. Y. Chow · L. M. Gan

4.1
Synthesis of Inorganic Nanoparticles in Inverse Microemulsion

For reactions in inverse microemulsions that involve the total confinement of

the reactant species within the dispersed water droplets, the exchange of reac-
tants by the coalescence of the two droplets take place prior to their chemical
reaction. The chemical reaction produces an (almost) insoluble product. The
reaction medium is first saturated with this product. When the saturation
exceeds a critical limit, nucleation occurs. Then the nuclei start to grow rapidly
and consume the reaction product leading to a decline in the supersaturation.
As soon as the supersaturation falls below the critical level, no further nuclea-
tion occurs, so only the existing particles grow beyond this point. If the time
period of nucleation is short in comparision to the growth period, rather
monodisperse particles are obtained.
Conceptually, these particle synthesis methods can be classified into three
main groups:
1. Double inverse microemulsion
2. Inverse microemulsion plus trigger
3. Inverse microemulsion plus a second reactant, as illustrated in Fig. 13

It is apparent that a number of chemical reactions involving aqueous, organic

and/or amphiphilic reagents may be accommodated in micellar systems through
some of these methods or a combination of them. It is important to note that,
in these processes, the reverse micelles not only serve as a passive host, but also
provide hydrophobic and electrostatic interactions between the micelles. The
reagents used have a profound influence on the chemical and physical proper-
ties of the resulting particles.
For the double inverse microemulsion technique, we use two separate micro-
emulsions: reactant A solubilized in a specific microemulsion composition, and
reactant B solubilized in another specific microemulsion composition. When
the two microemulsions are mixed, reactants A and B will then mix within the
microemulsion droplets, preventing sudden precipitation of the product AB.
If reactants A and B were to be pre-mixed without using the double micro-
emulsion technique, it would be difficult to control the particle sizes and to pre-
vent agglomeration of particles. This is one of the principles used to produce
nanoparticles with double inverse microemulsions (Fig. 13a). On the other
hand, nanoparticles can also be produced in inverse microemulsions by adding
a trigger (such as heat/light or a reducing/precipitating agent) to an inverse
microemulsion containing the primary reactant dissolved in the aqueous core
(Fig. 13b). Figure 13c shows oxide, hydroxide or carbonate precipitates formed
by passing gases such as oxygen, ammonia or carbon dioxide through an
inverse microemulsion containing soluble salts of the cations.
Since numerous papers have been published on the preparation of inorganic
nanomaterials by microemulsions, we do not intend to review all of them here.
As examples, we will describe how inverse microemulsions can be used to pre-
Microemulsion Polymerizations and Reactions 281

Fig. 13 Synthesis of nanoparticles in microemulsions using: a double inverse microemul-

sions; b inverse microemulsion plus a trigger; c inverse microemulsion plus reactant [122]

Table 2 Inorganic materials synthesized using inverse microemulsion

Material Inverse Investigation/results Reference

microemulsion
(surfactant-oil)

CdS NP-5+NP-9 and To study the synthesis of CdS using [141]

petroleum ether hydrothermal microemulsion
PbS-coated CdS NP-5 and To study the kinetic growth and [142]
petroleum ether nonlinear optical response
of PbS-coated CdS particles
Mn-doped ZnS NP-5+NP-9 and To compare the properties [143]
of Mn-ZnS particles prepared
via a conventional reaction,
a conventional reaction with
hydrothermal treatment,
a microemulsion, and a micro-
emulsion with hydrothermal
treatment
282 P. Y. Chow · L. M. Gan

Table 2 (continued)

Material Inverse Investigation/results Reference

microemulsion
(surfactant-oil)

NiFe2O4 NP5+NP-9 and To study and characterize NiFe2O4 [144]

cyclohexane
BaFe12O19 NP-5+NP-9 and Ultrafine, high coercivity BaFe12O19 [145]
cyclohexane was prepared via inverse micro-
emulsion and compared to that
prepared using the conventional
method
SiO2 NP-5+NP-9 and Influence of pH and concentration [153]
cyclohexane of sodium orthosilicate on the
silica particle size, morphology
and specific surface area
Cu-doped SiO2 AOT+SDS and To prepare and characterize Cu-SiO2 [134]
cyclohexane particles
Ru-Cu-doped AOT+SDS and The bimetallic Ru-Cu oxides [135]
SiO2 cyclohexane supported on silica showed high
catalytic conversion of N2O
Perovskite NP-5 and octane To prepare and characterize [155]
powders or petroleum perovskites
ether
Zirconia and NP-5+NP-9 and To produce zirconia, lead zirconate [158–161]
zirconate cyclohexane and lead zirconate titanate at much
lower calcination temperatures
Hydroxyapetite NP-5+NP-9 and To compare differences in the sizes, [162]
(HA) cyclohexane morphologies and specific surface
areas of HA powder prepared by
W/O microemulsion, bicontinuous
microemulsion and the emulsion
method
Hydroxyapetite KB6ZA and To synthesize HA powders in [163]
(HA) petroleum ether O/W emulsion using the nonionic
surfactant KB6ZA
Pt-Ru NP-5+NP-9 and To synthesize and characterize [164]
cyclohexane Pt-Ru catalyst for fuel cell
applications
Microemulsion Polymerizations and Reactions 283

pare semiconductors, nano-sized oxide particles, and hydroxapetite nano-

particles, as illustrated in Table 2. The attraction of using nonionic polyoxyethy-
lated alkyphenyl ether surfactants (like NP-5, NP-9, or mixture of NP-5/NP-9)
is that it allow us to formulate inverse microemulsions without the need for
cosurfactant; no undesirable counterions are needed. In addition, the various
sizes of the hydrophilic (oxyethylene) groups or/and the hydrophobic (alkyl)
groups provide flexibility during the selection of surfactant.

4.2
Materials Systems

4.2.1
Doped and Un-Doped CdS/ZnS Nanoparticles

Semiconductor nanoparticles have been extensively studied in recent years

owing to their strongly size-dependent optical properties. Among these nano-
materials, CdS and PbS are particularly attractive due to their nonlinear optical
behavior and unusual fluorescence or photoluminescence properties [136, 137].
A number of studies have been published recently regarding the preparation
of CdS, PbS and ZnS nanoparticles in inverse microemulsion systems [138–
143]. In these works, NP-5/NP-9 was the most commonly used surfactant and
petroleum ether the most commonly used oil. The aqueous phase for each
inverse microemulsion consisted of cadmium nitrate (0.1 M) and ammonia
sulfide (0.1 M) respectively. CdS was recovered from the mixture of double
microemulsions [141]. Electron microscopy revealed that the spherical par-
ticles were around 10–20 nm in diameter, as seen in Fig. 14.
The synthesis of PbS-coated CdS was conducted in a microemulsion system
similar to that mentioned above, except that a third inverse microemulsion,

Fig. 14 Transmission electron micrograph (TEM) of a CdS cluster synthesized at 30 °C [141]

284 P. Y. Chow · L. M. Gan

Fig. 15 Formation of PbS-coated CdS nanocomposite in microemulsion: a mixing of a micro-

emulsion containing Cd(NO3)2 aqueous solution with a microemulsion containing (NH4)2S
aqueous solution; b formation of CdS nanoparticle in microemulsion; c,d Pb2+ ions in a third
microemulsion replace the Cd2+ in the Cd-S band and diffuse through the PbS layer to form
the PbS shell [142]

containing lead nitrate (0.1 M) in the aqueous phase, was also needed [142].
This is to encourage the formation of PbS shells around the CdS particle cores,
as depicted Fig. 15. Pb2+ ions displaced some Cd2+ in the Cd–S bonds to form
Pb–S bonds on the surfaces of the CdS particles. The thicknesses of the PbS
shells in the PbS-coated nanocomposite could be controlled by changing the
Pb2+ ion concentration. Absorption measurements of the mixed systems as a
function of the concentration of lead nitrate solution indicated a continuous
trend toward the PbS spectrum with increasing Pb2+ concentration. The par-
ticle sizes of the CdS, PbS and CdS/PbS composites are shown in Fig. 16. The
resulting NLO properties of the CdS and PbS and CdS/PbS nanocomposite
particles were determined by a femtosecond Z-scan technique. A large refrac-
tive nonlinearity in these nanocomposite particles was observed due to the
optical Stark Effect and strong interfacial and inter-particle interactions. These
nanoparticles with large refractive nonlinearities may find application in opti-
cal devices, such as those for optical limiting and switching.
Mn-doped ZnS powders were also prepared by inverse microemulsion under
hydrothermal treatment at 120 °C [143]. This produced ultrafine and agglom-
erate-free particles 5–20 nm in diameter. In contrast, the particles prepared via
conventional aqueous solutions often coagulated, forming large aggregated
clusters. Furthermore, compared with Mn-doped ZnS materials synthesized
through conventional aqueous reactions, the nanoparticles prepared in mi-
croemulsion showed significantly enhanced photoluminescence. In particular,
the photoluminescence of particles prepared in microemulsion under hydro-
thermal treatment was found to be 60 times higher than that for material
Microemulsion Polymerizations and Reactions 285

Fig. 16 Transmission electron micrographs of a CdS, b PbS nanoparticles, and c PbS-coated

CdS nanocomposite [142]
286 P. Y. Chow · L. M. Gan

Fig. 17 Photoluminescence (PL) of ZnS:Mn2+ particles prepared by: a conventional reaction

at room temperature; b conventional reaction with hydrothermal treatment; c microemul-
sion at room temperature; d microemulsion with hydrothermal treatment, with an excita-
tion intensity of 4 W/cm2 [143]

obtained through the direct aqueous reaction at room temperature (as shown
in Fig. 17). This dramatic increase in photoluminescence yield is attributed to
the surface passivation of nanoparticles by the adsorption of surfactants, the
formation of sphelerite with cubic zinc blende structure, and Mn migration
into the interior lattice of the ZnS host.

4.2.2
Magnetic Ferrites

Nano-sized magnetic ferrite particles are the subject of intensive research be-
cause their physical properties are quite different from those of the bulk mate-
rial. The magnetic characteristics of particles used for recording media crucial-
ly depend on their sizes and shapes. So, the material used for high-quality
recording media should be ultrafine, chemically homogeneous, and stable, with
a narrow particle size distribution a predetermined shape. These requirements
demand a reliable and reproducible preparation technique.
The formation of microhomogeneous nanoparticles of nickel and barium
ferrite was carried out in a three-component microemulsion consisting of NP-
5/NP-9 as the surfactant, cyclohexane as the oil, and an aqueous solution of
Microemulsion Polymerizations and Reactions 287

Fig. 18 The saturation magnetization and intrinsic coercivity as a function of calcination

temperature for the microemulsion-derived barium ferrites [145]

nickel nitrate and ferric nitrate [144].An immediate precipitation was affected
by adding ammonia to the microemulsion, and the resulting amorphous pre-
cipitate was transformed into nickel ferrites by calcining at 600 °C. Discrete
nickel ferrite particles with polyhedron shapes and an average size range
from 10–20 nm were obtained. The sol-particles were identified as NiFe2O4 by
Mossbauer spectroscopy, electron diffraction and X-ray spectroscopy. Magnetic
measurement revealed a saturation magnetization value of 43.0 emu/g for the
sample calcined at 500 °C.
Barium ferrite particles were obtained by mixing a ferric ion-containing
microemulsion and a barium-containing microemulsion [145]. Precipitation of
the hydroxide precursor was affected by adding ammonia solution dropwise
into the microemulsions under stirring. The resulting barium ferrite particles
formed at a calcination temperature of 950 °C. The round particles synthesized
were well dispersed, although a limited degree of particle agglomeration was
also observed. The average size of the calcined particles was in the range of
100–200 nm, which was about ten times bigger than the average size of the
precursor particles. A high saturation magnetization of 69.74 emu/g and an
intrinsic coercivity of 5639 Oe were obtained for the powder calcined at 950 °C
(Fig. 18). These results are comparable to some of the best ever reported for fine
barium ferrite powders prepared via chemistry-based processing routes. The
much improved saturation magnetization and coercivity can be explained by
the high phase purity and well-defined crystallinity of BaFe12O19 developed in
the microemulsion-derived precursor when calcined at a temperature that is
high enough.
288 P. Y. Chow · L. M. Gan

4.2.3
Silica and Silica-Supported Ru-Cu Oxides

The vast majority of techniques published for the microemulsion-mediated

synthesis of silica [146–151] are based on the alkoxide sol-gel method. The oil-
soluble metal alkoxide was added to a microemulsion containing solubilized
water. The alkoxide must diffuse through the oil continuous phase to the water
pool where the hydrolysis and condensation reactions take place.
In our studies, the nano-sized silica was synthesized using cheap sodium
orthosilicate and sodium metasilicate rather than expensive alkoxysilanes. An
inverse microemulsion containing NP-5/NP-9 as surfactant and cylcohexane or
petroleum ether as the oil was used to carry out the hydrolysis and conden-
sation of sodium orthosilicate or metasilicate using an acidic medium [152,
153]. The spherical silica particles of size 10–20 nm were obtained in a system
using cyclohexane as the oil and with sodium orthosilicate of 0.01–0.1 M [152].
The particle size increased as the concentration of sodium orthosilicate and
the pH were increased. Silica particles prepared in basic conditions were more
uniform in size than those prepared in an acidic medium. But calcined silica
powders with larger specific areas (350–400 m2/g) were obtained for those pre-
pared in an acidic medium.
Much smaller silica particles of about 5–10 nm were formed by the con-
trolled hydrolysis and polymerization of sodium metasilicate in bicontinuous
microemulsions [153]. The bicontinuous microemulsion system consisted of
NP-5/NP-9, petroleum ether, and a high concentration of sodium metasilicate
solution (0.2 M). This example illustrates the feasibility of using bicontinuous
microemulsion to synthesize silica powders with high specific surface areas
(~400 m2/g) and good amorphous stabilities at 600 °C. The treatment of some
calcined silica powders with hexadecyltrimethylammonium hydroxide increas-
ed their specific surface areas still further, from 350 to 510 m2/g.
The double microemulsion-mediated process also provides a convenient
method for preparing a metal-containing silicate coating. The two microemul-
sion systems contained two common components: anionic surfactant AOT and
cyclohexane [134]. The difference was that the first microemulsion consisted
of an aqueous solution of sodium metasilicate (0.2 M) and 10 wt% SDS as the
co-surfactant, while the second microemulsion consisted of an aqueous solution
of copper nitrate (0.1 M) and 10 wt% SDS. The copper-ion microemulsion was
added to the silicate-ion microemulsion with constant stirring. After 8 h of gel-
lation, and ageing for an additional 24 h, copper nitrate crystals were identified
within the silicate network. Silica-copper composite powders with various cop-
per contents (4–20 wt%) and surface areas of 200–400 m2/g were synthesized.
Using a variation on the microemulsion-mediated process, catalytic bi-
metallic Ru-Cu oxides supported by silica were made through the controlled
hydrolysis/polymerization of sodium metasilicate, copper nitrate and ruthenium
chloride via a single microemulsion protocol [135]. With AOT and SDS-based
microemulsion, a high specific surface area (~400 m2/g) and uniform pore size
Microemulsion Polymerizations and Reactions 289

(~38 Å) of the produced catalyst can be maintained during catalytic reactions.

Moreover, tests of the catalytic activity indicated a high catalytic conversion of
N2O by the catalysts synthesized from the microemulsion process at a lower
temperature (~400 °C) than that prepared from traditional impregnation
processes. The uniform elemental distribution of RuO2 on the SiO2 produced
by this microemulsion process was established by XPS and SEM/line scanning.

4.2.4
Perovskites

The complex oxides that belong to the group known as perovskites have the
general formula ABO3, where the ionic charges on the metals can assume the
form A+B5+O3,A2+B4+O3 and A3+B3+O3. The first microemulsion-mediated pro-
cess to synthesize a perovskite used a double microemulsion protocol based on
a nonionic surfactant Genapol OX/decane/water system [154]. A microemul-
sion containing BaCl2 and TiCl4 in the dispersed phase was mixed with an
oxalic acid-containing microemulsion. However, X-ray diffraction could not
confirm the presence of the intermediate barium titanyl oxalate and the calcin-
ed solid did not yield barium titanate.
The double inverse microemulsion method was also used to synthesize per-
ovskite-type mixed metal oxides [155]. One microemulsion solution contained
nitrate salts of either Ba(NO3)2/Pb(NO3)2, La(NO3)3/Cu(NO3)2 or La(NO3)3/
Ni(NO3)2, and the other microemulsion contained ammonium oxalate or
oxalic acid as the precipitant. These metal oxalate particles of about 20 nm were
readily calcined into single phase perovskite-type BaPbO3, La2CuO4 and LaNiO3.
The calcinations required for the microemulsion-derived mixed oxalates were
100–250 °C below the temperatures used for the metal oxalates prepared by a
conventional aqueous solution precipitation method.

4.2.5
Zirconia, Lead Zirconate and Lead Zirconate Titanate

Most zirconia nanoparticles have been prepared by the microemulsion-medi-

ated alkoxide sol-gel method [156, 157] using zirconium tetrabutoxide as the
alkoxide precursor. The first step is to synthesize zirconium hydroxide pre-
cursor particles via double inverse microemulsions and then they are calcined
to zirconia [158]. Three types of processing routes have been used to prepare
fine lead zirconate (PbZrO3) powders: a conventional solid reaction, conven-
tional coprecipitation using either oxalic acid or ammonia solution as the pre-
cipitant, and microemulsion-refined coprecipitation using either oxalic acid or
ammonia solution. The microemulsion-derived precursors exhibited a much
lower formation temperature for the orthorhombic PbZrO3 phase [159] than
that of the conventionally coprecipitating precursors. Ammonia solution ap-
peared to be a better precipitant than oxalic acid for reducing the formation
temperature of PbZrO3. It was concluded that microemulsion-derived PbZrO3
290 P. Y. Chow · L. M. Gan

powders were much finer and has less particle-particle agglomeration. The
synthesis of PbZrO3 powders could also be performed via a polyaniline-medi-
ated microemulsion process [160].A small amount (~4 wt%) of polyaniline was
retained in the microemulsion-derived oxalate precursor by an in situ poly-
merization of lead oxalate and zirconate oxalate. The in situ polymerization of
aniline on the surfaces of the oxalate particles resulted in the formation of a
well-dispersed precursor powder. Upon calcination at 800 °C, ultrafine lead zir-
conate powder was obtained.
Low temperature synthesis of lead zirconate titanate (PZT) can also be obtain-
ed via a microemulsion process [161]. The microemulsion, containing cations
of lead zirconium and titanium in the aqueous phase, was coprecipitated as
hydroxide precursors by the addition of ammonium solution. Crystalline
tetragonal PZT powders were then obtained by calcining the precursors at a
temperature as low as 450 °C in air without forming any intermediate phases.

4.2.6
Hydroxyapetite (HA)

Hydroxyapetite (Ca10(PO4)6(OH)2) is the major constituent of human bone and

teeth. The strength of HA in almost all of its applications is largely determined
by its specific surface area, and so a nanocrystalline HA powder is required.
Ultrafine HA powders can be produced by reacting CaCl2 and (NH4)2HPO4
in bicontinuous microemulsion,W/O microemulsion and emulsion, which have
the same basic components of cyclohexane, nonionic surfactant (NP-5/NP-9)
and aqueous solution [162]. HA powder particle size, chemical homogeneity,
and the degree of particle agglomeration all depend on the reaction medium.
Both bicontinuous and W/O microemulsions lead to the formation of much
finer HA powders than those prepared from the emulsion composition. This is
because the precipitates formed in both bicontinuous and W/O microemul-
sions are limited in size by the dimensions of the nano-sized reaction domains.
In contrast, the aqueous droplets in the emulsion are much bigger, and there-
fore the HA particles obtained are twice as large as those formed in nano-sized
domains. It was found that the specific surface area of the powder prepared by
bicontinuous microemulsion was 86.03 m2/g, 76.64 m2/g by W/O microemul-
sion, and 42.63 m2/g by emulsion. Much smaller degrees of particle agglomera-
tion were observed for HA powders obtained by bicontinuous and W/O micro-
emulsions, but the powders prepared from emulsion showed a larger average
agglomerate size of 1.1 mm, which was twice the size observed from the former
two.
A similar study was performed on the formation of nanocrystalline HA in
nonionic surfactant emulsions [163]. Instead of using NP-5/NP-9 surfactant,
KB6ZA (nonionic surfactant which is a lauryl alcohol condensed with an
average of 6 mol of oxyethylene oxide) was used together with petroleum ether
as the oil phase to prepare HA powder in an O/W emulsion system. One of the
very apparent advantages of using O/W emulsion over W/O microemulsion is
Microemulsion Polymerizations and Reactions 291

the high production yield using smaller amounts of oil and surfactant. In this
study, HA was prepared by reacting CaCl2 and (NH4)2HPO4 in three reaction
systems: a conventional aqueous solution, a micellar solution containing
3.2 wt% of the nonionic surfactant KB6ZA and 1.0 M CaCl2 solution, and O/W
emulsions containing KB6ZA with varying amounts of petroleum ether and
1.0 M CaCl2 solution. For the emulsion system, the oil phase residing in the
hydrophobic cores of the oil-swollen micelles lead to a large expansion in the
micellar dimensions, in the form of emulsion droplets. This undoubtedly in-
creases the reaction sites between CaCl2 and (NH4)2HPO4 at the interfaces, as
shown schematically in Fig. 19. Moreover, the emulsion droplets are more stable
than the surfactant micelles and the average life-span of an emulsion droplet
is much longer than that of a micelle. All of these factors make an emulsion
system much more suitable than a micellar system for forming HA particles of
high crystallinity. The enhanced crystallinity of HA derived from the emulsion
composition derives from the complexation of Ca2+ by oxyethylene groups of
the nonionic surfactant. The resulting HA precursor powders underwent little
growth in crystallite and particle sizes upon calcination at 650 °C for 6 h.

4.2.7
PtRu/C Catalysts

PtRu alloys are currently the most active anode catalysts for the oxidation of
methanol or CO-contaminated H2, (like H2 derived from reformed methanol)
in low temperature solid polymer electrolyte fuel cells such as direct methanol
fuel cells (DMFC) or indirect methanol fuel cells (IMFC). High surface area
catalysts are generally prepared by co-impregnation,coprecipitation, absorbing
alloy colloids, or surface organometallic chemistry techniques. For both alloy
and oxide promoted catalytic systems it is important that Pt and the second
metal (or metal oxide) are in intimate contact. This close association of the
platinum and the co-catalyst can be difficult to achieve using conventional
catalyst preparation techniques because the active components may be depo-
sited at different sites on the support surface.As the preparation details control
the final composition, surface structure and morphology of the catalysts, it is
not surprising to learn that catalytic activity is strongly dependent on prep-
aration conditions.
Nano-sized PtRu catalysts supported on carbon have been synthesized
from inverse microemulsions and emulsions using H2PtCl6 (0.025 M)/RuCl3
(0.025 M)/NaOH (0.025 M) as the aqueous phase, cyclohexane as the oil phase,
and NP-5 or NP-9) as the surfactant, in the presence of carbon black suspended
in a mixture of cyclohexane and NP-5+NP-9 [164]. The titration of 10% HCHO
aqueous solution into the inverse microemulsions and emulsions resulted
in the formation of PtRu/C catalysts with average particle sizes of about 5 nm
and 20 nm respectively. The RuPt particles were identified by X-ray diffraction,
X-ray photoelectron, and BET techniques. All of the catalysts prepared show
characteristic diffraction peaks pertaining to the Pt fcc structure. XPS analysis
292 P. Y. Chow · L. M. Gan

Fig. 19 Micelle and emulsion droplets in CaCl2 aqueous solutions: The calcium-rich shell
in the adsorbed state, at the: a micelle-water interface; b emulsion droplet surface. The surfac-
tant molecules are represented by the zig-zag lines with Ca2+ heads
Microemulsion Polymerizations and Reactions 293

also revealed that the catalysts contained mostly Pt(0) and Ru(0), with a little
Pt(II), Pt(IV) and Ru(IV). The double microemulsion-derived PtRu/C catalysts
had higher electrocatalytic activities for methanol oxidation than that of the
emulsion-derived PtRu/C electrocatalyst. The peak current density for metha-
nol oxidation at PtRu/C obtained from the transparent inverse microemulsion
was about four times higher that of the emulsion-derived PtRu/C.

4.2.8
Polymer-Coated Inorganic Nanoparticles

Nanoparticles of polyaniline/BaSO4 composite, about 10–20 nm in diameter,

have also been synthesized by mixing double inverse microemulsions that have
some components of surfactant and oil, but different aqueous solutions [165].
The aqueous solution of one microemulsion should contain BaCl2 and aniline,
while the other contains K2S2O8 and H2SO4.After mixing both microemulsions,
nanoparticles of BaSO4 are formed instantaneously with the subsequent poly-
merization of aniline on the surface of BaSO4 particles, as verified by Ruther-
ford backscattering measurements. The conductivity of the polyaniline/BaSO4
nanocomposites increased almost linearly from 0.017 to 5 S/cm withincreasing
polyaniline content in the composites (which was varied from 5 to 22 wt%).
This result shows that a selection of double inverse microemulsion reactions for
producing some nanocomposites of polymer/inorganic materials can be con-
veniently carried out by a single-step microemulsion process.

5
Conclusions

The synthesis of polymer nanomaterials by microemulsion processes is reach-

ing maturity. Microlatexes of both hydrophilic and hydrophobic polymers can
now be produced by semi-continuous microemulsion processes yielding high
weight ratios of polymer to surfactant at a level attractive for to industry. How-
ever, the W/O microemulsion processes capable of producing inorganic nano-
materials with unique properties are not ready for large-scale production,
because large amounts of surfactant are still needed. Bicontinuous micro-
emulsion processes should be explored for inorganic nanomaterial production
since they require lower amounts of surfactant and give higher yields.
The recent development of using polymerizable surfactants in microemulsion
polymerizations has enabled the production of transparent solid polymers with
some nanostructure. Randomly distributed bicontinuous nanostructures of wa-
ter channels and polymer domains in solid polymers can be readily obtained
from the polymerization of bicontinuous microemulsions consisting of various
types of vinyl monomers and polymerizable surfactants with no allylic hydrogen.
These transparent polymers with inherent bicontinuous nanostructures
may be suitable for nanofiltration, selective permeable membranes for sepa-
294 P. Y. Chow · L. M. Gan

ration, or proton exchange membranes for fuel cell applications. The surface
characteristics of the nanostructured membranes can be modified further us-
ing suitable functional monomers. New polymerizable microemulsion systems
will be devised for specific copolymers and inorganic/polymer nanocompos-
ites. Some tailor-made nanostructure polymeric materials may therefore be ob-
tained by the polymerizable bicontinuous-microemulsion approach.

References
1. Zana R, Lang J (1987) In: Frieberg SE, Bothorel (eds) Microemulsions: structure and
dynamics. CRC, Boca Raton, FL, Ch 6
2. Biasia J, Clin B, Laolanne P (1987) In: Friberg SE, Bothorel P (eds) Microemulsions:
structure and dynamics. CRC, Boca Raton, FL, Ch 1
3. Frieberg SE, Bothorol P (1986) In: Friberg SE, Bothorel P (eds) Microemulsions: struc-
ture and dynamics. CRC, Boca Raton, FL
4. Winsor PA (1948) T Faraday Soc 44:376
5. Candau F (1999) In: Kumar P, Mittal KL (eds) Handbook of microemulsion science and
technology. Marcel Dekker, New York, Ch 22, p 679
6. Gan LM, Chew CH (2001) In: Nalwa HS (ed) Advanced functional molecules and
polymers. Gordon and Breach, New York, Ch 2, p 35
7. (a) Lopez-Quintela MA, Rivas J (1993) J Colloid Interf Sci 158:446; (b) Aikawa K, Kaneko
K, Tamura T, Fujitsu M, Ohbu K (1999) Colloids Surface A 150:95; (c) Zarur AJ,Ying JY
(2000) Nature 403:65; (d) Rymes J, Ehret G, Hilaire L, Boutonnet M, Jiratova K (2002)
Catal Today 75:297
8. Klier J, Tucker CJ, Kalantar TH, Green DP (2000) Adv Mater 18:1751
9. Gan LM, Chew CH (1996) In: Salamone (ed) Polymeric materials encyclopedia. CRC,
Boca Raton, FL, M4321
10. Candau F (1992) In: Paleos CM (ed) Polymerization in organized media. Gordon and
Breach, New York, p 215
11. Candau F (1998) In: Kumar P, Mittal KL (eds) Microemulsions: fundamental and
applied aspects. Marcel Dekker, New York
12. Herrera JR, Peralta RD, Lopez RG, Cesteros LC, Puig JE (2003) Polymer 44:1795
13. Candau F, Leong YS, Pouyet G, Candau SJ (1984) J Colloid Interf Sci 101:167
14. Candau F, Yong YS, Fitch RM (1985) J Polym Sci Pol Chem 23:193
15. Candau F (1987) In: Mark H, Bikales NM, Overberger CG, Menges G (eds) Encyclopedia
of polymer science and engineering. Wiley, New York, 9:718
16. Vaskova V, Juranicova V, Barton J (1990) Makromol Chem 191:717
17. Vaskova V, Hlouskova Z, Barton J, Juranicova V (1990) Makromol Chem 193:267
18. Lezovic M, Ogino K, Sato H, Capek I, Barton J (1998) Polymer Int 46:269
19. Barton J, Kawamoto S, Fujimoto K, Kawaguchi H, Capek I (2000) Polymer Int 49:358
20. Barton J, Capek I (2000) Macromolecules 33:5353
21. Candau F (1989) In: El-Nokaly M (ed) ACS Sym Ser 384, Ch 4
22. Canadu F, Anquetil JY (1998) In: Shah DO (ed) Micelles, microemulsions and mono-
layers. Marcel Dekker, New York, p 193
23. Candau F, Zekhnini Z, Durand JP (1987) Prog Colloid Polym Sci 73:33
24. Corpart JM, Candau F (1993) Colloid Polym Sci 271:1055
25. Corpart JM, Selb J, Candau F (1993) Polymer 34:3873
26. Candau F, Buchert P (1990) Colloids Surface 48:107
27. Candau F, Pabon M, Anqueti (1999) Colloids Surface A 153:47
Microemulsion Polymerizations and Reactions 295

28. Sosa N, Peralta RD, Lopez RG, Ramos LF, Katine I, Certeros C, Mendizabal E, Puig JE
(2001) Polymer 42:6923
29. Braun O, Selb J, Candau F (2001) Polymer 42:8499
30. Candau F, Braun O, Essler F, Stahler K, Selb J (2002) Macromol Symp 179:13
31. Stoffer JO, Bone T (1980) J Disper Sci Technol 1:37
32. Atik SS, Thomas JK (1981) J Am Chem Soc 103:4279
33. Tang HI, Johnson PL, Gulari E (1984) Polymer 25:1357
34. Kuo PL, Turro NJ, Tseng CM, El-Aasser MS, Vanderhoft JW (1987) Macromolecules
20:1216
35. Feng L, Ng KY (1990) Macromolecules 23:1048
36. Gan LM, Chew CH, Lye I (1992) Makromol Chem 193:1249
37. Gan LM, Chew CH, Frieberg SE (1983) J Macromol Sci Chem A 19:739
38. Ferrick MR, Murtagh J, Thomas JK (1989) Macromolecules 22:1515
39. Antonietti M, Bremser W, Muschenborn D, Rosenaur C, Schupp B, Schmidt M (1991)
Macromolecules 24:6636
40. Rodriguez-Guadarrama LA, Mendizabal E, Puig JE, Kaler EW (1993) J Appl Polym Sci
48:775
41. Gan LM, Chew CH, Ng SC, Loh SE (1993) Langmuir 9:2799
42. Gan LM, Chew CH, Lee KC, Ng SC (1993) Polymer 34:3064
43. Larpent C, Tados RF (1991) Colloid Polym Sci 269:1171
44. Full AP, Puig JE, Gron LU, Kaler EW, Minter JR, Meurey TH, Texter J (1992) Macro-
molecules 25:5157
45. Capek I, Potisk P (1995) Eur Polym J 31:1269
46. Capek I, Fouassier JP (1997) Eur Polym J 33:173
47. Capek I, Juranicova V (1996) J Polym Sci Pol Chem 34:575
48. Capek I, Juranicova V, Barton J, Asua JM, Ito K (1997) Polymer Int 43:1
49. Guo JS, Sudol ED, Vanderhoff JW, El-Aasser MS (1992) J Polym Sci Pol Chem 30:691
50. Puig JE, Perez-Luna VH, Macias ER, Rodriguez BE, Kaler EW (1993) Colloid Polym Sci
271:114
51. Co CC, Cotts P, Burauer S, deVries R, Kaler EW (2001) Macromolecules 34:3245
52. Kukulj D, Davis TP, Gilbert RG (1998) Macromolecules 31:994
53. Hentze HP, Kaler EW (2003) Curr Opin Coll Interf Sci 8(2):164–178
54. Mendizabal E, Flores J, Puig JE, Katime I, Lopez-Serrano F, Alvarez J (2000) Macromol
Chem Phys 201:1259
55. Co CC, Kaler EW (1998) Macromolecules 31:3203
56. Morgan JD, Kaler EW (1998) Macromolecules 31:3197
57. Sanghvi PG, Pokhriyal NK, Devis (2002) J Appl Polym Sci 84:1832
58. Gan LM, Chew CH, Lee KC, Ng SC (1994) Polymer 35:2659
59. Loh SE, Gan LM, Chew CH, Ng SC (1995) J Macromol Sci Pure A32:1681
60. Gan LM, Lee KC, Chew CH, Tok ES, Ng SC (1995) J Polym Sci Pol Chem 33:1161
61. Sutterline N, Kurth HJ, Markett G (1976) Angew Makromol Chem 177:1549
62. Aguiar A, Gonzales-Villegas S, Rabelero M, Mendizabal E, Puig JE (1999) Macromole-
cules 32:6767
63. Jang J, Ha H (2002) Langmuir 18:5613
64. Jang J, Lee K (2002) Chem Comm 1098
65. Gan LM, Lian N, Chew CH, Li GZ (1994) Langmuir 10:2197
66. Xu XJ, Siow KS, Wong MK, Gan LM (2001) Langmuir 17:4519
67. Ming W, Jones FN, Fu SK (1998) Polym Bull 40:749
68. Ming W, Jones FN, Fu SK (1998) Macromol Chem Phys 199:1075
69. (a) Xu XJ, Chew CH, Siow KS, Wong MK, Gan LM (1999) Langmuir 15:8067; (b) Chudej
J, Capek I (2002) Polymer 43:1681
296 P. Y. Chow · L. M. Gan

70. Dan Y, Yang YH, Chen SY (2002) J Polym Sci 85:2839

71. Barrere M, Silva SC, Balic R, Ganachaud F (2002) Langmuir 18:941
72. Xu XJ, Chow PY, Quek CH, Hng HH, Gan LM (2003) J Nanosci Nanotechno 3:3
73. Xu XJ, Siow KS, Wong MK, Gan LM (2001) Colloid Polym Sci 279:879
74. Xu XJ, Siow KS, Wong MK, Gan LM (2001) J Polym Sci Polym Chem 39:1634
75. Xu XY, Chow PY, Gan LM (2001) J Nanosci Nanotechno (in press)
76. Menger FM, Tsuno T, Hammond GS (1990) J Am Chem Soc 112:1263
77. Sáenz de Buruaga A, Capek I, de la Cal JC, Asua JM (1998) J Polym Sci Pol Chem 36:737
78. Antonietti M, Basten R, Gröhn F (1994) Langmuir 10:2498
79. Puig JE, Aguiar A, González-Villegas S (1999) Macromolecules 32:6767
80. Kaler EW, Co CC, de Vries R (2001) Macromolecules 34:3224
81. Kaler EW, Co CC, de Vries R (2001) Macromolecules 34:3233
82. Palani Raj WR, Sathav M, Cheung HM (1991) Langmuir 7:2586
83. Sasthav M, Cheung HM (1991) Langmuir 7:1378
84. Strey R, Lade O, Beizai K, Sottmann T (2000) Langmuir 16:4122
85. Gan LM, Chew CH (1983) J Disper Sci Technol 4:291
86. Palani Raj WR, Sasthav M, Cheung HM (1992) Langmuir 8:1931
87. Gan LM, Li TD, Chew CH, Teo WK (1995) Langmuir 11:3316
88. Gan LM, Li TD, Chew CH, Teo WK,Gan LH (1996) Langmuir 12:5863
89. Antonietti M, Hentze HP (1996) Colloid Polym Sci 274:696
90. Gan LM,Chieng TH, Chew CH, Ng SC, Pey KL (1996) Langmuir 12:319
91. (a) Gan LM, Li TD, Chew CH, Quek CH, Gan LH (1998) Langmuir 14:6068; (b) Guyot
A, Tauer K (1994) Adv Polym Sci 111:44
92. Gan LM, Chieng TH, Chew CH, Teo WK, Gan LH (1996) Langmuir 10:4022
93. Liu J, Gan LM, Chew CH, Teo WK, Gan LH (1997) Langmuir 13:6421
94. Gan LM, Chow PY, Chew CH, Ong CL, Wang J, Xu G (1999) Langmuir 15:3202
95. Xu G, Ong CL, Gan LM, Ong CK, Chan HSO (1999) J Phys Chem B 10: 7573
96. Gan LM, Xu W, Siow KS, Gao Z, Lee SY, Chow PY (1999) Langmuir 15: 4812
97. Gan LM, Chow PY, Han M, Liu ZL, Yeo E (2002) In: Proc 1st Int Conf on Materials
Processing for Properties and Performance, Singapore, 1–3 August 2002
98. Moy HY, Chow PY, Yu WL, Wong KMC, Yam VWW, Gan LM (2002) Chem Comm 9:982
99. Bonini M, Bardi U, Berti D, Neto C, Baglioni P (2002) J Phys Chem B 106:6178
100. Chow PY (2002) PhD Thesis, National University of Singapore, Singapore (to be pub-
lished)
101. Chow PY, Gan LM (2003) J Nanosci Nanotechno 4(1–2):197–202
102. Donescu D, Fusulan L, Petcu C,Vasilescu M, Deleann C. Udrea S (2002) Macromol Symp
179:315
103. Liu B, Li HP, Chew CH, Que WX, Lam YL, Kam CH, Gan LM, Xu GQ (2001) Mater Lett
51:461
104. Quek CH, Gan LM (work to be published)
105. Andreas RP, Averback RS, Brown WL, Brus LE, Goddard III WA, Kaldor A, Louie SG,
Moskovits M, Peercy PS, Riley SJ, Siegel RW, Spaepen F,Wang Y (1989) J Mater Res 4:704
106. Ichinose N, Ozaki Y, Kashu S (1988) Superfine particle technology. Springer, Berlin
Heidelberg New York
107. Haruta M, Delmon B (1986) J Chem Phys 83:859
108. Brinker CJ, Scherer GW (1990) In: Sol-gel science. Academic, New York
109. Lianos P, Thomas JK (1986) Chem Phys Lett 125:299
110. Lianos P, Thomas JK (1987) J Colloid Interf Sci 125:505
111. Barnickel P, Wokaun A, Sager W, Eicke HF (1992) J Colloid Interf Sci 148:80
112. Petit C, Jain TK, Billoudet F, Pileni MP (1994) Langmuir 10:4446
Microemulsion Polymerizations and Reactions 297

113. Pileni MP (1993) Adv Colloid Interf Sci 46:139

114. Pileni MP (2003) Nat Mater 2:145
115. Inouye K, Endo R, Otsuka Y, Miyashiro K, Kaneko K, Ishikawa T (1982) J Phys Chem
86:1465
116. Fendler JH (1987) Chem Rev 87:877
117. Matson DW, Fulton JL, Smith RD (1987) Mater Lett 6:31
118. Bandow S, Kimura K, Kon-on K, Kitahara A (1987) Jpn J Appl Phys 26:713
119. Kandori K, Shizuka N, Gobe M, Kon-on K, Kitahara A (1987) J Disper Sci Technol 8:477
120. Hou MJ, Shah DO (1988) Interfacial phenomena in biology and materials processing.
Elsevier, Amsterdam, p 443
121. Nagy JB (1989) Colloids Surface 35:201
122. Pillai V, Kumar P, Hou MJ, Ayyub P, Shah DO (1995) Adv Colloid Interf Sci 55:241
123. O’Sullivan EC, Patel RC, Ward AJI (1991) J Colloid Interf Sci 146:58
124. Friberg SE, Yang CC, Sjoblom J (1992) Langmuir 8:372
125. Friberg SE, Jones SM, Yang CC, Sjoblom J (1992) J Disper Sci Technol 13:65
126. Jones SM, Friberg SE (1992) J Disper Sci Technol 13:669
127. Pileni MP (1993) J Phys Chem 97:6961
128. O’Sullivan EC, Ward AJI, Budd T (1994) Langmuir 10:2985
129. Gan LM, Zhang K, Chew CH (1996) Colloid Surface A 110:199
130. Chabra V, Ayyub P, Chattopadhyay S, Maitra AN (1996) Mater Lett 26:21
131. Fang JY, Wang J, Ng SC, Chew CH, Gan LM (1997) Nanostruct Mater 8:499
132. Liu XY, Wang J, Gan LM, Ng SC, Ding J (1998) J Magn Magn Mater 184:344
133. Adair JH, Li T, Kido T, Havey K, Moon J, Mecholsky J, Morrone A, Talham DR, Ludwig
MN, Wang L (1998) Mater Sci Eng R23:139
134. Zhang K, Chew CH, Xu GQ, Wang J, LM Gan (1999) Langmuir 15:3056
135. Zhang K, Chew CH, Kawi S, Wang J, Gan LM (2000) Catal Lett 64:179
136. Liu SH, Qian XF, Yin J, Ma XD, Yuan JY, Zhu ZK (2003) J Phys Chem Solid 64:455
137. Wang DB, Yu DB, Mo MS, Liu XM, Qian YT (2003) Solid State Commun 125:475
138. Tata M, Banerjee S, John VT, Waguespack Y, McPhersond GL (1997) Colloid Surface A
127:39
139. Li HP, Liu B, Kam CH, Lam YL, Que WX, Gan LM, Chew CH, Xu GQ (2000) Opt Mater
14:321
140. Fu XA, Qutubuddin S (2001) Colloid Surface A 179:65
141. Liu B, Xu GQ, Gan LM, Chew CH, Li WS, Shen ZX (2001) J Appl Phys 89:1059
142. Liu B, Chew CH, Gan LM, Xu GQ, Li HP, Lam YL, Kam CH, Que WX (2001) J Mater Res
16:1644
143. Gan LM, Liu B, Chew CH, Xu SJ, Chua SJ, Loy GL, Xu GQ (1997) Langmuir 13:6427
144. Yee KL (1998) Masters Thesis, National University of Singapore, Singapore
145. Liu XY, Wang J, Gan LM, Ng SC, Ding J (1998) J Magn Magn Mater 184:344
146. Frieberg SE, Jones SM, Sjoblom J (1994) J Mater Synth Proces 2:1994
147. Jones SM, Frieberg SE (1995) J Non-Cryst Solids 181:39
148. Chang CL, Fogler HS (1996) AICHE J 42:3153
149. Chang CL, Fogler HS (1997) Langmuir 13:3295
150. Arriagada FJ, Osseo-Asare K (1992) Colloid Surface 69:105
151. Arriagada FJ, Osseo-Asare K (1995) J Colloid Interf Sci 170:8
152. Gan LM, Zhang K, Chew CH (1996) Colloid Surface A 110:199
153. Zhang K, Gan LM, Chew CH, Gan LH (1997) Mater Chem Phys 47:164
154. Schlag S, Eicke H-F, Mathys D, Guggenheim R (1994) Langmuir 10:3357
155. LM Gan, Zhang LH, Chan HSO, Chew CH, Loo BH (1996) J Mater Sci 31:1071
156. Kusakabe K, Yamaki T, Maeda H, Morooka S (1993) ACS Prepr 38:352
298 Microemulsion Polymerizations and Reactions

157. Kawai T, Fujino A, Kon-no K (1996) Colloid Surface 109:245

158. Wang J, Ee LS, Ng SC, Chew CH, Gan LM (1997) Mater Lett 30:119
159. Fang J, Wang J, Ng SC, Gan LM, Chew CH (1998) Ceram Int 24:507
160. Fang J, Wang J, Ng SC, Gan LM, Quek CH, Chew CH (1998) Mater Lett 36:179
161. Ee LS, Wang, Ng SC, Gan LM (1998) Mater Res Bull 33:1045
162. Lim GK, Wang J, Ng SC, Chew CH, Gan LM (1997) Biomaterials 18:1433
163. Lim GK, Wang J, Ng SC, Gan LM (1999) Langmuir 15:7472
164. Liu ZL, Lee JY, Han M, Chen WX, Gan LM (2002) J Mater Chem 12:2453
165. Gan LM, Zhang LH, Chan HSO, Chew CH (1995) Mater Chem Phys 40:94

Received: September 2003

Adv Polym Sci (2005) 175: 299–328
DOI 10.1007/b100118
© Springer-Verlag Berlin Heidelberg 2005

Dispersion Polymerization
Seigou Kawaguchi ( ) 1 · Koichi Ito2
1
Department of Polymer Science and Engineering, Faculty of Engineering,
Yamagata University, 4–3–16 Jonan, 992–8510 Yonezawa, Japan
[email protected]
2
Department of Materials Science, Toyohashi University of Technology,
1–1 Tempaku-cho, 441–8580 Toyohashi, Japan
[email protected]

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301

2 Microsphere Syntheses by Linear and Block Polymer Dispersants . . . . . . . 303

2.1 Functional Microspheres . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303
2.2 Living Dispersion Polymerization . . . . . . . . . . . . . . . . . . . . . . . . . 306
2.3 Microspheres from Non-Vinyl Monomers . . . . . . . . . . . . . . . . . . . . 307

3 Microsphere Syntheses by Reactive Dispersants, Macromonomers, Inimers,

and Transurfs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308

4 Particle Size Control in Dispersion (Co-)Polymerization . . . . . . . . . . . . 315

4.1 Theoretical Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315
4.2 Comparison of Experiment with Theory . . . . . . . . . . . . . . . . . . . . . 319

5 Chain Conformation of Grafted Polymer Chains at the Particle Surface . . . . 321

6 Conclusions and Future . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324

Abstract Dispersion polymerization is an attractive method for producing micron-size

monodisperse polymer particles in a single batch process. Great progress in this field has
been achieved over the past two decades. This article presents an overview of the recent
progress in the preparation of polymeric microspheres via dispersion polymerization in
organic media, focusing on the preparation of novel functional particles, the design of
microspheres using macromonomers, and on understanding mechanisms for the control of
particle size. Examples of functional microspheres obtained by dispersion polymerization
in the presence of linear polymers, block polymers, and macromonomers are tabulated, and
new developments are highlighted. Particle size control in dispersion polymerization in
the presence of macromonomers is discussed, and experimental results for poly(ethylene
oxide)-grafted particles are compared with theoretical expectations for ideal core-shell
particles.

Keywords Functional microsphere · Macromonomer · Block copolymer · Graft copolymer ·

Particle size control
300 S. Kawaguchi · K. Ito

Symbols and Abbreviations

AIBN 2,2¢-azobisisobutyronitrile
AAm acrylamide
ATR attenuated total reflection
BMA n-butyl methacrylate
Boc-AMST boc-p-aminostyrene
CMS chloromethylstyrene
Cs chain transfer constant
D mean separation between PEO anchor points
DSM dynamic swelling method
DVB divinylbenzene
EDMA ethylene dimethacrylate
EG ethylene glycol
ESCA electron spectroscopy
f initiator efficiency
FTIR Fourier transform infrared spectroscopy
GMA glycidyl methacrylate
GTP group transfer polymerization
HEMA 2-hydroxyethyl methacrylate
HLB hydrophobic-lipophilic-balance
HPC hydroxypropylcellulose
HPMA 2-hydroxypropyl methacrylate
[I]o initiator concentration
kd decomposition rate constant
kp propagation rate constant
kt termination rate constant
k2 diffusion-controlled rate constant for coalescence between similar-sized
particles
MD molecular weight of macromonomer
MMA methyl methacrylate
MW molecular weight
Mw weight average molecular weight
N number of particles per liter
n number of chains grafted onto surface
NA Avogadro’s number
NAD nonaqueous dispersion
NIPAM N-isopropyl methacrylate
NMR nuclear magnetic resonance
NVC N-vinyl carbazole
PAA poly(acrylic acid)
PANI polyaniline
PB poly(1,3-butadiene)
PBMA poly(n-butyl methacrylate)
PCL poly(e-caprolactone)
PDMAEMA poly(2-(dimethylamino)ethyl methacrylate
PDMS poly(dimethylsiloxane)
PDVB polydivinylbenezene
PE polyethylene
PEG poly(ethylene glycol)
PEO poly(ethylene oxide)
Dispersion Polymerization 301

PGMA poly(glycidyl methacrylate)

PHEM Apoly(2-hydroxyethyl methacrylate)
PHSA poly(12-hydroxystearic acid)
PIB polyisobutylene
PLMA poly(lauryl methacrylate)
PMA poly(methacrylic acid)
PMMA poly(methyl methacrylate)
PNIPAM poly(N-isopropylacrylamide)
POXZ polyoxazolines
PP polypropylene
PS polystyrene
P(PP-alt-E) poly(propylene-alt-ethylene)
PTBA poly(t-butyl acrylate)
PTBMA poly(t-butyl methacrylate)
PVA poly(vinyl alcohol)
PVAcA poly(N-vinylacetamide)
PVME poly(vinyl methyl ether)
PVP poly(vinylpyrrolidone)
P4VP poly(4-vinylpyridine)
ri reactivity ratio of i species
R radius of particle
Ç density
S surface area occupied by a dispersant chain
ST styrene
<S2> mean square radius of gyration
Scrit surface area occupied by a macromonomer chain at critical point
TDI toluene diisocyanate
TEMPO 2,2,6,6-tetramethylpiperidinyloxyl
4VBAC 4-vinylbenzyltrimethylammonium chloride
4VP 4-vinylpyridine
Vpy vinylpyrrolidone
q fractional conversion of monomer
qcrit fractional conversion of monomer at critical point
qD fractional conversion of macromonomer
qDcrit fractional conversion of macromonomer at critical point
WD weight of polymerized macromonomer
WDo initial weight of macromonomer
W weight of polymerized monomer
Wo initial weight of monomer

1
Introduction
Micron-size monodisperse polymeric microspheres are used in a wide variety
of applications, such as toners, instrument calibration standards, column pack-
ing materials for chromatography, spacers for liquid crystal displays, and bio-
medical and biochemical analysis [1–3]. Because of the commercial and scien-
tific interest in these particles, research into their preparation has been active
302 S. Kawaguchi · K. Ito

for the past two decades. Micron-size monodisperse particles were usually dif-
ficult to obtain because this size is in-between the diameter range of particles
produced by conventional emulsion polymerization (0.06–0.7 mm) in a batch
process and suspension polymerization (50–1000 mm). Vanderhoff et al. [4, 5]
used the successive seeding method to obtain micron-size monodisperse
polymer particles. The particles were also prepared by Ugelstad et al. [6, 7] by
means of the two-stage swelling method. Omi et al. [8] and Kamiyama [9] used
modified suspension polymerizations, and Okubo et al. [10] developed the
dynamic monomer swelling method (DSM).
Dispersion polymerization is an attractive and promising alternative to
other polymerization methods that affords micron-size monodisperse particles
in a single batch process. Dispersion polymerization may be defined as a type
of precipitation polymerization in which one carries out the polymerization of
a monomer in the presence of a suitable polymeric stabilizer soluble in the
reaction medium. The solvent selected as the reaction medium is a good sol-
vent for both the monomer and the steric stabilizer polymers, but a non-solvent
for the polymer being formed. Dispersion polymerization, therefore, involves
a homogeneous solution of monomer(s) with initiator and dispersant, in which
sterically stabilized polymer particles are formed by the precipitation of the
resulting polymers.As a continuous medium, the properties of the solvent also
change with increasing monomer conversion. Under favorable circumstances,
the polymerization can yield, in a batch step, polymer particles of 0.1–15 mm
in diameter, often of excellent monodispersity. This dispersant polymer can be
formed as a reactive, polymerizable macromonomer. It can be a block copoly-
mer in which one block has an affinity for the surface of the precipitated poly-
mer, or it can be a soluble polymer (a “stabilizer precursor”) to which grafting
is thought to occur during the polymerization reaction. In all instances, this sol-
uble dispersant polymer – a hairy layer – plays a crucial role in the dispersion
polymerization process. By adsorbing or becoming incorporated onto the
surface of the newly-formed precipitated polymers, it acts as a steric stabilizer,
directing the particle size and colloidal stability of the system. This feature of
dispersion polymerization is widely appreciated and well understood (Fig. 1).
Dispersion polymerization in organic hydrocarbon media was first devel-
oped by Osmond and coworkers at ICI [11]. They polymerized acrylic and
vinylic monomers in hydrocarbons with oil soluble polymer stabilizers to pro-
duce nonaqueous dispersions (NAD) of polymer particles. Later, Almog et al.
[12] extended the concept to dispersion polymerization in polar solvents
as a method of forming monodisperse polymeric microspheres. Ober et al.
[13–16], Tseng et al. [17], Okubo et al. [18, 19], and Paine et al. [20–24], among
other authors, studied this technique in order to control particle size and
achieve a narrow particle size distribution. A great deal of research has been
devoted to dispersion polymerization during past two decades, as reviewed by
Croucher and Winnik [25], Guyot and Tauer [26], Cawse [27], Pichot et al. [28],
Asua and Schoonbrood [29], and Ito et al. [30–32]. The present article is
intended to discuss state-of-the-art design of microspheres obtained by dis-
Dispersion Polymerization 303

Fig. 1 Schematic description of dispersion polymerization

persion polymerization, with particular attention paid to the preparation of

novel functional particles, the design of microspheres using the macromo-
nomer technique, and to mechanistic aspects for particle size control.

2
Microsphere Syntheses by Linear and Block Polymer Dispersants

2.1
Functional Microspheres

Most research into the study of dispersion polymerization involves common

vinyl monomers such as styrene, (meth)acrylates, and their copolymers with
stabilizers like polyvinylpyrrolidone (PVP) [33–40], poly(acrylic acid) (PAA)
[18, 41], poly(methacrylic acid) [42], or hydroxypropylcellulose (HPC) [43, 44]
in polar media (usually alcohols). However, dispersion polymerization is also
used widely to prepare functional microspheres in different media [45, 46].
Some recent examples of these preparations include the (co-)polymerization
of 2-hydroxyethyl methacrylate (HEMA) [47, 48], 4-vinylpyridine (4VP) [49],
glycidyl methacrylate (GMA) [50–53], acrylamide (AAm) [54, 55], chloro-
methylstyrene (CMS) [56, 57], vinylpyrrolidone (VPy) [58], Boc-p-amino-
styrene (Boc-AMST) [59], and N-vinyl carbazole (NVC) [60] (Table 1). Disper-
sion polymerization is usually carried out in organic liquids such as alcohols
and cyclohexane, or mixed solvent-nonsolvents such as 2-butanol-toluene,
alcohol-toluene, DMF-toluene, DMF-methanol, and ethanol-DMSO. In addition
to conventional PVP, PAA, and PHC as dispersant, poly(vinyl methyl ether)
(PVME) [54], partially hydrolyzed poly(vinyl alcohol) (hydrolysis=35%)
[61], and poly(2-(dimethylamino)ethyl methacrylate-b-butyl methacrylate)
304 S. Kawaguchi · K. Ito

Table 1 Examples of functional microspheres obtained by dispersion polymerization

Stabilizer Monomer(s) Medium Reference

Functional particles
PS, PMMA derivatives HEMA 2-Butanol/toluene [47]
Cellulose acetate HEMA Alcohol/toluene [48]
butyrate
PS-b-PB 4VP DMF/toluene [49]
PVP GMA MeOH/water or DMF [50]
PVP GMA/DVB Ethanol [51]
PVP ST/GMA Ethanol/water [52]
Cellulose acetate GMA DMF/methanol [53]
PVME AAm t-BuOH/water [54]
None AAm t-BuOH/water [55]
PVP CMS Ethanol/DMSO [56]
PAA CMS Methoxyethanol/MeOH [57]
PS-b-P(PP-alt-E) VPy/EDMA Cyclohexane [58]
PVP ST/Boc-AMST i-Propanol [59]
PVP ST/NVC Ethanol [60]
PVA-co-PVAc ST Methanol [61]
PDMAEMA-b-PBMA ST Alcohols [62]
Hybrid particles
Poly(amic acid) ST/4VBAC Ethanol/water [63]
PVP ST/polyimide prepolymer i-Propanol [64]
PVP GMA/iron oxide Alcohol/water [65]
Cellulose acetate HEMA/EDMA/iron oxide Alcohol/toluene [66]
butyrate
PVP HEMA/GMA/iron oxide Alcohol/toluene [67]
PVP ST/SiO2 Ethanol/water [68]
None 4VP/HPMA/SiO2 Water [69]
Crosslinked particles
PVP DVB Acetonitrile or Ethanol [70]
None DVB Acetonitrile [71]
None DVB/CMS Acetonitrile [72]
Chitosan NIPAM Acetic Acid [73]
PS-b-P(PP-alt-E) Oxazoline methacrylate Heptane [74]
PVP-Aerosol-OT ST/urethane acrylate Ethanol [75]
PVP ST/2,2¢-oxy-bisethanol Ethanol/heptane [76]
diacrylate
PVP MMA/EDMA Ethanol/water [77]
Dynamic swelling method (DSM)
None ST/DVB Ethanol/water [10, 78–83]
Dispersion Polymerization 305

(PDMAEMA-b-PBMA) [62] in alcohols and polystyrene-b-polybutadiene (PS-

b-PB) [49] and polystyrene-b-poly(propyrene-alt- ethylene) (PS-b-(PP-alt-E))
[58], linear PS [47], and PMMA [47] in hydrocarbon are used.
Dispersion polymerization in supercritical carbon dioxide (scCO2), pioneer-
ed by DeSimone and coworkers, has recently attracted considerable attention
as an environmentally friendly alternative to the use of organic solvents.A wide
range of monomers have been polymerized to produce the corresponding
microspheres in the presence of CO2-philic polymers and monomers, as re-
viewed in another chapter.
Core-shell polystyrene-polyimide high performance particles have been
successfully prepared by the dispersion copolymerization of styrene with vinyl-
benzyltrimethyl ammonium chloride (VBAC) in an ethanol-water medium
using an aromatic poly(amic acid) as stabilizer, followed by imidization with
acetic anhydride [63]. Micron-sized monodisperse polystyrene spheres impreg-
nated with polyimide prepolymer have also been prepared by the conventional
dispersion polymerization of styrene in a mixed solvent of isopropanol/2-
methoxyethanol in the presence of L-ascorbic acid as an antioxidant [64].
A recent interesting example is the preparation of organic/inorganic hybrid
particles by dispersion polymerization. Horák et al. [65–67] have prepared
cross-linked poly(2-hydroxyethyl methacrylate) (PHEMA)- and poly(glycidyl
methacrylate) (PGMA)-based magnetic microspheres by incorporating iron
oxide into cellulose acetate butyrate stabilizer or PVP. Polymer encapsulation
of small silica particles has also been achieved using the dispersion polymer-
ization of styrene in ethanol-water medium with PVP [68]. Silica-polymer
hybrid particles have also been prepared very recently by the precipitation
copolymerization of 4-vinylpyridine with 2-hydroxypropyl methacrylate
(HPMA) in water [69]. These organic/inorganic hybrid particles would be
expected to lead to a new generation of nanostructured materials with diverse
applications such as catalysts, electronic or phonic devices, and sensors.
Highly cross-linked monodisperse polydivinylbenzene (PDVB) microspheres
have been prepared in acetonitrile with or without PVP stabilizer [70–72].
Interestingly, PDVB microspheres are very stable without any stabilizer in aceto-
nitrile, possibly due to their highly crosslinking rigid surface. Other examples
of the syntheses of crosslinked microspheres [73–77] are listed in Table 1.
The dynamic swelling method (DSM) [10] has also been described for the
preparation of crosslinked microspheres with free vinyl groups [78]. Therefore,
polystyrene seed particles (1.9 mm) prepared by dispersion polymerization
are dispersed in ethanol-water (7/3, w/w) containing divinylbenzene (DVB),
benzoyl peroxide, and poly(vinyl alcohol) (PVA). The slow drop-wise addition
of water to the mixture causes the DVB phase to separate, and it is continuous-
ly imbibed by seed particles to produce relatively large swollen particles
(4.3 mm), which are then polymerized to afford the respective PS-PDVB com-
posite particles with free vinyl groups. DSM has recently been developed in
order to prepare hollow microspheres and various oddly-shaped polymer par-
ticles, including a rugby ball, red blood cells, or snowman structures [79–83].
306 S. Kawaguchi · K. Ito

2.2
Living Dispersion Polymerization

Living dispersion (co-) polymerization is interesting due to its straightforward

control of molecular weight with a narrow distribution, and also the function-
alization of living end groups that occurs. Living anionic dispersion poly-
merization of styrene or a-methylstyrene has been achieved in hydrocarbon
solvents with block copolymer stabilizers such as PS-b-poly(4-trimethylsilyl-
styrene) [84], PS-b-PB [85–87], PS-b-poly(4-tert-butylstyrene) [88, 89], poly-
(vinyl ethyl ether) [90], and PS-b-polyisobutylene (PS-b-PIB) [91], and PS-b-
P(PP-alt-E) [92]. Jenkins and coworkers [93] polymerized methyl methacrylate
in n-heptane via group transfer polymerization (GTP). Highly crosslinked mi-
crospheres have also been prepared by the living anionic copolymerization of
t-butylstyrene and divinylbenzene in heptane [94].
Living radical dispersion polymerization is a promising way to expand the
design and scope of functional polymer colloids to a wider range of other
monomers. The 2,2,6,6-tetramethyl-1-piperidinyloxy (TEMPO)-mediated liv-
ing radical dispersion polymerization of styrene has been carried out in pre-
sence of PS-b-P(PP-alt-E) in decane at 135 °C [95] or PVP in alcohol-water at
130 °C [96] in order to produce microspheres with a very broad size distribu-
tion, consisting of relatively low molecular weight polystyrene (Mw=104) with
Mw/Mn=1.1.

Table 2 Examples of microspheres obtained by living dispersion polymerization

Stabilizer Monomer Medium Reference

Living anionic dispersion polymerization

PS-poly(4-trimethylsilylstyrene) ST Hexane [84]
PS-b-PB ST Hexane [85, 86]
None ST/butadiene Pentane [87]
PS-b-Poly(4-tert-butylstyrene) ST Hexane [88, 89]
Poly(vinyl ethyl ether) a-methylstyrene Heptane [90]
PS-b-PIB ST Hexane [91]
PS-b-P(PP-alt-E) ST Hydrocarbon [92]
PS-b-P(PP-alt-E) MMA Heptane [93]
None t-butylstyrene/DVB Heptane [94]
Living radical dispersion polymerization
PS-b-P(PP-alt-E) ST Decane [95]
PVP ST Alcohol/water [96]
Ring opening dispersion polymerization
Poly(dodecyl methacrylate)-g- e-caprolactone Heptane/dioxane [97,98]
poly(e-caprolactone) and lactide
Dispersion Polymerization 307

Dispersion polymerization has also been applied to the ring opening poly-
merization of e-caprolactone and lactide in heptane-dioxane (4/1 v/v) with
poly(dodecyl methacrylate)-g-poly(e-caprolactone) as stabilizer [97]. Diethyl-
aluminium ethoxide and tin(II) 2-ethylhexanoate were used as initiators in
these two systems, respectively, to obtain functional microspheres with a
narrow particle size distribution and a narrow molecular weight distribution
[98]. Table 2 provides an overview of microspheres obtained by living disper-
sion polymerization.

2.3
Microspheres from Non-Vinyl Monomers

Microspheres have been prepared by the dispersion polymerization of mo-

nomers other than vinyl monomers, such as styrene and (meth-)acrylates.
Polyaniline (PANI) is one of the most frequently studied electrically conduct-
ing polymers. Since the paper by Armes and Aldissi [99] in 1989, there have
been numerous reports on the preparation of PANI dispersions by oxidative

Table 3 Examples of microspheres of monomers other than vinyl monomers obtained by

dispersion polymerization

Stabilizer Monomer Medium Reference

Oxidative polymerization
PVP Aniline Alcohol/water [100, 101]
PVME Aniline Alcohol/water [102, 103]
PVA Aniline Water/alcohol [104–106]
Poly(styrenesulfonic acid) Aniline Water [107]
Methycellulose Aniline Water/alcohol [108]
HPC Aniline Water [109]
PVA 3,5-Xylidine Water [110, 111]
Ethylhydroxycellulose Pyrrole Ethanol/water [112, 113]
PVME Pyrrole Ethanol/water [114]
Enzymatic polymerization
PVME, PVA, PEG Phenol Dioxane/phosphate [115, 116]
buffer
PVME p-Phenylphenol Dioxane/phosphate [115, 116]
buffer
Polyurethane
Poly(lauryl methacrylate) diol TDI/EG Paraffin oil [117]
Polyester
Poly(2-ethylhexyl acrylate-co- 4-Acetoxybenzoic – [118]
styrene-co-acrylic acid) acid + 2,6-aceto-
xynaphthoic acid
308 S. Kawaguchi · K. Ito

polymerization. Colloidally-stable submicron-sized PANI particles are pro-

duced in the aqueous-alcohol media by the dispersion polymerization of ani-
line in the presence of a suitable steric stabilizer such as PVP [100, 101], PVME
[102, 103], PVA [104–106], poly(styrenesulfonic acid) [107], methycellulose
[108], and HPC [109]. The oxidative dispersion polymerization of 3,5-xylidine
[110, 111] produced needle-shaped particles in water. Polypyrrole particles
were also prepared in aqueous ethanol using FeCl3 and ammonium persulfate
as the oxidant [112–114].
Uyama and Kobayashi et al. [115, 116] were first to prepare nearly mono-
disperse sub-micron polyphenol particles by the enzyme-catalyzed dispersion
polymerization of phenol and p-phenylphenol in a mixture of 1,4-dioxane and
phosphate buffer using a water-soluble polymer as stabilizer (such as PVME,
PVA, and PEG).
Polyurethane [117] and polyester [118] particles have also been prepared by
the dispersion polyaddition of ethylene glycol (EG) and toluene diisocyanate
(TDI) in paraffin, and the polycondensation of acid and ester at a high poly-
merization temperature, respectively. Table 3 provides an overview of micro-
spheres of monomers other than vinyl monomers obtained by dispersion poly-
merization.

3
Microsphere Syntheses by Reactive Dispersants, Macromonomers,
Inimers, and Transurfs

Macromonomers are polymers or oligomers with polymerizable end groups,

widely investigated for the preparation of functional polymers and polymer
microspheres by dispersion polymerization. For microspheres, the macro-
monomers should be designed to copolymerize with the main monomers in
such a way as to produce graft chains that serve as efficient stabilizers; in other
words, their main chain should be firmly bound to the particle surface and the
graft chains should extend into the polymerization medium.
Examples of dispersion polymerizations using macromonomers are sum-
marized in Table 4. Non-aqueous dispersion (NAD) polymerization of polar
Dispersion Polymerization 309

Table 4 Examples of dispersion copolymerization with reactive dispersants

Reactive dispersant Monomer Medium Reference

PHSA 1 MMA Hydrocarbon [11]

PLMA 2 MMA Hydrocarbon [11]
PE 3 MMA Dodecane, PE [119]
PDMS 4 ST Silicone oil [120]
POXZ 5 MMA MeOH/H2O [121]
POXZ 5 ST EtOH/H2O [122]
POXZ 5 CH2=CHNHCHO MeOH [123]
POXZ 6 MMA MeOH/H2O [124]
PEO 7 (m=1) MMA, ST EtOH/H2O [126, 127]
PEO 8 (m=3, 5, 7) ST Ethanol/H2O [128]
PEO 9a ST EtOH/H2O [129]
PEO 9a, 11 ST EtOH/H2O [130]
PEO 7 (m=1, 4, 7) ST, MMA, BMA MeOH/H2O [131–133]
PEO 9b (m=11) ST EtOH/H2O [134]
PEO 10 (m=2.6, 4.1, 6.8) ST MeOH/H2O [135]
PEO 9a, 9b (m=6, 10) ST, MMA MeOH/H2O [125, 137]
PEO 12 ST EtOH/H2O [136]
PEO 13 ST EtOH/H2O [138]
PVP 14 ST, MMA EtOH [139]
PVAcA 15 ST EtOH [140]
PVA 16 MMA EtOH/Water [141]
P4VP 17 ST EtOH [142]
PNIPAM 18 ST EtOH [143]
PTBMA 19 ST EtOH [144]
PAA 20 MMA EtOH/H2O [145]
PDMS 4 MMA, ST CO2 [146]
PCL 21 L,L-Lactide Heptane/dioxane [97, 98]
PMA 22 MMA EtOH/H2O [147]
PMA 23 ST MeOH/H2O [148]
PDMAEMA 24 ST Alcohols [150]
PEO 25 ST MeOH/H2O [151]
Inimer 26 ST, MMA Ethanol/H2O [155]
Transurf 27 ST Ethanol/H2O [156]

monomers was first carried out in aliphatic hydrocarbon media with the hy-
drophobic macromonomers 1 and 2 [11]. These were copolymerized with
MMA or other polar monomers to produce comb-graft copolymers, which
have limited solubility in pure aliphatic hydrocarbons but adequate solubility
in hydrocarbon-monomer mixtures. It is particularly effective in stabilizing
PMMA NAD particles. Polyethylene (PE) macromonomers 3 have been used
310 S. Kawaguchi · K. Ito

for the dispersion copolymerization of MMA in dodecane and in PE melts to

produce stable PMMA dispersions at a high temperature [119]. In the latter
case, nanocomposite materials in which submicron-sized fine PMMA particles
are uniformly dispersed in the PE bulk can be prepared during the copoly-
merization. Monodisperse PS particles (0.70 mm), prepared by the dispersion
copolymerization of styrene with the poly(dimethylsiloxane) PDMS macro-
monomer 4 in PDMS-diol (MW=7,500), was allowed (together with trimethylol-
propane) to react with hexamethylene diisocyanate using dibutyltin dilaurate
as a catalyst. The product of the polyaddition was a tough elastomeric com-
posite with polystyrene particles finely dispersed and strongly anchored in a
PDMS-polyurethane matrix [120].

This technique has been extended to polar media, especially alcohols and
their mixtures with water as a continuous phase.Kobayashi and Uyama et al.
[121– 124] reported that poly(2-oxazoline) macromonomers such as 5 and 6 are
very effective for dispersion copolymerization with styrene, MMA, and N-vinyl-
formamide in methanol, ethanol, and mixtures of these alcohols with water.
They reported that the particle size decreased with increasing initial macro-
monomer concentration and that graft-copolymerized poly(2-oxazoline) chains
are concentrated on the particle surface to act as steric stabilizers.
Dispersion copolymerizations that use the poly(ethylene oxide) (PEO)
macromonomers 7–13 in alcoholic media have been intensively studied by
Dispersion Polymerization 311

many researchers [125–138]. They produce nearly monodisperse polymeric

microspheres of submicron to micron sizes, covered with PEO chains on their
surface. Several factors that affect the particles’ size and the polymerization
kinetics have been studied. Theoretical models for particle nucleation in these
systems have also been developed and compared with the experimental ob-
servations, as will be discussed later.
312 S. Kawaguchi · K. Ito

Several other hydrophilic macromonomers including 14–20 have been ap-

plied to the dispersion polymerization [139–145]. These macromonomers were
synthesized by radical polymerization in the presence of appropriate chain
transfer agents, followed by transformation of the end group, as previously
summarized [30]. Akashi et al. [143] used poly(N-isopropylacrylyamide)
(PNIPAM) macromonomer 18 in ethanol and prepared thermosensitive micro-
spheres 0.4–1.2 mm in diameter consisting of a PS core and PNIPAM branches
on their surface. The particles are particularly useful for many biomedical
applications. Indeed, the particles have been reported to flocculate and change
in light transmittance with increasing temperature.
Dispersion Polymerization 313

DeSimone and his co-workers have intensively studied polymerization

reactions in an environmentally friendly solvent, CO2. In the presence of the
CO2-philic silicone-based macromonomer 4, relatively monodisperse micron-
sized polymer particles were obtained by the polymerization of MMA and
styrene in supercritical CO2, as shown in another chapter [146].
Sosnowski et al. [97, 98] have reported that uniform biodegradable poly-
meric particles with diameters of less than 5 mm can be prepared by ring-open-
ing dispersion polymerization of L,L-lactide in heptane-dioxane mixed solvent
in the presence of poly(dodecyl acrylate)-g-poly(e-caprolactone), which were
synthesized by the copolymerization of dodecyl acrylate with the poly(e-capro-
lactone) macromonomers 21. Note that the polymer particles consist of well-
defined poly(L,L-lactide) polymers with Mn≈1¥104 and Mw/Mn≈1.06.

The polyelectrolyte macromonomers 17, 20, 22, and 23 [142, 143, 147, 148]
were prepared and applied to dispersion copolymerizations to produce poly-
meric particles covered with polyelectrolyte chains. Evidently, the dependence
of the conformational properties of polyelectrolyte brush chains attached to
the latex surface on the pH, the degree of neutralization, and the salt concen-
tration have been the subject of growing experimental and theoretical effort.
314 S. Kawaguchi · K. Ito

Some novel water soluble macromonomers, 24, have been synthesized by the
oxyanionic polymerization [149] of 2-(dimethylamino)ethyl, 2-(diisopropyl-
amino)ethyl, and 2-(N-morpholino)ethyl methacrylate, and conducted to dis-
persion copolymerization of styrene in alcohol media [150]. Sufobetaine-based
macromonomer was prepared by the polymer reaction of 24 (R=CH3) with
propane sultone, and was found to be useful in the dispersion polymerization
of styrene even at high electrolyte levels (up to 1 M NaCl). Ito et al. [151] syn-
thesized new PEO macromonomers with a cationic charge at the w-end, 25, and
examined the influence of the charge on the particles’ size in dispersion copoly-
merization with styrene in alcohol media.

Poly(HEMA-co-MMA-g-PMMA) graft copolymer was also prepared with a

commercially available poly(methyl methacrylate) (PMMA) macromonomer,
HEMA, and MMA, and used as an efficient dispersant for the dispersion poly-
merization of styrene in ethanol [152].
Recently, Akashi and coworkers [153, 154] synthesized novel spherical
particles on which nano-projections are uniformly distributed over the
whole surface like “confetti” by the one-step dispersion terpolymerization
of acrylonitrile, styrene, and the PEO macromonomer 9a in ethanol/water
media. The control of nanoparticle morphology by a one-step synthetic pro-
cedure is important to self-organization at the polymer chain level, which is
a basis for the formation of biological nanoconstructs such as viruses and
organelles.
Dispersion polymerization in the presence of reactive surfactants includ-
ing surfmers, inisurfs and transurfs is also a versatile method for producing
functional microspheres [26]. For example, the macromonomeric azoinitiator
26 is an effective inisurf in the preparation of PS and PMMA particles [155].
Dispersion Polymerization 315

Similarly, the thiol-ended transurf 27 shows higher stabilizing efficiency than

PVP in the dispersion polymerization of styrene and MMA in water-ethanol
[156].

In all instances of the dispersion polymerization, amphiphilic graft copoly-

mers produced in a selective solvent for the branches play a crucial role.
Schematically, a microsphere obtained by copolymerization in this way with a
small amount of macromonomer has a core-shell structure as given in Fig. 2,
with the core occupied by the insoluble substrate polymer chains and the shell
by the soluble, graft-copolymerized macromonomer chains. The backbone
chains of the graft copolymers, which must be insoluble in the medium, serve
as the anchors into the core. The following section presents general criteria
for the size control of polymeric microspheres by dispersion copolymerization
using macromonomers.

Fig. 2 Schematic description of dispersion copolymerization of styrene with macro-

monomer

4
Particle Size Control in Dispersion (Co-)Polymerization

4.1
Theoretical Model

According to the aggregative and coagulative nucleation theories [157] which

all originally derive from the homogeneous nucleation theory [158], the most
important point when determining particle number is the instant at which
sterically-stabilized particles form [159].After this point, coagulation between
similar-sized particles no longer occurs, and the number of particles present
in the reaction medium remains constant. Sufficient particle stabilization may
be achieved with physically adsorbed stabilizers in the case of linear soluble
polymers, copolymers, and block copolymers. For covalent bonding, polymer-
ization is performed in the presence of macromonomer stabilizers, which
copolymerize with monomer(s) to graft stabilizer on the particles’ surface [26,
31]. The graft copolymer may be also produced by the chain transfer reaction
of a propagating radical with a soluble polymer chain.
316 S. Kawaguchi · K. Ito

Fig. 3 Schematic model for the particle nucleation and growth of sterically-stabilized
particles in dispersion polymerization

As shown in Fig. 3, the dispersion polymerization is considered to proceed

as follows:
1. Before polymerization, the reaction mixture dissolves completely into the
continuous phase.
2. When the reaction mixture is heated, free radicals are formed by initiator de-
composition and grow in the continuous phase to produce linear oligomer,
polymers, and/or graft copolymers. The solubility of these polymers is a
function of their molecular weight (MW) and the composition of the graft
copolymers. Polymers with a MW larger than a certain critical value precipi-
tate and begin to coagulate to form unstable particles.
3. These particles coagulate on contact, and the coagulation among them
continues until sterically-stabilized particles form.
4. This point is referred to as the critical point, and it occurs when all of the
particles contain sufficient stabilizer polymer chains on the surface to pro-
vide colloidal stability.
5. After this point, no new nuclei or particles are formed and the particles may
grow both by the diffusive capture of oligomers and the coagulation of very
small unstable particles (nuclei, precursors) produced in the continuous
phase and by the polymerization of the monomer included within the
particles until all of the monomer is consumed. The total number of such
sterically-stabilized particles remains constant so that their size is only a
function of the amount of polymer produced.
Dispersion Polymerization 317

In 1990, Paine [24] first developed a multibin kinetic model for the aggregation
of precipitated radicals or unstabilized particles in dispersion polymerization,
based on diffusion-controlled particle aggregation, which was described by
Smoluchowski [160] and Frenklach [161]. The stabilized particles whose sur-
faces are completely covered with grafted polyvinylpyrrolidone (PVP) chains
do not aggregate. This was the first model that quantitatively simulated the role
of the stabilizer molecules at the particle formation stage.Assuming ideal core-
shell structures in which the main monomer forms the core, and dispersant the
shell, one can readily establish a basic relationship between microsphere size
and weights of main core polymer and stabilizer polymer attached on the
particle surface, as given by the following equations:
4
W = Woq = 3 p R3 ÇN (1)
3
and
WDoqD NAS
nS = 08 = 4p R2 (2)
NMD

Here, W (g/l) is the weight of monomer polymerized, Wo is the weight of

monomer feed, q is the fractional conversion of monomer, R is the particle core
radius (cm), Ç is the core polymer density, N is the particle number per liter, n
is the number of dispersant molecules on one particle, S is the area of particle
surface covered by one dispersant molecule, WDo is the weight (in g/l) of dis-
persants in feed, MD is the molecular weight of dispersant, qD is the fraction
of adsorbed or grafted dispersant on particle surface, and NA is Avogadro’s
number. Combining Eqs. 1 and 2 produces a universal relationship between the
particle radius and the extent of polymerization for the microspheres obtained
by an ideally grafted steric stabilizer:
3Woq
R = 009 (3)


S
ÇNAWDoqD 6
MD
In spite of different mechanisms of particle formation and growth, Eq. 3
predicts larger particles with increasing amounts of main polymer, and with
lower surface coverage by individual dispersant molecules (S/MD), but smaller
particles with increasing amounts of dispersant. Also, S/MD increases with the
solvency of the dispersion medium, and hence the particle radius should be
smaller in media that are good solvents for the dispersant.
One problem is the estimation of qD and S with respect to the monomer
conversion (q) and the attachment of the dispersant to particles. For adsorbed
dispersants, qD is determined by the partition of dispersant between the par-
ticle surface (4pR2N) and the continuous medium. For chemical grafting via
chain transfer, Paine [24] derived the equation
318 S. Kawaguchi · K. Ito

2 1 1 1





1 3Wo 32 MD 22 0.386k2 62 kt 12
22
R= q 32 (4)
ÇNA
7 CSWDoScrit
06 4pkt
02 2kd f [I]o
04

where Cs is the chain transfer constant to dispersant polymer chains, Scrit is S

at critical point, k2 is the diffusion-controlled rate constant for coalescence
between similar-sized particles (M–1s–1), kp is the propagation rate constant
(M–1s–1), kt is the termination rate constant (M–1s–1), [I]o is the initiator con-
centration (M), and f is the initiator efficiency and kd is the initiator decompo-
sition rate constant (s–1).
Equation 4 was found to explain particle size data fairly well, with reason-
able kinetic and coverage parameter values (k’s and Scrit), in the dispersion
polymerization of styrene in ethanol with PVP dispersant [24]. Many other
dispersion polymerization systems with homopolymer dispersants appear to
be explained by Eq. 4, except for the frequently observed direct particle size
dependence on initiator concentration [27].
For dispersion polymerization with macromonomer stabilizers, we use Eq. 5
and Eq. 6 [31, 32, 131–133].
2 1 1 1





1 3Wo 32 MDr1 22 0.386k2 62 kt 12
22
R= q 32 (5)
ÇNA
7 WDoScrit
03 4pkp
02 2kd f [I]o
04

1 1 1 1






– 21 3Wo 32 MDScrit 22 4pkp 62 2kd f [I]o 12
22 q
S= q 3 (6)
Ç
7NA rW
02
1 Do 0.386k2
02 04
kt qD
4

Here r1 is the reactivity ratio (=qcrit/qDcrit) of the monomer (M1) in a copoly-

merization with the macromonomer (M2) at critical stabilization.
In Eqs. 4 and 5, one sees that the radius of the latex particle follows simple
scaling relationships with the key parameters in the system: q1/3, [monomer]o2/3,
[dispersant]o–1/2, [initiator]o–1/12, where []o means initial concentration. These
equations predict that the particle size and stabilization are determined by the
magnitude of r1. In addition, looking at Eq. 6, it is apparent that the surface area
occupied by a stabilizer chain follows q–1/3, and in the case of azeotropic copoly-
merization, q=qD. This means that the chain conformation for the grafts on the
latex particle will change with grafting density. The S value is closely related
to the conformation of a single polymer chain as a stabilizer grafted onto the
surface of a latex particle.According to de Gennes’“mushroom” model [162]for
a polymer grafted to a noninteracting surface, the polymer chain occupies a vol-
ume determined by its mean-squared radius of gyration <S2>.When the surface
becomes crowded with chains, additional energy is needed to deform the poly-
mer mushrooms into brushes. When the particle surfaces are completely cov-
ered with random coils of the polymer, they are also sterically-stabilized against
coagulation with other particles. One, therefore, defines Scrit as the maximum
surface area occupied by a single polymer chain in the continuous phase.
Dispersion Polymerization 319

4.2
Comparison of Experiment with Theory

For dispersion copolymerization with PEO macromonomers, the power law

exponents in Eq. 5 have been experimentally determined and compared, as
summarized in Table 5. Initial monomer concentration has a major influence
on the final particle radius. The experimental power law exponents (0.82–1.02)
are usually significantly larger than those in Eqs. 4 and 5, except for 0.63 for
styrene as a monomer with 7 (m=4, n=45). This is likely to be due to a solvency
effect of the monomer. The values of the exponents for the macromonomer and
initiator concentration dependences in the polymerization of hydrophobic
monomer, styrene and n-butyl methacrylate are in good agreement with those
from Eq. 5. Figure 4a shows a comparison of Eq. 5 with the particle radius
obtained for the dispersion copolymerization of styrene with the PEO macro-
monomer 7 (m=4, n=45), in methanol-water medium (9/1 v/v). One sees that
the experimental particle radius is quantitatively described by the model with
reasonable constants, q=1, Ç=1.05 g cm–3, NA=6.02¥1023, k2=109 L mol–1 s–1,
kp=352 L mol–1 s–1, kt=6.1¥107 L mol–1 s–1, kd=3.2¥10–7 s–1, f=1, and Scrit/r1=
10 nm2, r1=1.
In remarkable contrast, unusually high exponent values (~1.2) have been
obtained in the dispersion copolymerization of a polar monomer, MMA in
methanol-water (8:2 and 7:3 v/v) media. The value of the exponent drops to
0.51 when the water content is increased to higher than 40%, as shown in

Table 5 Values of the exponents in R=K [Monomer] a[Macromonomer] b[Initiator]

c
(see also Eq. 5) for dispersion copolymerizations with macromonomers

Macromonomer Monomer Medium a b c Reference

Theory (Eq. 5) – – 0.67 –0.50 –0.083 –

7, m=4, n=45 ST MeOH/H2O (9:1) 0.63 –0.52 –0.068 [131]
7, m=1, 4, 7, BMA MeOH/H2O (8:2) 0.82 –0.54 –0.10 [132]
n=53, 110
9b, m=11, n=40 ST EtOH/H2O (9:1) 1.02 –0.60 –0.090 [134]
10, m=2.6, n=41 ST MeOH/H2O (9:1) – –0.47 – [135]
10, m=4.1, n=41 ST MeOH/H2O (9:1) – –0.40 – [135]
10, m=2.6, n=41 ST MeOH/H2O (9:1) – –0.59 – [135]
8, m=3,5,6, n=50 ST MeOH/H2O (9:1) – –0.50 – [128]
4, n=130 ST Silicone oil – –0.40 – [120]
(M=2700)
7, m=1, n=45 MMA MeOH/H2O (8:2) – –1.17 – [133]
7, m=1, n=45 MMA MeOH/H2O (7:3) 0.85 –1.15 –0.030 [133]
7, m=1, n=45 MMA MeOH/H2O (6:4) – –0.51 – [133]
7, m=1, n=45 MMA MeOH/H2O (5:5) – –0.52 – [133]
7, m=1, n=36 ST/AN Ethanol/H2O (8:2) – –0.68 – [153]
320 S. Kawaguchi · K. Ito

Fig. 4 Double logarithmic plots of average particle radius (R) as a function of weight (WDo)
of PEO macromonomer. a Styrene 7 (m=4 and n=45), Wo=100 g/L, [I]0=0.0122 mol/L, q=1
at 60 °C. The straight line is a theoretical curve calculated from Eq. 5 with the parameters in
the text. b MMA 7 (m=1 and n=45), Wo=100 g/L, [I]0=0.012 mol/L, q=1 at 60 °C in methanol:
water=8:2 (empty triangles), =7:3 (empty circles), =6:4 (empty squares), and =5:5 (filled
squares)

Fig. 4b. The significant change in the value of the exponent with the polarity of
the continuous phase cannot be simply explained by the current model, so
further refinements [163–165] are needed.
The criteria for designing PEO macromonomers to be used as efficient dis-
persants in dispersion polymerization have been thoroughly studied by Ito
et al. [128, 131, 135, 151], who examined the effects of the length of spacer (m),
the degree of polymerization of PEO chain (n) of 7, and charge group at the
w-end of 25 on particle size.
Dispersion Polymerization 321

5
Chain Conformation of Grafted Polymer Chains at the Particle Surface

Polymer adsorption has been a subject of both theoretical and experimental

interest, because the adsorption behavior of the polymer at the solid-liquid
interface is strongly connected with many technologically important processes
such as flocculation, adhesion, coating, and lubrication in addition to the col-
loidal stabilization already discussed above. This subject has been recently
reviewed by Cosgrove, Griffiths [166], Fleer et al. [167], and Kawaguchi and
Takahashi [168]. Among a variety of adsorbed polymers, two are of particular
interest with respect to macromonomers. One is the adsorption of comb and
graft copolymers with highly grafted chain density onto the solid surface,
which is referred to as “brush adsorption”, as illustrated in (a) of Fig. 5.Another
is the attachment of the double bond of the macromonomer with the solid
surface by chemical reaction, which is referred to as “terminally-attached

Fig. 5 Schematic representation of the possible conformations of adsorbed (co)polymers

prepared using the macromonomer technique. a Brush adsorption of graft copolymer; b ter-
minally-attached adsorption, and; c the mushroom-brush transition for strongly overlap-
ping chains as proposed by de Gennes [162] and Alexander [170]
322 S. Kawaguchi · K. Ito

adsorption”, as illustrated in (b) of Fig. 5. The conformational properties of

these grafted polymer chains have been the subject of growing attention, from
the point of view of the “mushroom-brush” transition proposed by de Gennes
[162, 169] and Alexander [170], as shown in (c) of Fig. 5.
While there have been many studies on the conformational properties of
terminally-attached polymer chains, prepared either by the adsorption of block
copolymers in a selective solvent onto a solid surface [171] or by the reaction
of a solid surface with reactive groups of polymer [172], little has been reported
for graft copolymer chains prepared from macromonomers. Cairns et al. [173]
carried out a SANS study of a non-aqueous dispersion system comprised
of deuterated PMMA latex grafted with poly(12-hydroxystearic acid), 1. The
thickness of the layer was found to correspond to about 2/3 of the extended
chain length. Comb-like PEO gradient surfaces were grafted onto low density
PE sheets by corona discharge treatment followed by the homopolymerization
of the PEO macromonomers 9b (n=1, 5 and 10) and the gradient PEO concen-
tration at the surface was characterized by measuring the water contact angle,
by FTIR-ATR, and by ESCA [174]. The gradient surfaces can be used to inves-
tigate the interactions between biological species and the surface PEO chains.
Hadziioannou and coworkers [175] prepared a terminally-attached cationic
polyelectrolyte brush on a gold-coated Si-wafer by end-grafting styryl-termi-
nated poly(vinylpyridine) macromonomer, followed by quarternarization with
methyl iodide. The surface was characterized by means of scanning force micro-
scopy, ellipsometry, and FTIR-ATR.
Wu et al. [176] studied the surface properties of PS and PMMA microspheres
stabilized by the PEO macromonomer 7 (m=1) using dynamic light scatter-
ing, and claimed that for PMMA microspheres the surface area occupied by
a PEO molecule is nearly twice as large as that for PS microspheres, assuming
that 100% macromonomer is copolymerized to attach to the latex surface.
However, this is not the case for styrene copolymerization with PEO macro-
monomers in which only 10% PEO macromonomer was copolymerized
[131]. In contrast, it was confirmed that 100% of PEO macromonomers
were copolymerized for the MMA and BMA dispersion copolymerization [132,
133].
1H NMR studies have been carried out for the dispersion copolymerization

of BMA with the PEO macromonomer 7 (m=7), in a deuterated methanol-water

medium [177]. The fractional composition and surface-grafted PEO concen-
tration were monitored as functions of conversion and particle size. In Fig. 6,
the mobile fraction of PEO chains incorporated into the particles is plotted
against the interchain spacing D, as shown in (c) of Fig. 5, which can be calcu-
lated using particle size values and conversions. One sees that the values of the
mobile fraction increase sharply with decreasing D in the region of D<1.6 nm
and become constant below D=1.4 nm. The radius of gyration of the PEO chain
coil in methanol is calculated to be 1.6 nm, which corresponds to the D value
at which a sharp increase in mobile fraction occurs. This result may suggest that
the onset of a pancake-to-brush transition of grafted chains at the interface
Dispersion Polymerization 323

Fig. 6 Plots of mobile fraction of surface anchored PEO chains against the estimated mean
separation D between PEO anchor points on the surface of the particles. The D values were
calculated from the particle size and number, assuming that all PEO chains were located at
the surface

occurs when, say, D≈<S2>1/2, as expected by theory [162, 169, 170]. This subject
should lead to a better understanding of the true nature of the steric stabiliza-
tion that exists in many dispersion systems.

6
Conclusions and Future

A wide variety of polymer microspheres can be made by dispersion polymer-

ization. A key component in all of these systems is the stabilizer (dispersant)
both during particle formation and for the stability of the resulting colloidal
particles. Functionality can be introduced into colloidal particles in various
ways: by copolymerization of functional monomers (like HEMA), or incorpo-
ration of functional dispersants, initiators, chain transfer agents, or macro-
monomers. Many different types of macromonomer are prepared and used to
prepare functional microspheres.Amphiphilic macromonomers provide a par-
ticularly versatile component in these systems, being the source of both sta-
bilizer and functional residue. They act as stabilizer because they are covalently
grafted onto the particles’ surface by copolymerization with main monomers,
and form tightly bound hairy shells on the particles’ surface.
The experimental details of dispersion polymerization with various poly-
meric dispersants and macromonomers are fairly well established. A basic
expression for particle size control has also been derived for the formation of
clear-cut core-shell particles based on highly incompatible core-shells such as
polystyrene-PVP and polystyrene-PEO. However, results deviate considerably
from theory in compatible polymers such as PMMA with PEO macromonomer.
The detailed structures of the “hairy” shells need to be discovered in order
to better understand the exact mechanism of their formation and stabilizing
function.
324 S. Kawaguchi · K. Ito

References
1. Ugelstad J, Mørk PC, Mfutakanba HR, Soleimany E, Nordhuus I, Schmid R, Berge A,
Ellingsen T, Aune O, Nustad K (1983) In: Poehlein GW, Ottewill RH, Goodwin JW (eds)
Science and technology of polymer colloids. Nijhoff, Boston, MA
2. Ober CK, Lok KP (1986) US Pat 4617249
3. Ugelstad J, Berge A, Ellingsen T, Schmid R, Nilsen T-N, Mørk PC, Stenstad P, Hornes E,
Olsvik Ø (1992) Prog Polym Sci 17:87
4. Vanderhoff JV, El-Aasser MS, Micale FJ, Sudol ED, Tseng CM, Silwanowicz A, Kornfeld
DM, Vicente FA (1984) J Dispers Sci Technol 5:231
5. Vanderhoff JV, El-Aasser MS, Micale FJ, Sudol ED, Tseng CM, Silwanowicz A, Sheu HR,
Kornfeld DM (1986) Polym Mater Sci Eng 54:1986
6. Ugelstad J, Mørk PC, Kaggerud KH, Ellingsen T, Berge A (1980) Adv Coll Interfac Sci
13:101
7. Ugelstad J, Mørk PC, Kaggerud KH, Ellingsen T, Khan AA (1982) In: Piirma (ed) Emul-
sion polymerization. Academic, New York, Ch 11
8. Omi S, Katami K, Yamamoto A, Iso M (1994) J Appl Polym Sci 51:1
9. Kamiyama M, Koyama K, Matsuda H, Sano Y (1993) J Appl Polym Sci 50:107
10. Okubo M, Shiozaki M, Tsujihiro M, Tsukada Y (1991) Coll Polym Sci 269:222
11. Barret KEJ (ed) (1975) Dispersion polymerization in organic media. Wiley, London
12. Almog Y, Reich S, Levy M (1982) Brit Polym J 14:131
13. Lok KP, Ober CK (1985) Can J Chem 63:209
14. Ober CK, Lok KP, Hair ML (1985) J Polym Sci Pol Lett 23:103
15. Ober CK, Hair ML (1987) J Polym Sci Pol Chem 25:1395
16. Ober CK, Lok KP (1987) Macromolecules 20:268
17. Tseng CM, Lu YY, El-Aasser MS, Vanderhoff JW (1986) J Polym Sci Pol Chem 24:2995
18. Okubo M, Ikegami K, Yamamoto Y (1989) Colloid Polym Sci 267:193
19. Okubo M, Katayama Y, Yamamoto Y (1991) Colloid Polym Sci 269:217
20. Paine AJ, McNulty J (1990) J Polym Sci Pol Chem 28:2569
21. Paine AJ (1990) J Colloid Interf Sci 138:157
22. Paine AJ (1990) J Polym Sci Pol Chem 28:2485
23. Paine AJ, Luymes W, McNulty (1990) Macromolecules 23:3104
24. Paine AJ (1990) Macromolecules 23:3109
25. Croucher MD, Winnik MA (1990) In: Candau F, Ottewill RH (eds) An introduction to
polymer colloids. Kluwer Academic, Dordrecht, p 35
26. Guyot A, Tauer K (1994) Adv Polym Sci 111:43
27. Cawse JL (1997) In: Lovell PA, El-Aasser MS (eds) Emulsion polymerization and emul-
sion polymers. Wiley, Chichester, UK, p 743
28. Pichot C, Delair T, Elaissari A (1997) In: Asua JM (ed) Polymeric dispersions: principles
and applications. Kluwer Academic, Dordrecht, p 515
29. Asua JM, Schoonbrood HAS (1998) Acta Polym 49:671
30. Ito K (1998) Prog Polym Sci 23:581
31. Ito K, Kawaguchi S (1999) Adv Polym Sci 142:129
32. Ito K, Cao J, Kawaguchi S (2002) In: Arshady R, Guyot A (eds) Functional colloids (MML
Series, Vol 4). Citus, London, p 109
33. Sáenz J, Asua J M (1995) J Polym Sci Pol Chem 33:1511
34. Sáenz J, Asua J M (1995) J Polym Sci Pol Chem 34:1977
35. Bamnolker H, Margel S (1996) J Polym Sci Pol Chem 34:1857
36. Horák D, Švec F, Fréchet JMJ (1995) J Polym Sci Pol Chem 33:2961
37. Kiatkamjornwong S, Kongsupapsiri C (2000) Polym Int 49:1395
Dispersion Polymerization 325

38. Yang W, Yang D, Hu J, Wang C, Fu S (2001) J Polym Sci Pol Chem 39:555
39. Takahashi K, Nagai K (1996) Polymer 37:1257
40. Shen S, Sudol ED, El-Aasser MS (1994) J Polym Sci Pol Chem 32:1087
41. Horák D, Švec F, Fréchet JMJ (1995) J Polym Sci Pol Chem 33:2329
42. Takahashi K, Uyama H, Kobayashi S (1998) J Macromol Sci Pure A35:1473
43. Ho C-H, Chen S-A, Amiridis MD, Zee JWV (1997) J Polym Sci Pol Chem 35:2907
44. Chen Y, Yang H-W (1992) J Polym Sci Pol Chem 30:2765
45. Arhady R (1999) In: Arshady R (ed) Microspheres, microcapsules & liposomes, vol 1.
Citus, London, p 11
46. Margel S (1999) In: Arshady R (ed) Microspheres, microcapsules & liposomes, vol 2.
Citus, London, p 11
47. Takahashi K, Miyanori S, Uyama H, Kobayashi S (1996) J Polym Sci Pol Chem 34:175
48. Horák D (1999) J Polym Sci Pol Chem 37:3785
49. Takahashi K, Miyamori S, Uyama H, Kobayashi S (1997) Macromol Rapid Commun
18:471
50. Takahashi K, Uyama H, Kobayashi S (1998) Polym J 30:684
51. Park KY, Jeong WW, Suh KD (2003) J Macromol Sci A50:617
52. Yang W, Hu J, Tao Z, Li L, Wang C, Fu S (1999) Colloid Polym Sci 277:446
53. Horák A, Shapoval P (2000) J Polym Sci Pol Chem 38:3855
54. Ray B, Mandal BM (1999) J Polym Sci Pol Chem 37:493
55. Lee K, Lee S, Song B, Lee D (2000) Polymer (Korea) 24:629
56. Margel S, Nov E, Fisher I (1991) J Polym Sci Pol Chem 29:347
57. Bahar T, Tuncel A (1999) Polym Eng Sci 39:1849
58. Horák A, Kryštůfek M, Spĕváček J (2000) J Polym Sci Pol Chem 38:653
59. Covolan VL, D’Antone S, Ruggeri G, Chiellini E (2000) Macromolecules 33:6685
60. Yang W, Zhou H, Tao Z, Hu J, Wang C, Fu S (2000) J Macromol Sci Pure A37:659
61. Dawkins JV, Neep DJ, Shaw PL (1994) Polymer 35:5366
62. Baines FL, Dionisio S, Billingham NC, Armes SP (1996) Macromolecules 29:3096
63. Watanabe S, Ueno K, Kudoh K, Murata M, Masuda Y (2000) Macromol Rapid Commun
21:1323
64. Omi S, Saito M, Hashimoto S, Nagai M, Ma G-H (1998) J Appl Polym Sci 68:897
65. Horák D (2001) J Polym Sci Pol Chem 39:3707
66. Horák D, Boháček J, Šubrt M (2000) J Polym Sci Pol Chem 38:1161
67. Horák D, Semenyuk N, Lednicky F (2000) J Polym Sci Pol Chem 41:1848
68. Bourgeat-Lami E, Lang J (1998) J Colloid Interf Sci 197:293
69. Percy MJ, Michailidou V, Armes SP, Perruchot C, Watts JF, Greaves SJ (2003) Langmuir
19:2072
70. (a) Li K, Stöver HDH (1993) J Polym Sci Pol Chem 31:2473; (b) Li K, Stöver HDH (1993)
J Polym Sci Pol Chem 31:3257
71. (a) Downey JS, Frank RS, Li W-H, Stöver HDH (1999) Macromolecules 32:2838;
(b) Li W-H, Stöver HDH (2000) Macromolecules 33:4354
72. Li WH, Li LK, Stöver HDH (1999) J Polym Sci Pol Chem 37:2295
73. Lee C-F, Wen C-J, Chiu Q-Y (2003) J Polym Sci Pol Chem 41:2063
74. Hölderle M, Bar G, Mülhaupt R (1997) J Polym Sci Pol Chem 35:2539
75. Kim J-W, Suh K-D (1998) Colloid Polym Sci 276:878
76. Huang J, Zhang H, Hou J, Jiang P (2002) React Funct Polym 53:1
77. Huang J-X, Yuan X-Y, Yu X-L, Zhang H-T (2003) Polym Int 52:819
78. Okubo M, Nakagawa T (1992) Colloid Polym Sci 270:853
79. Okubo M, Minami H (1996) Macromol Symp 101:509
80. Okubo M, Minami H (1996) Colloid Polym Sci 274:433
326 S. Kawaguchi · K. Ito

81. Okubo M, Konishi Y, Minami H (1998) Colloid Polym Sci 276:638

82. Okubo M, Minami H (1997) Colloid Polym Sci 275:992
83. Okubo M, Yamashita T, Minami H, Konish Y (1998) Colloid Polym Sci 276:887
84. Schneider M, Mülhaupt R (1994) Polym Bull 32:545
85. Awan MA, Dimonie VL, El-Aasser MS (1996) J Polym Sci Pol Chem 34:2633
86. Awan MA, Dimonie VL, El-Aasser MS (1996) J Polym Sci Pol Chem 34:2651
87. Tausendfreund I, Bandermann F, Siesler HW, Kleimann M (2002) Polymer 43:7085
88. Kim J, Jeong SY, Kim KU, Ahn YH, Quirk RP (1996) J Polym Sci Pol Chem 34:3277
89. Murray JG, Schwab FC (1982) Ind Eng Chem Prod Res Dev 21:93
90. Stampa GB (1970) J Appl Polym Sci 14:1227
91. Keki S, Deak G, Daroczi L, Kuki A, Zsuga M (1999) Macromol Symp 157:217
92. Schwab FC, Murray JG (1985) Anionic dispersion polymerization of styrene. In:
Culbertson BM, McGrath JE (eds) Advances in polymer synthesis. Plenum, New York
93. Jenkins AD, Maxfield D, dos Santos CG, Walton DRM, Stejskal J, Kratochvil P (1992)
Makromol Chem Rapid Commun 13:61
94. Okay O, Funke W (1990) Macromolecules 23:2623
95. Holderle M, Baumert M, Mülhaupt R (1997) Macromolecules 30:3420
96. Gabaston LI, Jackson RA, Armes SP (1998) Macromolecules 31:2888
97. Sosnowski S, Gadzinowski M, Slomokowski S, Penczek S (1994) J Bioact Compat Pol
9:345
98. Sosnowski S, Gadzinowski M, Slomokowski S (1996) Macromolecules 29:4556
99. Armes SP, Aldissi M (1989) J Chem Soc Chem Commun 88
100. Eisazadeh H, Gilmore KJ, Hodgson AJ, Spinks G, Wallace GG (1995) Colloid Surface A
103:281
101. Eisazadeh H, Gilmore KJ, Wallace GG (1995) Polym Int 37:87
102. Banerjee P, Mandal BM (1995) Synth Met 74:257
103. Banerjee P, Digar ML, Bhattacharyya SN, Madal BM (1995) Langmuir 11:2414
104. Armes SP, Aldissi M, Hawley M, Berry JG, Gottesfeld S (1991) Langmuir 7:1447
105. Nagaoka T, Nakano H, Suyama T, Ogura K, Oyama M, Okazaki S (1997) Anal Chem
69:1030
106. Stejskal J, Kratochvíl P, Gospodinova N, Terlemezyan L, Mokreva P (1992) Polymer
33:4857
107. Shannon K, Fernandez JE (1994) J Chem Soc Chem Commun 643
108. Chattopadhyay D, Mandal BM (1996) Langmuir 12:1585
109. Stejskal J, Špírková M, Riede A, Helmstedt M, Mokreva P, Prokeš J (1999) Polymer
40:2847
110. Okubo M, Masuda T, Mukai T (1998) Colloid Polym Sci 276:96
111. Minami H, Okubo M, Murakami K, Hirano S (2000) J Polym Sci Pol Chem 38:4238
112. Mandal TK, Mandal BM (1995) Polym Commun 36:1911
113. Mandal TK, Mandal BM (1999) J Polym Sci Pol Chem 37:3723
114. Digar ML, Bhattacharyya SN, Mandal BM (1994) Polymer 35:377
115. Uyama H, Kurioka H, Kobayashi S (1995) Chem Lett 795
116. Kobayashi S, Uyama H, Ohmae M (2001) Bull Chem Soc Jpn 74:613
117. Ramanathan LS, Shukla PG, Sivaram S (1998) Pure Appl Chem 70:1295
118. Bunn A, Griffin BP, MacDonald WA, Rance DG (1992) Polymer 33:3066
119. Kawaguchi S, Okada T, Tano K, Ito K (2000) Des Monomers Polym 3:263
120. Uchida T, Kawaguchi S, Ito K (2002) Des Monomers Polym 5:285
121. Kobayashi S, Uyama H, Choi JH, Matsumoto Y (1991) Proc Jpn Acad Ser B 67:140
122. (a) Kobayashi S, Uyama H, Lee SW, Matsumoto Y (1993) J Polym Sci Pol Chem 31:3133;
(b) Shimano Y, Sato K, Kobayashi S (1995) J Polym Sci Pol Chem 33:2715
Dispersion Polymerization 327

123. Uyama H, Kato H, Kobayashi S (1993) Chem Lett 261

124. (a) Kobayashi S, Uyama H, Narita Y (1992) Makromol Chem Rapid Commun 13:337; (b)
Koyabashi S, Uyama H (1993) Kobunshi Ronbunshu 50:209
125. Furuhashi H, Kawaguchi S, Itsuno S, Ito K (1997) Colloid Polym Sci 275:227
126. Akashi M, Chao D, Yashima E, Miyauchi N (1990) Appl Polym Sci 39:2027
127. Capek I, Riza M, Akashi M (1992) Polym J 24:959
128. (a) Shen R, Akiyana C, Senyo T, Ito K (2003) C R Chim 6:1329; (b) Shen R, Senyo T,
Akiyana C, Atago Y, Ito K (2003) Polymer 44:3221
129. Prestige C, Tadros ThF (1988) J Colloid Interf Sci 124:660
130. (a) Capek I, Riza M, Akashi M (1992) Makromol Chem 193:2843; (b) Riza M, Capek I,
Kishida A, Akashi M (1993) Angew Makromol Chem 206:69; (c) Capek I, Riza M,
Akashi M (1997) J Polym Sci Pol Chem 35:3131; (d) Capek I, Nguyen SH, Berek D (2000)
Polymer 41:7011
131. Nugroho M, Kawaguchi S, Ito K (1995) Macromolecular Reports A32:593
132. Kawaguchi S, Winnik MA, Ito K (1995) Macromolecules 28:1159
133. (a) Ito K (1994) Kobunshi Kako 30:510; (b) Hattori T, Nugroho MB, Kawaguchi S, Ito K
(1996) Polym Prepr Jpn 45:157
134. (a) Liu J, Gan LM, Chew CH, Quek CH, Gan LH (1997) J Polym Sci Pol Chem 35:3575;
(b) Liu J, Chew CH, Wong SY, Gan LM, Lin J, Tan KL (1998) Polymer 39:283
135. Imai H, Kawaguchi S, Ito K (2003) Polym J 35:528
136. (a) Lacroix-Desmages P, Guyot A (1996) Macromolecules 29:4508; (b) Lacroix-
Desmages P, Guyot A (1996) Colloid Polym Sci 274:1129
137. (a) Akashi M, Yanagi T, Yashima E, Miyauchi N (1989) J Polym Sci Pol Chem 27:3521;
(b) Serizawa T, Takehara S, Akashi M (2000) Macromolecules 33:1759; (c) Chen M-Q,
Serizawa T, Kishida A, Akashi M (1999) J Polym Sci Pol Chem 37:2155
138. Shay JS, English RJ, Spontak RJ, Balik CM, Khan SA (2000) Macromolecules 33:6664
139. Akashi M, Yanagi T, Yahsima E, Miyauchi N (1989) J Polym Sci Pol Chem 27:3521
140. Iwasaki I,Yashima E,Akashi M, Miyauchi N, Marumo K (1991) Polym Prepr Jpn 40:2597
141. (a) Ishizu K, Tahara N (1996) Polymer 37:1729; (b) Ishizu K, Yamashita M, Ichimura A
(1997) Macromol Rapid Commun 18:639
142. Riza M, Tokura S, Kishida A, Akashi M (1993) Polym Prepr Jpn 42:4617
143. Chen MQ, Kishida A, Akashi M (1996) J Polym Sci Pol Chem 34:2213
144. Riza M, Tokura S, Iwasaki M, Yashima E, Kishida A, Akashi M (1995) J Polym Sci Pol
Chem 33:1219
145. Ishizu K, Yamashita M, Ichimura A (1997) Polymer 38:5471
146. Shaffer KA, Jones TA, Canelas DA, Desimone JM,Wilkinson SP (1996) Macromolecules
29:2704
147. Ishizu K, Tahara N (1996) Polymer 37:2853
148. Obayashi N, Kawaguchi S, Ito K (1997) Polym Prepr Jpn 46:162
149. Nagasaki Y, Sato Y, Kato M (1997) Macromol Rapid Commun 18:827
150. Lascelles SF, Malet F, Mayada R, Billingham NC, Armes SP (1999) Macromolecules
32:2462
151. (a) Senyo T, Atago Y, Kiang H, Shen R, Ito K (2003) Polym J 35:513; (b) Senyo T, Atago
Y, Shen R, Ito K (2003) Polym Prepr Jpn 52:1224
152. Fujioka M, Ma G-H, Du Y-Z, Ogino K, Nagai M, Omi S (2003) J Polym Sci Pol Chem
41:1788
153. Chen M-Q, Kaneko T, Chen C-H, Akashi M (2001) Chem Lett 1306
154. Kaneko T, Hamada K, Chen M-Q, Akashi M (2004) Macromolecules 37:501
155. (a) Yilidiz U, Hazer B, Capek I (1995) Angew Makromol Chem 231:135; (b) Yilidiz U,
Hazer B (2000) Polymer 41:539
328 Dispersion Polymerization

156. (a) Bourgeat-Lami E, Guyot A (1997) Colloid Polym Sci 275:716; (b) Schipper ETWM,
Sindt O, Hamaide T, Lacroix-Desmazes P, Müller B, Huyot A, van den Enden MJWA,
Vidal F, van Es JJGS, German AL, Montaya Goñi AM, Sherrington DC, Schoonbrood
HAS, Asua JM, Sjöberg M (1998) Colloid Polym Sci 276:402
157. Feeney PJ, Napper DH, Gilbert RG (1987) Macromolecules 20:2922
158. Fitch RM, Tsai CH (1971) In: Fitch RM (ed) Polymer colloids. Plenum, New York
159. Gilbert RG (1995) In: Emulsion polymerization: A mechanistic approach. Academic,
New York
160. Smoluchowski MV (1917) Z Phys Chem 192:129
161. Frenklach M (1985) J Colloid Interf Sci 108:237
162. de Gennes PG (1980) Macromolecules 13:1069
163. Yasuda M,Yokoyama H, Seki H, Ogino H, Ishimi K, Ishikawa H (2001) Macromol Theor
Simul 10:54
164. Yasuda M, Seki H,Yokoyama H, Ogino H, Ishimi K, Ishikawa H (2001) Macromolecules
34:3261
165. Lacroix-Desmazes P, Guillot J (1998) J Polym Sci Pol Phys 36:325
166. Cosgrove T Griffiths PC (1992) Adv Colloid Interfac 42:175
167. Fleer GJ, Cohen Stuart MA, Scheutjens JMHM, Cosgrove T, Vincent B (1993) In: Poly-
mers at interfaces. Chapman & Hall, London
168. Kawaguchi M, Takahashi A (1992) Adv Colloid Interfac 37:219
169. de Gennes PG (1987) Adv Colloid Interfac 278:189
170. Alexander S (1977) J Phys (Paris) 38:983
171. Auroy P, Auvray L, Leger LL (1991) Physica A 172:269
172. Taunton HJ, Toprakcioglu C, Fetters LJ, Klein J (1990) Macromolecules 23:571
173. Cairns RJR, Ottewill RH, Osmond DWJ, Wagstaff I (1976) J Colloid Interf Sci 54:45
174. Jeong BJ, Lee JH, Lee HB (1996) J Colloid Interf Sci 178:757
175. Werts MPL, van der Vegte EW, Hadziioannou G (1997) Langmuir 13:4939
176. Wu C, Akashi M, Chen MQ (1997) Macromolecules 30:2187
177. Kawaguchi S, Winnik MA, Ito K (1996) Macromolecules 29:4465

Received: March 2004

Adv Polym Sci (2005) 175: 329–346
DOI 10.1007/b100114
© Springer-Verlag Berlin Heidelberg 2005

Heterogeneous Polymerization of Fluoroolefins

in Supercritical Carbon Dioxide
Karen A. Kennedy 1 · George W. Roberts 1 · Joseph M. DeSimone ( ) 1, 2
1
Department of Chemical Engineering, North Carolina State University, Raleigh,
NC 27695–7905, USA
[email protected], [email protected]
2 Department of Chemistry, University of North Carolina at Chapel Hill, Chapel Hill,
NC 27599–3290, USA
[email protected]

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 330

2 Conventional Polymerization of Fluoroolefins . . . . . . . . . . . . . . . . 331

2.1 Tetrafluoroethylene . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332
2.2 Vinylidene Fluoride . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333
2.3 Copolymers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334

3 Supercritical Carbon Dioxide as a Polymerization Medium . . . . . . . . . 335

3.1 Properties of Supercritical Carbon Dioxide . . . . . . . . . . . . . . . . . . 335
3.2 Advantages of Using Supercritical Carbon Dioxide . . . . . . . . . . . . . . 336

4 Precipitation Polymerization in Supercritical Carbon Dioxide . . . . . . . 338

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338
4.2 Polymerization of Tetrafluoroethylene . . . . . . . . . . . . . . . . . . . . 338
4.3 Polymerization of Vinylidene Fluoride . . . . . . . . . . . . . . . . . . . . 339
4.3.1 Continuous Polymerization . . . . . . . . . . . . . . . . . . . . . . . . . . 340
4.3.1.1 Polymer Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342
4.4 Synthesis of Copolymers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343

5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345

Abstract In recent years, carbon dioxide has been investigated for its potential to replace the
aqueous and organic solvents used in polymerization processes. Carbon dioxide has several
benefits, including being environmentally benign, a tunable solvent, resistant to chain trans-
fer, and of low viscosity which facilitates high initiator efficiencies and fluid handling. Homo-
and copolymerization of fluoroolefins such as tetrafluoroethylene (TFE), vinylidene fluoride
(VF2), hexafluoropropylene (HFP), and perfluorinated vinyl ethers in carbon dioxide are
particularly interesting because of the significant waste reduction and the elimination of po-
tentially harmful processing agents. This review focuses on recent developments in poly-
merizing fluoroolefins via heterogeneous polymerization in carbon dioxide. At least one
CO2-based polymerization process has recently been commercialized, and additional re-
search is underway to understand the kinetics associated with polymerizing fluoroolefins
330 K. A. Kennedy et al.

in carbon dioxide. As researchers improve their understanding of CO2-based polymeriza-

tions, and as industry continues to look for more economically and environmentally-sound
alternatives to existing processes, it will become more and more common to employ carbon
dioxide as a polymerization medium.

Keywords Supercritical carbon dioxide · Fluoropolymers · Tetrafluoroethylene ·

Vinylidene fluoride

Abbreviations
AIBN 2,2¢-azobis(isobutyronitrile)
APFO ammonium perfluorooctanoate
BPPP Bis(perfluoro-2-N-propoxypropionyl) peroxide
C8 ammonium perfluorooctanoate
CMC Critical micelle concentration
CSTR Continuous stirred tank reactor
CTFE Chlorotetrafluoroethylene
DEPDC Diethyl peroxydicarbonate
EPA Environmental Protection Agency
FEP Poly(tetrafluoroethylene-co-hexafluoropropylene)
FMG Fluoropolymers Manufacturers Group
FTIR Fourier transform infrared spectroscopy
HFP Hexafluoropropylene
LDPE Low-density polyethylene
Mn Number-average molecular weight
MAC Maleic acid
MAN Maleic anhydride
MFI Melt flow index
MWD Molecular weight distribution
NMR Nuclear magnetic resonance spectroscopy
PFA Poly(tetrafluoroethylene-co-perfluoro(propyl vinyl ether))
PMVE Perfluoro(methyl vinyl ether)
PPVE Perfluoro(propyl vinyl ether)
PTFE Poly(tetrafluoroethylene)
PVDF Poly(vinylidene fluoride)
scCO2 Supercritical carbon dioxide
SCF Supercritical fluid
TFE Tetrafluoroethylene
VF Vinyl fluoride
VF2 Vinylidene fluoride

1
Introduction

The first application of supercritical fluids (SCFs) to the polymer industry

occurred in the 1930s, with the development of a free-radical bulk polymer-
ization of supercritical ethylene to produce low-density polyethylene (LDPE)
[1, 2]. LDPE is synthesized by a high-pressure process, at 180–300 °C and 1000–
Heterogeneous Polymerization of Fluoroolefins in Supercritical Carbon Dioxide 331

Table 1 Critical temperatures and pressures of substances useful as supercritical fluids [9]

Critical temperature Tc (°C) Critical pressure Pc (bar)

Carbon Dioxide (CO2) 31.1 73.8

Water (H2O) 374.4 221.2
Ethane (C2H6) 32.4 48.8
Propane (C3H8) 96.8 42.5
Ammonia (NH3) 132.4 113.5
Nitrous Oxide (N2O) 36.6 72.4
Fluoroform (CHF3) 26.3 48.6

3000 bar, in which the polymer is soluble in supercritical ethylene monomer

[3]. However, inert SCFs have only recently been exploited in a wide range of
polymer applications. Research with supercritical carbon dioxide (scCO2) has
outpaced research with other SCFs because carbon dioxide is inexpensive, has
a critical point (Table 1) at conditions lower than that of most other SCFs (in-
creasing the range of tunability), and is environmentally benign. Supercritical
carbon dioxide has found particular application as a polymerization medium
because of its inertness to free radicals [4] (in other words, there is no chain
transfer to CO2), and because of CO2’s high initiator efficiencies. DuPont has
built a commercial-scale facility for the production of tetrafluoroethylene-
based polymers, including TeflonTM FEP (a copolymer of tetrafluoroethylene
and hexafluoropropylene), in carbon dioxide [5]. Significant research [6–8] into
the use of scCO2 as a polymerization medium has been reported. This review
will focus on recent developments in the polymerization of fluoroolefins, such
as tetrafluoroethylene, via heterogeneous polymerization in scCO2. The
particular challenges and rewards associated with polymerizing fluoroolefins
in CO2 will be discussed.

2
Conventional Polymerization of Fluoroolefins

The major commercial fluoropolymers are made by homopolymerization of

tetrafluoroethylene (TFE), chlorotrifluoroethylene (CTFE), vinylidene fluoride
(VF2), and vinyl fluoride (VF), or by co-polymerization of these monomers
with hexafluoropropylene (HFP), perfluoro(propyl vinyl ether) (PPVE), per-
fluoro(methyl vinyl ether) (PMVE), or ethylene. The polymers are formed by
free-radical polymerization in water or fluorinated solvents.
Currently, there is concern about the use of ammonium perfluorooctanoate
(APFO), also known as “C8”, which is necessary for the manufacture of fluorin-
ated plastics and elastomers in water. C8 is a perfluorinated anionic surfactant
used as a dispersing agent in the polymerization and copolymerization of many
fluoropolymers, including poly(tetrafluoroethylene) (PTFE), poly(vinylidene
332 K. A. Kennedy et al.

fluoride) (PVDF), poly(tetrafluoroethylene-co-hexafluoropropylene) (FEP),

poly(tetrafluoroethylene-co-perfluoro(propyl vinyl ether)) (PFA), and poly-
(tetrafluoroethylene-co-perfluoro(methyl vinyl ether)) [10]. It is necessary
to use a perfluorinated anionic surfactant for two important reasons. First,
fluorinated anionic surfactants have the right interfacial properties to disperse
hydrophobic particles like fluoropolymers in water. Second, fluorinated carbon-
centered radicals are extremely electrophilic and will react (via chain transfer)
with any hydrocarbon-based species in solution, including the surfactant.
Hence, the surfactant chosen needs to be hydrogen free.
Earlier this year, the Environmental Protection Agency (EPA) released a
preliminary risk assessment of C8 and other fluorochemicals [10].Although the
EPA refers to C8 as a persistent organic pollutant, there have been no reported
adverse health affects in humans. However, because of these potential environ-
mental and health problems, 3 M, who was the leading manufacturer of C8,
exited the C8 business [11]. DuPont and other members of the Fluoropolymers
Manufacturers Group (FMG) have conducted research over the last 30 years to
identify alternatives to C8 for economic reasons, and because C8 persists in
the body. However, no viable, alternative dispersing agent has been identified.
Carbon dioxide is the only known alternative that can eliminate the need for
C8 in the manufacturing process.
Polymers made from TFE and VF2 dominate the fluoropolymer market.
Existing methods for synthesizing these polymers will be reviewed in the
following sections.

2.1
Tetrafluoroethylene

Poly(tetrafluoroethylene) is a fluorinated analog of polyethylene with chains

that are arranged into a helical twisting conformation in which the carbon
backbone is shielded by fluorine atoms. This results in a polymer with superior
chemical resistance. Additionally, PTFE has high thermal stability and a very
low coefficient of friction. This makes PTFE an ideal polymer to use as an anti-
stick coating and to use with aggressive chemicals. In 2001, the U.S. consumed
39.7 million pounds of PTFE – it was the most widely consumed of any fluoro-
polymer [12]. PTFE is sold as a granular resin, fine powder, or an aqueous dis-
persion for use in a variety of applications, including coatings, chemical proces-
sing equipment, and wire insulation.
PTFE is synthesized via either suspension or dispersion polymerization
in water, depending on the particle size required. Both processes are run at
70–120 °C and 1–6 MPa using a water-soluble initiator such as ammonium
persulfate [13]. In the dispersion polymerization, a dispersing agent, such as
ammonium perfluorooctanoate (C8), and mild agitation are used to produce
a stable dispersion of fine PTFE particles of average particle size 0.2 mm. Hydro-
carbon wax can be added to prevent coagulation of the particles. Because the
amount of dispersing agent is typically less than the critical micelle concen-
Heterogeneous Polymerization of Fluoroolefins in Supercritical Carbon Dioxide 333

tration (CMC), the polymerization is not a true emulsion polymerization. How-

ever, the polymerization does possess characteristics of an emulsion polymer-
ization.After the polymerization is complete, the particles may be precipitated
to form a fine powder or the dispersion may be concentrated for direct use as
a coating. Fine powder PTFE is commonly mixed with a lubricant such as
kerosene to perform paste extrusion of tubing and wire insulation.
In suspension polymerization, little or no dispersing agent and vigorous
agitation is used to produce a coagulated granular resin with particles 1–2 mm
in size. The particles are dried and ground to varying particle sizes for mold-
ing or ram extrusion into precision parts.
Conventional wisdom suggests that PTFE homopolymer cannot be melt-
processed to form molded parts due to its extremely high molecular weight
and high melt viscosity. Grades of PTFE, known as micropowders, with low
molecular weight and low melt viscosity are typically used as additives in coat-
ings and thermoplastics [14]. However, these grades cannot be melt-processed
due to their brittleness. Tervoort et al [15] recently reported that PTFE with
medium viscosities (melt flow rate 0.2–2.6 g/10 min) and bimodal molecular
weight distributions can be melt processed. PTFE grades with high and low
viscosity are blended in a twin-screw extruder to form a blend exhibiting a
pseudobimodal molecular weight distribution. The blend retains the mechan-
ical strength of the high viscosity material, but the presence of the low viscosity
material enables melt-processing of the material into a variety of parts and also
suggests the possible recycling of PTFE.

2.2
Vinylidene Fluoride

Poly(vinylidene fluoride) is a semi-crystalline polymer that is produced com-

mercially by a free-radical emulsion or suspension polymerization of vinylidene
fluoride (CH2=CF2) in water [16–18]. PVDF exhibits a unique combination of
mechanical and electrical properties due to the alternating spatial arrangement
of CH2 and CF2 groups along the polymer backbone. PVDF is the fluoro-
polymer consumed in the greatest quantities aside from PTFE [12], because of
its (1) high mechanical and impact strength, (2) resistance to environmental
stress, (3) ease of melt-processability, and (4) low cost compared to other
fluoropolymers. PVDF is used in architectural coatings (~40%), semiconductor
manufacture and chemical processing (~40%), and wire and cable insulation
(~20%) [19]. A small, but increasing amount of PVDF is used as electrodes in
lithium batteries [12, 20].
Polymerization of vinylidene fluoride by emulsion or suspension polymer-
ization in water is conducted at conditions of 10–130 °C and 10–200 bar. In the
emulsion polymerization, either water-soluble peroxides or monomer-soluble
peroxy or organic peroxides are used as initiators [17]. Fluorinated surfactants,
such as ammonium perfluorooctanoate, are used as dispersing agents. Chain
transfer agents, such as acetone, chloroform, or trichlorofluoromethane, may be
334 K. A. Kennedy et al.

used to control the molecular weight [13]. The resulting latex of dispersed
particles is coagulated, washed, and spray dried to form a fine powder. The
powder consists of spherical agglomerates 0.2–0.5 mm in diameter.
The suspension polymerization is conducted using monomer-soluble per-
oxy initiators.Water-soluble polymers, such as poly(vinyl alcohol), are typically
used as suspending agents to reduce the coalescence of the polymer particles
[17].A slurry of polymer particles 30–100 mm in diameter is formed during the
polymerization. The particles are washed and dried before further processing.
The effect of polymerization conditions on the polymer morphology in
PVDF has been studied [16, 21]. Regioisomer defects in the form of reversed
head-to-tail addition are common in the polymerization of vinyl monomers
[22]. These defects affect the crystallinity and final properties of the polymer.
The fraction of defects in PVDF ranges from 3.5 to 6 mol%, increasing with
reaction temperature [22]. The a-phase crystal is the most common crystal
observed in PVDF. However, an increase in defects leads to a greater amount of
b-phase crystals in the polymer [16, 23], which gives PVDF its piezoelectric
properties. An increase in the amount of b crystals formed was observed for
different initiating systems [23].Additionally, enhancement of the b crystals in
PVDF is achieved by applying tensile forces or mechanical stretching to the
polymer [24]. The g and d crystal phases may be generated from the a-phase
using heat or an electric field, respectively. The g-phase crystal has also been
formed by crystallization under high pressure [25].
Commercial PVDF is sold in powder form or may be compression molded
into pellets. Depending on the polymerization conditions, differing properties
of PVDF may be produced. PVDF exhibiting a multimodal molecular weight
distribution has been reported due to its improved melt flow characteristics
[26] and its use in lithium batteries [24]. Dohany [26] reports the formation of
a distinct bimodal molecular weight distribution that is strongly dependent on
the addition of the initiator to the reaction. The initiator is continuously fed to
the reaction until about 50% of the total monomer feed is added to the reactor.
At that point, the addition of the initiator is stopped and the remaining
monomer is continuously fed to the reactor. The high molecular weight peak
measured by gel permeation chromatography is more than 30% of the overall
molecular weight distribution. The high molecular weight portion allows pro-
cessing of the polymer at high shear rates for improved productivity. The use
of these polymers in lithium batteries is largely the result of their relatively
high defect structure, which promotes ionic conductivity in the electrochem-
ical cell [24].

2.3
Copolymers

PVDF and PTFE have wide application in industry due to their strength, chem-
ical and wear resistance, and dielectric properties. However,VF2 and TFE may
be copolymerized with comonomers such as hexafluoropropylene (HFP) and
Heterogeneous Polymerization of Fluoroolefins in Supercritical Carbon Dioxide 335

vinyl ethers to modify the properties of the polymer. For example, copolymer-
izing VF2 with HFP improves the flexibility of the polymer without compro-
mising the electrical properties. Therefore,VF2-HFP copolymers are used more
commonly in cable insulation than VF2 homopolymer. Copolymers of TFE
with perfluoroalkyl vinyl ethers have properties very similar to the TFE homo-
polymer except that the copolymer may be processed by injection molding
because of its lower melt viscosity. Copolymers of TFE and VF2 can also be
made to form fluorinated elastomers and perfluorinated elastomers when cure
site monomers are used. Important examples of such elastomers include
terpolymers of VF2, TFE, and either HFP or PMVE along with a cure site
monomer such as bromotetrafluorobutene [19].
Copolymers are produced by aqueous copolymerization of the monomers
in a manner similar to the homopolymerizations. Nonaqueous polymerizations
of TFE with PPVE may be conducted in water or in fluorinated solvents at lower
temperatures (30–60 °C) using a soluble organic initiator such as perfluoro-
propionyl peroxide. The molecular weight is controlled by addition of chain
transfer agents such as methanol in nonaqeuous polymerizations or hydrogen
in aqeous polymerizations.

3
Supercritical Carbon Dioxide as a Polymerization Medium

3.1
Properties of Supercritical Carbon Dioxide

Supercritical fluids possess characteristics that make them interesting for use
as polymerization media.A supercritical fluid exists at temperatures and pres-
sures above its critical values. In the supercritical state, the fluid exhibits phys-
ical and transport properties intermediate between the gaseous and liquid
state. This is illustrated in Table 2. SCFs have liquid-like densities, but gas-like
diffusivities. These intermediate properties can provide advantages over liquid-
based processes. In particular, the higher diffusivities of SCFs reduce mass
transfer limitations in diffusion-controlled processes.Additionally, lower energy
is required for processing the supercritical fluid because its viscosity is lower
than that of most liquids, and because the need to vaporize large quantities of
liquid is avoided.

Table 2 Comparison of typical SCF, liquid, and gas properties [27]

Liquid SCF Gas

Density (g/cm3) 1 0.1–0.5 10–3

Viscosity (Pa s) 10–3 10–4–10–5 10–5
Diffusivity (cm2/s) 10–5 10–3 10–1
336 K. A. Kennedy et al.

Fig. 1 Density versus pressure isotherms for carbon dioxide [28]

SCFs have a tunable density that may offer further advantages in reaction
and processing applications. This tunability is illustrated in Fig. 1 for carbon
dioxide. Near the critical point, even small changes in the temperature or
pressure of carbon dioxide dramatically affect its density. Similarly, the viscos-
ity, dielectric constant, and diffusivity are also tunable parameters, which allows
specific control of systems involving supercritical fluids.
Carbon dioxide is inexpensive and widely available. The main source of
carbon dioxide is from chemical manufacturing. For example, it is a major
byproduct in the production of ammonia, ethanol, and hydrogen. Finally, car-
bon dioxide is non-toxic, nonflammable, and easily recycled.

3.2
Advantages of Using Supercritical Carbon Dioxide

Supercritical carbon dioxide is a very good solvent for small molecules, but a
poor solvent for most high molecular weight polymers at mild conditions
(T<100 °C, P<350 bar). Amorphous fluoropolymers and silicones are the only
polymers known to be soluble in CO2 at mild conditions [6]. This difference in
solubilities is an advantage for CO2-based polymerizations, as it can be used to
reduce the energy requirements necessary to separate and purify a polymer
after synthesis. Consider, for example, a batch precipitation polymerization in
Heterogeneous Polymerization of Fluoroolefins in Supercritical Carbon Dioxide 337

CO2. Initially, the reactor contains a homogeneous solution of monomer and

initiator (and perhaps a chain transfer agent) in CO2. As the polymerization
proceeds, the polymer precipitates to form a separate phase. When the polym-
erization is complete, the polymer is separated from the reaction mixture by a
conventional fluid/solid separation process (sedimentation, filtration, centri-
fugation). The fluid phase contains unreacted monomer and initiator, which
can be recycled. When the polymer is depressurized, the CO2 escapes, leaving
a dry polymer. This contrasts with water-based processes in which the polymer
must be dried to remove water. The drying step can be quite energy intensive.
Additionally, CO2 can remove residual monomer from the polymer by super-
critical fluid extraction. Significant energy savings may be realized in a CO2-
based polymerization.
Other advantages of CO2-based polymerizations are that there is no chain
transfer to the solvent, and that the production of unstable end groups can be
dramatically reduced. Guan et al [4] studied the decomposition of 2,2¢-azobis-
(isobutyronitrile) (AIBN) in scCO2. It was found that initiator efficiencies
greater than 80% were possible due to the low viscosity of CO2 and negligible
solvent cage effects. Additionally, analysis of the decomposition products
showed that there was no chain transfer to CO2.
Reactive end groups, such as acyl fluoride and carboxylic acid end groups,
are commonly formed in the aqueous polymerization of fluoroolefins, espe-
cially the copolymerization of TFE with HFP and PPVE [13]. These unstable
end groups may decompose during melt processing of the polymer and cause
defects in the final application. Generally, finishing steps are required after
the polymerization to remove these end groups by hydrolysis or fluorination.
Using CO2 as a polymerization medium eliminates the need for these finishing
steps and additional control of the polymerization is achieved with the absence
of chain transfer to solvent.
The feasibility of using scCO2 in polymerizations has been demonstrated by
the commercial production of TeflonTM in CO2 [29] using methods developed
by DeSimone and Romack [30–35]. Polymerization in carbon dioxide does not
require the use of C8, which has received negative publicity for being persistent
in the blood. DuPont has built a $40 million development facility for the
production of Teflon FEP using CO2-based technologies [5]. Poly(vinylidene
fluoride) (PVDF) is another fluoropolymer that may be manufactured in su-
percritical carbon dioxide. DeSimone and co-workers [36, 37] reported the first
continuous polymerization of vinylidene fluoride in scCO2. Subsequently,
Solvay Corporation submitted a patent application on a process for making
poly(vinylidene fluoride) in scCO2 [38]. As researchers continue to improve
their understanding of CO2-based polymer applications, and as industry con-
tinues to look for more economical and environmentally-sound alternatives to
existing processes, the use of supercritical carbon dioxide in commercial poly-
mer reactions and processing will increase.
338 K. A. Kennedy et al.

4
Precipitation Polymerization in Supercritical Carbon Dioxide

4.1
Introduction

The first reported polymerization of fluoroolefins in carbon dioxide was by

Fukui and coworkers [39, 40]. Tetrafluoroethylene, chlorotrifluoroethylene, and
other fluoroolefins were polymerized in the presence of CO2 using ionizing
radiation [39, 40] and free-radical initiators [40]. DeSimone and coworkers
reported the homogeneous telomerization of tetrafluoroethylene [41] and
vinylidene fluoride [42] in CO2 using AIBN as an initiator. The kinetics of AIBN
decomposition in CO2 is well understood [4]. However, peroxide initiators are
preferred over azo initiators for producing stable endgroups in fluoroolefins
[43]. The peroxy initiators bis(perfluoro-2-N-propoxypropionyl) peroxide
(BPPP) and diethyl peroxydicarbonate (DEPDC) have had the greatest appli-
cation in the heterogeneous polymerization of fluoroolefins in CO2.
DeSimone and coworkers have studied the decomposition kinetics of BPPP
[44] and DEPDC [45] in CO2. The decomposition of DEPDC was studied [45]
by a technique in which the decomposition rate constant, kD, and the initiator
efficiency, f, were determined from a single set of experiments using a con-
tinuous stirred tank reactor (CSTR). It was found that the decomposition was
first order and that the rate constant was independent of the pressure. The
activation energy of the decomposition rate constant in CO2 was consistent
with those in other solvents. It was concluded that the nature of the solvent did
not affect the decomposition kinetics. The decomposition kinetics of BPPP [44]
in CO2 and fluorinated solvents was measured using FTIR. Based on results
from the effect of viscosity on the observed decomposition rate constant, it was
concluded that the decomposition mechanism was a single-bond homolysis.
Additionally, it was concluded that the decomposition rate constants in CO2
and fluorinated solvents should be similar. The studies of BPPP and DEPDC
decomposition in CO2 set the stage for a better understanding of the overall
kinetics of polymerizations in CO2. The next section contains a discussion
of recent research into the polymerization of fluoroolefins in CO2 using BPPP,
DEPDC, and ionizing radiation.

4.2
Polymerization of Tetrafluoroethylene

Several alternative polymerization media have been proposed for reducing the
amount of unstable end groups in poly(tetrafluoroethylene). These include
chlorofluorocarbons, which are detrimental to the environment, perfluoro-
carbons, hydrofluorocarbons, and perfluoroalkyl sulfide acids, which are all
expensive. Supercritical carbon dioxide has been identified as a viable alter-
native to aqueous and fluorocarbon reaction media [31]. Furthermore, mixing
Heterogeneous Polymerization of Fluoroolefins in Supercritical Carbon Dioxide 339

tetrafluoroethylene with carbon dioxide increases the safety of handling the

monomer [46]. Pure tetrafluoroethylene (TFE) is susceptible to autopolymer-
ization and may be explosive without proper handling. Storing and shipping
TFE in a mixture with CO2 makes handling safer.
The polymerization of TFE in CO2 [31–35, 41, 47] has been studied exten-
sively. Polymerizations were typically run at 35 °C using BPPP as an initiator.
This particular initiator was selected to produce stable end groups in the poly-
mer made in CO2. The selection of initiators for polymerization of fluoromono-
mers in CO2 that would produce stable end groups had not been reported pre-
viously. The reaction produced up to a 100% yield and a 160,000 g/mol
molecular weight polymer. This work set the stage for commercialization of a
CO2-based process for making TFE-based polymers by DuPont [5]. DuPont has
started to sell melt-processible polymers with enhanced performance capabil-
ities [29] that are polymerized in a CO2-based process.

4.3
Polymerization of Vinylidene Fluoride

The current means for commercial production of poly(vinylidene fluoride)

generates large amounts of wastewater due to water-based suspension or emul-
sion polymerization. Additionally, large amounts of energy are required to
remove the water from the polymer before downstream processing. The use of
scCO2 as a polymerization medium for vinylidene fluoride has been previously
demonstrated in batch polymerizations [42, 48]. A CO2-based system has the
advantage of eliminating wastewater generation during the polymerization.
Because the polymer is readily separated from CO2, significant energy savings
may be realized. Recently, Charpentier et al have developed a system for poly-
merizing vinyl monomers, including vinylidene fluoride, in scCO2 using a con-
tinuous stirred tank reactor [36]. The technical viability of continuously poly-
merizing vinylidene fluoride in scCO2 was demonstrated with this system.
Furthermore, Solvay Corporation has recently submitted a patent application
for the continuous polymerization of vinylidene fluoride and other halo-
genated monomers in scCO2 [38].
The first demonstration of vinylidene fluoride polymerization in scCO2 was
the homogeneous telomerization of vinylidene fluoride with perfluorobutyl
iodide and AIBN [42]. Several batch studies [48, 49] then were conducted to
identify appropriate initiators for the CO2-based synthesis of high molecular
weight PVDF. Kipp [48] used several different peroxy and azo initiators and
found that the azo initiators, AIBN and 2,2¢-azobis(2,4-dimethyl-methoxy-
pentanitrile), were ineffective in producing significant amounts of PVDF.
Amongst the peroxy initiators studied, Kipp found that BPPP was very effec-
tive as an initiator for vinylidene fluoride. Number-average molecular weights
(Mn) upwards of 24,000 g/mol and conversions of up to 85% were reported.
Brothers [49] found that dimethyl(2,2¢-azobisisobutyrate) was better than other
azo initiators, such as AIBN, for initiating the polymerization of vinylidene
340 K. A. Kennedy et al.

fluoride in CO2. Brothers reported reaction yields of 22.8%, compared to only

2.6% for an AIBN-initiated system.
Galia and coworkers [50] used g-radiation to initiate the polymerization of
VF2 in CO2. Reactions were run at 20–40 °C at pressures less than 25 MPa.
Number-average molecular weights as high as 607,000 g/mol were observed in
polymerizations with initial monomer concentrations ranging from 3.4 to
4.7 mol/L. The polydispersity ranged from 2.5 to 8.8. Conversions were 20–42%.
At the highest monomer concentration studied (6.2 mol/L), a conversion of
73% was observed. Galia concluded from rheological measurements of the
polymers that increasing the monomer concentration increased the degree of
branching and crosslinking in the polymer. The effect of temperature and
system density on the kinetics and polymer properties was also studied. There
was no observable impact of the system density on the kinetics or polymer
properties. However, the polymer properties were very sensitive to the reaction
temperature.
Recently, Charpentier et al. [36] demonstrated the synthesis of PVDF
by a continuous precipitation polymerization in scCO2, using diethyl per-
oxydicarbonate (DEPDC) as an initiator. Low molecular weight polymers
(Mn<20,000 g/mol) were reported in this first work. However, subsequent
research has yielded PVDF with number-average molecular weights up-
wards of 79,000 g/mol [51, 52]. PVDF made in the continuous CO2-based sys-
tem had molecular weights comparable to commercial polymers, but also
exhibited unique properties not observed in PVDF made by conventional
processes.

4.3.1
Continuous Polymerization

In a typical continuous polymerization [36, 51–54], vinylidene fluoride (VF2),

carbon dioxide, and initiator are continuously fed to a continuous stirred tank
reactor (CSTR). The initiator is dissolved in a solvent such as (other solvents
possible) Freon-113 and the ratio of initiator to monomer was typically less
than 0.01 g/g. This initial reaction mixture is homogeneous over the full range
of monomer concentrations evaluated (up to 6 M or 50 wt% VF2 in CO2).As the
reaction proceeds, the polymer precipitates from the fluid phase to form a solid
polymer phase. The effluent exiting the reactor consists of polymer particles
suspended in a fluid mixture of carbon dioxide, unreacted monomer, and un-
reacted initiator. The effluent is passed through a heat exchanger to quench the
reaction. The polymer is collected in one of three filters, depending on whether
the reactor is running at unsteady state or steady state. Finally, after the poly-
mer is separated, the effluent is vented into a hood. Conceivably, the effluent
may be recycled back to the reactor.A schematic of the polymerization system
is shown in Fig. 2. Typical conversions in the reactor are 7–25%. A range of
reaction conditions were studied at 30–80 °C, 130–280 bar, and with average
residence times of 12– 50 minutes.
Heterogeneous Polymerization of Fluoroolefins in Supercritical Carbon Dioxide 341

Fig. 2 Continuous polymerization apparatus: A, CO2 cylinder; B, monomer; C, initiator

solution; D, continuous syringe pumps; E, syringe pumps; F1, steady-state filter; G, thermo-
stated autoclave; H, static mixer; I, chiller/heater unit; J, effluent cooler; K, gas chromato-
graph; V1, V2, four-way valves; V4, three-way valves; V5, V6, two-way valves; V7, heated
control valve [51]

A kinetic model based on homogeneous polymerization was developed to

describe the polymerization in CO2 [51, 54]. A model based on the reaction
scheme in Fig. 3 adequately described the polymerization rates and the poly-
dispersity of the polymer. Monomer inhibition was incorporated into the
model to account for the observed deviation from first-order kinetics. However,
imperfect mixing of the higher viscosity medium is an alternative explana-
tion. It was concluded that termination was by combination, for three reasons.
First, there was no existing literature to support termination by dispropor-
tionation for PVDF. Second, the polydispersity was approximately 1.5 at low
monomer concentrations. Third, NMR studies showed no evidence of un-
saturation.
Chain transfer to the polymer was proposed to account for the broadening
of the molecular weight distribution at high monomer concentrations. How-
ever, as discussed in the next section, the model failed to predict the bimodal
character that is unique to PVDF polymerized in scCO2 at high monomer
concentrations.
342 K. A. Kennedy et al.

Fig. 3 Kinetic scheme to model the continuous polymerization of vinylidene fluoride in

supercritical carbon oxide [51]

4.3.1.1
Polymer Properties

A comparison was made between the properties of PVDF synthesized in CO2

and a commercial PVDF sample. The polymers synthesized in CO2 had mole-
cular weights similar to the commercial polymer and the melt flow indices
(MFIs) were similar [54]. For PVDF synthesized in CO2 with a weight-average
molecular weight of 150,000 g/mol, the MFI was 2.6 g/10 min. The MFI for a
commercial sample with a weight-average molecular weight greater than
195,000 was 1.4 g/10 min.
Unique PVDF properties have been observed after polymerization in super-
critical carbon dioxide [38, 51–53, 55] at certain conditions. The polymer syn-
thesized in supercritical carbon dioxide exhibits a bimodal molecular weight
distribution (MWD), as illustrated in Fig. 4 [52]. At molar VF2 feed concen-
trations less than about 1.9 M, the polymer has a unimodal distribution, at the
conditions of Fig. 4. As the monomer concentration is increased, the distribu-
tion becomes broader and bimodal. Changes in temperature, pressure, and re-
sidence time also have an effect on the MWD [51, 52]. In Fig. 4, t is the average
residence time (the reactor volume divided by the inlet volumetric flow rate).
Several hypotheses were investigated to explain the bimodal character of the
polymer. These include: (1) imperfect mixing, (2) long chain branching, and (3)
heterogeneous polymerization. A heterogeneous polymerization is character-
ized by polymerization in both the fluid and precipitated polymer phases, so
that the phase equilibrium is an important element of kinetics and polymer
properties. It was stated that neither imperfect mixing nor a homogeneous
polymerization with long chain branching due to chain transfer to polymer
could explain the bimodal nature of the CO2-polymerized PVDF [51, 52]. How-
ever, it was recently demonstrated that a combination of chain branching and
heterogeneous polymerization could produce the bimodal MWDs observed
experimentally [56].
Heterogeneous Polymerization of Fluoroolefins in Supercritical Carbon Dioxide 343

Fig. 4 Effect of inlet monomer concentration on molecular distributions. Polymerization

conditions are 75 °C, 4000 psig, t =20–22 min, [I]o=2.8–3.3 mmol/L [52]

4.4
Synthesis of Copolymers

Copolymerizations of TFE [31, 47, 57–61] and VF2 [57, 58, 61, 62] in scCO2 have
been reported. Romack and coworkers [47, 60] did early work to perform
precipitation polymerizations of TFE with HFP and TFE with perfluoropropyl
vinyl ether (PPVE) at 35 °C and moderate pressures (<133 bar).A chain transfer
agent, methanol, was necessary to limit the molecular weight to a range of com-
mercial interest. Hydrogen chloride and hydrogen bromide have also been re-
ported as effective chain transfer agents for these polymerizations in CO2 [58].
Romack and coworkers concluded that the significantly higher molecular
weights produced in CO2 were due to the fact that chain transfer and b-scission
reactions occurred to a lesser extent in scCO2, especially at the lower tempera-
tures used. These undesirable reactions are commonplace in the copolymeriza-
tion of TFE/PPVE conducted at elevated temperatures in conventional solvents.
The limited b-scission reactions were attributed to the enhanced diffusion of
TFE into the polymer phase to improve the rates of radical cross-propagation
with TFE. It was demonstrated that conducting the polymerization in CO2 at
lower temperatures improved the effectiveness of the propagation reaction over
that of the chain transfer reactions to produce high molecular weight polymers.
The research group of Shoichet [57, 59] also has demonstrated that using
CO2 as a polymerization medium can improve the effectiveness of the propaga-
tion reaction over chain transfer reactions in fluoroolefin synthesis. Shoichet’s
344 K. A. Kennedy et al.

group polymerized vinyl acetate with fluoroolefins in CO2 with [59] and with-
out [57] surfactants. The copolymers formed in CO2 were linear, compared to
the highly branched copolymers formed by aqueous polymerization. They con-
cluded that because radical hydrogen abstraction from vinyl acetate was signi-
ficantly reduced in CO2, branching of the polymer was practically eliminated.
There has been research on new, novel fluoroolefin-based copolymers that
benefit from synthesis in scCO2. In 2000,Wheland and Brothers [61] described
novel copolymers of fluoroolefins with maleic anhydride (MAN) and maleic
acid (MAC) for use as coatings and as adhesives or compatibilizing agents for
fluoropolymers. Currently, maleic anhydride must be grafted onto existing
fluoropolymers to achieve the desired properties. Maleic acid does not readily
copolymerize with fluoroolefins in the presence of water. Therefore, polymer-
ization of MAN or MAC with fluoroolefins in CO2 provides a way to synthesize
the desired fluoropolymer directly. There is an interest in using 157 nm photo-
resists in microchip manufacturing to create smaller and smaller features
in circuitry. Zannoni and DeSimone [62] have copolymerized fluoroolefins
with norbornene and a norbornene analog for use as 193 nm and 157 nm pho-
toresists. Ultimately, these photoresists may be synthesized, processed and de-
veloped entirely in CO2, with dramatically reduced solvent use, revolutionizing
the manufacture of microchips.

5
Conclusions

Supercritical carbon dioxide shows promise as a polymerization medium for

the synthesis of fluoroplastics and fluoroelastomers. CO2 offers the benefit of
an environmentally friendly process that reduces the amount of waste gener-
ated and energy required when processing polymers. Additionally, polymer-
ization kinetics and polymer properties may be enhanced by polymerization
in CO2. DuPont has already commercialized a process for making fluoropoly-
mers in CO2. Use of CO2 as a polymerization medium for certain grades of
fluoropolymers eliminates the need for C8, which has received a lot of contro-
versy in recent years. It remains to be seen whether many of the current grades
of fluoropolymers can be synthesized in carbon dioxide. For PTFE, to date only
granular forms of PTFE have been achieved. It is important to find a way to syn-
thesize fine powders of PTFE, which should be feasible through process vari-
able manipulations, and other fluoropolymers in CO2 to eliminate the C8 prob-
lem. As researchers continue to enhance their understanding of CO2-based
polymer applications, and as industry continues to look for more economically
and environmentally-sound alternatives to existing processes, supercritical car-
bon dioxide should become a technology of choice for new and improved com-
mercial polymerization processes.
Heterogeneous Polymerization of Fluoroolefins in Supercritical Carbon Dioxide 345

References
1. Swallow JC (1960) The history of polyethene. In: Renfrew A, Morgan P (eds) Polyethene:
The technology and uses of ethylene polymers. Interscience, New York, p 1
2. Kirby CF, McHugh MA (1999) Chem Rev 99:565
3. Folie B, Radosz M (1995) Ind Eng Chem Res 34:1501
4. Guan Z, Combes JR, Menceloglu YZ, DeSimone JM (1993) Macromolecules 26:2663
5. McCoy M (1999) Chem Eng News 77:10
6. Canelas DA, DeSimone JM (1997) Adv Polym Sci 133:103
7. Cooper AI (2000) J Mater Chem 10:207
8. Kendall JL, Canelas DA, Young JL, DeSimone JM (1999) Chem Rev 99:543
9. Manivannan G, Sawan SP (1998) The supercritical state. In: McHardy J, Sawan SP (eds)
Supercritical fluid cleaning: fundamentals, technology and applications. Noyes, West-
wood, NJ, p 1
10. EPA (2003) Federal Register: Perfluorooctanoic acid (PFOA) and fluorinated telomers.
US EPA, Washington, DC (http://www.epa.gov/opptintr/pfoa/)
11. Lee J (2003) Chemical might pose health risk to younger women and girls. New York
Times, New York
12. Ring K-L, Kalin T, Kishi A (2002) Fluoropolymers. Chemical economics handbook. SRI
International, Menlo Park, CA
13. Feiring AE (1994) Fluoroplastics. In: Banks RE, Smart BE, Tatlow JC (eds) Organofluorine
chemistry: Principles and commercial applications. Plenum, New York, p 339
14. Gangal SV (1985) Polytetrafluoroethylene, homopolymer of tetrafluoroethylene. In:
Mark HF, Bikales NM, Overberger CG, Menges G (eds) Encyclopedia of polymer science
and engineering, vol 16. Wiley, New York, p 577
15. Tervoort T, Visjager J, Graf B, Smith P (2000) Macromolecules 33:6460
16. Lovinger AJ (1982) Poly(vinylidene fluoride). In: Bassett DC (ed) Developments in crystal-
line polymers, vol 1. Applied Science, London, p 195
17. Dohany JE (1994) Poly(vinylidene fluoride). In: Kroschwitz JI (ed) Kirk-Othmer encyclo-
pedia of chemical technology, vol 11. Wiley, New York, p 694
18. Seiler DA (1997) PVDF in the chemical process industry. In: Scheirs J (ed) Modern fluoro-
polymers: High performance polymers for diverse applications. Wiley, Chichester, UK,
p 487
19. Scheirs J (2001) Fluoropolymers: Technology, markets, and trends. Rapra Technology
Limited, Shawbury, UK
20. Atofina Chemicals (2003) Internet brochure: Kynar PVDF for lithium batteries. Atofina
Chemicals, Philadelphia, PA (http://www.atofinachemicals.com/kynarglobal/kynar-
literature.cfm)
21. Kochervinskii VV, Danilyuk TY, Madorskaya LY (1986) Polym Sci USSR 28:690
22. Dohany JE, Humphrey JS (1985) Vinylidene fluoride polymers. In: Mark HF, Kroschwitz
JI (eds) Encyclopedia of polymer science and engineering, vol 17.Wiley, New York, p 532
23. Jo SM, Lee WS, Oh HJ, Park S, Ahn BS, Park KY (1999) Polymer (Korea) 23:800
24. Wille RA, Burchill MT (1998) WO 9838687
25. Doll WW, Lando JB (1968) J Macromol Sci Phys 2:219
26. Dohany JE (1978) US 4076929
27. Savage PE, Gopalan S, Mizan TI, Martino CJ, Brock EE (1995) AIChE J 41:1723
28. DeSimone JM (2002) Science 297:799
29. DuPont (2002) Press Release: DuPont introduces fluoropolymers made with supercritical
CO2 technology. DuPont, Wilmington, DE
30. DeSimone JM (1996) US 5496901
346 Heterogeneous Polymerization of Fluoroolefins in Supercritical Carbon Dioxide

31. DeSimone JM, Romack T (1997) US 5618894

32. DeSimone JM, Romack T (1997) US 5674957
33. DeSimone JM, Romack T (1999) US 5981673
34. DeSimone JM, Romack T (1999) US 5939501
35. DeSimone JM, Romack T (1999) US 5939502
36. Charpentier PA, Kennedy KA, DeSimone JM, Roberts GW (1999) Macromolecules
32:5973
37. DeSimone JM, Charpentier PA, Roberts GW (2001) WO 0134667
38. Blaude J-M, Chauvier J-M, Martin R, Bienfait C (2002) WO 02055567
39. Fukui K, Kagiya T, Yokota H (1970) JP 45003390
40. Fukui K, Kagiai T, Yokota H (1971) JP 46011031
41. Romack TJ, Combes JR, DeSimone JM (1995) Macromolecules 28:1724
42. Combes JR, Guan Z, DeSimone JM (1994) Macromolecules 27:865
43. Hintzer K, Lohr G (1997) Melt processable tetrafluoroethylene-perfluoropropylvinyl
ether copolymers (PFA). In: Scheirs J (ed) Modern fluoropolymers: High performance
polymers for diverse applications. Wiley, Chichester, UK, p 223
44. Bunyard WC, Kadla JF, DeSimone JM (2001) J Am Chem Soc 123:7199
45. Charpentier PA, DeSimone JM, Roberts GW (2000) Chem Eng Sci 55:5341
46. Bramer DJV, Shiflett MB, Yokozeki A (1994) US 5345013
47. Romack TJ (1997) Polymerization of fluoroolefins in liquid and supercritical carbon
dioxide (free radical). PhD Thesis, University of North Carolina at Chapel Hill, Chapel
Hill, NC
48. Kipp BE (1997) The synthesis of fluoropolymers in carbon dioxide and carbon dioxide/
aqueous systems. PhD Thesis, University of North Carolina at Chapel Hill, Chapel Hill,
NC
49. Brothers PD (2000) US 6103844
50. Galia A, Caputo G, Spadaro G, Filardo G (2002) Ind Eng Chem Res 41:5934
51. Saraf MK, Gerard S,Wojcinski.II LM, Charpentier PA, DeSimone JM, Roberts GW (2002)
Macromolecules 35:7976
52. Saraf MK (2001) Polymerization of vinylidene fluoride in supercritical carbon dioxide:
Molecular weight distribution. Masters Thesis, North Carolina State University, Raleigh,
NC
53. DeSimone JM, Charpentier PA, Roberts GW (2001) WO 0190206
54. Charpentier PA, DeSimone JM, Roberts GW (2000) Ind Eng Chem Res 39:4588
55. Saraf MK,Wojcinski LM, Kennedy KA, Gerard S, Charpentier PA, DeSimone JM, Roberts
GW (2002) Macromolecular Symp 182:119
56. Saraf MK (2002) Personal communication
57. Baradie B, Shoichet MS (2002) Macromolecules 35:3569
58. Farnham WB, Wheland RC (2001) WO 0146275
59. Lousenberg RD, Shoichet MS (2000) Macromolecules 33:1682
60. Romack TJ, DeSimone JM, Treat TA (1995) Macromolecules 28:8429
61. Wheland RC, Brothers PD (2000) US 6107423
62. Zannoni LA, DeSimone JM (2002) Polym Mater Sci Eng 87:197

Received: April 2004

Author Index Volumes 101 – 175
Author Index Volumes 1–100 see Volume 100

de, Abajo, J. and de la Campa, J. G.: Processable Aromatic Polyimides. Vol. 140, pp. 23–60.
Abetz, V. see Förster, S.: Vol. 166, pp. 173–210.
Adolf, D. B. see Ediger, M. D.: Vol. 116, pp. 73–110.
Aharoni, S. M. and Edwards, S. F.: Rigid Polymer Networks. Vol. 118, pp. 1–231.
Albertsson, A.-C. and Varma, I. K.: Aliphatic Polyesters: Synthesis, Properties and Appli-
cations. Vol. 157, pp. 99–138.
Albertsson, A.-C. see Edlund, U.: Vol. 157, pp. 53–98.
Albertsson, A.-C. see Söderqvist Lindblad, M.: Vol. 157, pp. 139–161.
Albertsson, A.-C. see Stridsberg, K. M.: Vol. 157, pp. 27–51.
Albertsson, A.-C. see Al-Malaika, S.: Vol. 169, pp. 177–199.
Al-Malaika, S.: Perspectives in Stabilisation of Polyolefins. Vol. 169, pp. 121–150.
Améduri, B., Boutevin, B. and Gramain, P.: Synthesis of Block Copolymers by Radical
Polymerization and Telomerization. Vol. 127, pp. 87–142.
Améduri, B. and Boutevin, B.: Synthesis and Properties of Fluorinated Telechelic Monodis-
persed Compounds. Vol. 102, pp. 133–170.
Amselem, S. see Domb, A. J.: Vol. 107, pp. 93–142.
Andrady, A. L.: Wavelenght Sensitivity in Polymer Photodegradation. Vol. 128, pp. 47–94.
Andreis, M. and Koenig, J. L.: Application of Nitrogen–15 NMR to Polymers. Vol. 124,
pp. 191–238.
Angiolini, L. see Carlini, C.: Vol. 123, pp. 127–214.
Anjum, N. see Gupta, B.: Vol. 162, pp. 37–63.
Anseth, K. S., Newman, S. M. and Bowman, C. N.: Polymeric Dental Composites: Properties
and Reaction Behavior of Multimethacrylate Dental Restorations. Vol. 122, pp. 177–218.
Antonietti, M. see Cölfen, H.: Vol. 150, pp. 67–187.
Armitage, B. A. see O’Brien, D. F.: Vol. 126, pp. 53–58.
Arndt, M. see Kaminski, W.: Vol. 127, pp. 143–187.
Arnold Jr., F. E. and Arnold, F. E.: Rigid-Rod Polymers and Molecular Composites. Vol. 117,
pp. 257–296.
Arora, M. see Kumar, M. N. V. R.: Vol. 160, pp. 45–118.
Arshady, R.: Polymer Synthesis via Activated Esters: A New Dimension of Creativity in
Macromolecular Chemistry. Vol. 111, pp. 1–42.
Auer, S. and Frenkel, D.: Numerical Simulation of Crystal Nucleation in Colloids. Vol. 173,
pp. 149–208.

Bahar, I., Erman, B. and Monnerie, L.: Effect of Molecular Structure on Local Chain
Dynamics: Analytical Approaches and Computational Methods. Vol. 116, pp. 145–206.
Ballauff, M. see Dingenouts, N.: Vol. 144, pp. 1–48.
Ballauff, M. see Holm, C.: Vol. 166, pp. 1–27.
Ballauff, M. see Rühe, J.: Vol. 165, pp. 79–150.
348 Author Index Volumes 101–175

Baltá-Calleja, F. J., González Arche, A., Ezquerra, T. A., Santa Cruz, C., Batallón, F., Frick, B.
and López Cabarcos, E.: Structure and Properties of Ferroelectric Copolymers of
Poly(vinylidene) Fluoride. Vol. 108, pp. 1–48.
Barnes, M. D. see Otaigbe, J.U.: Vol. 154, pp. 1–86.
Barshtein, G. R. and Sabsai, O. Y.: Compositions with Mineralorganic Fillers. Vol. 101,
pp. 1–28.
Baschnagel, J., Binder, K., Doruker, P., Gusev, A. A., Hahn, O., Kremer, K., Mattice, W. L.,
Müller-Plathe, F., Murat, M., Paul, W., Santos, S., Sutter, U. W. and Tries, V.: Bridging
the Gap Between Atomistic and Coarse-Grained Models of Polymers: Status and
Perspectives. Vol. 152, pp. 41–156.
Batallán, F. see Baltá-Calleja, F. J.: Vol. 108, pp. 1–48.
Batog, A. E., Pet’ko, I.P. and Penczek, P.: Aliphatic-Cycloaliphatic Epoxy Compounds and
Polymers. Vol. 144, pp. 49–114.
Barton, J. see Hunkeler, D.: Vol. 112, pp. 115–134.
Bell, C. L. and Peppas, N. A.: Biomedical Membranes from Hydrogels and Interpolymer
Complexes. Vol. 122, pp. 125–176.
Bellon-Maurel, A. see Calmon-Decriaud, A.: Vol. 135, pp. 207–226.
Bennett, D. E. see O’Brien, D. F.: Vol. 126, pp. 53–84.
Berry, G. C.: Static and Dynamic Light Scattering on Moderately Concentraded Solutions:
Isotropic Solutions of Flexible and Rodlike Chains and Nematic Solutions of Rodlike
Chains. Vol. 114, pp. 233–290.
Bershtein, V. A. and Ryzhov, V. A.: Far Infrared Spectroscopy of Polymers. Vol. 114,
pp. 43–122.
Bhargava R., Wang S.-Q. and Koenig J. L: FTIR Microspectroscopy of Polymeric Systems.
Vol. 163, pp. 137–191.
Biesalski, M.: see Rühe, J.: Vol. 165, pp. 79–150.
Bigg, D. M.: Thermal Conductivity of Heterophase Polymer Compositions. Vol. 119,
pp. 1–30.
Binder, K.: Phase Transitions in Polymer Blends and Block Copolymer Melts: Some Recent
Developments. Vol. 112, pp. 115–134.
Binder, K.: Phase Transitions of Polymer Blends and Block Copolymer Melts in Thin Films.
Vol. 138, pp. 1–90.
Binder, K. see Baschnagel, J.: Vol. 152, pp. 41–156.
Binder, K., Müller, M., Virnau, P. and González MacDowell, L.: Polymer+Solvent Systems:
Phase Diagrams, Interface Free Energies, and Nucleation. Vol. 173, pp. 1–104.
Bird, R. B. see Curtiss, C. F.: Vol. 125, pp. 1–102.
Biswas, M. and Mukherjee, A.: Synthesis and Evaluation of Metal-Containing Polymers.
Vol. 115, pp. 89–124.
Biswas, M. and Sinha Ray, S.: Recent Progress in Synthesis and Evaluation of Polymer-
Montmorillonite Nanocomposites. Vol. 155, pp. 167–221.
Bogdal, D., Penczek, P., Pielichowski, J. and Prociak, A.: Microwave Assisted Synthesis,
Crosslinking, and Processing of Polymeric Materials. Vol. 163, pp. 193–263.
Bohrisch, J., Eisenbach, C.D., Jaeger, W., Mori H., Müller A.H.E., Rehahn, M., Schaller, C.,
Traser, S. and Wittmeyer, P.: New Polyelectrolyte Architectures. Vol. 165, pp. 1–41.
Bolze, J. see Dingenouts, N.: Vol. 144, pp. 1–48.
Bosshard, C.: see Gubler, U.: Vol. 158, pp. 123–190.
Boutevin, B. and Robin, J. J.: Synthesis and Properties of Fluorinated Diols. Vol. 102.
pp. 105–132.
Boutevin, B. see Amédouri, B.: Vol. 102, pp. 133–170.
Boutevin, B. see Améduri, B.: Vol. 127, pp. 87–142.
Bowman, C. N. see Anseth, K. S.: Vol. 122, pp. 177–218.
Author Index Volumes 101–175 349

Boyd, R. H.: Prediction of Polymer Crystal Structures and Properties. Vol. 116, pp. 1–26.
Briber, R. M. see Hedrick, J. L.: Vol. 141, pp. 1–44.
Bronnikov, S. V., Vettegren, V. I. and Frenkel, S. Y.: Kinetics of Deformation and Relaxation
in Highly Oriented Polymers. Vol. 125, pp. 103–146.
Brown, H. R. see Creton, C.: Vol. 156, pp. 53–135.
Bruza, K. J. see Kirchhoff, R. A.: Vol. 117, pp. 1–66.
Budkowski, A.: Interfacial Phenomena in Thin Polymer Films: Phase Coexistence and
Segregation. Vol. 148, pp. 1–112.
Burban, J. H. see Cussler, E. L.: Vol. 110, pp. 67–80.
Burchard,W.: Solution Properties of Branched Macromolecules. Vol. 143, pp. 113–194.
Butté, A. see Schork, F. J.: Vol. 175, pp. 129–255.

Calmon-Decriaud, A., Bellon-Maurel, V., Silvestre, F.: Standard Methods for Testing the
Aerobic Biodegradation of Polymeric Materials.Vol 135, pp. 207–226.
Cameron, N. R. and Sherrington, D. C.: High Internal Phase Emulsions (HIPEs)-Structure,
Properties and Use in Polymer Preparation. Vol. 126, pp. 163–214.
de la Campa, J. G. see de Abajo, J.: Vol. 140, pp. 23–60.
Candau, F. see Hunkeler, D.: Vol. 112, pp. 115–134.
Canelas, D. A. and DeSimone, J. M.: Polymerizations in Liquid and Supercritical Carbon
Dioxide. Vol. 133, pp. 103–140.
Canva, M. and Stegeman, G. I.: Quadratic Parametric Interactions in Organic Waveguides.
Vol. 158, pp. 87–121.
Capek, I.: Kinetics of the Free-Radical Emulsion Polymerization of Vinyl Chloride.Vol. 120,
pp. 135–206.
Capek, I.: Radical Polymerization of Polyoxyethylene Macromonomers in Disperse
Systems. Vol. 145, pp. 1–56.
Capek, I.: Radical Polymerization of Polyoxyethylene Macromonomers in Disperse
Systems. Vol. 146, pp. 1–56.
Capek, I. and Chern, C.-S.: Radical Polymerization in Direct Mini-Emulsion Systems.
Vol. 155, pp. 101–166.
Cappella, B. see Munz, M.: Vol. 164, pp. 87–210.
Carlesso, G. see Prokop, A.: Vol. 160, pp. 119–174.
Carlini, C. and Angiolini, L.: Polymers as Free Radical Photoinitiators. Vol. 123, pp. 127–
214.
Carter, K. R. see Hedrick, J. L.: Vol. 141, pp. 1–44.
Casas-Vazquez, J. see Jou, D.: Vol. 120, pp. 207–266.
Chandrasekhar, V.: Polymer Solid Electrolytes: Synthesis and Structure. Vol. 135, pp.
139–206.
Chang, J. Y. see Han, M. J.: Vol. 153, pp. 1–36.
Chang, T.: Recent Advances in Liquid Chromatography Analysis of Synthetic Polymers.
Vol. 163, pp. 1–60.
Charleux, B. and Faust R.: Synthesis of Branched Polymers by Cationic Polymerization.
Vol. 142, pp. 1–70.
Chen, P. see Jaffe, M.: Vol. 117, pp. 297–328.
Chern, C.-S. see Capek, I.: Vol. 155, pp. 101–166.
Chevolot, Y. see Mathieu, H. J.: Vol. 162, pp. 1–35.
Choe, E.-W. see Jaffe, M.: Vol. 117, pp. 297–328.
Chow, P. Y. and Gan, L. M.: Microemulsion Polymerizations and Reactions. Vol. 175,
pp. 257–298.
Chow, T. S.: Glassy State Relaxation and Deformation in Polymers. Vol. 103, pp. 149–190.
Chujo, Y. see Uemura, T.: Vol. 167, pp. 81–106.
350 Author Index Volumes 101–175

Chung, S.-J. see Lin, T.-C.: Vol. 161, pp. 157–193.

Chung, T.-S. see Jaffe, M.: Vol. 117, pp. 297–328.
Cölfen, H. and Antonietti, M.: Field-Flow Fractionation Techniques for Polymer and
Colloid Analysis. Vol. 150, pp. 67–187.
Colmenero J. see Richter, D.: Vol. 174, in press
Comanita, B. see Roovers, J.: Vol. 142, pp. 179–228.
Connell, J. W. see Hergenrother, P. M.: Vol. 117, pp. 67–110.
Creton, C., Kramer, E. J., Brown, H. R. and Hui, C.-Y.: Adhesion and Fracture of Interfaces
Between Immiscible Polymers: From the Molecular to the Continuum Scale. Vol. 156,
pp. 53–135.
Criado-Sancho, M. see Jou, D.: Vol. 120, pp. 207–266.
Curro, J. G. see Schweizer, K. S.: Vol. 116, pp. 319–378.
Curtiss, C. F. and Bird, R. B.: Statistical Mechanics of Transport Phenomena: Polymeric
Liquid Mixtures. Vol. 125, pp. 1–102.
Cussler, E. L.,Wang, K. L. and Burban, J. H.: Hydrogels as Separation Agents. Vol. 110,
pp. 67–80.

Dalton, L.: Nonlinear Optical Polymeric Materials: From Chromophore Design to Commer-
cial Applications. Vol. 158, pp. 1–86.
Dautzenberg, H. see Holm, C.: Vol. 166, pp. 113–171.
Davidson, J. M. see Prokop, A.: Vol. 160, pp. 119–174.
Desai, S. M. and Singh, R. P.: Surface Modification of Polyethylene. Vol. 169, pp. 231–293.
DeSimone, J. M. see Canelas D. A.: Vol. 133, pp. 103–140.
DeSimone, J. M. see Kennedy, K. A.: Vol. 175, pp. 329–346.
DiMari, S. see Prokop, A.: Vol. 136, pp. 1–52.
Dimonie, M. V. see Hunkeler, D.: Vol. 112, pp. 115–134.
Dingenouts, N., Bolze, J., Pötschke, D. and Ballauf, M.: Analysis of Polymer Latexes by Small-
Angle X-Ray Scattering. Vol. 144, pp. 1–48.
Dodd, L. R. and Theodorou, D. N.: Atomistic Monte Carlo Simulation and Continuum Mean
Field Theory of the Structure and Equation of State Properties of Alkane and Polymer
Melts. Vol. 116, pp. 249–282.
Doelker, E.: Cellulose Derivatives. Vol. 107, pp. 199–266.
Dolden, J. G.: Calculation of a Mesogenic Index with Emphasis Upon LC-Polyimides.
Vol. 141, pp. 189–245.
Domb, A. J., Amselem, S., Shah, J. and Maniar, M.: Polyanhydrides: Synthesis and Char-
acterization. Vol. 107, pp. 93–142.
Domb, A. J. see Kumar, M. N. V. R.: Vol. 160, pp. 45118.
Doruker, P. see Baschnagel, J.: Vol. 152, pp. 41–156.
Dubois, P. see Mecerreyes, D.: Vol. 147, pp. 1–60.
Dubrovskii, S. A. see Kazanskii, K. S.: Vol. 104, pp. 97–134.
Dunkin, I. R. see Steinke, J.: Vol. 123, pp. 81–126.
Dunson, D. L. see McGrath, J. E.: Vol. 140, pp. 61–106.
Dziezok, P. see Rühe, J.: Vol. 165, pp. 79–150.

Eastmond, G. C.: Poly(e-caprolactone) Blends. Vol. 149, pp. 59–223.

Economy, J. and Goranov, K.: Thermotropic Liquid Crystalline Polymers for High Per-
formance Applications. Vol. 117, pp. 221–256.
Ediger, M. D. and Adolf, D. B.: Brownian Dynamics Simulations of Local Polymer Dy-
namics. Vol. 116, pp. 73–110.
Edlund, U. and Albertsson, A.-C.: Degradable Polymer Microspheres for Controlled Drug
Delivery. Vol. 157, pp. 53–98.
Author Index Volumes 101–175 351

Edwards, S. F. see Aharoni, S. M.: Vol. 118, pp. 1–231.

Eisenbach, C. D. see Bohrisch, J.: Vol. 165, pp. 1–41.
Endo, T. see Yagci, Y.: Vol. 127, pp. 59–86.
Engelhardt, H. and Grosche, O.: Capillary Electrophoresis in Polymer Analysis. Vol. 150,
pp. 189–217.
Engelhardt, H. and Martin, H.: Characterization of Synthetic Polyelectrolytes by Capillary
Electrophoretic Methods. Vol. 165, pp. 211–247.
Eriksson, P. see Jacobson, K.: Vol. 169, pp. 151–176.
Erman, B. see Bahar, I.: Vol. 116, pp. 145–206.
Eschner, M. see Spange, S.: Vol. 165, pp. 43–78.
Estel, K. see Spange, S.: Vol. 165, pp. 43–78.
Ewen, B. and Richter, D.: Neutron Spin Echo Investigations on the Segmental Dynamics of
Polymers in Melts, Networks and Solutions. Vol. 134, pp. 1–130.
Ezquerra, T. A. see Baltá-Calleja, F. J.: Vol. 108, pp. 1–48.

Fatkullin, N. see Kimmich, R.: Vol. 170, pp. 1–113.

Faust, R. see Charleux, B.: Vol. 142, pp. 1–70.
Faust, R. see Kwon, Y.: Vol. 167, pp. 107–135.
Fekete, E. see Pukánszky, B.: Vol. 139, pp. 109–154.
Fendler, J. H.: Membrane-Mimetic Approach to Advanced Materials. Vol. 113, pp. 1–209.
Fetters, L. J. see Xu, Z.: Vol. 120, pp. 1–50.
Fontenot, K. see Schork, F. J.: Vol. 175, pp. 129–255.
Förster, S., Abetz, V. and Müller, A. H. E.: Polyelectrolyte Block Copolymer Micelles.Vol. 166,
pp. 173–210.
Förster, S. and Schmidt, M.: Polyelectrolytes in Solution. Vol. 120, pp. 51–134.
Freire, J. J.: Conformational Properties of Branched Polymers: Theory and Simulations.
Vol. 143, pp. 35–112.
Frenkel, S. Y. see Bronnikov, S.V.: Vol. 125, pp. 103–146.
Frick, B. see Baltá-Calleja, F. J.: Vol. 108, pp. 1–48.
Fridman, M. L.: see Terent’eva, J. P.: Vol. 101, pp. 29–64.
Fukui, K. see Otaigbe, J. U.: Vol. 154, pp. 1–86.
Funke, W.: Microgels-Intramolecularly Crosslinked Macromolecules with a Globular Struc-
ture. Vol. 136, pp. 137–232.
Furusho, Y. see Takata, T.: Vol. 171, pp. 1–75.

Galina, H.: Mean-Field Kinetic Modeling of Polymerization: The Smoluchowski Coagula-

tion Equation. Vol. 137, pp. 135–172.
Gan, L. M. see Chow, P. Y.: Vol. 175, pp. 257–298.
Ganesh, K. see Kishore, K.: Vol. 121, pp. 81–122.
Gaw, K. O. and Kakimoto, M.: Polyimide-Epoxy Composites. Vol. 140, pp. 107–136.
Geckeler, K. E. see Rivas, B.: Vol. 102, pp. 171–188.
Geckeler, K. E.: Soluble Polymer Supports for Liquid-Phase Synthesis. Vol. 121, pp. 31–80.
Gedde, U. W. and Mattozzi, A.: Polyethylene Morphology. Vol. 169, pp. 29–73.
Gehrke, S. H.: Synthesis, Equilibrium Swelling, Kinetics Permeability and Applications of
Environmentally Responsive Gels. Vol. 110, pp. 81–144.
de Gennes, P.-G.: Flexible Polymers in Nanopores. Vol. 138, pp. 91–106.
Georgiou, S.: Laser Cleaning Methodologies of Polymer Substrates. Vol. 168, pp. 1–49.
Geuss, M. see Munz, M.: Vol. 164, pp. 87–210.
Giannelis, E. P., Krishnamoorti, R. and Manias, E.: Polymer-Silicate Nanocomposites:
Model Systems for Confined Polymers and Polymer Brushes. Vol. 138, pp. 107–148.
Godovsky, D. Y.: Device Applications of Polymer-Nanocomposites. Vol. 153, pp. 163–205.
352 Author Index Volumes 101–175

Godovsky, D. Y.: Electron Behavior and Magnetic Properties Polymer-Nanocomposites.

Vol. 119, pp. 79–122.
González Arche, A. see Baltá-Calleja, F. J.: Vol. 108, pp. 1–48.
Goranov, K. see Economy, J.: Vol. 117, pp. 221–256.
Gramain, P. see Améduri, B.: Vol. 127, pp. 87–142.
Grest, G. S.: Normal and Shear Forces Between Polymer Brushes. Vol. 138, pp. 149–184.
Grigorescu, G. and Kulicke, W.-M.: Prediction of Viscoelastic Properties and Shear Stability
of Polymers in Solution. Vol. 152, p. 1–40.
Gröhn, F. see Rühe, J.: Vol. 165, pp. 79–150.
Grosberg, A. and Nechaev, S.: Polymer Topology. Vol. 106, pp. 1–30.
Grosche, O. see Engelhardt, H.: Vol. 150, pp. 189–217.
Grubbs, R., Risse, W. and Novac, B.: The Development of Well-defined Catalysts for Ring-
Opening Olefin Metathesis. Vol. 102, pp. 47–72.
Gubler, U. and Bosshard, C.: Molecular Design for Third-Order Nonlinear Optics. Vol. 158,
pp. 123–190.
van Gunsteren, W. F. see Gusev, A. A.: Vol. 116, pp. 207–248.
Gupta, B., Anjum, N.: Plasma and Radiation-Induced Graft Modification of Polymers for
Biomedical Applications. Vol. 162, pp. 37–63.
Gusev, A. A., Müller-Plathe, F., van Gunsteren, W. F. and Suter, U. W.: Dynamics of Small
Molecules in Bulk Polymers. Vol. 116, pp. 207–248.
Gusev, A. A. see Baschnagel, J.: Vol. 152, pp. 41–156.
Guillot, J. see Hunkeler, D.: Vol. 112, pp. 115–134.
Guyot, A. and Tauer, K.: Reactive Surfactants in Emulsion Polymerization. Vol. 111,
pp. 43–66.

Hadjichristidis, N., Pispas, S., Pitsikalis, M., Iatrou, H. and Vlahos, C.: Asymmetric Star Poly-
mers Synthesis and Properties. Vol. 142, pp. 71–128.
Hadjichristidis, N. see Xu, Z.: Vol. 120, pp. 1–50.
Hadjichristidis, N. see Pitsikalis, M.: Vol. 135, pp. 1–138.
Hahn, O. see Baschnagel, J.: Vol. 152, pp. 41–156.
Hakkarainen, M.: Aliphatic Polyesters: Abiotic and Biotic Degradation and Degradation
Products. Vol. 157, pp. 1–26.
Hakkarainen, M. and Albertsson, A.-C.: Environmental Degradation of Polyethylene. Vol.
169, pp. 177–199.
Hall, H. K. see Penelle, J.: Vol. 102, pp. 73–104.
Hamley, I.W.: Crystallization in Block Copolymers. Vol. 148, pp. 113–138.
Hammouda, B.: SANS from Homogeneous Polymer Mixtures: A Unified Overview.Vol. 106,
pp. 87–134.
Han, M. J. and Chang, J. Y.: Polynucleotide Analogues. Vol. 153, pp. 1–36.
Harada, A.: Design and Construction of Supramolecular Architectures Consisting of
Cyclodextrins and Polymers. Vol. 133, pp. 141–192.
Haralson, M. A. see Prokop, A.: Vol. 136, pp. 1–52.
Hassan, C. M. and Peppas, N. A.: Structure and Applications of Poly(vinyl alcohol) Hydro-
gels Produced by Conventional Crosslinking or by Freezing/Thawing Methods.Vol. 153,
pp. 37–65.
Hawker, C. J.: Dentritic and Hyperbranched Macromolecules Precisely Controlled Macro-
molecular Architectures. Vol. 147, pp. 113–160.
Hawker, C. J. see Hedrick, J. L.: Vol. 141, pp. 1–44.
He, G. S. see Lin, T.-C.: Vol. 161, pp. 157–193.
Hedrick, J. L., Carter, K. R., Labadie, J. W., Miller, R. D., Volksen, W., Hawker, C. J., Yoon, D.Y.,
Russell, T. P., McGrath, J. E. and Briber, R. M.: Nanoporous Polyimides. Vol. 141, pp. 1–44.
Author Index Volumes 101–175 353

Hedrick, J. L., Labadie, J. W., Volksen, W. and Hilborn, J. G.: Nanoscopically Engineered Poly-
imides. Vol. 147, pp. 61–112.
Hedrick, J. L. see Hergenrother, P. M.: Vol. 117, pp. 67–110.
Hedrick, J. L. see Kiefer, J.: Vol. 147, pp. 161–247.
Hedrick, J. L. see McGrath, J. E.: Vol. 140, pp. 61–106.
Heine, D. R., Grest, G. S. and Curro, J. G.: Structure of Polymer Melts and Blends: Comparison
of Integral Equation theory and Computer Sumulation. Vol. 173, pp. 209–249.
Heinrich, G. and Klüppel, M.: Recent Advances in the Theory of Filler Networking in
Elastomers. Vol. 160, pp. 1–44.
Heller, J.: Poly (Ortho Esters). Vol. 107, pp. 41–92.
Helm, C. A.: see Möhwald, H.: Vol. 165, pp. 151–175.
Hemielec, A. A. see Hunkeler, D.: Vol. 112, pp. 115–134.
Hergenrother, P. M., Connell, J. W., Labadie, J. W. and Hedrick, J. L.: Poly(arylene ether)s
Containing Heterocyclic Units. Vol. 117, pp. 67–110.
Hernández-Barajas, J. see Wandrey, C.: Vol. 145, pp. 123–182.
Hervet, H. see Léger, L.: Vol. 138, pp. 185–226.
Hilborn, J. G. see Hedrick, J. L.: Vol. 147, pp. 61–112.
Hilborn, J. G. see Kiefer, J.: Vol. 147, pp. 161–247.
Hiramatsu, N. see Matsushige, M.: Vol. 125, pp. 147–186.
Hirasa, O. see Suzuki, M.: Vol. 110, pp. 241–262.
Hirotsu, S.: Coexistence of Phases and the Nature of First-Order Transition in Poly-N-iso-
propylacrylamide Gels. Vol. 110, pp. 1–26.
Höcker, H. see Klee, D.: Vol. 149, pp. 1–57.
Holm, C., Hofmann, T., Joanny, J. F., Kremer, K., Netz, R. R., Reineker, P., Seidel, C., Vilgis, T. A.
and Winkler, R. G.: Polyelectrolyte Theory. Vol. 166, pp. 67–111.
Holm, C., Rehahn, M., Oppermann, W. and Ballauff, M.: Stiff-Chain Polyelectrolytes. Vol.
166, pp. 1–27.
Hornsby, P.: Rheology, Compounding and Processing of Filled Thermoplastics. Vol. 139,
pp. 155–216.
Houbenov, N. see Rühe, J.: Vol. 165, pp. 79–150.
Huber, K. see Volk, N.: Vol. 166, pp. 29–65.
Hugenberg, N. see Rühe, J.: Vol. 165, pp. 79–150.
Hui, C.-Y. see Creton, C.: Vol. 156, pp. 53–135.
Hult, A., Johansson, M. and Malmström, E.: Hyperbranched Polymers. Vol. 143, pp. 1–34.
Hünenberger, P. H.: Thermostat Algorithms for Molecular-Dynamics Simulations. Vol. 173,
pp. 105–147.
Hunkeler, D., Candau, F., Pichot, C., Hemielec, A. E., Xie, T. Y., Barton, J., Vaskova, V., Guillot, J.,
Dimonie, M. V. and Reichert, K. H.: Heterophase Polymerization: A Physical and Kinetic
Comparision and Categorization. Vol. 112, pp. 115–134.
Hunkeler, D. see Macko, T.: Vol. 163, pp. 61–136.
Hunkeler, D. see Prokop, A.: Vol. 136, pp. 1–52; 53–74.
Hunkeler, D. see Wandrey, C.: Vol. 145, pp. 123–182.

Iatrou, H. see Hadjichristidis, N.: Vol. 142, pp. 71–128.

Ichikawa, T. see Yoshida, H.: Vol. 105, pp. 3–36.
Ihara, E. see Yasuda, H.: Vol. 133, pp. 53–102.
Ikada, Y. see Uyama,Y.: Vol. 137, pp. 1–40.
Ikehara, T. see Jinnuai, H.: Vol. 170, pp. 115–167.
Ilavsky, M.: Effect on Phase Transition on Swelling and Mechanical Behavior of Synthetic
Hydrogels. Vol. 109, pp. 173–206.
Imai, Y.: Rapid Synthesis of Polyimides from Nylon-Salt Monomers. Vol. 140, pp. 1–23.
354 Author Index Volumes 101–175

Inomata, H. see Saito, S.: Vol. 106, pp. 207–232.

Inoue, S. see Sugimoto, H.: Vol. 146, pp. 39–120.
Irie, M.: Stimuli-Responsive Poly(N-isopropylacrylamide), Photo- and Chemical-Induced
Phase Transitions. Vol. 110, pp. 49–66.
Ise, N. see Matsuoka, H.: Vol. 114, pp. 187–232.
Ito, H.: Chemical Amplification Resists for Microlithography. Vol. 172, pp. 37–245.
Ito, K., Kawaguchi, S.: Poly(macronomers), Homo- and Copolymerization. Vol. 142, pp.
129–178.
Ito, K. see Kawaguchi, S.: Vol. 175, pp. 299–328.
Ito, Y. see Suginome, M.: Vol. 171, pp. 77–136.
Ivanov, A. E. see Zubov, V. P.: Vol. 104, pp. 135–176.

Jacob, S. and Kennedy, J.: Synthesis, Characterization and Properties of OCTA-ARM

Polyisobutylene-Based Star Polymers. Vol. 146, pp. 1–38.
Jacobson, K., Eriksson, P., Reitberger, T. and Stenberg, B.: Chemiluminescence as a Tool for
Polyolefin. Vol. 169, pp. 151–176.
Jaeger, W. see Bohrisch, J.: Vol. 165, pp. 1–41.
Jaffe, M., Chen, P., Choe, E.-W., Chung, T.-S. and Makhija, S.: High Performance Polymer
Blends. Vol. 117, pp. 297–328.
Jancar, J.: Structure-Property Relationships in Thermoplastic Matrices. Vol. 139, pp. 1–66.
Jen, A. K-Y. see Kajzar, F.: Vol. 161, pp. 1–85.
Jerome, R. see Mecerreyes, D.: Vol. 147, pp. 1–60.
Jiang, M., Li, M., Xiang, M. and Zhou, H.: Interpolymer Complexation and Miscibility and
Enhancement by Hydrogen Bonding. Vol. 146, pp. 121–194.
Jin, J. see Shim, H.-K.: Vol. 158, pp. 191–241.
Jinnai, H., Nishikawa, Y., Ikehara, T. and Nishi, T.: Emerging Technologies for the 3D
Analysis of Polymer Structures. Vol. 170, pp. 115–167.
Jo, W. H. and Yang, J. S.: Molecular Simulation Approaches for Multiphase Polymer Systems.
Vol. 156, pp. 1–52.
Joanny, J.-F. see Holm, C.: Vol. 166, pp. 67–111.
Joanny, J.-F. see Thünemann, A. F.: Vol. 166, pp. 113–171.
Johannsmann, D. see Rühe, J.: Vol. 165, pp. 79–150.
Johansson, M. see Hult, A.: Vol. 143, pp. 1–34.
Joos-Müller, B. see Funke, W.: Vol. 136, pp. 137–232.
Jou, D., Casas-Vazquez, J. and Criado-Sancho, M.: Thermodynamics of Polymer Solutions
under Flow: Phase Separation and Polymer Degradation. Vol. 120, pp. 207–266.

Kaetsu, I.: Radiation Synthesis of Polymeric Materials for Biomedical and Biochemical
Applications. Vol. 105, pp. 81–98.
Kaji, K. see Kanaya, T.: Vol. 154, pp. 87–141.
Kajzar, F., Lee, K.-S. and Jen, A. K.-Y.: Polymeric Materials and their Orientation Techniques
for Second-Order Nonlinear Optics. Vol. 161, pp. 1–85.
Kakimoto, M. see Gaw, K. O.: Vol. 140, pp. 107–136.
Kaminski, W. and Arndt, M.: Metallocenes for Polymer Catalysis. Vol. 127, pp. 143–187.
Kammer, H. W., Kressler, H. and Kummerloewe, C.: Phase Behavior of Polymer Blends –
Effects of Thermodynamics and Rheology. Vol. 106, pp. 31–86.
Kanaya, T. and Kaji, K.: Dynamcis in the Glassy State and Near the Glass Transition of
Amorphous Polymers as Studied by Neutron Scattering. Vol. 154, pp. 87–141.
Kandyrin, L. B. and Kuleznev, V. N.: The Dependence of Viscosity on the Composition of
Concentrated Dispersions and the Free Volume Concept of Disperse Systems. Vol. 103,
pp. 103–148.
Author Index Volumes 101–175 355

Kaneko, M. see Ramaraj, R.: Vol. 123, pp. 215–242.

Kang, E. T., Neoh, K. G. and Tan, K. L.: X-Ray Photoelectron Spectroscopic Studies of
Electroactive Polymers. Vol. 106, pp. 135–190.
Karlsson, S. see Söderqvist Lindblad, M.: Vol. 157, pp. 139–161.
Karlsson, S.: Recycled Polyolefins. Material Properties and Means for Quality Determina-
tion. Vol. 169, pp. 201–229.
Kato, K. see Uyama,Y.: Vol. 137, pp. 1–40.
Kautek, W. see Krüger, J.: Vol. 168, pp. 247–290.
Kawaguchi, S. see Ito, K.: Vol. 142, pp. 129–178.
Kawaguchi, S. and Ito, K.: Dispersion Polymerization. Vol. 175, pp. 299–328.
Kawata, S. see Sun, H.-B.: Vol. 170, pp. 169–273.
Kazanskii, K. S. and Dubrovskii, S. A.: Chemistry and Physics of Agricultural Hydrogels.
Vol. 104, pp. 97–134.
Kennedy, J. P. see Jacob, S.: Vol. 146, pp. 1–38.
Kennedy, J. P. see Majoros, I.: Vol. 112, pp. 1–113.
Kennedy, K. A., Roberts, G. W. and DeSimone, J. M.: Heterogeneous Polymerization of
Fluoroolefins in Supercritical Carbon Dioxide. Vol. 175, pp. 329–346.
Khokhlov, A., Starodybtzev, S. and Vasilevskaya, V.: Conformational Transitions of Polymer
Gels: Theory and Experiment. Vol. 109, pp. 121–172.
Kiefer, J., Hedrick J. L. and Hiborn, J. G.: Macroporous Thermosets by Chemically Induced
Phase Separation. Vol. 147, pp. 161–247.
Kihara, N. see Takata, T.: Vol. 171, pp. 1–75.
Kilian, H. G. and Pieper, T.: Packing of Chain Segments. A Method for Describing X-Ray
Patterns of Crystalline, Liquid Crystalline and Non-Crystalline Polymers. Vol. 108,
pp. 49–90.
Kim, J. see Quirk, R. P.: Vol. 153, pp. 67–162.
Kim, K.-S. see Lin, T.-C.: Vol. 161, pp. 157–193.
Kimmich, R. and Fatkullin, N.: Polymer Chain Dynamics and NMR. Vol. 170, pp. 1–113.
Kippelen, B. and Peyghambarian, N.: Photorefractive Polymers and their Applications.
Vol. 161, pp. 87–156.
Kirchhoff, R. A. and Bruza, K. J.: Polymers from Benzocyclobutenes. Vol. 117, pp. 1–66.
Kishore, K. and Ganesh, K.: Polymers Containing Disulfide, Tetrasulfide, Diselenide and
Ditelluride Linkages in the Main Chain. Vol. 121, pp. 81–122.
Kitamaru, R.: Phase Structure of Polyethylene and Other Crystalline Polymers by Solid-
State 13C/MNR. Vol. 137, pp 41–102.
Klee, D. and Höcker, H.: Polymers for Biomedical Applications: Improvement of the
Interface Compatibility. Vol. 149, pp. 1–57.
Klier, J. see Scranton, A. B.: Vol. 122, pp. 1–54.
v. Klitzing, R. and Tieke, B.: Polyelectrolyte Membranes. Vol. 165, pp. 177–210.
Klüppel, M.: The Role of Disorder in Filler Reinforcement of Elastomers on Various Length
Scales. Vol. 164, pp. 1–86.
Klüppel, M. see Heinrich, G.: Vol. 160, pp 1–44.
Knuuttila, H., Lehtinen, A. and Nummila-Pakarinen, A.: Advanced Polyethylene Techno-
logies – Controlled Material Properties. Vol. 169, pp. 13–27.
Kobayashi, S., Shoda, S. and Uyama, H.: Enzymatic Polymerization and Oligomerization.
Vol. 121, pp. 1–30.
Köhler, W. and Schäfer, R.: Polymer Analysis by Thermal-Diffusion Forced Rayleigh
Scattering. Vol. 151, pp. 1–59.
Koenig, J. L. see Bhargava, R.: Vol. 163, pp. 137–191.
Koenig, J. L. see Andreis, M.: Vol. 124, pp. 191–238.
Koike, T.: Viscoelastic Behavior of Epoxy Resins Before Crosslinking. Vol. 148, pp. 139–188.
356 Author Index Volumes 101–175

Kokko, E. see Löfgren, B.: Vol. 169, pp. 1–12.

Kokufuta, E.: Novel Applications for Stimulus-Sensitive Polymer Gels in the Preparation of
Functional Immobilized Biocatalysts. Vol. 110, pp. 157–178.
Konno, M. see Saito, S.: Vol. 109, pp. 207–232.
Konradi, R. see Rühe, J.: Vol. 165, pp. 79–150.
Kopecek, J. see Putnam, D.: Vol. 122, pp. 55–124.
Koßmehl, G. see Schopf, G.: Vol. 129, pp. 1–145.
Kozlov, E. see Prokop, A.: Vol. 160, pp. 119–174.
Kramer, E. J. see Creton, C.: Vol. 156, pp. 53–135.
Kremer, K. see Baschnagel, J.: Vol. 152, pp. 41–156.
Kremer, K. see Holm, C.: Vol. 166, pp. 67–111.
Kressler, J. see Kammer, H. W.: Vol. 106, pp. 31–86.
Kricheldorf, H. R.: Liquid-Cristalline Polyimides. Vol. 141, pp. 83–188.
Krishnamoorti, R. see Giannelis, E. P.: Vol. 138, pp. 107–148.
Krüger, J. and Kautek, W.: Ultrashort Pulse Laser Interaction with Dielectrics and Poly-
mers, Vol. 168, pp. 247–290.
Kuchanov, S. I.: Modern Aspects of Quantitative Theory of Free-Radical Copolymerization.
Vol. 103, pp. 1–102.
Kuchanov, S. I.: Principles of Quantitive Description of Chemical Structure of Synthetic
Polymers. Vol. 152, p. 157–202.
Kudaibergennow, S. E.: Recent Advances in Studying of Synthetic Polyampholytes in
Solutions. Vol. 144, pp. 115–198.
Kuleznev, V. N. see Kandyrin, L. B.: Vol. 103, pp. 103–148.
Kulichkhin, S. G. see Malkin, A. Y.: Vol. 101, pp. 217–258.
Kulicke, W.-M. see Grigorescu, G.: Vol. 152, p. 1–40.
Kumar, M. N. V. R., Kumar, N., Domb, A. J. and Arora, M.: Pharmaceutical Polymeric
Controlled Drug Delivery Systems. Vol. 160, pp. 45–118.
Kumar, N. see Kumar M. N. V. R.: Vol. 160, pp. 45–118.
Kummerloewe, C. see Kammer, H. W.: Vol. 106, pp. 31–86.
Kuznetsova, N. P. see Samsonov, G. V.: Vol. 104, pp. 1–50.
Kwon, Y. and Faust, R.: Synthesis of Polyisobutylene-Based Block Copolymers with
Precisely Controlled Architecture by Living Cationic Polymerization. Vol. 167, pp.
107–135.

Labadie, J. W. see Hergenrother, P. M.: Vol. 117, pp. 67–110.

Labadie, J. W. see Hedrick, J. L.: Vol. 141, pp. 1–44.
Labadie, J. W. see Hedrick, J. L.: Vol. 147, pp. 61–112.
Lamparski, H. G. see O’Brien, D. F.: Vol. 126, pp. 53–84.
Laschewsky, A.: Molecular Concepts, Self-Organisation and Properties of Polysoaps.
Vol. 124, pp. 1–86.
Laso, M. see Leontidis, E.: Vol. 116, pp. 283–318.
Lazár, M. and Rychl, R.: Oxidation of Hydrocarbon Polymers. Vol. 102, pp. 189–222.
Lechowicz, J. see Galina, H.: Vol. 137, pp. 135–172.
Léger, L., Raphaël, E. and Hervet, H.: Surface-Anchored Polymer Chains: Their Role in
Adhesion and Friction. Vol. 138, pp. 185–226.
Lenz, R. W.: Biodegradable Polymers. Vol. 107, pp. 1–40.
Leontidis, E., de Pablo, J. J., Laso, M. and Suter, U. W.: A Critical Evaluation of Novel
Algorithms for the Off-Lattice Monte Carlo Simulation of Condensed Polymer Phases.
Vol. 116, pp. 283–318.
Lee, B. see Quirk, R. P.: Vol. 153, pp. 67–162.
Lee, K.-S. see Kajzar, F.: Vol. 161, pp. 1–85.
Author Index Volumes 101–175 357

Lee, Y. see Quirk, R. P: Vol. 153, pp. 67–162.

Lehtinen, A. see Knuuttila, H.: Vol. 169, pp. 13–27.
Leónard, D. see Mathieu, H. J.: Vol. 162, pp. 1–35.
Lesec, J. see Viovy, J.-L.: Vol. 114, pp. 1–42.
Li, M. see Jiang, M.: Vol. 146, pp. 121–194.
Liang, G. L. see Sumpter, B. G.: Vol. 116, pp. 27–72.
Lienert, K.-W.: Poly(ester-imide)s for Industrial Use. Vol. 141, pp. 45–82.
Lin, J. and Sherrington, D. C.: Recent Developments in the Synthesis, Thermo-
stability and Liquid Crystal Properties of Aromatic Polyamides. Vol. 111, pp. 177–
220.
Lin, T.-C., Chung, S.-J., Kim, K.-S., Wang, X., He, G. S., Swiatkiewicz, J., Pudavar, H. E. and
Prasad, P. N.: Organics and Polymers with High Two-Photon Activities and their
Applications. Vol. 161, pp. 157–193.
Lippert, T.: Laser Application of Polymers. Vol. 168, pp. 51–246.
Liu, Y. see Söderqvist Lindblad, M.: Vol. 157, pp. 139–161.
López Cabarcos, E. see Baltá-Calleja, F. J.: Vol. 108, pp. 1–48.
Löfgren, B., Kokko, E. and Seppälä, J.: Specific Structures Enabled by Metallocene Catalysis
in Polyethenes. Vol. 169, pp. 1–12.
Löwen, H. see Thünemann, A. F.: Vol. 166, pp. 113–171.
Luo, Y. see Schork, F. J.: Vol. 175, pp. 129–255.

Macko, T. and Hunkeler, D.: Liquid Chromatography under Critical and Limiting Condi-
tions: A Survey of Experimental Systems for Synthetic Polymers. Vol. 163, pp. 61–136.
Majoros, I., Nagy, A. and Kennedy, J. P.: Conventional and Living Carbocationic Polymeriza-
tions United. I. A Comprehensive Model and New Diagnostic Method to Probe the
Mechanism of Homopolymerizations. Vol. 112, pp. 1–113.
Makhija, S. see Jaffe, M.: Vol. 117, pp. 297–328.
Malmström, E. see Hult, A.: Vol. 143, pp. 1–34.
Malkin, A. Y. and Kulichkhin, S. G.: Rheokinetics of Curing. Vol. 101, pp. 217–258.
Maniar, M. see Domb, A. J.: Vol. 107, pp. 93–142.
Manias, E. see Giannelis, E. P.: Vol. 138, pp. 107–148.
Martin, H. see Engelhardt, H.: Vol. 165, pp. 211–247.
Marty, J. D. and Mauzac, M.: Molecular Imprinting: State of the Art and Perspectives.
Vol. 172, pp. 1–35.
Mashima, K., Nakayama, Y. and Nakamura, A.: Recent Trends in Polymerization of a-Olefins
Catalyzed by Organometallic Complexes of Early Transition Metals. Vol. 133, pp. 1–52.
Mathew, D. see Reghunadhan Nair, C.P.: Vol. 155, pp. 1–99.
Mathieu, H. J., Chevolot, Y, Ruiz-Taylor, L. and Leónard, D.: Engineering and Character-
ization of Polymer Surfaces for Biomedical Applications. Vol. 162, pp. 1–35.
Matsumoto, A.: Free-Radical Crosslinking Polymerization and Copolymerization of
Multivinyl Compounds. Vol. 123, pp. 41–80.
Matsumoto, A. see Otsu, T.: Vol. 136, pp. 75–138.
Matsuoka, H. and Ise, N.: Small-Angle and Ultra-Small Angle Scattering Study of the
Ordered Structure in Polyelectrolyte Solutions and Colloidal Dispersions. Vol. 114,
pp. 187–232.
Matsushige, K., Hiramatsu, N. and Okabe, H.: Ultrasonic Spectroscopy for Polymeric
Materials. Vol. 125, pp. 147–186.
Mattice, W. L. see Rehahn, M.: Vol. 131/132, pp. 1–475.
Mattice, W. L. see Baschnagel, J.: Vol. 152, pp. 41–156.
Mattozzi, A. see Gedde, U. W.: Vol. 169, pp. 29–73.
Mauzac, M. see Marty, J. D.: Vol. 172, pp. 1–35.
358 Author Index Volumes 101–175

Mays, W. see Xu, Z.: Vol. 120, pp. 1–50.

Mays, J. W. see Pitsikalis, M.: Vol. 135, pp. 1–138.
McGrath, J. E. see Hedrick, J. L.: Vol. 141, pp. 1–44.
McGrath, J. E., Dunson, D. L. and Hedrick, J. L.: Synthesis and Characterization of Segment-
ed Polyimide-Polyorganosiloxane Copolymers. Vol. 140, pp. 61–106.
McLeish, T. C. B. and Milner, S. T.: Entangled Dynamics and Melt Flow of Branched
Polymers. Vol. 143, pp. 195–256.
Mecerreyes, D., Dubois, P. and Jerome, R.: Novel Macromolecular Architectures Based on
Aliphatic Polyesters: Relevance of the Coordination-Insertion Ring-Opening Polymeri-
zation. Vol. 147, pp. 1–60.
Mecham, S. J. see McGrath, J. E.: Vol. 140, pp. 61–106.
Menzel, H. see Möhwald, H.: Vol. 165, pp. 151–175.
Meyer, T. see Spange, S.: Vol. 165, pp. 43–78.
Mikos, A. G. see Thomson, R. C.: Vol. 122, pp. 245–274.
Milner, S. T. see McLeish, T. C. B.: Vol. 143, pp. 195–256.
Mison, P. and Sillion, B.: Thermosetting Oligomers Containing Maleimides and Nadiimides
End-Groups. Vol. 140, pp. 137–180.
Miyasaka, K.: PVA-Iodine Complexes: Formation, Structure and Properties. Vol. 108.
pp. 91–130.
Miller, R. D. see Hedrick, J. L.: Vol. 141, pp. 1–44.
Minko, S. see Rühe, J.: Vol. 165, pp. 79–150.
Möhwald, H., Menzel, H., Helm, C. A. and Stamm, M.: Lipid and Polyampholyte Monolayers
to Study Polyelectrolyte Interactions and Structure at Interfaces. Vol. 165, pp. 151–175.
Monkenbusch, M. see Richter, D.: Vol. 174, in press.
Monnerie, L. see Bahar, I.: Vol. 116, pp. 145–206.
Mori, H. see Bohrisch, J.: Vol. 165, pp. 1–41.
Morishima, Y.: Photoinduced Electron Transfer in Amphiphilic Polyelectrolyte Systems.
Vol. 104, pp. 51–96.
Morton M. see Quirk, R. P: Vol. 153, pp. 67–162.
Motornov, M. see Rühe, J.: Vol. 165, pp. 79–150.
Mours, M. see Winter, H. H.: Vol. 134, pp. 165–234.
Müllen, K. see Scherf, U.: Vol. 123, pp. 1–40.
Müller, A. H. E. see Bohrisch, J.: Vol. 165, pp. 1–41.
Müller, A. H. E. see Förster, S.: Vol. 166, pp. 173–210.
Müller, M. see Thünemann, A. F.: Vol. 166, pp. 113–171.
Müller-Plathe, F. see Gusev, A. A.: Vol. 116, pp. 207–248.
Müller-Plathe, F. see Baschnagel, J.: Vol. 152, p. 41–156.
Mukerherjee, A. see Biswas, M.: Vol. 115, pp. 89–124.
Munz, M., Cappella, B., Sturm, H., Geuss, M. and Schulz, E.: Materials Contrasts and Nano-
lithography Techniques in Scanning Force Microscopy (SFM) and their Application to
Polymers and Polymer Composites. Vol. 164, pp. 87–210.
Murat, M. see Baschnagel, J.: Vol. 152, p. 41–156.
Mylnikov, V.: Photoconducting Polymers. Vol. 115, pp. 1–88.

Nagy, A. see Majoros, I.: Vol. 112, pp. 1–11.

Naka, K. see Uemura, T.: Vol. 167, pp. 81–106.
Nakamura, A. see Mashima, K.: Vol. 133, pp. 1–52.
Nakayama, Y. see Mashima, K.: Vol. 133, pp. 1–52.
Narasinham, B. and Peppas, N. A.: The Physics of Polymer Dissolution: Modeling Ap-
proaches and Experimental Behavior. Vol. 128, pp. 157–208.
Nechaev, S. see Grosberg, A.: Vol. 106, pp. 1–30.
Author Index Volumes 101–175 359

Neoh, K. G. see Kang, E. T.: Vol. 106, pp. 135–190.

Netz, R.R. see Holm, C.: Vol. 166, pp. 67–111.
Netz, R.R. see Rühe, J.: Vol. 165, pp. 79–150.
Newman, S. M. see Anseth, K. S.: Vol. 122, pp. 177–218.
Nijenhuis, K. te: Thermoreversible Networks. Vol. 130, pp. 1–252.
Ninan, K. N. see Reghunadhan Nair, C.P.: Vol. 155, pp. 1–99.
Nishi, T. see Jinnai, H.: Vol. 170, pp. 115–167.
Nishikawa, Y. see Jinnai, H.: Vol. 170, pp. 115–167.
Noid, D. W. see Otaigbe, J. U.: Vol. 154, pp. 1–86.
Noid, D. W. see Sumpter, B. G.: Vol. 116, pp. 27–72.
Nomura, M., Tobita, H. and Suzuki, K.: Emulsion Polymerization: Kinetic and Mechanistic
Aspects. Vol. 175, pp. 1–128.
Novac, B. see Grubbs, R.: Vol. 102, pp. 47–72.
Novikov, V. V. see Privalko, V. P.: Vol. 119, pp. 31–78.
Nummila-Pakarinen, A. see Knuuttila, H.: Vol. 169, pp. 13–27.

O’Brien, D. F., Armitage, B. A., Bennett, D. E. and Lamparski, H. G.: Polymerization and
Domain Formation in Lipid Assemblies. Vol. 126, pp. 53–84.
Ogasawara, M.: Application of Pulse Radiolysis to the Study of Polymers and Polymeri-
zations. Vol. 105, pp. 37–80.
Okabe, H. see Matsushige, K.: Vol. 125, pp. 147–186.
Okada, M.: Ring-Opening Polymerization of Bicyclic and Spiro Compounds. Reactivities
and Polymerization Mechanisms. Vol. 102, pp. 1–46.
Okano, T.: Molecular Design of Temperature-Responsive Polymers as Intelligent Materials.
Vol. 110, pp. 179–198.
Okay, O. see Funke, W.: Vol. 136, pp. 137–232.
Onuki, A.: Theory of Phase Transition in Polymer Gels. Vol. 109, pp. 63–120.
Oppermann, W. see Holm, C.: Vol. 166, pp. 1–27.
Oppermann, W. see Volk, N.: Vol. 166, pp. 29–65.
Osad’ko, I. S.: Selective Spectroscopy of Chromophore Doped Polymers and Glasses.
Vol. 114, pp. 123–186.
Osakada, K., Takeuchi, D.: Coordination Polymerization of Dienes, Allenes, and Methyl-
enecycloalkanes. Vol. 171, pp. 137–194.
Otaigbe, J. U., Barnes, M. D., Fukui, K., Sumpter, B. G. and Noid, D. W.: Generation, Charac-
terization, and Modeling of Polymer Micro- and Nano-Particles. Vol. 154, pp. 1–86.
Otsu, T. and Matsumoto, A.: Controlled Synthesis of Polymers Using the Iniferter Tech-
nique: Developments in Living Radical Polymerization. Vol. 136, pp. 75–138.

de Pablo, J. J. see Leontidis, E.: Vol. 116, pp. 283–318.

Padias, A. B. see Penelle, J.: Vol. 102, pp. 73–104.
Pascault, J.-P. see Williams, R. J. J.: Vol. 128, pp. 95–156.
Pasch, H.: Analysis of Complex Polymers by Interaction Chromatography. Vol. 128,
pp. 1–46.
Pasch, H.: Hyphenated Techniques in Liquid Chromatography of Polymers. Vol. 150,
pp. 1–66.
Paul, W. see Baschnagel, J.: Vol. 152, p. 41–156.
Penczek, P. see Batog, A. E.: Vol. 144, pp. 49–114.
Penczek, P. see Bogdal, D.: Vol. 163, pp. 193–263.
Penelle, J., Hall, H. K., Padias, A. B. and Tanaka, H.: Captodative Olefins in Polymer
Chemistry. Vol. 102, pp. 73–104.
360 Author Index Volumes 101–175

Peppas, N. A. see Bell, C. L.: Vol. 122, pp. 125–176.

Peppas, N. A. see Hassan, C. M.: Vol. 153, pp. 37–65.
Peppas, N. A. see Narasimhan, B.: Vol. 128, pp. 157–208.
Pet’ko, I. P. see Batog, A. E.: Vol. 144, pp. 49–114.
Pheyghambarian, N. see Kippelen, B.: Vol. 161, pp. 87–156.
Pichot, C. see Hunkeler, D.: Vol. 112, pp. 115–134.
Pielichowski, J. see Bogdal, D.: Vol. 163, pp. 193–263.
Pieper, T. see Kilian, H. G.: Vol. 108, pp. 49–90.
Pispas, S. see Pitsikalis, M.: Vol. 135, pp. 1–138.
Pispas, S. see Hadjichristidis, N.: Vol. 142, pp. 71–128.
Pitsikalis, M., Pispas, S., Mays, J. W. and Hadjichristidis, N.: Nonlinear Block Copolymer
Architectures. Vol. 135, pp. 1–138.
Pitsikalis, M. see Hadjichristidis, N.: Vol. 142, pp. 71–128.
Pleul, D. see Spange, S.: Vol. 165, pp. 43–78.
Plummer, C. J. G.: Microdeformation and Fracture in Bulk Polyolefins. Vol. 169, pp. 75–119.
Pötschke, D. see Dingenouts, N.: Vol 144, pp. 1–48.
Pokrovskii, V. N.: The Mesoscopic Theory of the Slow Relaxation of Linear Macromolecules.
Vol. 154, pp. 143–219.
Pospíśil, J.: Functionalized Oligomers and Polymers as Stabilizers for Conventional
Polymers. Vol. 101, pp. 65–168.
Pospíśil, J.: Aromatic and Heterocyclic Amines in Polymer Stabilization.Vol. 124, pp. 87–190.
Powers, A. C. see Prokop, A.: Vol. 136, pp. 53–74.
Prasad, P. N. see Lin, T.-C.: Vol. 161, pp. 157–193.
Priddy, D. B.: Recent Advances in Styrene Polymerization. Vol. 111, pp. 67–114.
Priddy, D. B.: Thermal Discoloration Chemistry of Styrene-co-Acrylonitrile. Vol. 121,
pp. 123–154.
Privalko, V. P. and Novikov, V. V.: Model Treatments of the Heat Conductivity of Hetero-
geneous Polymers. Vol. 119, pp 31–78.
Prociak, A. see Bogdal, D.: Vol. 163, pp. 193–263.
Prokop, A., Hunkeler, D., DiMari, S., Haralson, M. A. and Wang, T. G.: Water Soluble Poly-
mers for Immunoisolation I: Complex Coacervation and Cytotoxicity.Vol. 136, pp. 1–52.
Prokop, A., Hunkeler, D., Powers, A. C., Whitesell, R. R. and Wang, T. G.: Water Soluble Poly-
mers for Immunoisolation II: Evaluation of Multicomponent Microencapsulation
Systems. Vol. 136, pp. 53–74.
Prokop, A., Kozlov, E., Carlesso, G. and Davidsen, J. M.: Hydrogel-Based Colloidal Polymeric
System for Protein and Drug Delivery: Physical and Chemical Characterization,
Permeability Control and Applications. Vol. 160, pp. 119–174.
Pruitt, L. A.: The Effects of Radiation on the Structural and Mechanical Properties of
Medical Polymers. Vol. 162, pp. 65–95.
Pudavar, H. E. see Lin, T.-C.: Vol. 161, pp. 157–193.
Pukánszky, B. and Fekete, E.: Adhesion and Surface Modification. Vol. 139, pp. 109–154.
Putnam, D. and Kopecek, J.: Polymer Conjugates with Anticancer Acitivity. Vol. 122,
pp. 55–124.

Quirk, R. P., Yoo, T., Lee, Y., M., Kim, J. and Lee, B.: Applications of 1,1-Diphenylethylene
Chemistry in Anionic Synthesis of Polymers with Controlled Structures. Vol. 153,
pp. 67–162.

Ramaraj, R. and Kaneko, M.: Metal Complex in Polymer Membrane as a Model for
Photosynthetic Oxygen Evolving Center. Vol. 123, pp. 215–242.
Rangarajan, B. see Scranton, A. B.: Vol. 122, pp. 1–54.
Author Index Volumes 101–175 361

Ranucci, E. see Söderqvist Lindblad, M.: Vol. 157, pp. 139–161.

Raphaël, E. see Léger, L.: Vol. 138, pp. 185–226.
Reddinger, J. L. and Reynolds, J. R.: Molecular Engineering of p-Conjugated Polymers.
Vol. 145, pp. 57–122.
Reghunadhan Nair, C. P., Mathew, D. and Ninan, K. N.: Cyanate Ester Resins, Recent
Developments. Vol. 155, pp. 1–99.
Reichert, K. H. see Hunkeler, D.: Vol. 112, pp. 115–134.
Rehahn, M., Mattice, W. L. and Suter, U. W.: Rotational Isomeric State Models in Macro-
molecular Systems. Vol. 131/132, pp. 1–475.
Rehahn, M. see Bohrisch, J.: Vol. 165, pp. 1–41.
Rehahn, M. see Holm, C.: Vol. 166, pp. 1–27.
Reineker, P. see Holm, C.: Vol. 166, pp. 67–111.
Reitberger, T. see Jacobson, K.: Vol. 169, pp. 151–176.
Reynolds, J. R. see Reddinger, J. L.: Vol. 145, pp. 57–122.
Richter, D. see Ewen, B.: Vol. 134, pp. 1–130.
Richter, D., Monkenbusch, M. and Colmenero J.: Neutron Spin Echo in Polymer Systems.
Vol. 174, in press
Risse, W. see Grubbs, R.: Vol. 102, pp. 47–72.
Rivas, B. L. and Geckeler, K. E.: Synthesis and Metal Complexation of Poly(ethyleneimine)
and Derivatives. Vol. 102, pp. 171–188.
Roberts, G. W. see Kennedy, K. A.: Vol. 175, pp. 329–346.
Robin, J.J.: The Use of Ozone in the Synthesis of New Polymers and the Modification of
Polymers. Vol. 167, pp. 35–79.
Robin, J. J. see Boutevin, B.: Vol. 102, pp. 105–132.
Roe, R.-J.: MD Simulation Study of Glass Transition and Short Time Dynamics in Polymer
Liquids. Vol. 116, pp. 111–114.
Roovers, J. and Comanita, B.: Dendrimers and Dendrimer-Polymer Hybrids. Vol. 142,
pp. 179–228.
Rothon, R. N.: Mineral Fillers in Thermoplastics: Filler Manufacture and Characterisation.
Vol. 139, pp. 67–108.
Rozenberg, B. A. see Williams, R. J. J.: Vol. 128, pp. 95–156.
Rühe, J., Ballauff, M., Biesalski, M., Dziezok, P., Gröhn, F., Johannsmann, D., Houbenov, N.,
Hugenberg, N., Konradi, R., Minko, S., Motornov, M., Netz, R. R., Schmidt, M., Seidel, C.,
Stamm, M., Stephan, T., Usov, D. and Zhang, H.: Polyelectrolyte Brushes.Vol. 165, pp. 79–150.
Ruckenstein, E.: Concentrated Emulsion Polymerization. Vol. 127, pp. 1–58.
Ruiz-Taylor, L. see Mathieu, H. J.: Vol. 162, pp. 1–35.
Rusanov, A. L.: Novel Bis (Naphtalic Anhydrides) and Their Polyheteroarylenes with
Improved Processability. Vol. 111, pp. 115–176.
Russel, T. P. see Hedrick, J. L.: Vol. 141, pp. 1–44.
Russum, J. P. see Schork, F. J.: Vol. 175, pp. 129–255.
Rychly, J. see Lazár, M.: Vol. 102, pp. 189–222.
Ryner, M. see Stridsberg, K. M.: Vol. 157, pp. 27–51.
Ryzhov, V. A. see Bershtein, V. A.: Vol. 114, pp. 43–122.

Sabsai, O. Y. see Barshtein, G. R.: Vol. 101, pp. 1–28.

Saburov, V. V. see Zubov, V. P.: Vol. 104, pp. 135–176.
Saito, S., Konno, M. and Inomata, H.: Volume Phase Transition of N-Alkylacrylamide Gels.
Vol. 109, pp. 207–232.
Samsonov, G. V. and Kuznetsova, N. P.: Crosslinked Polyelectrolytes in Biology. Vol. 104,
pp. 1–50.
Santa Cruz, C. see Baltá-Calleja, F. J.: Vol. 108, pp. 1–48.
362 Author Index Volumes 101–175

Santos, S. see Baschnagel, J.: Vol. 152, p. 41–156.

Sato, T. and Teramoto, A.: Concentrated Solutions of Liquid-Christalline Polymers.Vol. 126,
pp. 85–162.
Schaller, C. see Bohrisch, J.: Vol. 165, pp. 1–41.
Schäfer R. see Köhler, W.: Vol. 151, pp. 1–59.
Scherf, U. and Müllen, K.: The Synthesis of Ladder Polymers. Vol. 123, pp. 1–40.
Schmidt, M. see Förster, S.: Vol. 120, pp. 51–134.
Schmidt, M. see Rühe, J.: Vol. 165, pp. 79–150.
Schmidt, M. see Volk, N.: Vol. 166, pp. 29–65.
Scholz, M.: Effects of Ion Radiation on Cells and Tissues. Vol. 162, pp. 97–158.
Schopf, G. and Koßmehl, G.: Polythiophenes – Electrically Conductive Polymers. Vol. 129,
pp. 1–145.
Schork, F. J., Luo, Y., Smulders, W., Russum, J. P., Butté, A. and Fontenot, K.: Miniemulsion
Polymerization. Vol. 175, pp. 127–255.
Schulz, E. see Munz, M.: Vol. 164, pp. 97–210.
Seppälä, J. see Löfgren, B.: Vol. 169, pp. 1–12.
Sturm, H. see Munz, M.: Vol. 164, pp. 87–210.
Schweizer, K. S.: Prism Theory of the Structure, Thermodynamics, and Phase Transitions
of Polymer Liquids and Alloys. Vol. 116, pp. 319–378.
Scranton, A. B., Rangarajan, B. and Klier, J.: Biomedical Applications of Polyelectrolytes.
Vol. 122, pp. 1–54.
Sefton, M. V. and Stevenson, W. T. K.: Microencapsulation of Live Animal Cells Using
Polycrylates. Vol. 107, pp. 143–198.
Seidel, C. see Holm, C.: Vol. 166, pp. 67–111.
Seidel, C. see Rühe, J.: Vol. 165, pp. 79–150.
Shamanin, V. V.: Bases of the Axiomatic Theory of Addition Polymerization. Vol. 112,
pp. 135–180.
Sheiko, S. S.: Imaging of Polymers Using Scanning Force Microscopy: From Super-
structures to Individual Molecules. Vol. 151, pp. 61–174.
Sherrington, D. C. see Cameron, N. R.,Vol. 126, pp. 163–214.
Sherrington, D. C. see Lin, J.: Vol. 111, pp. 177–220.
Sherrington, D. C. see Steinke, J.: Vol. 123, pp. 81–126.
Shibayama, M. see Tanaka, T.: Vol. 109, pp. 1–62.
Shiga, T.: Deformation and Viscoelastic Behavior of Polymer Gels in Electric Fields.
Vol. 134, pp. 131–164.
Shim, H.-K. and Jin, J.: Light-Emitting Characteristics of Conjugated Polymers. Vol. 158,
pp. 191–241.
Shoda, S. see Kobayashi, S.: Vol. 121, pp. 1–30.
Siegel, R. A.: Hydrophobic Weak Polyelectrolyte Gels: Studies of Swelling Equilibria and
Kinetics. Vol. 109, pp. 233–268.
Silvestre, F. see Calmon-Decriaud, A.: Vol. 207, pp. 207–226.
Sillion, B. see Mison, P.: Vol. 140, pp. 137–180.
Simon, F. see Spange, S.: Vol. 165, pp. 43–78.
Singh, R. P. see Sivaram, S.: Vol. 101, pp. 169–216.
Singh, R. P. see Desai, S. M.: Vol. 169, pp. 231–293.
Sinha Ray, S. see Biswas, M: Vol. 155, pp. 167–221.
Sivaram, S. and Singh, R. P.: Degradation and Stabilization of Ethylene-Propylene Co-
polymers and Their Blends: A Critical Review. Vol. 101, pp. 169–216.
Smulders, W. see Schork, F. J.: Vol. 175, pp. 129–255.
Söderqvist Lindblad, M., Liu, Y., Albertsson, A.-C., Ranucci, E. and Karlsson, S.: Polymer
from Renewable Resources. Vol. 157, pp. 139–161.
Author Index Volumes 101–175 363

Spange, S., Meyer, T., Voigt, I., Eschner, M., Estel, K., Pleul, D. and Simon, F.: Poly(Vinyl-
formamide-co-Vinylamine)/Inorganic Oxid Hybrid Materials. Vol. 165, pp. 43–78.
Stamm, M. see Möhwald, H.: Vol. 165, pp. 151–175.
Stamm, M. see Rühe, J.: Vol. 165, pp. 79–150.
Starodybtzev, S. see Khokhlov, A.: Vol. 109, pp. 121–172.
Stegeman, G. I. see Canva, M.: Vol. 158, pp. 87–121.
Steinke, J., Sherrington, D. C. and Dunkin, I. R.: Imprinting of Synthetic Polymers Using
Molecular Templates. Vol. 123, pp. 81–126.
Stenberg, B. see Jacobson, K.: Vol. 169, pp. 151–176.
Stenzenberger, H. D.: Addition Polyimides. Vol. 117, pp. 165–220.
Stephan, T. see Rühe, J.: Vol. 165, pp. 79–150.
Stevenson,W. T. K. see Sefton, M. V.: Vol. 107, pp. 143–198.
Stridsberg, K. M., Ryner, M. and Albertsson, A.-C.: Controlled Ring-Opening Polymer-
ization: Polymers with Designed Macromoleculars Architecture. Vol. 157, pp. 27–51.
Sturm, H. see Munz, M.: Vol. 164, pp. 87–210.
Suematsu, K.: Recent Progress of Gel Theory: Ring, Excluded Volume, and Dimension.
Vol. 156, pp. 136–214.
Sugimoto, H. and Inoue, S.: Polymerization by Metalloporphyrin and Related Complexes.
Vol. 146, pp. 39–120.
Suginome, M. and Ito, Y.: Transition Metal-Mediated Polymerization of Isocyanides. Vol.
171, pp. 77–136.
Sumpter, B. G., Noid, D. W., Liang, G. L. and Wunderlich, B.: Atomistic Dynamics of
Macromolecular Crystals. Vol. 116, pp. 27–72.
Sumpter, B. G. see Otaigbe, J. U.: Vol. 154, pp. 1–86.
Sun, H.-B. and Kawata, S.: Two-Photon Photopolymerization and 3D Lithographic Micro-
fabrication. Vol. 170, pp. 169–273.
Suter, U. W. see Gusev, A. A.: Vol. 116, pp. 207–248.
Suter, U. W. see Leontidis, E.: Vol. 116, pp. 283–318.
Suter, U. W. see Rehahn, M.: Vol. 131/132, pp. 1–475.
Suter, U. W. see Baschnagel, J.: Vol. 152, p. 41–156.
Suzuki, A.: Phase Transition in Gels of Sub-Millimeter Size Induced by Interaction with
Stimuli. Vol. 110, pp. 199–240.
Suzuki, A. and Hirasa, O.: An Approach to Artifical Muscle by Polymer Gels due to Micro-
Phase Separation. Vol. 110, pp. 241–262.
Suzuki, K. see Nomura, M.: Vol. 175, pp. 1–128.
Swiatkiewicz, J. see Lin, T.-C.: Vol. 161, pp. 157–193.

Tagawa, S.: Radiation Effects on Ion Beams on Polymers. Vol. 105, pp. 99–116.
Takata, T., Kihara, N. and Furusho, Y.: Polyrotaxanes and Polycatenanes: Recent Advances
in Syntheses and Applications of Polymers Comprising of Interlocked Structures.
Vol. 171, pp. 1–75.
Takeuchi, D. see Osakada, K.: Vol. 171, pp. 137–194.
Tan, K. L. see Kang, E. T.: Vol. 106, pp. 135–190.
Tanaka, H. and Shibayama, M.: Phase Transition and Related Phenomena of Polymer Gels.
Vol. 109, pp. 1–62.
Tanaka, T. see Penelle, J.: Vol. 102, pp. 73–104.
Tauer, K. see Guyot, A.: Vol. 111, pp. 43–66.
Teramoto, A. see Sato, T.: Vol. 126, pp. 85–162.
Terent’eva, J. P. and Fridman, M. L.: Compositions Based on Aminoresins. Vol. 101,
pp. 29–64.
Theodorou, D. N. see Dodd, L. R.: Vol. 116, pp. 249–282.
364 Author Index Volumes 101–175

Thomson, R. C., Wake, M. C., Yaszemski, M. J. and Mikos, A. G.: Biodegradable Polymer
Scaffolds to Regenerate Organs. Vol. 122, pp. 245–274.
Thünemann, A. F., Müller, M., Dautzenberg, H., Joanny, J.-F. and Löwen, H.: Polyelectrolyte
complexes. Vol. 166, pp. 113–171.
Tieke, B. see v. Klitzing, R.: Vol. 165, pp. 177–210.
Tobita, H. see Nomura, M.: Vol. 175, pp. 1–128.
Tokita, M.: Friction Between Polymer Networks of Gels and Solvent. Vol. 110, pp. 27–48.
Traser, S. see Bohrisch, J.: Vol. 165, pp. 1–41.
Tries, V. see Baschnagel, J.: Vol. 152, p. 41–156.
Tsuruta, T.: Contemporary Topics in Polymeric Materials for Biomedical Applications.
Vol. 126, pp. 1–52.

Uemura, T., Naka, K. and Chujo, Y.: Functional Macromolecules with Electron-Donating
Dithiafulvene Unit. Vol. 167, pp. 81–106.
Usov, D. see Rühe, J.: Vol. 165, pp. 79–150.
Uyama, H. see Kobayashi, S.: Vol. 121, pp. 1–30.
Uyama, Y: Surface Modification of Polymers by Grafting. Vol. 137, pp. 1–40.

Varma, I. K. see Albertsson, A.-C.: Vol. 157, pp. 99–138.

Vasilevskaya, V. see Khokhlov, A.: Vol. 109, pp. 121–172.
Vaskova, V. see Hunkeler, D.: Vol.: 112, pp. 115–134.
Verdugo, P.: Polymer Gel Phase Transition in Condensation-Decondensation of Secretory
Products. Vol. 110, pp. 145–156.
Vettegren, V. I. see Bronnikov, S. V.: Vol. 125, pp. 103–146.
Vilgis, T. A. see Holm, C.: Vol. 166, pp. 67–111.
Viovy, J.-L. and Lesec, J.: Separation of Macromolecules in Gels: Permeation Chromato-
graphy and Electrophoresis. Vol. 114, pp. 1–42.
Vlahos, C. see Hadjichristidis, N.: Vol. 142, pp. 71–128.
Voigt, I. see Spange, S.: Vol. 165, pp. 43–78.
Volk, N., Vollmer, D., Schmidt, M., Oppermann, W. and Huber, K.: Conformation and Phase
Diagrams of Flexible Polyelectrolytes. Vol. 166, pp. 29–65.
Volksen, W.: Condensation Polyimides: Synthesis, Solution Behavior, and Imidization
Characteristics. Vol. 117, pp. 111–164.
Volksen, W. see Hedrick, J. L.: Vol. 141, pp. 1–44.
Volksen, W. see Hedrick, J. L.: Vol. 147, pp. 61–112.
Vollmer, D. see Volk, N.: Vol. 166, pp. 29–65.

Wake, M. C. see Thomson, R. C.: Vol. 122, pp. 245–274.

Wandrey C., Hernández-Barajas, J. and Hunkeler, D.: Diallyldimethylammonium Chloride
and its Polymers. Vol. 145, pp. 123–182.
Wang, K. L. see Cussler, E. L.: Vol. 110, pp. 67–80.
Wang, S.-Q.: Molecular Transitions and Dynamics at Polymer/Wall Interfaces: Origins of
Flow Instabilities and Wall Slip. Vol. 138, pp. 227–276.
Wang, S.-Q. see Bhargava, R.: Vol. 163, pp. 137–191.
Wang, T. G. see Prokop, A.: Vol. 136, pp. 1–52; 53–74.
Wang, X. see Lin, T.-C.: Vol. 161, pp. 157–193.
Webster, O. W.: Group Transfer Polymerization: Mechanism and Comparison with Other
Methods of Controlled Polymerization of Acrylic Monomers. Vol. 167, pp. 1–34.
Whitesell, R. R. see Prokop, A.: Vol. 136, pp. 53–74.
Williams, R. J. J., Rozenberg, B. A. and Pascault, J.-P.: Reaction Induced Phase Separation in
Modified Thermosetting Polymers. Vol. 128, pp. 95–156.
Author Index Volumes 101–175 365

Winkler, R. G. see Holm, C.: Vol. 166, pp. 67–111.

Winter, H. H. and Mours, M.: Rheology of Polymers Near Liquid-Solid Transitions.Vol. 134,
pp. 165–234.
Wittmeyer, P. see Bohrisch, J.: Vol. 165, pp. 1–41.
Wu, C.: Laser Light Scattering Characterization of Special Intractable Macromolecules in
Solution. Vol 137, pp. 103–134.
Wunderlich, B. see Sumpter, B. G.: Vol. 116, pp. 27–72.

Xiang, M. see Jiang, M.: Vol. 146, pp. 121–194.

Xie, T. Y. see Hunkeler, D.: Vol. 112, pp. 115–134.
Xu, Z., Hadjichristidis, N., Fetters, L. J. and Mays, J. W.: Structure/Chain-Flexibility Relation-
ships of Polymers. Vol. 120, pp. 1–50.

Yagci, Y. and Endo, T.: N-Benzyl and N-Alkoxy Pyridium Salts as Thermal and Photo-
chemical Initiators for Cationic Polymerization. Vol. 127, pp. 59–86.
Yannas, I. V.: Tissue Regeneration Templates Based on Collagen-Glycosaminoglycan Co-
polymers. Vol. 122, pp. 219–244.
Yang, J. S. see Jo, W. H.: Vol. 156, pp. 1–52.
Yamaoka, H.: Polymer Materials for Fusion Reactors. Vol. 105, pp. 117–144.
Yasuda, H. and Ihara, E.: Rare Earth Metal-Initiated Living Polymerizations of Polar and
Nonpolar Monomers. Vol. 133, pp. 53–102.
Yaszemski, M. J. see Thomson, R. C.: Vol. 122, pp. 245–274.
Yoo, T. see Quirk, R. P.: Vol. 153, pp. 67–162.
Yoon, D. Y. see Hedrick, J. L.: Vol. 141, pp. 1–44.
Yoshida, H. and Ichikawa, T.: Electron Spin Studies of Free Radicals in Irradiated Polymers.
Vol. 105, pp. 3–36.

Zhang, H. see Rühe, J.: Vol. 165, pp. 79–150.

Zhang, Y.: Synchrotron Radiation Direct Photo Etching of Polymers. Vol. 168, pp. 291–340.
Zhou, H. see Jiang, M.: Vol. 146, pp. 121–194.
Zubov, V . P., Ivanov, A. E. and Saburov, V. V.: Polymer-Coated Adsorbents for the Separation
of Biopolymers and Particles. Vol. 104, pp. 135–176.
Subject Index

Absorption efficiency factor 9 – / ZnS nanoparticles, dopes/undoped

Acoustic intensity 65 283
Acrylamide 261 Cetyl alcohol 146
Additives 66 Chain entry 84
Adsorption, terminally-attached 321 Chain transfer 91
Agitation 74 – –, retardive 211, 213
AIBA 46 Chain transfer agents 72, 154, 343
AIBN 338 Chemical potential 239
Alkoxyamine 217 CO2, supercritical 329, 335
Alkyd 209 CO2-philic polymers 305
AOT 260 Coagulation 31
APFO 331 –, shear-induced 81
Ascorbic acid 222 Coagulative nucleation 22, 27
ATRP 223 Coagulum 80
Average rate coefficient 20 Coalescence 161
Coalescing aid 212
Blend 208 Colloidal stability 235
Block copolymers 299 Colloids 131
BPO 220 Competition technique 82
BPPP, decomposition kinetics 338 Conversion, limiting 211
Branching 91, 92 Copolymerization 195, 197
– density 95 Copolymers, amphiphilic 244
– – distribution 96 –, branched 344
–, long-chain 94 –, composition 197, 200
Brownian motion 144 Costabilizer 135, 144, 145
Brush adsorption 321 Couette-Taylor vortex flow reactor,
Bulk polymerization 133 continuous 115
n-Butyl methacrylate 157 Critical micelle concentration 134, 332
Butylacrylate 263 Crosslinking 103
–, oxidative 210
C8 331 Crosslinking density 104
Cage effect 206 Crystallization 245
Calorimetry, reaction 54 CSTR 174, 340
Camphorsulfonic acid 222 Cumen hydroperoxide 58
Carbon black 216 b-Cyclodextrin 79
Carboxylic monomer 45
Carboxymethylcellulose 134 DBMA 46
Catsurf 243 Degenerative transfer 230
CdS 277, 281 DEPDC 338
368 Subject Index

Desorption, radical 85 Hollow-fiber feeding, microemulsion

Desorption rate coefficeint 17 268
Diffusion-controlled entry 8 Homogeneous nucleation 22, 25, 27, 139
Diphenylpicrylhydrazol 178 Homogenization 149
Dispersion polymerization 299 Hybrid miniemulsion polymerization
– –, enzyme-catalyzed 308 208
– –, living 306 Hydrogen abstraction 211, 212
Distribution, most probable 84, 88, 90 Hydroxyapetite 282, 290
–, power-law 101
Dithiocarbonyl 229 Imprinting, molecular 246
Dithioesters 228 Impurities 66
DLVO theory 171 Inden 52
Dodecyl mercaptan 155, 186 Inhibitor 180
Double bonds, pendant 103 Inisurf 242
– –, terminal 91 Initiation, ultrasonic 64
DPPH 180 Initiator type 57
Droplet nucleation 22, 26, 141 Initiators, oil-soluble 57
Droplet stability 159 –, water-soluble 57
Interfacial polymerization 204
Elastomers 331 Interfacial tension 161
Emulsifier 22 Irradiation, ultrasonic 63
–, anionic 33
–, mixed 33 KPS 13
–, nonionic 32 Kraton 214
Emulsion polymerization 1, 134
– –, continuous 1 Lactone 245
Encapsulation 305 Laplace pressure 164
EPA 332 Lauroyl peroxide (LPO) 58, 155
Epoxy latex 214 Lauryl methacrylate 154
Ethylene, supercritical 330 LCST 262
Exhaustive extraction 210 Lead zirconate 289
Extraction 210 Lead zirconate titanate 289
Liquid crystals 152
Ferrites 276, 286 Lithium batteries 333
Ferrofluid 276
Finishing initiator 158 Macroemulsion 135, 160
Fluoroalkyl acrylates 244 Macromonomer 299
Fluoroolefins, heterogeneous poly- MADQUAT 262
merization 329 MAETAC 205
Fremy’s salt 71 Maleic anhydride 344
Fuchs stability ratio 171 Mass-transfer 78
Mayo Lewis equation 195
Gel effect 112, 143 Melt flow indices 342
Glass effect 144 Membranes, functional 258
Glass transition 211 –, ion-conductive 272
Graft copolymers 208, 299, 321 –, proton-exchange 274
Grafting 210 Mercaptan 68
Methanol fuel cells 291
HDQ 70 Micellar nucleation 22, 139
HEMA 261, 269, 303 Microemulsion 135, 160, 257
Hexadecane 146, 184 MicroFluidizer 149
Subject Index 369

Microlatexes 260, 266 Particle growth 2, 36, 42

Microspheres, functional 303 Particle size 91
–, hollow 305 Particles, sterically-stabilized 315
–, thermosensitive 312 PB 52
Microsuspension polymerization 134 PbS 277, 281
Miniemulsion 160 PDVB 305
– polymerization 129, 134 PE 309
MMA 44, 260 PEDB 235
Molecular weight distribution 81 PEPDTA 235
– – –, bimodal 100, 106, 333 Perfluoropropionyl peroxide 335
– – –, instantaneous 83 Perovskites 282, 289
Monomer, cationic 204 Persistent radical effect 218
–, hydrophilic 207 Phase transfer agent 79
Monomer concentartion 47 Photoresists 344
Monomer partitioning 48 Piezoelectric properties 334
Monomer swelling method, dynamic Plug flow reactors 173
(DSM) 302 PMMA 265, 268, 310, 314
Monomer transfer constants 92 PNIPAM 262
Monomer transport 177, 194 Polyaddition 244
Monomer-starvation 200 Polyaniline 245, 307
Monte Carlo (MC) simulation 81 – / BaSO4 293
Morton equation 161 Polycondensation 243
Multiple steady states 176 Polydispersity index 86, 183
Mushroom-brush transition 322 Polyelectrolyte macromonomers 313
Polyester 308
NAD polymerization 308 – resin 213
Nanoencapsulation 215 Polyethylene, low-density (LDPE) 330
Nanoparticles 204, 245 Polymer transfer reactions 93, 97
–, inorganic 280 Polymerization, anionic 242
NMCRP 216 –, cationic 243
NMRP 223 –, dispersed-phase 133
Nonlinear polymers 2 Polypropylene 267
Norbornene 245, 344 Polypyrrole 308
Nucleation 22, 139 Polyurethane 213, 308
–, enhanced 156 Polyvinyl alcohol 134, 150
–, robust 177 Power-law distribution 101
Nucleation theory, homogeneous 315 Prepolymer 214
Number fraction distribution 88, 90 Pre-reactor 113, 117
Promoting effect 69
Oil-in-water microemulsion 258 Propagation-controlled entry 11
Oligomeric control 238 PSD 183
Oligomers 232, 235, 238 Pseudobulk polymerization kinetics 89
OMPU 213 Pseudo-homopolymerization approach
Oscillations 112 20, 42
Ostwald ripening 144, 161 Pseudo-kinetic rate constant method 42
Oxygen 67, 74 PTFE 334
PtRu/C catalysts 291
Pair radicals 58 Pulsed packed column reactor 114
Pancake-to-brush transition 322 Pulsed sieve plate column reactor 114
Particle formation 22 Pulsed tubular reactor 114
– –, secondary 30, 31 PVDF 334, 337
 370 Subject Index

PVF, commercial production 339 Sulfobetaine-based macromonomers 314

PVP 303 Supercritical fluids 335
Superswelling 232, 234
Quasi-steady-state approximation 142 Surfactants, nonionic 150
–, polymeric 14, 35
Radical capture efficiency 9, 23 –, reactive 34, 314
Radical desorption 16 Surfers 270
Radical entry 7, 84 Surfmers 314
Radical segregation 138 Suspension polymerization 133, 333
Radicals, single 58 Swelling, partial 48
RAFT 230 –, two-stage 302
Rate coefficient, average 42
Rate-determining step 77 Tank-in-series model 116
Reaction calorimetry 54 Taylor number 116
Reactivity ratio 195, 318 TBC 70
Reactor, tubular 113 TDBP 91
Regioisomer defects 334 Teflon 331
Residence time distribution (RTD) 114 TEMPO 13, 220
Retardation 210, 211 Termination, bimolecular 85, 88, 89, 138
Retarder 179 –, reversible 224
Ripening time 188 Tetrafluoroethylene 329
ROMP 245 Tetrahydrofurfuryl methacrylate 263
ROP 243, 307 TFE, explosive 339
Ru(dip)3Cl2 274 Titanium dioxide 216
Ru-Cu oxides, silica 288 Transurfs 314
–, thiol-ended 315
SAN 53 TREM LF-40 34
Saturation swelling 50 Trommsdorff-Norrish effect 112
SDBS 220
Seeding reactor 113 Ugelstad plot 37
Segregation 219
Semibatch 200 Vinyl acetate 266, 344
Shear stability 189 Vinyl chloride 244
Slope-and-intercept method 12 Vinyl versatate 194
Smith/Ewart theory 6 4-Vinylbenzene sulfonic acid lithium salt
Sodium 1,4-bis(2-ethylhexyl)sulfosuccinate 272
260 Vinylidene fluoride 330
Sodium nitrite 178 VOC 212
Solution polymerization 133
Sonication 148 Water-in-oil microemulsion 258
Space effects, limited 97 Weight fraction distributions 89, 93
Split-feed operation 117 Winsor systems 260
Stearyl methacrylate 151, 154
Steric hindrance 212 Xanthates 228, 232
Stirred-tank reactor 109
Stirring 74 Zirconia 282, 289
Stokes law creaming 144 ZnS 281

View publication stats