Successful use of RAFT techniques in seeded emulsion polymerization of styrene: Living character, RAFT agent transport, and rate of polymerization

Article abstract of DOI:10.1021/ma011840g

Reversible addition-fragmentation chain transfer (RAFT) polymerization techniques are successfully used to control molecular weight and polydispersity in the seeded emulsion polymerization of styrene. A novel technique was used to assist the transport int

Full text of DOI:10.1021/ma011840g

Macromolecules 2002, 35, 5417-5425
5417
Successful Use of RAFT Techniques in Seeded Emulsion Polymerization
of Styrene: Living Character, RAFT Agent Transport, and Rate of
Polymerization
Stu a r t W. P r escott,†,‡,§ Ma th ew J . Ba lla r d ,‡ Ezio Rizza r d o,‡,§ a n d
Rober t G. Gilber t*,†
Key Centre for Polymer Colloids, School of Chemistry, Sydney University, NSW 2006, Australia;
CSIRO Molecular Science, Private Bag 10, Clayton South VIC 3169, Australia; and Cooperative
Research Centre for Polymers, 32 Business Park Drive, Notting Hill VIC 3168, Australia
Received October 23, 2001; Revised Manuscript Received February 28, 2002
ABSTRACT: Reversible addition-fragmentation chain transfer (RAFT) polymerization techniques are
successfully used to control molecular weight and polydispersity in the seeded emulsion polymerization
of styrene. A novel technique was used to assist the transport into the seed particles of a RAFT agent
that is of very low water solubility, ensuring the RAFT agent was localized in the particle phase. The
polymer produced in these experiments was seen to have low molecular weight polydispersity (1.1 <
Mh w/Mh n < 1.4), and the molecular weight could be controlled by the amount of RAFT agent used. The
procedure permitted living polymerization to be performed in an emulsion polymerization system.
Importantly, the RAFT agent had no adverse effect on latex stability, and unusual amounts of coagulum
were not observed. Reductions in the rate of polymerization (measured by dilatometry) were observed
along with significant inhibition periods, for which possible explanations and remedies are discussed.
This appears to be the first instance where good living character has been demonstrated in a true emulsion
polymerization of styrene while maintaining good colloidal stability.
In tr od u ction
Review of P r eviou s Stu d ies. As previous studies
have shown,5,12 the additional physical events involved
in an emulsion polymerization experiment (such as the
entry of an oligomeric radical into a particle and the
complexities of particle formation) make the already
intricate process of RAFT polymerization significantly
more complex.
A survey of the current work in this field shows that,
other than a single report of an ab initio polymerization
of butyl methacrylate,1 all published experiments have
exhibited one or more of the following problems: (i) poor
colloidal stability,7,11 (ii) poor control of the number-
average molecular weight,5,7,11 or (iii) poor control of the
polydispersity.3,5,6,11,13 The work presented here does not
suffer from any of these problems.
It should be noted that the poor control of the
polydispersity in some studies is due to the use of
xanthates as the RAFT agent.5,6,13 While xanthates are
easily included in emulsion polymerizations (with no
significant colloidal stability problems), they are unable
to produce polymers of low polydispersity even in the
most ideal conditions. Indeed, the use of slightly surface-
active RAFT agents (such as the MADIX agents of
Smulders et al.14 or the amide-functionalized RAFT
agent of Uzulina et al.7) appears to be important to the
success of the experiment; the use of highly water-
insoluble RAFT agents without facilitating RAFT agent
transport has so far been unsuccessful.11 The water-
soluble RAFT agent used by Uzulina et al.7 was also
quite unsuccessful, presumably due to significant and
detrimental effects on the aqueous phase chemistry that
is so important in an emulsion polymerization.15,16 The
mechanistic implications of this observation will be
investigated further in the Discussion.
Reversible addition-fragmentation chain transfer
(RAFT) polymerization techniques are well developed
for solution and bulk polymerization systems,1-4 and
various results have been reported for its application
in emulsion3,5-7 and miniemulsion8-10 systems. The use
of RAFT in emulsion polymerization systems has, until
recently, been quite unsuccessful; previous studies have
reported high levels of coagulum, thick red layers (phase
separation) forming during the course of the polymer-
ization, and very slow polymerization rates.7,11 Indeed,
these results have led to speculation that RAFT cannot
be made to work in dispersed systems, although recent
results in miniemulsions8,10 and the results reported
here for emulsion polymerizations would tend to suggest
otherwise.
Here, the successful use of RAFT in a true emulsion
polymerization is reported. (The word “true” is used to
distinguish this work as being on (seeded) emulsion
polymerizations, where the latex particles do not arise
from droplets, as distinct from miniemulsion polymer-
izations.) The use of seeded emulsion polymerization in
this study opens the way to measuring important kinetic
parameters for these systems such as the rate coef-
ficients for the entry of radicals into particles.
In addition, it is hoped that these systems will achieve
faster rates of polymerization while maintaining good
temperature control and, most importantly, would have
the well-known significant environmental advantages
of emulsion polymerization compared to solution sys-
tems.
† Sydney University.
RAF T Agen t Tr a n sp or t. One of the most significant
difficulties in the application of RAFT to emulsion
polymerizations is the transport of the RAFT agent into
‡ CSIRO Molecular Science.
§ Cooperative Research Centre for Polymers.
* Author for correspondence.
10.1021/ma011840g CCC: $22.00 © 2002 American Chemical Society
Published on Web 06/04/2002

5418 Prescott et al.
Macromolecules, Vol. 35, No. 14, 2002
Sch em e 1. RAF T Agen t 2-P h en ylp r op -2-yl
p h en yld ith ioa ceta te (P P P DTA)
the latex particles. RAFT agents are often quite water-
insoluble, making this process very slow; while the
synthesis of water-soluble RAFT agents is possible,
previous studies have shown significant inhibition from
such RAFT agents.7 Once the RAFT agent has been
dissolved in the monomer and then added to the water
phase, the kinetic and thermodynamic barriers to the
transport of the RAFT agent into the particles could be
quite large.
With analogy to styrene dimer, Monteiro et al.11
estimated that, for RAFT agents similar to the one used
here, the rate of transport should be quite fast (∼106
s-1). However, as the water solubility of the RAFT agent
is often less than the purity of the prepared compound,
accurate estimates for the rate of transport cannot be
prepared.
pectation is be no means a certainty as other authors
have noted in other RAFT/emulsion studies.7,11 Emul-
sion polymerization kinetics can be quite different from
those in bulk due to compartmentalization effects (isola-
tion of the radicals in particles), and moreover, it has
already been observed that xanthate RAFT agents
display some surface activity that can result in very
different kinetics from what is seen in a corresponding
emulsion polymerization system.14
It is important to note that even if the kinetic rate of
transport is much faster than the rate of propagation
(e.g., in this case), it is still possible for thermodynamic
barriers to prevent sufficient RAFT agent transport. As
the RAFT agent will initially partition with approxi-
mately equal concentration between the monomer drop-
let and particle phases, a significant amount of RAFT
agent would be present in the monomer droplets. This
could allow RAFT polymerization within the droplets,
making oligomeric RAFT adducts (dormant chains), that
would be unable to transport across the aqueous phase
and result in a viscous, highly colored layer as the
monomer droplets evaporate.11
An alternative approach to transporting the RAFT
agent from monomer droplets to the locus of polymer-
ization (i.e., the particles) is therefore desirable. Addi-
tion of an organic cosolvent (such as acetone) is known
to facilitate transport of hydrophobic species in the
aqueous phase in emulsion polymerization reactions.
This allows the RAFT agent to be transported to the
preformed particles of a seed latex before the onset of
polymerization, allowing the kinetics of the actual RAFT
mechanism to be studied with reference to the kinetics
of particle growth, uncomplicated by particle formation
events.
Seed ed E m u lsion P olym er iza t ion . The use of a
preformed latex, or seed, on which subsequent polymer
may be grown is a common technique in both academic
studies and industrial processes.15 In this work, a seed
latex is used to eliminate the complexities of particle
formation, allowing the effect of the RAFT agent on
polymerization to be more easily studied.15 Coupling
seeded experiments with the acetone transport tech-
nique reported here also obviate the need to address
issues surrounding RAFT agent transport to the par-
ticles during the reaction.
In this study, a small polystyrene seed was used, and
the new, living polymer formed was also polystyrene.
The use of the same polymer for the seed and new
polymer obviates the need to make assumptions about
polymer miscibility that “heteroseeded emulsion polym-
erization” entails.17 As the molecular weights of the new
and old polymer used in this study are quite distinct, it
is sufficient to use GPC signal subtraction as detailed
below to measure the molecular weight distribution
(MWD) of the new polymer.
A seeded emulsion polymerization experiment may
start in either interval II (monomer droplets present)
or interval III (no monomer droplets), depending on the
ratio between the amount of monomer added and the
amount of seed polymer present (strictly also the volume
of water in the system, although for styrene this effect
is negligible). Most of the emulsion polymerization
experiments reported here were carried out in interval
III, although successful RAFT/emulsion experiments
were also carried out in interval II.
RAF T Tech n iqu es. The RAFT process has been well
described elsewhere.2-4 The RAFT agent 2-phenylprop-
2-ylphenyldithioacetate (1, PPPDTA) was used in this
study, shown in Scheme 1. This RAFT agent was chosen
as it has previously been seen to produce good results
in bulk and solution polymerization experiments with
styrene, it is quite water insoluble (thus it should not
affect the aqueous-phase chemistry which is so impor-
tant for emulsion polymerization15), and preliminary
work for this study was quite promising.
Exp er im en ta l Section
Ma ter ia ls. All materials were used as received from Aldrich
with the following exceptions. Inhibitors were removed from
the styrene using a commercial inhibitor removal column
(Aldrich), carbon disulfide was purified with several freeze-
thaw cycles, and R-methylstyrene was purified with a basic
alumina column. Demineralized water was used throughout
the experiments.
Syn th esis of RAF T Agen t. The RAFT agent PPPDTA was
synthesized using a method analogous to the synthesis of
2-phenylprop-2-yl dithiobenzoate described elsewhere.3 The
synthetic route used here was developed by Thang18 and has
been separately described by Quinn et al.19,20
Benzyl chloride (40 g) was added dropwise over 20 min to
magnesium turnings (7.7 g) and dry diethyl ether (150 cm3,
redistilled over sodium wire) under a nitrogen atmosphere at
4 °C. The mixture was stirred for 1 h and then heated to reflux
for 30 min. Carbon disulfide (24 g) was added dropwise with
stirring over 15 min to the chilled mixture. The reaction
mixture was maintained at 0 °C for a further 3 h. The mixture
was poured into ice water (150 cm3), and the aqueous layer
was extracted with diethyl ether (2 × 50 cm3). The aqueous
layer was acidified with cold hydrochloric acid solution (10%
w/w) to pH ∼2 and extracted with several portions of ether.
The combined ether portions were dried over anhydrous
magnesium sulfate. After removal of the drying agent and
evaporation of the ether, the crude phenyldithioacetic acid
(32.2 g, 64.0% yield) was isolated as a yellow oil that later
The overall objective of the present paper is to test
the expectation that use of an acetone transport system,
together with seeded emulsion polymerization, would
lead to the same sort of living character as is seen for a
corresponding solution RAFT polymerization. This ex-
1
solidified to a yellow/orange solid. H NMR (CDCl3): δ (ppm)
4.29 (s, 2H), 5.90 (br s, 1H) and 7.30-7.43 (m, 5H).
Phenyldithioacetic acid (30.4 g), R-methylstyrene (21.4 g),
and carbon tetrachloride (36 cm3) were combined under
nitrogen and heated at 70 °C for 18 h. After removal of the

Macromolecules, Vol. 35, No. 14, 2002
Seeded Emulsion Polymerization of Styrene 5419
carbon tetrachloride solvent, the crude product was dissolved
in the minimum amount of boiling hexane/ether (80:20) and
then recrystallized at 2 °C. The purified product was isolated
by filtration and a second crop of crystals obtained by the same
recrystallization method. (No attempt was made to obtain
further crops of crystals.) RAFT agent 2-phenylprop-2-yl
phenyldithioacetate (PPPDTA) was isolated as orange needles
(21.4 g, mp 56 °C, 23.7% yield on benzyl chloride, estimated
same experiment performed at 85 °C showed a reduction in
absorbance of 30% after 1 h. Since in these seeded emulsion
polymerization experiments reported here the RAFT agent is
not solublized into the water phase with ethanol (instead of
partitioning strongly into the particle phase), the hydrolysis
of PPPDTA would not be expected to be problematic in this
study.
Seed ed Em u lsion P olym er iza tion s. The seeded emulsion
polymerization experiments used the diluted seed prepared
by the acetone transport technique described above. Note that
for the control experiments the seeds were still subjected to
the acetone transport technique, but without the RAFT agent.
In a typical experiment, the latex from the acetone transport
technique (65 g, 3.1% solids) was combined with sodium
hydrogencarbonate (33 mg), sodium dodecyl sulfate (0.20 g),
water (35 g), and styrene (4.0 g). Sodium persulfate solution
(1 cm3, 0.042 mol dm-3) was added once thermal equilibrium
had been reached. The buffer, initiator, and surfactant were
all added as aqueous solutions, and the mixture was stirred
throughout the process. The reactions were carried out at 50
°C, and the particle number (calculated using the particle size)
1
purity 99%). H NMR (CDCl3): δ (ppm) 1.92 (s, 6H), 4.20 (s,
2H), 7.20-7.40 (m, 8H) and 7.47 (d, 2H). 13C NMR (CDCl3): δ
(ppm) 27.9, 56.3, 59.2, 126.6, 126.8, 127.0, 128.4, 128.9, 137.1,
143.9, and 223.3 (CdS).
An attempt was made to estimate the water solubility of
PPPDTA as follows. The absorption maximum at 311 nm
(characteristic of CdS) was monitored using a Cary 5 UV-
vis spectrometer (Varian) in both phases of an equilibrated
solution of PPPDTA in an octanol-water mixture for a range
of added amounts of PPPDTA. The concentration of chro-
mophores in the water-phase component was always less than
0.5% of that in the organic phase. This is less than the level
of minor impurities in the PPPDTA sample (e.g., other species
containing CdS arising from the preparation); the only
quantitative statement that can be made about the amount
of PPPDTA in the water phase is that it is less than 0.5% of
that in the particles, i.e., less than 1 × 10-4 M.
in each reaction was about 1 × 1017 dm-3 (range 9.6 × 1016
-
1.2 × 1017 dm-3).
Dilatometry. The dilatometry experiments were carried out
in a jacketed glass vessel approximately 100 cm3 in volume.
The temperature was controlled using a water bath. The cold
reaction mixture (without monomer and initiator) was thor-
oughly degassed under vacuum, backfilling with nitrogen.
Once the reaction mixture had reached thermal equilibrium
at the reaction temperature, the degassed initiator solution
was added and the capillary topped up with decane. The height
of the meniscus in the capillary was followed using a computer-
controlled tracking device.
P r ep a r a tion of P olystyr en e Seed . Sodium dodecyl sul-
fate (10 g) and sodium hydrogencarbonate (0.5 g) were dis-
solved in water (0.73 dm3) in a 1 dm3 jacketed glass reactor at
85 °C. The reaction mixture was stirred throughout using a
three-blade turbine stirrer at 300 rpm with a constant flow of
nitrogen through the reactor. After thermal equilibrium had
been reached, styrene (250 g) was added, and the mixture was
allowed to equilibrate for a further 15 min. Sodium persulfate
solution (12 cm3, 0.882 mol dm-3) was then added. The mixture
was allowed to react for 31/2 h, yielding a 25% solids latex with
particle radius rjw ) 39 nm, rjw/rjn ) 1.18, particle number
concentration Nc ) 2.2 × 1018 (calculated from rjn), molecular
weight Mh n ) 7.4 × 104, and and Mh w/Mh n ) 3.4.
Rate data were extracted from the meniscus height data
using the densities for styrene and polystyrene from Hawkett
et al.22,23 Conversion of the final latex from each dilatometry
experiment was determined by gravimetry and MWDs by gel
permeation chromatography (GPC).
Sampled Reactions. These reactions were performed at the
250 cm3 scale as a direct scale-up of the dilatometry recipe
presented above. As degassing under vacuum was not practi-
cal, nitrogen was bubbled through the reaction mixture for at
least 1 h before polymerization commenced, and the reaction
was carried out with a slight positive pressure of nitrogen. The
reaction mixture was stirred using a three-blade turbine
stirrer at 300 rpm. Samples of around 5 cm3 were taken
periodically from the reactor using a syringe. The samples were
quenched using hydroquinone and placed in an ice bath.
Conversion and MWD data were collected for each of the
samples.
La tex Ch a r a cter iza tion . Particle Sizing. For each of the
latexes prepared, the particle size distribution was measured
by a combination of Fraunhofer light scattering and polariza-
tion intensity differential scattering techniques using a Coulter
LS230 instrument (lower size limit 40 nm diameter).24 For
each of the seeded experiments, the number-average diameter
(between 84 and 100 nm) agreed with the predicted size to
within 4%. Additionally, the shape and broadness of the
distribution were unchanged, and neither new nucleation nor
coagulum was observable. There were no systematic differ-
ences between RAFT and non-RAFT particle sizes.
The accuracy of the Coulter LS230 sizer for particles of this
type was compared to that of a Matec CHDF (capillary
hydrodynamic fractionator) using two other latexes (rjw ) 55
and 180 nm by CHDF).25 The values of rjw and rjn obtained by
CHDF were within 5% of those obtained by the Coulter LS230.
Molecular Weights. Molecular weight distributions were
measured by GPC. Samples of dried latex were dissolved in
THF (1 mg cm-3) and filtered using a 0.2 µm PTFE filter.
Analyses were carried out using four PLgel columns (three
Mixed-C columns and one Mixed-E column, Polymer Labs).
Injection volumes of 100 µL were used with a Waters 410
differential refractive index detector and a flow rate of 1 cm3
min-1. Cubic calibration curves were generated using 13
polystyrene standards (Polymer Labs) with molecular weights
Aceton e Tr a n sp or t Tech n iqu e. In a typical preparation,
seed latex (10 g), water (55 g), and acetone (20 g) were
combined with stirring. RAFT agent (30 mg) was added to this
mixture as fine crystals. The mixture was allowed to stir for
24-72 h. After this time, the crystals of RAFT agent were no
longer visible. Liquid (acetone/water mixture, 30 g) was
stripped from the mixture using a rotary evaporator at 30 °C
over 45 min, and the mixture was made back up to volume
with more water (10 g).
Deter m in a tion of Aceton e Resid u e. At the conclusion
of the seed preparation by the acetone transport technique
detailed above, some residual acetone may be left in the latex
despite the fact that acetone/water mixtures are zeotropic.21
To estimate the level of residual acetone, an acetone/water
mixture (same proportions as above) was subjected to rotary
evaporation at 30 °C over 45 min. Additionally, a series of
seven acetone/water standards ranging from 65 to 130 ppm
were prepared gravimetrically. Using a Cary 5 UV-vis
spectrometer (Varian), the carbonyl absorbance of acetone (265
nm) was measured for the standards and unknown. The
standards gave a linear calibration to high precision (R2
>
0.999), indicating that the acetone level in the seed latex was
92 ppm. When this seed was further diluted as described
below, the actual polymerization reactions were carried out
in the presence of 60 ppm acetone.
Hyd r olysis of RAF T Agen t. The dithiocarbonyl functional
group of the RAFT agents may be expected to hydrolyze given
suitable reaction conditions. A solution of RAFT agent (9 ×
10-5 mol dm-3) in ethanol/water mixture (50% w/w) was
prepared and separated into nine aliquots. The samples were
heated to 50 °C (the reaction temperature for the polymeri-
zation experiments described below) for periods between 30
min and 5 h. The hydrolysis of the dithiocarbonyl group was
observed spectroscopically by comparing the peak absorbance
(312 nm) of the CdS group of an unheated sample to each of
the heated samples using a Cary 5 UV-vis spectrometer
(Varian). Over 5 h, the absorbance was reduced by 5%. The

5420 Prescott et al.
Macromolecules, Vol. 35, No. 14, 2002
ranging from 264 to 2.56 × 106. GPC signal analyses were
performed using the software GPC for Windows (Chemware).
GPC Signal Subtraction. For the seeded emulsion polym-
erization experiments, the molecular weight distribution that
pertained to the new polymer formed was isolated from the
raw GPC data using a method similar to that of Clay et al.26
In this treatment, we use the raw GPC data and look only at
the differences between new and old polymer, rather than
using the calculated number distribution and pseudo-instan-
taneous distributions.
Consider two polymer samples, seed and new, that would
give GPC DRI signals, gseed and gnew, normalized to unit area.
When samples of the polymer are mixed using masses mseed
and mnew, the resulting (normalized) GPC signal, g, will be
(mseed + mnew)g ) mseedgseed + mnewgnew
(1)
given that the DRI response is linear in the mass of polymer
in the detector.27 Note that this also assumes that the
calibration of the GPC is unchanged (e.g., when the samples
are run sequentially).
Given that the experimentally measured signal will be (mseed
+ mnew)g, which may be normalized to unit are to yield g, the
signal from just one of the components may be calculated to
be
F igu r e 1. Mh n determined experimentally plotted against Mh n
predicted by eq 4. While there is some scatter around the line
of slope 1, good molecular weight control has been achieved.
Each of the points on this plot has molecular weight polydis-
persity Mh w/Mh n less than 1.4 and in most cases 1.25 or better.
Molecu la r Weigh t Con tr ol. One of the features that
characterizes a living system is the ability to control the
molecular weight of the polymer produced. In the case
of RAFT polymerization, this control is on the basis of
the ratio [monomer]/[RAFT], providing that the transfer
constant, Ctr is greater than two.3,31 For a given mass
of monomer, mmon, and a given amount of RAFT agent,
nRAFT, in the system, this relationship may be formalized
as
mnew
mseed
gnew ) g -
gseed
(2)
mseed + mnew
(mseed + mnew
)
Further taking into account the formation of the new polymer
as a function of the mass of monomer added to the system,
m
mon, and the fractional conversion for the sample, x
xmmon
mseed
Mh pnred
)
(4)
g′new ) g -
gseed
(3)
nRAFT
mseed + xmmon
where x is the conversion at which the sample was taken
and Mh pnred is the target Mh n. This relation holds for bulk,
solution, and emulsion polymerizations in both intervals
2 and 3 (presence and absence of monomer droplets),
with the proviso in the case of interval 2 polymerizations
that the RAFT agent is at the locus of polymerization
(i.e., in the particles) and that significant uncontrolled
droplet polymerization is absent. For good control of the
molecular weight, the value of Mh n should also be lower
in a RAFT system than in a control reaction.
A number of Mh n values determined in this study are
shown in Figure 1. It may be seen that the control of
Mh n in these systems is quite good, although there is
some scatter around the 45° line. Importantly, each of
these samples shown in this plot has molecular weight
polydispersity, Mh w/Mh n, less than 1.4 and in most samples
1.25 or better. The data shown include experiments
carried out in both intervals II and III, along with a
wide range of [RAFT]/[I] and [monomer]/[RAFT] condi-
tions.
The Mh n and polydispersity values for the control
reactions could not be determined as the polymer formed
had significant amounts of material in excess of the
exclusion limit of the columns (molecular weight around
6 × 106). This implies a significant molecular weight
reduction compared to polymer prepared in the same
conditions in the absence of RAFT agents.
The molecular weights reported here are similar to
those generated in other seeded RAFT/emulsion systems
that showed poor molecular weight polydispersity,7,11
indicating that previously reported failures were not due
to an insurmountable limitation of RAFT/emulsion
systems.
where g′new is an unnormalized GPC trace from the new
polymer.
Signal subtraction may thus be performed using eq 3 to yield
a GPC trace reflecting solely the contribution of the new
polymer. It is important to note that even when the values of
the coefficients in eq 3 were varied by (10%, the values of
Mh n, Mh w, and the polydispersity index (Mh w/Mh n) were not
significantly changed for the experiments reported here.
GPC Signal Debroadening. In some low-conversion samples
it was evident from the GPC trace that the GPC columns had
separated the discrete populations of oligomers. In this case,
it is reasonable to undertake a deconvolution of the signal to
attempt to remove the effect of column broadening to clarify
the oligomer distribution. Between 75 and 100 debroadening
cycles were applied, using a sharp solvent peak as the basis
for the deconvolution, following of Tung et al.28-30
Resu lts a n d Discu ssion
Based on work by other authors, the first criterion
for success of RAFT in an emulsion system is the
colloidal stability of the latex during the course of the
polymerization.7,11 In the polymerization experiments
performed in this work, coagulum was rarely observed,
and in the cases when some polymer was left as a vortex
plug in the dilatometer, the amount of coagulum was
insignificant (less than 0.5% of the total polymer).
Moreover, highly colored creamed oils observed on the
top of the reaction mixtures, as reported by other
authors,7,11 were not seen at any stage during the
polymerization experiments.
While some reduction in the rate of polymerization
was observed (see below), the reaction was not retarded
to the same degree as has been observed in previous
studies.7

Macromolecules, Vol. 35, No. 14, 2002
Seeded Emulsion Polymerization of Styrene 5421
ing column-debroadening algorithms to this trace, the
debroadened MWD in Figure 3b is obtained. Since the
presence of the oligomeric species is already evident
from the original MWD, the debroadening algorithm
provides a guide to the eye to assist in the estimation
of the spacing between the peaks. Indeed, the spacing
observed is around 100 in molecular weight, consistent
with the assignment of these peaks to oligomeric spe-
cies. As this peak is not evident in subsequent traces
(even at higher magnifications than presented here),
and it lies on the Mh n vs conversion curve shown in
Figure 4, it can be concluded that these molecules are
oligomeric species with living character (presumably the
dormant form of the growing chain in the RAFT mech-
anism) that have subsequently grown in molecular
weight.
While Figure 4 shows a linear increase of Mh n as a
function of conversion as expected for a living system,
the linear fit for the higher initiator experiment does
not pass through the origin (i.e., predicted Mh n values
are consistently higher than the experimental value at
all conversions). This effect is consistent with increased
termination in the system, although the effect does not
appear to be large.
Ra te of P olym er iza tion . One common characteristic
of living polymerization systems is a reduction in the
rate of polymerization, although the extent of this effect
appears to be RAFT-agent dependent.3 In the case of
RAFT/emulsion systems, previous studies have shown
significant reduction in the rate of polymerization.11,14
The systems presented here are no different; the addi-
tion of RAFT agent to the system leads to both an
increase in the inhibition period and a retardation in
the polymerization rate.
In this system, inhibition periods of 45 min to 3 h were
seen, accompanied by a reduction in the rate of polym-
erization to between 30 and 50% of its previous value,
as shown in Figure 5.
F igu r e 2. Molecular weight distributions for a seeded po-
lymerization at [I] ) 0.446 mM, [RAFT]/[I] ) 2.75, and
[monomer]/[RAFT] ) 324. MWD traces were rescaled so that
the seed polymer peak was the same height in each plot.
Evolu tion of th e Molecu la r Weigh t Distr ibu tion .
For a RAFT agent with a transfer constant Ctr . 1, such
as the one used here, evidence of living character in a
homopolymerization is seen in the linear increase of Mh n
with conversion. From the sampled reactions, the evolu-
tion of the molecular weight distribution may be ob-
served. In each of the molecular weight distributions
presented here, the contribution to the signal of seed
polymer is still present, although this was subtracted
for the calculation of Mh n and Mh w as described above.
The evolution of the MWD follows the expected
behavior for a living polymerization system, rather than
following the typical progress of a normal emulsion
polymerization experiment. In each experiment, the
molecular weight distribution shows a comparatively
sharp peak of new polymer superposed onto the seed
polymer peak. Two representative MWD evolutions are
shown in Figure 2 and Figure 3. The value of Mh n from
each of these experiments is plotted against conversion
in Figure 4, showing a linear increase of Mh n as a
function of conversion.
It is interesting to note from this plot that increasing
the initiator concentration by a factor of about 9 appears
to remove these detrimental effects. Moreover, at this
higher initiator concentration, good molecular weight
control and narrow polydispersity are still seen, as seen
in Figures 3 and 4, although it should lead to the
It is interesting to note the low molecular weight
region of the lowest conversion sample in Figure 3.
Enlarged in Figure 3b, this sample shows the presence
of polymer with molecular weight around 1000. Apply-
F igu r e 3. (a) Molecular weight distributions for a seeded polymerization at [I] ) 1.39 mM, [RAFT]/[I] ) 0.93, and [monomer]/
[RAFT] ) 360. (b) Low molecular weight region of the 17% conversion sample showing the original data (dotted line) and the
debroadened data (solid line). The debroadened data indicated the peaks are around 100 separation in molecular weight. MWD
traces were rescaled so that the seed polymer peak was the same height in each plot.

5422 Prescott et al.
Macromolecules, Vol. 35, No. 14, 2002
Inhibition and Retardation. The causes of the in-
creased induction period and retardation are unclear at
this stage. While RAFT techniques in solution polym-
erization are known to sometimes produce a small
induction period and a reduction in the polymerization
rate (depending on the RAFT agent used),2,3 the induc-
tion periods and retardation seen in emulsion systems
are more significant. It appears likely that these effects
have both a chemical and physical cause.
Evidence from solution polymerization experiments
indicates that there is a portion of the inhibition period
that may be linked to the chemical nature of the cumyl
radical.3 It has been postulated that the cumyl radical
may be quite poor at reinitiating by analogy with the
behavior of poly(R-methylstyrene),3 although previous
radical addition studies do not support this conclu-
sion.32,33 However, the reaction of cumyl radicals with
the original RAFT agent in preference to the reaction
with monomer may go some way to explaining the
retardation seen in solution polymerization.3 Indeed, the
observed retardation with cumyl-based RAFT agents is
significantly greater than with cyanoisopropyl-based
RAFT agents.3
F igu r e 4. Mh n as a function of conversion for the seeded
polymerizations shown in Figures 2 and 3. Each of the points
on this plot has molecular weight polydispersity 1.2 < Mh w/Mh n
< 1.4. While both sets of points lie on straight lines as
expected, the plot for the higher initiator concentration system
does not pass through the origin.
Other authors have postulated that the exit of the
RAFT leaving group (the cumyl radical in this study)
from the latex particle is the primary cause of the
inhibition period.11,14 This is certainly a distinct pos-
sibility, as the high diffusion coefficient of the cumyl
radical allows it to exit quite rapidly,11 and for that
reason we now examine this in detail from a theoretical
viewpoint. Following the method described by Monteiro
et al.,11 it is possible to express the probability that the
cumyl radical will exit the latex particle in terms of the
first-order rate coefficient for the desorption of the
radical, kdR, and the rate coefficient for the addition of
the radical to monomer, kaRdd
:
kdR
kdR + kRaddCp
P(exit) )
(5)
F igu r e 5. Conversion vs time plots showing the effect of
adding the RAFT agent to the emulsion polymerization system
at varying initiator concentrations: 1.32 mM (dotted line),
0.436 mM (solid line), and 0.144 mM (dashed line). In each
case, the addition of the RAFT agent leads to a longer
inhibition period and slower rate of polymerization compared
to the control reaction.
where Cp is the concentration of monomer in the
particle.
It is important to note that there are three possible
processes that the cumyl radical may undergo (with the
monomer, M, and the particle):
production of more dead chains.1 Using higher initiator
concentrations may provide a suitable way of overcom-
ing the twin problems of the inhibition period and
reduction in polymerization rate.
Mech a n istic Discu ssion . While this study has
shown that RAFT/emulsion systems may be successfully
used to make controlled molecular weights, there are
still some difficulties to be addressed. In particular, the
mechanisms involved in the inhibition, retardation, and
RAFT agent transport are unknown.
addition to monomer: R• + M f RM•
exit from particle: R•s(particle) f
(6)
R• + particle (7)
addition to RAFT: R• + RAFT h RAFT + R (8)
However, the reaction shown in eq 8 does not actually
change the concentration of R• in the particle. This
reaction may thus be ignored in the model, allowing the
event scheme to be reduced to the processes in eqs 6
and 7, yielding eq 5 from the (pseudo)-first-order rate
coefficients. A full mathematical exposition of this
argument is set out in the Appendix.
While it is desirable to compare the rates of polym-
erization in these RAFT/emulsion experiments to previ-
ous studies (such as those of Monteiro et al.5,11), it
should be recognized that the rate of polymerization is
a poor measure of the performance for an emulsion
polymerization. Instead, the quantity that should be
compared is the average number of propagating (not
bipolymeric intermediates) radicals per particle, nj,
which removes variations due to minor changes in
particle size and particle number concentration.15 As
discussed below, even measuring nj is a somewhat
fruitless task until new models to describe RAFT/
emulsion behavior are developed.
The desorption of a radical may be expressed as a
diffusive process:34
3DRwCRw
kdR
)
(9)
2
rs CRp

Macromolecules, Vol. 35, No. 14, 2002
Seeded Emulsion Polymerization of Styrene 5423
Ta ble 1. P a r a m eter s (fr om a Typ ica l Exp er im en t) for
Ca lcu la tin g th e P r oba bility of Exit of a Cu m yl Ra d ica l
fr om a La tex P a r ticle, P (exit)
radical exiting is quite low. This tends to suggest that
the influence of radical exit may be less than first
supposed.
parameter
value
Nevertheless, a small amount of exit will lead to a
reduction in polymerization rate,36 as there are the
possibilities of aqueous phase termination between the
cumyl radical and other radicals and reentry of the
cumyl radical into a growing particle leading to rapid
termination. It is important to note that the exit of
radicals only has the potential to explain retardation
or inhibition in the first few percent of conversion, as
after that time all the initial RAFT agent has been
converted to the dormant species on a chain end.
Obviously, to understand these effects, more experi-
mental work accompanied by suitable modeling of these
systems is required.
Another possibility for the retardation is some surface
activity of the RAFT agent; this effect has been used to
explain retardation in RAFT/emulsion systems using
xanthates as the RAFT agent.14 This surface excess of
transfer agent is thought to cause an increase in the
exit rate coefficient and a decrease in the entry rate
coefficient. However, it is felt that the present case with
1, PPPDTA, is unlikely to show significant surface
activity.
rs/nm
45
CRw/CRp
6.2 × 10-4
1.6 × 10-5
5.5
Dw/cm2 s-1
Cp/mol dm-3
Ta ble 2. Ra te Coefficien t for th e Ad d ition of a Cu m yl
Ra d ica l on to Styr en e a n d th e P r oba bility of Exit of a
Cu m yl Ra d ica l fr om a La tex P a r ticle, P (exit)
kRadd/dm3 mol-1 s-1
P(exit)
1.3 × 103 a
0.19
0.098
2.86 × 103 b
Based on the method described by Monteiro et al.;11 see text.
a
Using the experimental data from Fischer et al.32,33
b
where Dw is the diffusion coefficient of the small radical
in water, rs is the swollen radius of the particle, and CR
w
and CRp are the equilibrium concentrations of the
cumyl radical in the water and particle phases, respec-
tively. It will be seen that these CRw and CpR terms can
prove rather difficult to estimate, depending on the
monomer and radical involved.
Monteiro et al.11 implicitly assumed that the cumyl
radical is similar enough to styrene so that the equi-
librium concentration of cumyl radicals may be ex-
pressed as a mole fraction of the equilibrium concen-
tration of styrene molecules. This is in line with the
method of Ugelstad et al.15,34 for the determination of
the rate coefficient for the desorption of the monomeric
radical. Writing the mole fraction of cumyl radicals on
Two additional mechanisms for retardation in RAFT
systems have been recently proposed. Both mechanisms
involve the intermediate bipolymeric radical (Pn-S-•-
CZ-S-Pm) in the RAFT equilibration reaction. Barner-
Kowollik et al.20,37 reported that kinetic simulations of
bulk RAFT reactions indicated that this radical may be
long-lived enough so as to cause significant amounts of
retardation. Monteiro et al.12 suggested that this radical
may become involved in a termination reaction with
other polymeric radicals in the reaction, thus retarding
polymerization.
While both of these mechanisms offer insights into
the retardation seen in bulk and solution polymeriza-
tion, they do not explain why these effects are worse in
emulsion polymerization. Now, for styrene, the particle
size of the system studied here is such that it should
follow “zero-one” kinetics:15 entry of a radical from the
aqueous phase into a particle already containing a
growing radical results in virtually instantaneous ter-
mination. It is worth noting that the intermediate
radical termination mechanism12 would not lead to
retardation in an emulsion polymerization following
zero-one kinetics; the compartmentalization of radicals
means that an oligomeric radical has to enter the
particle before the intermediate radical termination
reaction could take place, but the “zero-one” assump-
tion15 indicates that termination should take place
pseudo-instantaneously even in the absence of RAFT
and this mechanism.
RAFT Agent Transport. As discussed in the Introduc-
tion, it is postulated here that the one of the principal
problems experienced in other studies is the transport
of RAFT agent to the locus of polymerization. These
problems may be either kinetic or thermodynamic in
origin. The experimental results presented here and
those of other recent studies of RAFT/emulsion systems
are consistent with this postulate.
Thanks to the work of many authors, a general
pattern may now be observed in the literature. It has
been suggested that RAFT agents with some surface
activity7,14 may be used in emulsion polymerization
without considerable difficulty. This may be mechanisti-
styrene radicals as øR, CR and CpR may be written as
w
CRw ) øRCw
CRp ) øRCp
(10)
(11)
The ratio CRw/CRp for the cumyl radical in the presence
of styrene may thus be approximated as
CRw øRCw Cw
)
)
(12)
CpR
øRCp
Cp
where Cw and Cp are the concentrations of styrene in
the water and particle phases, respectively. While this
approach would appear to be valid for a cumyl radical
in a styrene/polystyrene matrix, its applicability to other
non-styrene-like radicals (such as the ethyl methacry-
late radical11) in a styrene/polystyrene matrix is unclear.
Additionally, Dw will be similar to that of styrene.
Monteiro et al.11 further suggested that a reasonable
approximation for kRadd was 5kspty, where kspty is the rate
coefficient for the propagation for styrene radicals from
pulsed laser polymerization data.35 Alternative values
for kRadd may be obtained from the experimentally
determined Arrhenius parameters of Fischer et al.32,33
The values for the parameters used in the calculation
of P(exit) are summarized in Tables 1 and 2, along with
the calculated values of P(exit). It is interesting to note
that the value of P(exit) is quite sensitive to the
parameter kRadd. Even so, in both of the estimates of
P(exit) prepared here, the actual probability of a cumyl

5424 Prescott et al.
Macromolecules, Vol. 35, No. 14, 2002
cally justified by considering the effects of the RAFT
agent on the aqueous-phase chemistry and the transport
of the RAFT agent from monomer droplets to the locus
of polymerization (the particles):
(i) If a RAFT agent is quite hydrophobic, its water
solubility is small enough (especially after one or more
addition-fragmentation steps have occurred within a
monomer droplet) so as to prevent it from reaching the
particles11 early enough in the reaction to commence
controlled polymerization (as all RAFT agents should
undergo their first fragmentation quite early in the
reaction, or else molecular weight control will be poor3).
(ii) If a RAFT agent is quite water-soluble, consider-
able chain transfer will occur in the water phase. Hence,
it will take quite some time for z-meric species (oligo-
radicals that are capable of entering a particle15,38) to
be formed. In such a situation, the RAFT agent becomes
an effective inhibitor.7
(iii) If a RAFT agent is surface active,5,6,13,14 it may
have enough water solubility to transport across the
aqueous phase from the droplets to the particles, but
the relatively low water solubility associated with its
surface activity will keep it out of the aqueous phase
chemistry. Even still, partitioning of the RAFT agent
between droplets and particles may lead to poor control
of Mh n.5
F igu r e 6. First three steps of the event scheme for the cumyl
radical in a styrene/polystyrene particle, described by in eqs
6-8.
chemical and physical aspects of inhibition and retarda-
tion to be more easily elucidated.
Ack n ow led gm en t . Dr. San Thang (CSIRO) is
thanked for describing the synthetic route to the RAFT
agent used here and Ms. Binh Pham (KCPC) is thanked
for performing the CHDF experiments. The ever-helpful
comments of Dr. Chris Fellows (KCPC) throughout the
preparation of this manuscript are also appreciated. The
assistance of Ms. Amanda Finlay (CSIRO), Mr. Heng
Taing (CSIRO), Mr. David Sangster (KCPC), and Dr.
Hank De Bruyn (KCPC) with various experimental
techniques is gratefully acknowledged. The Key Centre
for Polymer Colloids is established and supported under
the Australian Research Council’s Research Centres
Program.
While experimental proof of these suggestions is still
quite limited, all available data on RAFT/emulsion
polymerizations are consistent with these postulates.
It may be expected that the RAFT agent could diffuse
back out of the particles once polymerization has
started; however, the relatively fast rate of addition of
styrenic radicals to PPPDTA20,37 should mean that all
the PPPDTA is consumed to form polystyryl-PPPDTA
adducts (the dormant species) within the first few
percent conversion.31 Moreover, as the water solubility
of PPPDTA is immeasurably low (below the impurity
level of the sample prepared here), the RAFT agent
would not be expected to have a significant impact on
the entry of radicals into latex particles.15,38
Ap p en d ix: Sim p lifica tion of P (exit) Mod el
In calculating the probability of a process occurring,
complications arise within event models when one or
more reactions are reversible or have some products
that are also reactants. It will be shown that while these
make diagrams of the event model much more complex,
they may be readily simplified and that the additional
reactions may often be neglected.
Let us consider the specific example of cumyl radicals
and styrene monomer in a polystyrene particle. The
three processes that may occur are shown in eqs 6-8
in the text. The event model for this system is quite
complicated; in the first instance it should be drawn as
shown in Figure 6 where the series continues ad
infinitum.
Con clu sion s
The key result of this work is that RAFT techniques
can be successfully applied to seeded emulsion polym-
erization systems, producing controlled molecular ar-
chitectures and retaining living character. Evidence of
the living character of these systems was seen in the
control of the final molecular weight based on the
amount of RAFT agent used, the evolution of the
molecular weight distribution as a function of conver-
sion, and the low polydispersities of the polymers
obtained.
Some detrimental effects of the use of RAFT agents
in emulsion polymerization systems are also noted,
namely significant inhibition periods and a reduction
in the rate of polymerization. These difficulties were
overcome by using higher levels of initiator; this did not
appear to compromise living character.
Ma t h em a t ica l Abst r a ct ion a n d Sim p lifica t ion .
First, it is necessary to define some notation for this
exercise. For simplicity of notation, let us commence by
using events “A”, “B”, and “C” rather than the more
complicated chemical descriptors “exit”, “add to M”, and
“add to RAFT”.
Next, define the number of A events at step i to be
ni(A), the total number of A events be n(A), and the
probability of an A event occurring P(A), with similar
definitions for B and C events. The final value for P(A)
is thus
n(A)
P(A) )
(A1)
n(A) + n(B) + n(C)
The extension of the acetone transport technique
outlined here to other RAFT agents and other mono-
mers is a natural next step in the implementation of
RAFT/emulsion techniques. Further experimental and
concomitant modeling work is required to fully under-
stand the inhibition and retardation effects seen in
RAFT/emulsion systems. Using the experimental tech-
niques described here decouples RAFT agent transport
from other RAFT/emulsion behaviors, allowing the
which may be reexpressed in terms of infinite series:
∞
n (A)
∑
i
i)1
P(A) )
(A2)
∞
∞
n (A) + n (B) + n (C)
∑
∑
i
i
∞
i)1
i)1

Macromolecules, Vol. 35, No. 14, 2002
Seeded Emulsion Polymerization of Styrene 5425
(3) Moad, G.; Chiefari, J .; Chong, Y. K.; Krstina, J .; Mayadunne,
R. T. A.; Postma, A.; Rizzardo, E.; Thang, S. H. Polym. Int.
2000, 49, 993.
(4) Rizzardo, E.; Chiefari, J .; Chong, B. Y. K.; Ercole, F.; Krstina,
J .; J effery, J .; Le, T. P. T.; Mayadunne, R. T. A.; Meijs, G. F.;
Moad, C. L.; Moad, G.; Thang, S. H. Macromol. Symp. 1999,
143, 291.
F igu r e 7. Mathematically simplified event scheme for the
abstracted system.
The numerator of eq A2 may be rewritten in terms of
the ni(C):
(5) Monteiro, M. J .; de Barbeyrac, J . Macromolecules 2001, 34,
4416.
(6) Monteiro, M. J .; Sjo¨berg, M.; van der Vlist, J .; Gottgens, C.
M. J . Polym. Sci., Polym. Chem. 2000, 38, 4206.
(7) Uzulina, I.; Kanagasabapathy, S.; Claverie, J . Macromol.
Symp. 2000, 150, 33.
(8) Vosloo, J . J .; De Wet-Roos, D.; Tonge, M. P.; Sanderson, R.
D. Macromolecules, in press.
∞
ni(A) ) P(A)n + P(A)n (C) + P(A)n (C) + ... (A3)
∑
1
2
i)1
(9) de Brouwer, H.; Tsavalas, J . G.; Schork, F. J .; Monteiro, M.
J . Macromolecules 2000, 33, 9239.
(10) Ballard, M. J .; Rizzardo, E.; Taing, H. C. Personal com-
munication, 2000.
(11) Monteiro, M. J .; Hodgson, M.; De Brouwer, H. J . Polym. Sci.,
Part A: Polym. Chem. 2000, 38, 3864.
(12) Monteiro, M. J .; de Brouwer, H. Macromolecules 2001, 34,
349.
which in turn may be expressed as the sum of a
geometric progression, since ni(C) ) n[P(C)]i:
∞
∞
n (A) ) P(A)n
P(C)i
(A4)
∑
∑
i
i)1
i)0
(13) Charmot, D.; Corpart, P.; Adam, H.; Zard, S. Z.; Biadatti, T.;
Bouhadir, G. Macromol. Symp. 2000, 150, 23.
(14) Smulders, W.; Gilbert, R. G.; Monteiro, M. J . Macromolecules,
in press.
(15) Gilbert, R. G. Emulsion Polymerization: A Mechanistic
Approach; Academic: London, 1995.
(16) Prescott, S. W.; Fellows, C. M.; Gilbert, R. G. Macromol.
Theory Simul. 2002, 11, 163.
(17) Schoonbrood, H. A. S.; German, A. L.; Gilbert, R. G. Macro-
molecules 1995, 28, 34.
(18) Thang, S. H. Personal communication, 1999.
(19) Quinn, J . F.; Rizzardo, E.; Davis, T. P. Chem. Commun. 2001,
1044.
(20) Barner-Kowollik, C.; Quinn, J . F.; Nguyen, T. L. U.; Heuts,
J . P. A.; Davis, T. P. Macromolecules 2001, 34, 7849.
(21) CRC Handbook of Chemistry and Physics, 81st ed.; Lide, D.
R., Ed.; CRC Press: Boca Raton, FL, 2000.
A similar expression for n(B) in terms of P(B), n, and
P(C) may be readily obtained.
Since the allowable range for P(C) is 0 e P(C) < 1
(the case P(C) ) 1 is a trivial case in which pathways A
and B do not actually exist), the two limit expressions
may be readily evaluated:
∞
1
P(C)i )
(A5)
(A6)
∑
1 - P(C)
n∞(C) ) 0
i)0
This allows eq A2 to be simplified to be
(22) Hawkett, B. S.; Napper, D. H.; Gilbert, R. G. J . Chem. Soc.,
Faraday Trans. 1 1980, 76, 1323.
(23) Hawkett, B. S.; Napper, D. H.; Gilbert, R. G. J . Chem. Soc.,
Faraday Trans. 1 1981, 77, 2395.
P(A)n
(24) Bott, S. E.; Hart, W. H. In Particle Size Distribution II:
Assessment and Characterization; Provder, T., Ed.; American
Chemical Society: Washington, DC, 1991; Vol. 472.
(25) Pham, B. T. T. Personal communication, 2000.
(26) Clay, P. A.; Gilbert, R. G.; Russell, G. T. Macromolecules 1997,
30, 1935.
1 - P(C)
P(A) )
(A7)
P(A)n
P(B)n
+
+ 0
1 - P(C) 1 - P(C)
(27) Ballard, M. J . Ph.D., University of Sydney, 1983.
(28) Tung, L. H. J . Appl. Polym. Sci. 1966, 10, 375.
(29) Tung, L. H. J . Appl. Polym. Sci. 1966, 10, 1271.
(30) Tung, L. H.; Moore, J . C.; Knight, G. W. J . Appl. Polym. Sci.
1966, 10, 1261.
(31) Mu¨ller, A. H. E.; Zhuang, R.; Yan, D.; Litvinenko, G.
Macromolecules 1995, 28, 4326.
Recognizing that by definition P(A)n ) n1(A) and P(B)-
n ) n1(B), this may be rewritten as
n1(A)
P(A) )
(A8)
n1(A) + n1(B)
(32) Walbiner, M.; Wu, J . Q.; Fischer, H. Helv. Chim. Acta 1995,
78, 910.
which is equivalent redrawing the original event model
as shown in Figure 7. This may be readily applied to
the specific case of the RAFT polymerization outlined
above.
(33) Fischer, H.; Radom, L. Angew. Chem., Int. Ed. 2001, 40, 1340.
(34) Ugelstad, J .; Hansen, F. K. Rubber Chem. Technol. 1976, 49,
536.
(35) Buback, M.; Gilbert, R. G.; Hutchinson, R. A.; Klumperman,
B.; Kuchta, F.-D.; Manders, B. G.; O’Driscoll, K. F.; Russell,
G. T.; Schweer, J . Macromol. Chem. Phys. 1995, 196, 3267.
(36) Lichti, G.; Sangster, D. F.; Whang, B. C. Y.; Napper, D. H.;
Gilbert, R. G. J . Chem. Soc., Faraday Trans. 1 1982, 78, 2129.
(37) Barner-Kowollik, C.; Quinn, J . F.; Morsley, D. R.; Davis, T.
P. J . Polym. Sci., Polym. Chem. 2001, 39, 1353.
(38) Maxwell, I. A.; Morrison, B. R.; Napper, D. H.; Gilbert, R. G.
Macromolecules 1991, 24, 1629.
Refer en ces a n d Notes
(1) Chiefari, J .; Chong, Y. K.; Ercole, F.; Krstina, J .; J effery, J .;
Le, T. P. T.; Mayadunne, R. T. A.; Meijs, G. F.; Moad, C. L.;
Moad, G.; Rizzardo, E.; Thang, S. H. Macromolecules 1998,
31, 5559.
(2) Rizzardo, E.; Chiefari, J .; Mayadunne, R.; Moad, G.; Thang,
S. H. ACS Symp. Ser. 2000, 768, 278.
MA011840G