Surface structure and composition of narrowly-distributed functional polystyrene particles prepared by dispersion polymerization with poly(l-glutamic acid) macromonomer as stabilizer

Article abstract of DOI:10.1016/j.polymer.2015.06.022

A novel macromonomer composed of poly(α-l-glutamic acid) was used as a stabilizer for dispersion polymerization of styrene in DMF-water medium with AIBN initiator, giving narrowly-distributed functional polystyrene particles on which the poly(α-l-glutamic acid) was grafted. The resultant particles had 0.54-2.12 μm in size and 0.2-2.6 residue/nm² in surface density and showed a pH-responsive colloidal behavior associated with a helix-coil transformation of the surface poly(α-l-glutamic acid). Not only the particle size but also the surface density were controlled with macromonomer concentration, macromonomer length, DMF composition, and styrene concentration, while no consistent trend for AIBN concentration was observed. A gel-permeation-chromatography curve of the particles was separated into three components. We tentatively identify the origin of each component and propose a possibility that unstable particles, which were generated even after the growing particles were stabilized, took an important role in particle growth and size distribution of the resultant particles.

Full text of DOI:10.1016/j.polymer.2015.06.022

Polymer 70 (2015) 183e193
Contents lists available at ScienceDirect
Polymer
journal homepage: www.elsevier.com/locate/polymer
Surface structure and composition of narrowly-distributed functional
polystyrene particles prepared by dispersion polymerization with
poly(
L-glutamic acid) macromonomer as stabilizer
*
Tomomichi Itoh , Tetsuo Tamamitsu, Hiroaki Shimomoto, Eiji Ihara
Department of Materials Science and Biotechnology, Graduate School of Science and Engineering, Ehime University, 3 Bunkyo-cho, Matsuyama 790-8577,
Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 24 April 2015
Received in revised form
A novel macromonomer composed of poly(a-L-glutamic acid) was used as a stabilizer for dispersion
polymerization of styrene in DMFewater medium with AIBN initiator, giving narrowly-distributed
functional polystyrene particles on which the poly( -glutamic acid) was grafted. The resultant parti-
cles had 0.54e2.12 m in size and 0.2e2.6 residue/nm in surface density and showed a pH-responsive
colloidal behavior associated with a helixecoil transformation of the surface poly( -glutamic acid). Not
a-L
11 June 2015
2
m
Accepted 12 June 2015
Available online 20 June 2015
a-L
only the particle size but also the surface density were controlled with macromonomer concentration,
macromonomer length, DMF composition, and styrene concentration, while no consistent trend for AIBN
concentration was observed. A gel-permeation-chromatography curve of the particles was separated into
three components. We tentatively identify the origin of each component and propose a possibility that
unstable particles, which were generated even after the growing particles were stabilized, took an
important role in particle growth and size distribution of the resultant particles.
Keywords:
Macromonomer
Dispersion polymerization
Poly(L-glutamic acid)
©
2015 Elsevier Ltd. All rights reserved.
1
. Introduction
steric stabilizer adsorbed on the surface, narrowly-distributed
stimuli-responsive polymer particles have been prepared by us-
ing functional stabilizers such as poly[(2-dimethylamino)ethyl
methacrylate] [7,13,25], poly(N-isopropyl acrylamide) [6,23,24],
and polypeptide [15e18]. In particular, use of a macromonomer as a
stabilizer is attractive to prepare functional polymer particles
because of a wide variety of its synthetic routes. For example,
Armes et al. synthesized a poly[(2-dimethylamino)ethyl methac-
rylate] macromonomer by oxyanionic polymerization [25], and
Akashi et al. prepared a poly(N-isopropyl acrylamide) macro-
monomer by free radical polymerization with 2-mercaptoethanol
as a chain-transfer agent [23]. Postpolymerization modification
was also available for preparation of a poly(N-isopropyl acrylamide)
macromonomer by combination of RAFT polymerization and
Michael addition of 3-(acryloyloxy)-2-hydroxypropyl methacrylate
There has been a growing interest in functional particles
because of their fascinating properties for application in various
fields such as biomedical analysis, electronics, and chemical in-
dustry. Particle surface has a strong effect on particle function, so
that functionalization and precise control of the particle surface are
demanded. Dispersion polymerization is a useful technique to
prepare polymer particles with well-defined size, morphology, and
surface structure. In order to achieve colloidal stability, the
dispersion polymerization requires a steric stabilizer which is
covalently attached to particle surface to provide a sterically-
stabilizing effect. A variety of homopolymers [1e5], macro-
initiators [6,7], graft copolymers [8e11], block copolymers [11e17],
and macromonomers [18e26] have been reported as effective
stabilizers as well as macromolecular chain-transfer agents for
reversible additionefragmentation chain transfer (RAFT) poly-
merization in recent studies [11,27e29]. Taking advantage of the
[24]. In addition, Ono and Tomita prepared a poly(sodium
a,b-
aspartate) macromonomer by polycondensation of aspartic acid
followed by a reaction with acryloyl chloride [18]. These macro-
monomers successfully gave uniform-sized surface-functional
particles.
For advanced applications of the particles, not only surface
functionalization but also control of size, surface structure, and
*
Corresponding author.
E-mail address: itoh@ehime-u.ac.jp (T. Itoh).
http://dx.doi.org/10.1016/j.polymer.2015.06.022
032-3861/© 2015 Elsevier Ltd. All rights reserved.
0

184
T. Itoh et al. / Polymer 70 (2015) 183e193
composition of functional polymer particles are important subjects.
These subjects are strongly related to formation mechanism of
narrowly-distributed particles, that of surface structure, and poly-
merization kinetics. Particularly, adsorption behavior of a stabilizer
is a key element to control size and surface structure of particles.
The particle formation involving two periods of stabilization and
growth is explained as follows [2,19]. Dispersion polymerization
starts from a homogeneous solution of monomer, initiator, stabi-
lizer, and dispersion medium, while oligomeric radicals and dead
polymers are generated and aggregated to form unstable particles
which coalesce each other in the dispersion medium. When a
minimum amount of the stabilizer required for colloidal stability is
adsorbed, the unstable particles stop the coalescence and then no
further stable particle is newly formed. The stabilization is
completed within a few percent conversion [3,4], and, in the sub-
sequent period of particle growth, each particle has an equal chance
to grow. Further stabilizer chains are adsorbed on the particles
which enlarge their surface area along with their growth. Unless
the particle surface becomes scarce or overcrowded with the sta-
bilizer, the dispersion polymerization gives narrowly-distributed
polymer particles. Therefore, final size of the resultant particles is
determined by adsorption behavior of the stabilizer in the very
early stage, while surface structure reflects the stabilizer adsorption
in the growth period.
In order to precisely control particle function, it is of great
importance to study surface structure of functional particles. For a
macromonomer-stabilizing dispersion polymerization, Kawaguchi
and Ito reported an excellent model where particle size and surface
structure were theoretically controlled with monomer concentra-
tion, macromonomer concentration, molecular weight of macro-
monomer, and initiator concentration [19]. The theoretical
exponents for the particle size were validated with experimental
data. However, those for surface structure have not been numeri-
cally compared since no experimental value has not been available.
Previously, we carried out a dispersion polymerization stabilized by
components will lead us to useful information of the polymeriza-
tion kinetics in dispersion polymerization.
We try to focus on fine control of size, surface structure, and
composition for functional polymer particles prepared by disper-
sion polymerization stabilized with a functional macromonomer.
As an example, we use a polypeptide-based macromonmer as a
functional stabilizer for the study of the particle control. Poly-
peptides are typical multi-functional polymers having stimuli re-
sponsibility, biocompatibility, biodegradability, and side-chain
functionality. Therefore, polypeptide-functional polymer particles
are promising candidates for a broad range of application of
stimuli-responsive material [16e18,33e35]. First, in this study, a
novel functional macromonomer of poly(
(mPLGA ) was synthesized by ring-opening polymerization of an
amino acid N-carboxy anhydride, which is a controlled polymeri-
zation to give well-defined polypeptide [36e38]. Then, the mPLGA
a-L-glutamic acid)
n
n
was used as a stabilizer for dispersion polymerization of styrene to
give pH-responsive narrowly-distributed polystyrene particles. The
use of the well-defined mPLGA
structure of the resultant particles on which the mPLGA
grafted. Dependences of particle size and surface density on several
conditions of initial concentration of the mPLGA , chain length of
the mPLGA , medium composition, styrene concentration, and
n
enabled us to clarify the surface
n
was
n
n
initiator concentration were investigated and compared with the
theoretical exponents. Finally, we try to discuss the relation be-
tween polymerization kinetics and formation mechanism based on
compositions of the surface-functional particles.
2. Experimental section
2.1. Materials
N,N-Dimethylformamide (DMF; Nacalai Tesque, Kyoto, Japan,
99%) was dried over calcium hydride and distilled before use.
Styrene (Nacalai Tesque, 98%) was washed with 5 wt% aqueous
sodium hydroxide, dried over magnesium sulfate then calcium
a polystyrene-block-poly(L-glutamic acid) or a polystyrene-block-
0
poly( -lysine), and measured the amount of surface-adsorbed sta-
L
hydride, and distilled before use. 2,2 -Azoisobutyronitrile (AIBN;
bilizer by a pH titration to provide a power law dependence on
polymerization conditions for surface structure [16,17]. This tech-
nique can be applied to a macromonomer-stabilizing dispersion
polymerization.
Nacalai Tesque, 98%), benzylalcohol (Wako Pure Chemical In-
dustries, Osaka, Japan, 99%), tert-buthylamine (Tokyo Chemical
Industry, Tokyo, Japan, 98%), p-tert-butylcatechol (Nacalai Tesque,
98%), 4-chloromethylstyrene (Tokyo Chemical Industry, 90%),
diethyl ether (Wako Pure Chemical Industries, 99%), ethanol
Polymerization kinetics of dispersion polymerization is another
important subject for the precise control of functional particles.
However, the kinetics is still controversial because of following
mechanistic complexities. During particle growth, bulk and solu-
tion polymerizations take place in particle phase and in dispersion
medium, respectively, in the same period. Oligomeric radicals
generated in the dispersion medium cause bimolecular termination
or transfer into the particle phase. Although stable particles are not
newly generated under the colloidal stability, unstable particles are
still generated in the dispersion medium and incorporated into the
stable ones before stabilized as secondary particles. For a poly(N-
vinylpyrrolidone)-stabilizing dispersion polymerization, Paine et al.
reported that molecular weight of particles had an inverse corre-
lation to particle size because, in the case of small particles, large
surface area of total particles could cause efficient capture of
solution-initiated oligomeric radicals [1]. In contrast, El-Aasser
et al. found that polymerization rate was not dependent on parti-
cle size but on medium composition due to partitioning behavior of
initiator, oligomeric radicals, and monomer between dispersion
medium and particle phase [30e32]. While they discussed a value
of molecular weight, we found that a GPC curve of the resultant
particles was successfully separated into three components in the
study of a block-copolymer-stabilizing dispersion polymerization
(Wako Pure Chemical Industries, 99%), L-glutamic acid (Wako Pure
Chemical Industries, 99%), hydrazine monohydrate (Nacalai Tes-
que, 80%), 25% hydrogen bromideeacetic acid solution (Wako
Pure Chemical Industries), lithium chloride (Nacalai Tesque, 98%),
methanol (Wako Pure Chemical Industries, 99%), phthalimide
potassium salt (Nacalai Tesque, 98%), sodium hydroxide (Nacalai
Tesque, 97%), trifluoroacetic acid (Nacalai Tesque, 99%), tri-
phosgene (Tokyo Chemical Industry, 98%), and urea (Nacalai Tes-
que, 99%) were used as received.
2.2. Methods
1
1
H nuclear magnetic resonance ( H NMR; 400 MHz) spectra
were recorded in a Bruker Avance 400 spectrometer (Bruker,
1
Rheinstein, Germany). The H NMR measurements were taken in
20% trifluoroacetic acidechloroform-d mixture, dimethylsulfoxide-
6 2
d , or 20% sodium hydroxideeD O solution. Helix content (fH,NMR)
a
was determined from the ratio of signal intensities assigned to C H
protons of residues in a helix at 4.1 p.p.m. to those in the other
conformation in 4.4e4.2 p.p.m [39].
Gel-permeation-chromatography (GPC) measurements were
carried out by a Jasco-Borwin system (version 1.50; Jasco, Tokyo,
Japan) equipped with Waters styragel HR4 THF and HR2 THF
[16]. We have expected that a careful consideration of particle

T. Itoh et al. / Polymer 70 (2015) 183e193
185
ꢀ
calibrated with six polystyrene standards (M
n
¼ 660 000, 210 000,
(80 mL). After stirring for 4 h at 50 C, the solution was evaporated
and then the residue was dissolved in chloroform. The solution was
washed with 5% sodium hydroxide aqueous solution and then
water for several times, dried over magnesium sulfate, filtrated, and
evaporated. Purification of the residue was performed by recrys-
5
0 000, 23 000, 5700, and 2400) using tetrahydrofuran as an
eluent. Analysis of a GPC curve was performed by using Origin Pro
software equipped with a peak-fitting module (version 7.5J; Ori-
ginLab, Northampton, MA).
Scanning-electron-microscopic studies were conducted using a
JEOL JSM-5310 (Japan Electron Optics Laboratory, Tokyo, Japan).
tallization from methanol solution for several times to give 4-
1
vinylbenzyl phthalamide as white crystals (yield: 23 g, 67%).
NMR (DMSO-d , 400 MHz),
phthalimide), 7.4e7.2 (m, 4H, aromatic H for benzyl), 6.7 (dd, 1H,
eCH ]CHe), 5.8e5.2 (d, 2H, eCH ]CHe), 4.8 (s, 2H, eCH eNe).
H
n
Average diameter of particles (D ) and particle size distribution
6
d (p.p.m.): 7.9e7.8 (m 4H, aromatic H for
(D
w
/D ) were determined by following equations:
n
P
2
2
2
n
Di ;
The 4-vinylbenzyl phthalimide (5.1 g, 19 mmol), 80% hydrazine
monohydrate (1.6 g, 32 mmol), and p-tert-butylcatechol (50 mg,
i¼1
D
n
¼
n
0.30 mmol) were dissolved in ethanol (50 mL). After reflux for 4 h,
P
n
4
D
i
3
D
i
the solution was adjusted to pH 1e2 by adding 1 M hydrochloric
acid, filtrated, and then adjusted to pH 10e11 with 1 M sodium
hydroxide aqueous solution. The resultant was extracted to diethyl
ether for several times. The organic phase was washed with water,
dried over magnesium sulfate, filtrated, and evaporated. The crude
i¼1
D
w
¼ P
;
n
i¼1
where D
i
is the diameter of the ith particle measured by using Mac-
View software (version 4; Mountech, Tokyo, Japan) and n is the
number of particles (>100).
oil was further purified by vacuum distillation under calcium oxide
1
to give a clear oil (yield: 1.6 g, 62%). H NMR (CDCl
d
3
, 400 MHz),
]CHe),
Circular dichroism spectra were recorded on a Jasco J-820
spectrometer using a 1 mm jacked quartz cell in the wavelength
(p.p.m.): 7.4e7.2 (m, 4H, aromatic H), 6.7 (dd, 1H, eCH
2
ꢀ
5.8e5.2 (d, 2H, eCH ]CHe), 3.8 (s, 2H, eCH eNe).
2
2
region of 195e250 nm at 0 C. Helix content (fH,CD) was determined
by the following equation:
2.4. Synthesis of poly(g-benzyl-L-glutamate) macromonomer
ꢁ
½qꢂ
2
22
fH;CD
¼
;
g
-Benzyl-
L
-glutamate N-carboxy anhydride (14 g, 52 mmol)
4
0000
222 is the molar ellipticity (deg cm2 dmol ) at 222 nm
which was prepared from
phosgene [41], 4-vinylbenzyl amine (70 mg, 0.53 mmol), and urea
(140 mg) were dissolved in dry DMF (150 mL) under nitrogen at-
L-glutamic acid benzyl ester with tri-
ꢁ1
where [
40].
Titrations of adsorbed PLGA onto particle surface were carried
q
]
[
ꢀ
mosphere. After stirring for 7 days at 0 C, the solution was poured
out by a TOADKK AUT-701 automatic titrator (TOA-DKK, Tokyo,
Japan). In a typical case, polystyrene particles (100 mg) were added
to 5.0 mM aqueous sodium hydroxide (8 mL) and titrated with
into 70/30 (wt/wt) methanolewater. The precipitant was dissolved
in chloroform, purified by reprecipitation for several times, and
thoroughly dried under vacuum (yield: 9.0 g, 80%). H NMR (20%
1
1
0 mM hydrochloric acid. On the assumption that a chain terminal
of the mPLGA is located at the particle surface, surface density of
glutamic acid residue ( ), surface area occupied by one PLGA chain
S), and the ratio of the mPLGA consumed for surface coverage to
trifluoroacetic acideCDCl , 400 MHz),
d
(p.p.m.): 7.9 (NH), 7.3 (ar-
3
n
L-
omatic H), 6.7e6.6 (CH ]CHe), 5.7 and 5.2 (CH ]CHe), 5.0
2
2
a
g
b
s
2 2 2
(benzyl CH ), 4.6 (C H), 2.5 (C H ), 2.1 and 1.9 (C H ).
1
(
n
Degree of polymerization (n) was determined from the H NMR
signal intensity ratio of the backbone C H proton at 4.6 p.p.m. to
that of the C-terminal vinyl proton at 5.7 p.p.m. Molecular weight
a
that used for the reaction mixture (
following equations:
q) were determined from the
w n
distribution (M /M ) was determined by a GPC calibrated with
polystyrene standards using DMF solution of 8 M lithium bromide
as an eluent.
CHCl
n
T
D rV N
A
s ¼
;
6W
p
S ¼ ns
;
2.5. Synthesis of poly(
a
-
L
-glutamic acid) macromonomer (mPLGAn)
-benzyl-L-
To an ice-cooled DMF solution (20 mL) of the poly(
g
CHClVTMn;PLGA
glutamate) macromonomer (3.0 g; n ¼ 100) and lithium chloride
q ¼
;
nW
p
YCPLGA
(8.5 mg), 9.4 M sodium hydroxide aqueous solution (40 mL) was
ꢀ
added dropwise. After stirring for 1.5 h at 0 C, the solution was
where CHCl is the concentration of the hydrochloric acid used for
the titration (10 mM), is the density of the polystyrene particles
is the difference of the titer between the two in-
flections in the titration curve, N is the Avogadro number, W is the
weight of particle feed for titration, n is the number of repeating
units in the PLGA segment, Mn,PLGA is the molecular weight of the
added to an ice-cooled 1 M hydrochloric acid (900 mL) and placed
ꢀ
r
at 4 C overnight. The precipitant was collected by filtration and
ꢁ
3
1
(
1.05 g cm ), V
T
thoroughly dried under vacuum (yield: 1.2 g, 66%). H NMR (20%
NaOHeD
(CH ]CHe), 5.9 and 5.4 (CH
2.1 and 1.9 (C H
2
O, 400 MHz), d (p.p.m.): 7.5 and 7.3 (aromatic H), 6.8
A
p
a
g
2
2 2
]CHe), 4.3 and 4.1 (C H), 2.3 (C H ),
b
2
).
n
mPLGA , Y is the yield of the polymer particles, and CPLGA is the
initial concentration of the mPLGA
the reaction mixture.
n
relative to styrene monomer in
2.6. Dispersion polymerization of styrene
In a typical run, styrene (300 mg, 2.9 mmol), AIBN 5 mg
0.03 mmol), and mPLGA100 (15 mg) were dissolved in 75/25 (wt/
(
2.3. Synthesis of 4-vinylbenzyl amine
wt) DMFewater mixture (5.4 g) in a 16 mm-diameter glass tube.
The solution was degassed by several freeze-pump-thaw cycles and
4
-Vinylbenzyl amine was synthesized as reported by Heise et al.
34] Phthalimide potassium (26 g, 140 mmol) and 4-
chloromethylstyrene (20 g, 130 mmol) were dissolved in DMF
sealed off under vacuum. The tube was immersed in an oil bath at
ꢀ
[
70 C for 17 h. The resultant polymer was filtrated over a 0.45
mm
membrane filter, washed with methanol repeatedly, and dried

186
T. Itoh et al. / Polymer 70 (2015) 183e193
ꢀ
under vacuum at 40 C to give polystyrene particles (yield: 270 mg,
Table 1
Synthesis of poly(g-benzyl-L-glutamate) macromonomers.
8
9%).
Run
[BLG-NCA]
0
/[VBA]
0
Time (d)
Yield (%)
M
w
/M
n
n
1
2
3
10/1
100/1
500/1
1
7
7
73
80
81
1.10
1.12
1.39
15
100
550
3
. Results and discussion
3.1. Synthesis and characterization of mPLGA
n
ꢀ
Polymerization conditions: BLG-NCA (10 wt%); urea (0.1 wt%); 0 C; dry DMF. Ab-
breviations: BLG-NCA,
vinylbenzyl amine;
g
-benzyl-
L-glutamate N-carboxy anhydride; VBA, 4-
A mPLGA
poly( -benzyl-
opening polymerization of
n
was synthesized as shown in Scheme 1. First, a
-glutamate) macromonomer was prepared by ring-
-benzyl- -glutamate N-carboxy anhy-
M
w
/M molecular weight distribution determined by a
n
,
g
L
polystyrene-calibrated gel-permeation chromatography; n, degrees of polymeriza-
tion determined by 1H NMR.
g
L
dride (BLG-NCA) with 0.01 equivalent of 4-vinylbenzyl amine as an
ꢀ
initiator at 0 C for 7 days in dry DMF containing 1 M urea (run 2 in
1
O
H
e
Table 1). H NMR spectra in Fig. 1a, which was measured in a 20%
a
a)
a
b
c
a
NH C CH N
trifluoroacetic acidechloroform-d solution, shows signals assigned
to protons in a poly(g-benzyl-L-glutamate) and a C-terminal
d
b
CH2
CH2
n
f
g
6
.5 6.0 5.5
vinylbenzyl group. Degree of polymerization (n) was determined
C
i
d,h
O
O
from a ratio of signal intensity of 4.6 and 5.7 p.p.m., which were
c,i
g
a
CH
2
h
assigned to backbone C H and C-terminal vinyl protons, respec-
e
f
tively, to be n ¼ 100. Molecular weight distribution was measured
by polystyrene-calibrated gel-permeation-chromatography (GPC)
to give M
BLG-NCA]
Table 1.
The side-chain benzyl ester group of the poly(
glutamate) macromonomer was subjected to hydrolysis under an
w
/M
n
¼ 1.12. Other polymerization results with different
8
6
4
2
0
chemical shift (ppm)
[
0
/[4-vinylbenzyl amine]
0
feed ratios are summarized in
O
H
b)
c’
e’
a’ c’
g
-benzyl-
L-
NH C CH N
d’
g’
HDO
a’ a’
b’
f’CH2
g ’C H2
n
d’ b’
1
alkaline condition to yield a mPLGA
n
. Fig. 1b shows H NMR spec-
7
.5 7.0 6.5 6.0 5.5
C
trum of the product. The signal of side-chain benzyl protons
completely disappeared, while other signals for a PLGA and the C-
terminal vinylbenzyl group remained intact. The degree of poly-
e’
O
OH
f’
merization of the mPLGA
quantitative deprotection of the side-chain group without apparent
n
was kept at n ¼ 100. This indicates
8
6
4
2
0
cleavage in the peptide backbone. Other mPLGA
50 were also successfully prepared.
A solvent effect on a mPLGA
NMR spectroscopy, where a C H proton signal is useful to deter-
mine the conformation of polypeptides [39]. We measured H NMR
n
s with n ¼ 15 and
chemical shift (ppm)
5
Fig. 1. 1H NMR spectra for a) poly(
g-benzyl-L-glutamate) macromonomer with n ¼ 100
1
n
conformation was studied with H
a
in 20% trifluoroacetic acideCDCl
3
2
and b) mPLGA100 in 20% NaOHeD O.
1
of the mPLGA100 in 75/25 (wt/wt) DMF-d
including 10 wt% styrene at 70 C, which is a condition of a
dispersion polymerization in this study. As shown in Fig. 2a, the
7
eD
2
O mixed solvent
ꢀ
a)
b)
2
0000
pH 10
7
mPLGA100 clearly showed a signal at 4.1 p.p.m., which is assignable
4
.6
4.4
4.2
4.0
0
a
to a C H proton of a
L-glutamic acid residue in an
a
-helix. In addi-
0
tion, a signal at 4.4 p.p.m. and an intermediate broad one at
.3e4.2 p.p.m., which are accounted for formation of a random coil
pH 4
4
-
20000
and for an interconversion between helical and coiled forms,
respectively, were observed because of rather short chain of the
200
220
240
8
6
4
2
wavelength/nm
1
chemical shift (ppm)
mPLGA100 [39]. From the H NMR signal intensities, helix content of
the mPLGA100 (fH,NMR) was determined to be 57%. The fH,NMR value
1
Fig. 2. a) H NMR spectra for the mPLGA100 in 75/25 (wt/wt) DMF-d
7
eD
2
O including
did not change when the DMF-d
7
eD
2
O composition was varied in
10 wt% styrene at 70 ꢀC. b) Circular dichroism of the mPLGA
in aqueous solution at
100
the range of 75/25e85/15, indicating no apparent effect of the DMF
composition on the mPLGA100 conformation.
pH 4, 7, and 10, respectively.
A helixecoil transformation of PLGA induced by a pH change in
an aqueous solution has been reported [40]. The transformation is
caused by carboxylation of the side-chain groups which leads to an
increase in solubility of the side chain and electrostatic repulsion
between neighboring glutamate units. To investigate the pH-
n
Scheme 1. Synthesis of mPLGA .

T. Itoh et al. / Polymer 70 (2015) 183e193
187
responsive property of the mPLGA100, circular dichroism measure-
ments were performed on its aqueous solutions. Two minima at 208
and 222 nm were observed at pH 4 as shown in Fig. 2b, indicating
formation of an
a-helical structure. Helix content from the circular
dichroism (fH,CD) was determined from the molar ellipticity per
2
ꢁ1
amino acid at 222 nm ([
In contrast, only one minimum at 196 nm was observed at pH 7 and
0, indicating formation of a random-coiled structure under these
q
]
222) of ꢁ20,000 deg cm dmol to be 50%.
1
conditions. The observed pH-induced helixecoil transformation of
the mPLGA100 was in accordance with that for a PLGA homopolymer
[40], but the smaller fH,CD value could be ascribed to rather low
molecular weight of the mPLGA100
.
3.2. Dispersion polymerization of styrene with mPLGA100 and
characterization of the resultant particles
As an example, a dispersion polymerization was conducted from
a homogeneous solution of styrene (5 wt%), the mPLGA100 (0.25 wt
%
), AIBN (0.08 wt%; [styrene]
0
/[AIBN]
0
¼ 100/1), and 75/25 (wt/wt)
DMFewater medium at 70 C. After 17 h, narrowly-distributed
spherical polystyrene particles with D m and D
¼ 1.24
¼ 1.01 were obtained at 93% yield as shown in Fig. 3. In contrast,
Fig. 4. The PLGA-functionalized polystyrene particles dispersed in aqueous solutions
of hydrochloric acid (pH 2), sodium chloride (pH 7), and sodium hydroxide (pH 12),
and placed overnight, respectively. The colloidal stability remained at pH 7 and 12,
while sedimentation was observed at pH 2.
ꢀ
n
m
w
/
D
n
massive coagulum was observed when a PLGA homopolymer with
n ¼ 100, which was prepared by ring-opening polymerization of
BLG-NCA with tert-butylamine as an initiator, was used in place of
the mPLGA100, indicating that the C-terminal vinylbenzyl group
played an important role in the formation of the spherical
morphology. It is apparent that a graft copolymer composed of a
polystyrene anchor (stem) and the PLGA100 segment (graft) was
generated and adsorbed on the polystyrene particles during the
dispersion polymerization to achieve the colloidal stability.
Fig. 2b. This indicates that the novel macromonomer, mPLGA100
,
was an effective functional stabilizer for the dispersion polymeri-
zation to prepare the stimuli-responsive polymer particles.
The amount of surface-attached PLGA chains was measured by a
pH titration, where the polystyrene particles were added to
aqueous sodium hydroxide and back titrated with hydrochloric
acid. Two inflections of neutralization of free sodium hydroxide and
titration of side-chain carboxylic groups in the PLGA were observed.
We briefly checked the function of the polystyrene particles. The
resultant particles were dispersed in aqueous solutions of hydro-
chloric acid (pH 2), sodium chloride (pH 7), and sodium hydroxide
From the difference of the two titers, surface density of
acid residues ( ) was determined to be
¼ 0.76 residue/nm and,
hence, the surface area occupied by one PLGA segment (S) with
L-glutamic
2
s
s
(
pH 12) by ultrasonication, respectively, and stood still overnight.
2
n ¼ 100 was S ¼ 130 nm /chain. On the assumption of an ideal
helical conformation with the length of 0.15 nm per amino acid
residue, random distribution of the -helical PLGA chains with
n ¼ 100 requires S ¼ 180 nm /chain. When the surface PLGA
segment includes random-coiled residues in addition to the -he-
lix, which is the most probable from the fH,NMR ¼ 57% in the 75/25
DMF-d eD O solution as described above, even larger area is
a-
Colloidal stability remained at pH 7 and 12, while sedimentation of
the particles was observed at pH 2 as shown in Fig. 4. The colloidal
stability is attributed to electrostatic repulsion among the particles
caused by carboxylation of side-chain groups in the PLGA segment
a
2
a
(
pK
a
¼ 4), supporting the presence of the mPLGA100 graft chains on
the particle surface. The pH-responsive colloidal behavior is related
7
2
to the pH-dependent helixecoil transformation of the mPLGA100 in
required for the random distribution of such a PLGA chain with
mixed conformation since the long axis of the PLGA in a random
coil should be elongated by the possible formation of a polyproline
1
II or a 2.5 -helix with 0.30e0.35 nm per residue [42,43]. Therefore,
it is conceivable that the PLGA segments lied with some over-
lapping on the polystyrene particles.
From the pH titration, the ratio of the mPLGA100 consumed for
surface coverage to that used for the reaction mixture is calculated
to be
q
¼ 2.3%. We expected that the low
q value was attributed to
the PLGA fraction embedded into the particle core and/or that
dissolved into the dispersion medium. In order to check the
embedded fraction, an infrared measurement for the resultant
particles was carried out. However, any signals for the PLGA fraction
were not observed, indicating that the embedded fraction was
negligible. For the dissolved fraction, the dispersion medium of the
reaction and the wash solvent were collected and thoroughly dried.
Non-volatile residue was obtained at 6% yield, which was some-
what larger than the weight fraction of the used mPLGA100 at 5 wt%
1
to the styrene monomer. A H NMR measurement of the residue
indicated signals assigned to the protons in the PLGA along with
unidentified impurities. From the integrals of the proton signals,
the extent of the PLGA fraction based on the total proton areas was
Fig. 3. A scanning-electron-microscopic photograph for polystyrene particles prepared
by dispersion polymerization of styrene with mPLGA100 as a stabilizer.
74%. It is suggested that the most of the PLGA fraction which was

188
T. Itoh et al. / Polymer 70 (2015) 183e193
2
not grafted on the resultant particles remained in the dispersion
medium.
in the increase in surface density from
along with twice increase in total particle surface.
s
¼ 0.2 to 1.3 residue/nm ,
An effect of chain length of the macromonomer was investi-
gated by using 0.5 wt% mPLGA
50 in addition to the mPLGA100. The increase in the n value
resulted in an enlargement of the particles from D
¼ 1.06 to
.99 m, indicating that the use of a larger mPLGA produced a
smaller number of stable particles. Because weight percentage of
the mPLGA concentration was the same at 0.50 wt%, mole per-
centage of the mPLGA550 was 34 times smaller than that of the
mPLGA15, which is responsible for the decrease in the stable par-
ticles. In addition, it is reasonable to assume that the large soluble
PLGA segment enhanced solubility of the graft copolymer in the
dispersion medium and, thus, weakened its adsorption. Although
total surface area of the particles decreased by half, a much larger
n
s with repeating unit of n ¼ 15 and
3
.3. Control of particle size and surface structure
5
n
We study effects of several conditions of dispersion polymeri-
1
m
n
zation on particle size and surface density to compare the experi-
mental data with the theoretical model [19]. For particle size, the
theoretical exponents of macromonomer concentration, macro-
monomer length, and monomer concentration have been reported
to be ꢁ0.50, 0.50, and 0.67, respectively, whereas those for surface
area occupied by one macromonomer chain, S, have been described
to be ꢁ0.50, 0.50, and 0.33, respectively. In this study, however, we
discuss a formation mechanism which is strongly related to surface
amount or surface density rather than exclusion volume of a sta-
bilizer chain on particles [2,19]. Moreover, when chain length of a
macromonomeric stabilizer is varied, comparison of the S may not
be appropriate for our investigation. Therefore, we use surface
n
decrease in the surface-coverage efficiency from
caused by the increase in the n value, resulting in 4 times dilution in
the surface density of the -glutamic acid. This indicates that the
q
¼ 4 to 0.6% was
L
weak adsorption of the graft copolymer continued until the end of
the dispersion polymerization.
density of
L-glutamic acid residue,
s, rather than S in this study.
Because of the inverse relation between S and
s
values, S ¼ n/ , the
s
From the results described above, dependency of the particle
size on the macromonomer is represented to be
theoretical exponents of macromonomer concentration, macro-
monomer length, and monomer concentration for surface density
are calculated to be 0.50, ꢁ0.50, and ꢁ0.33, respectively.
ꢁ
0.13 0.19
D
n
f [mPLGA100
]
0
n
. Our experimental exponents are far
from the theoretical values of ꢁ0.50 and 0.50 for macromonomer
concentration and length, respectively [19]. For surface density, in
3.3.1. Time evolution
0.61 ꢁ0.63
contrast, the dependency is
s
f [mPLGA100
]
0
n
, indicating a
Dispersion polymerization was performed between 1 and 17 h
stronger effect in the experiment than that in the theoretical
model with exponents of 0.50 and ꢁ0.50, respectively [19]. Since
particle size and surface density reflected adsorption behavior of
graft copolymers at the stabilization and at the end of the
dispersion polymerization, respectively, the experimental values
indicate that the influence of the macromonomer became greater
as the dispersion polymerization proceeded in the growth period.
From the extremely-small value of surface coverage efficiency less
in a homogeneous solution of styrene (10 wt%), mPLGA100 (0.5 wt
%
DMFewater medium at 70 C. As the polymerization proceeded,
particle size increased from 0.54 to 1.36 mm accompanied by an
increase in the yield from 5 to 93 wt% as listed in Table 2. However,
the number of particles prepared from 1 g of styrene (N) was not
varied at N ¼ 0.6 ꢃ 10 e0.7 ꢃ 10 , clearly indicating that the
polystyrene particles were formed in the manner of stabilization
and growth, which is the well-accepted mechanism for dispersion
polymerization [2,19]. The stabilization was achieved within 5% of
the monomer conversion and then each particle had an equal
chance to grow with further monomer conversion, giving
narrowly-distributed polystyrene particles. Hence, it is reasonable
that the particle size and the surface density provide us with in-
formation of the stabilizer adsorption in the stabilization and the
growth periods, respectively.
), AIBN (0.16 wt%; [styrene]
0
/[AIBN]
0
¼ 100/1), and 75/25
ꢀ
12
12
than
q
¼ 4%, a large amount of unreacted macromonomers or
soluble graft copolymers remained in the dispersion medium,
even at the end of the dispersion polymerization. Two reasons are
assumable for the strong effect in the later stage of the dispersion
polymerization. One is that the relative concentration of the
mPLGA
n
to remaining styrene monomer increased with the
monomer conversion to lead to enhanced generation of the graft
copolymer. The other is that the styrene conversion reduced the
solubility of the polystyrene anchor in the medium, so that a part
of the dissolved graft copolymers were adsorbed on the particles.
3
.3.2. Macromonomer concentration and length
Initial concentration of the mPLGA100 was varied in the range of
.05e1.0 wt%, while the other conditions of styrene (10 wt%), AIBN
In contrast to the small
ciency over
oxide)-macromonomer-stabilizing dispersion polymerizations
[20e22]. Further experimental studies are necessary to investi-
gate the effect of macromonomeric stabilizers on the particle size
and the surface density.
q, several values of surface coverage effi-
¼ 10e100% have been reported for poly(ethylene
q
0
ꢀ
(
0.16 wt%), 75/25 DMFewater medium, 17 h, and 70 C were fixed.
As listed in Table 3, the increase in the mPLGA100 concentration
caused a decrease in the particle size from D
accompanied by 8 times increase in the number of the particles.
The 20 times increase in the amount of the used mPLGA100 resulted
n
¼ 1.98 to 0.98
mm
3.3.3. Solvent effect
A solvent effect derived from medium composition on the
dispersion polymerization was investigated. As the DMF composi-
tion increased from 75 to 85 wt%, an increase in the particle size
along with a decrease in the number of particles was observed,
while surface-coverage efficiency remained at a constant value of
Table 2
Effect of polymerization time on the dispersion polymerization of styrene in the
presence of mPLGA100
.
Run
Time (h)
Yield (%)
D
n
(
m
m)
D
w
/D
n
N (1012
)
q
¼ 2% as listed in Table 4. We expected that the particles could be
4
5
6
1
6
17
5
87
93
0.54
1.27
1.36
1.05
1.07
1.02
0.60
0.70
0.67
controlled by a solvent-induced conformational change of the
mPLGA100. In reality, the conformation was not affected by the
small increase in the DMF composition, which was observed in the
Polymerization conditions: styrene (10 wt%); mPLGA100 (0.5 wt%); AIBN (0.16 wt%;
1
same range of the DMF-d
7
eD
2
O composition as described in the H
ꢀ
[styrene]
0
/[AIBN]
0
¼ 100/1); 75/25 DMFewater medium; 70 C. Abbreviations: D
n
,
NMR study. Therefore, the increase in the particle size is attributed
to the solvency of the dispersion medium in the early stage of the
dispersion polymerization where most of the styrene monomer
number average diameter; D
ticles prepared from 1 g of styrene monomer; M , number average molecular weight
n
of polystyrene forming particle core.
w n
/D , particle size distribution; N, the number of par-

T. Itoh et al. / Polymer 70 (2015) 183e193
189
Table 3
n
Effect of mPLGA on polystyrene particles prepared by dispersion polymerization of styrene.
Run
mPLGA
n
(wt%)
n
Yield (%)
D
n
(
m
m)
D
w
/D
n
N (1012
)
s
(residue/nm2)
q (%)
7
8
9
1
1
0.05
0.50
1.0
0.50
0.50
100
100
100
15
94
92
89
92
92
1.98
1.34
0.98
1.06
1.99
1.08
1.01
1.02
1.01
1.04
0.22
0.70
1.7
1.4
0.21
0.20
0.82
1.3
1.8
0.45
2.4
1.5
1.6
4.2
0.6
0
1
550
ꢀ
Polymerization conditions: styrene (10 wt%); AIBN (0.16 wt%; [styrene]
0
/[AIBN]
0
¼ 100/1); 75/25 DMFewater medium 70 C; 17 h. Abbreviations: D
n
, number average
, surface-coverage
diameter; D /D , particle size distribution; N, the number of particles prepared from 1 g of styrene monomer;
w
n
s
, surface density of
L
-glutamic acid residues;
q
efficiency. Polystyrene particles in run 8 and run 6 (Table 2) were prepared under the same condition but in different batches.
Table 4
Solvent effect on polystyrene particles prepared by dispersion polymerization of styrene in the presence of mPLGA100
.
Run
DMFewater (wt/wt)
Styrene (wt%)
Yield (%)
D
n
(
m
m)
D /D
w n
N (1012
)
s
(residue/nm2)
q (%)
1
1
1
1
2
3
4
5
80/20
85/15
75/25
75/25
10
10
5
90
78
93
92
1.54
2.12
1.24
1.96
1.01
1.05
1.01
1.06
0.45
0.15
0.83
0.20
0.94
2.6
0.76
1.2
1.5
2.5
2.3
1.8
15
ꢀ
Polymerization conditions: mPLGA100 (5 wt% to styrene); AIBN (1.6 wt% to styrene; [styrene]
0
/[AIBN]
0
¼ 100/1); 70 C; 17 h. Abbreviations: D
n
, number average diameter; D
w
/
D
n
, particle size distribution; N, the number of particles prepared from 1 g of styrene monomer;
s
, surface density of
L
-glutamic acid residues; q, surface-coverage efficiency.
remained unreacted. Solubility parameters of polystyrene, styrene,
initiator concentration of ꢁ0.083 for the both parameters [19],
0.55
1
/2
ꢁ0.39
0
and DMF have been reported to be 17e18, 19.0, and 24.8 MPa
,
giving D
n
f [styrene]
0
and
s
f [styrene]
. Both the exponents
respectively [44], indicating that the solvency of styrene to the
polystyrene anchor was stronger than that of DMF. One possible
explanation is that the styrene monomer amplified the solvency
difference between 75/25 and 85/15 DMFewater mediums. The
increase in the DMF in the presence of 10 wt% styrene monomer
could make the polystyrene anchor more soluble leading to a weak
adsorption of the graft copolymer, which caused frequent coales-
cence of unstable particles in the stabilization period. In contrast,
are in agreement with the theoretical values.
3.3.4. Radical concentration
Initial concentration of AIBN was varied at 0.03, 0.16, and
0.80 wt% ([styrene] /[AIBN] ¼ 500/1, 100/1, and 20/1, respectively)
0
0
and results are listed in Table 5. The theoretical exponent for the
size dependence on the initiator concentration is ꢁ0.083 [19]. In
this study, no consistent trend was observed in particle size,
although rather small particles were obtained at the intermediate
AIBN concentration of 0.16 wt% (run 8 in Table 3). We previously
studied a dispersion polymerization stabilized by a block copol-
ymer in which a polystyrene anchor was attached to a PLGA
segment beforehand, and found no apparent effect of AIBN con-
centration on particle size [16]. Therefore, the AIBN effect in this
study is attributed to its role in generation of the polystyrene an-
chor on the mPLGA100. The surface-coverage efficiencies of the
resultant particles prepared with 0.03 and 0.80 wt% AIBN were
smaller than that with 0.16 wt% AIBN. The large particles and the
low surface-coverage efficiency indicate weak adsorption of the
graft copolymer on the particles throughout the dispersion poly-
merization. The initial AIBN concentration would be related to
chain length of the polystyrene anchor. In other words, a large
amount of AIBN would give a small anchor to result in the weak
adsorption of the graft copolymer. At the lowest concentration of
the initial AIBN, on the other hand, it is assumable that a graft
copolymer was composed of a large polystyrene anchor and more
than one PLGA100 segment, which enhanced solubility of the
copolymer to the dispersion medium, giving the small number of
stable particles and the low efficiency of surface coverage.
the constant q value indicates that the small increase in the DMF did
not markedly change the solubility of the graft copolymers at the
end of the dispersion polymerization where the styrene as an
amplifier was completely consumed. The enlargement of the par-
ticles with the constant surface-coverage efficiency resulted in an
2
increase in surface density from
s
¼ 0.82 to 2.6 residue/nm . It is
interesting that the surface density was positively correlated to the
particle size, while an inverse correlation was observed when the
mPLGA100 concentration was varied. The dependences of the par-
ticle size and the surface density on the DMF composition are
0
.60
1.8
represented to be D
n
f [DMF]
and s f [DMF] .
It is known that styrene monomer gives a solvent effect on
dispersion polymerization [1]. Initial concentration of styrene was
varied in the range of 5e15 wt%, while relative concentration of
AIBN and mPLGA100 were kept at 1.6 and 5 wt% to the monomer,
respectively. The effect of styrene was similar to that of DMF as
expected; that is, an increase in the initial styrene concentration
caused increases in particle size and surface density as listed in
Table 4. When styrene monomer was completely consumed, the
initial styrene concentration did not have any effect on the medium
solvency, so that the constant value of the surface coverage effi-
ciency at
q
¼ 2% is reasonable. The theoretical exponents of
monomer concentration have been reported to be 0.67 and ꢁ0.33
for particle size and surface density, respectively [19]. Note that the
theoretical values reflect contribution of the monomer concentra-
tion without any effect of macromonomer and initiator concen-
trations. Contrary to the reported model, the concentrations of
mPLGA100 and AIBN were varied in proportion to the styrene con-
centration in this study. For the comparison, the power law de-
pendences are derived by a correction of the experimental
exponent of the macromonomer concentration of ꢁ0.13 and 0.61
for the size and the surface, respectively, and the theoretical one of
3
.3.5. Comparison with a block-copolymeric stabilizer
In addition to a macromonomeric stabilizer, use of a block
copolymer is another attractive technique for preparation of finely-
controlled functional particles. Previously, we reported a dispersion
polymerization in the presence of a polystyrene-block-poly(L-glu-
tamic acid) as a stabilizer, which also gave PLGA-grafted poly-
styrene particles [16]. The block-copolymer-stabilizing particles
had smaller size in the range of D
surface-coverage efficiency at
¼ 5e20%, indicating that the
polystyrene segment enhances adsorption to the particles in both
n
¼ 0.48e1.36
mm with larger
q

190
T. Itoh et al. / Polymer 70 (2015) 183e193
Table 5
Effect of AIBN concentration on polystyrene particles prepared by dispersion polymerization of styrene in the presence of mPLGA100
.
Run
AIBN (wt%)
Yield (%)
D
n
(
m
m)
D
w
/D
n
N (1012
)
s
(residue/nm2)
q (%)
1
1
6
7
0.03
0.80
90
85
1.82
1.79
1.06
1.02
0.27
0.27
0.53
0.96
0.7
1.3
ꢀ
Polymerization conditions: styrene (10 wt%); mPLGA100 (0.5 wt%); 75/25 DMFewater medium; 70 C; 17 h. Abbreviations: D
n w n
, number average diameter; D /D , particle size
distribution; N, the number of particles prepared from 1 g of styrene monomer; , surface density of -glutamic acid residues; q
s
L
, surface-coverage efficiency.
stabilization and growth periods as an anchor. It is known that
block-copolymer-stabilizing dispersion polymerizations provide
submicron-sized polymer particles in many cases. One of advan-
tages for a block-copolymer system is that the adsorption behavior
of a block copolymer is tunable by changing the molecular weight
of the anchor segment, leading to a possibility for highly-precise
control of particle size and surface structure [17]. For the
Hence, several components are expected to be prepared by
different kinetics.
3.4.1. Three components of polystyrene particles
A GPC measurement was carried out for the polystyrene parti-
cles in run 8 to show a broad curve in Fig. 5a, indicating
M
n
¼ 52 000. By using a peak-fitting software, the curve was
mPLGA
n
-stabilizing dispersion polymerization, on the other hand,
separated into four Gaussians having M
n
¼ 530 000, 220 000,
the anchor segment was generated in the course of a dispersion
polymerization to give micron-sized particles, so that the anchor
segment is not tunable. However, the concurrent generation leads
82 000, and 18 000 with 18, 45, 31, and 6% of their volume fractions,
respectively. However, the highest molecular weight component, a'
in Fig. 5a, should be incorporated into the second highest one, a,
since the used column of the GPC measurement had an exclusion
n
to another advantage that mPLGA will be available for dispersion
polymerization of various monomers because the anchor and the
particle core are perfectly compatible due to their same chemical
structure. In this study, we have extended the control range of the
size and the surface structure to prepare uniform-sized surface-
limit of M
n
¼ 500 000 to lead to a shoulder at 10.6 mL of the elution
volume. For the particles with rather low molecular weight in runs
12, 13, and 17, the GPC curves were separated into three Gaussians
below the exclusion limit. Hence, it is concluded that the poly-
styrene particles were composed of three components with high,
middle, and low molecular weight corresponding to a (þa'), b, and
c, respectively. This result is not specific to the macromonomer-
stabilizing dispersion polymerization, because we previously
observed similar three components in polystyrene particles pre-
pared by a block-copolymer-stabilizing dispersion polymerization
[16], while assignment of these components had been remained as
an issue to be solved.
In dispersion polymerization, it has been generally explained
that initiation and propagation take place both in particle phase
and in dispersion medium, leading to the following mechanisms;
(i) bulk polymerization is initiated and propagated in a particle, (ii)
an oligomeric radical generated in dispersion medium is trans-
ferred into a particle and then bulk polymerization proceeds, and
(iii) a dead polymer chain is generated in dispersion medium and
captured by a particle [2]. In general, bulk polymerization gives
polymers with high molecular weight since highly-viscose solid
phase suppresses bimolecular termination (gel-effect), while so-
lution polymerization provides small polymers. Therefore, it is
reasonable to assume that the mechanism (iii) was responsible for
the generation of the low-molecular-weight component c, while
the other components were synthesized by bulk polymerization
with the mechanisms (i) and (ii). In particle phase, bulk polymer-
ization was initiated by thermal decomposition of AIBN in a particle
or by capture of a solution-initiated oligomeric radical. One can
n
functional particles by the mPLGA -stabilizing dispersion poly-
merization as well as the PLGA-block-copolymer-stabilizing one.
Accordingly, we have established a full control of particle size,
surface density, and surface PLGA length for the functional poly-
styrene particles by changing the polymerization conditions in the
n
dispersion polymerization stabilized by the mPLGA . For example,
we realized two-fold control of the size of particles while the sur-
face density was kept, which was observed in runs 9 and 15. The
particles in runs 11 and 17 had similar size and surface density but
different length of the surface PLGA. For a macromonomer-
stabilizing dispersion polymerization, to our knowledge, this is
the first example to obtain experimental values of power law de-
pendences for surface structure, while those for particle size have
been compared to the theoretical values [19]. However, regulation
of size distribution remains as a matter to be solved since some
particles showed rather wide distribution up to D
w
/D
n
¼ 1.08. Two
factors have been accepted for preparation of widely-distributed
polymer particles [2], which can be applied to the results in this
study. One is that lack of effective stabilizers on particles causes
particle aggregation in the growth period, which accounts for
widely-distributed particles with rather low surface density with
2
s
¼ 0.2e0.5 residue/nm in runs 7, 11, and 16. The other is that,
when the particle surface is overcrowded after reaching the limited
capacity, the excess stabilizer newly stabilizes an unstable particle
as a secondary particle, which is responsible for other widely-
distributed particles with highly-dense surface at
residues/nm in runs 13 and 15, respectively. However, the particles
s
¼ 2.6 and 1.2
2
2
with rather high surface density with
runs 9 and 10, respectively, had a narrow size distribution within
/D
s
¼ 1.3 and 1.8 residue/nm in
a)
b)
a
D
w
n
¼ 1.02, indicating that wide distribution was not always
accompanied by dense surface. This issue is discussed further in the
following section along with a detailed mechanism of the particle
formation.
b
a’
c
3
.4. Polymerization kinetics and growth mechanism
1
0
12 14 16 18
10 12 14 16 18
elution volume (mL)
elution volume (mL)
In addition to the particle size and the surface structure, particle
composition will give us useful information of a polymerization
mechanism in the particle-growth period. It is widely accepted that
polymer chains are generated both inside and outside of a particle.
Fig. 5. GPC curves of the polystyrene particles. Polymerization condition: styrene
(
10 wt%); mPLGA100 (5 wt% to styrene); AIBN (1.6 wt% to styrene); 75/25 DMFewater
ꢀ
medium; 70 C; a) 17 h and b) 1 h.

T. Itoh et al. / Polymer 70 (2015) 183e193
191
expect that high probability of bimolecular termination near the
particle surface, which was exposed to oligomeric radicals in
dispersion medium, could give rather small polymer. However,
progress with both the mechanisms (i) and (ii) in one stable particle
would result in a continuous change of molecular weight to give a
single peak which is most probably assigned to the high-molecular-
weight component. The middle-molecular-weight one was clearly
separable from the high component, implying a different mode of
bulk polymerization.
Another possible mechanism of the particle growth is an
involvement of small unstable particles [2,31]. Such unstable par-
ticles could be formed in dispersion medium even in the growth
period and subsequently incorporated into large stable particles
before stabilized as secondary particles. Although isolation of the
unstable particles was very difficult, related particles could be
formed in the early stage of the dispersion polymerization. For the
particles prepared at 5% conversion in run 4, the presence of two
components were indicated from the GPC curves in Fig. 5b, which
are assignable to polystyrene generated by the mechanisms (ii) and
parameters of monomer, initiator, medium, temperature, and
polymerization time were the same, the inverse relation of the
volume fractions between the high- and the middle-molecular-
weight components would be attributed to the particle size. It
has been described that oligomeric radicals are ineffectively
captured by large particles because their total surface area is small
[1]. We expand this hypothesis into other two modes. One is the
incorporation efficiency of the unstable particles by the stable
particles and the other is the capture efficiency of the oligomeric
radicals by the unstable particles. When stable particles were large,
unstable particles could stay in the dispersion medium for a long
time to enlarge their size by coalescence until they were incorpo-
rated into the stable particles, leading to, in turn, a shrink of total
surface of the unstable particles, a reduction of the capture effi-
ciency of the oligomeric radicals by the unstable particles, and a
decrease in the volume fraction of the middle-molecular-weight
component. Whereas the stable particles would have the same
effect of particle size on the capture efficiency of the oligomeric
radicals, it was reported that unstable particles were an order of
magnitude smaller size than stable ones [2], which could enhance
the size effect of the unstable particles on the capture of the olig-
omeric radicals.
The size effect on the incorporation efficiency of the unstable
particles into the stable ones is possibly related to achievement of
the narrowly-distributed particle size with the dense surface,
which was observed in runs 9 and 10. It has been accepted that
graft copolymers are liable to be adsorbed on an unstable particle to
cause a secondary stabilization when surface density of a stable
particle is close to the acceptable limit [2]. In contrast, it is
assumable that the secondary particle is not generated when the
unstable particle is incorporated into the stable one faster than the
secondary stabilization, which could be achieved when the stable
particle is small. The resultant particles in runs 9 and 10 were rather
small, supporting this assumption. Consequently, we suggest a
possibility that shortening retention time of an unstable particle in
dispersion medium suppress the close relation between secondary
stabilization and dense surface.
(
iii) since solution polymerization was still dominant at the stage.
The particles were obtained shortly after the particle stabilization,
so that the character of the unstable particles would substantially
remain. This leads us to a following possible explanation for the
generation of the middle-molecular-weight component as sche-
matically illustrated in Fig. 6. In the growth period, small unstable
particles were formed and coalesced each other as those in the
stabilization period. The unstable particles captured oligomeric
radicals in addition to dead polymers and initiated bulk polymer-
ization inside. The unstable particles was small enough for effective
capture of further oligomeric radicals. This increased a probability
of bimolecular termination in the unstable particles, providing
smaller polymers than the high-molecular-weight component.
Subsequently, the unstable particles were incorporated to inject the
middle-molecular-weight component into the stable ones. From
the 31% fraction of the middle component, we suggest that the
unstable particles should significantly participate in the particle
growth.
Assuming the participation of the unstable particles in the
particle-growth mechanism, one can explain the effects of the
initial concentration of DMF, styrene, and AIBN on the three modes
of the polymerization in the growth period. With increasing the
DMF composition from 75 to 85 wt%, the low-molecular-weight
component increased from 6 to 40%, while the other components
decreased. This would be caused by oligomeric radicals highly-
soluble to the 85/15 DMFewater medium in the presence of sty-
rene monomer, where the oligomeric radicals would increase the
probability of their bimolecular termination in the solution phase
rather than that of their migration into stable and unstable
particles.
3
.4.2. Control of the components
For the dispersion polymerizations in Tables 3e5, their poly-
merization conditions had a strong effect on volume fractions of the
three components as plotted in Fig. 7. When particle size was
enlarged by a decrease in the initial concentration of mPLGA100 and
n
an increase in the mPLGA length, volume fractions of the high- and
the middle-molecular-weight components increased and
decreased, respectively, while that of the low-molecular-weight
component was kept lower than 13%. Since the kinetic
When the initial concentration of styrene monomer increased
from 5 to 15 wt% with keeping the relative concentration of
mPLGA100 and AIBN to the monomer, the volume fraction of the
middle-molecular-weight component increased by 20%, accompa-
nied by a decrease in the high one. Since the polymerization system
was concentrated, the increased oligomeric radicals were possibly
liable to be trapped by the concentrated unstable particles in the
medium before incorporated into the stable ones. Whereas the
fraction of the middle-molecular-weight component was inversely
unstable particle
stable
particle
n
related to the particle size in the study of mPLGA effect, a positive
relation was observed here. The effect of the initial styrene con-
centration could be dominant over the dependence of the incor-
poration efficiency on particle size.
Fig. 6. A schematic illustration for the proposal mechanism of particle growth
involving unstable particles which are formed by capture of oligomeric radicals and
dead polymers. Bulk polymerization which is initiated by an oligomeric radical is
terminated by effective capture of another radical, giving the middle-molecular-weight
component. Then, the unstable particles are incorporated into the large stable ones.
The increase in the initial AIBN concentration from 0.03 to
0.80 wt% resulted in drastic increases in the middle- and the low-
molecular-weight components by 40 and 30%, respectively, which

192
T. Itoh et al. / Polymer 70 (2015) 183e193
Fig. 7. Dependence of volume fraction of high- (closed squares), middle- (closed circles), and low-molecular-weight components (closed triangles) on polymerization conditions.
is in agreement with our previous report for a block-copolymer-
stabilizing dispersion polymerization [16]. A plausive explanation
is that a large amount of oligomeric radicals were generated in the
dispersion medium to give a large number of unstable particles and
bimolecular-terminated polystyrenes.
The three components of the particles are reasonably explained
by assuming the participation of the unstable particles in the
particle-growth mechanism, although further studies are necessary
to validate this assumption. Since polymerization kinetics has a
strong effect on formation of particle core, we expect that the re-
sults will contribute to fine control of surface-functional polymer
particles with non-spherical morphology, crosslink, multicompo-
nent core, and so on, as well as functional spherical particles in the
near future.
References
[
1] A.J. Paine, W. Luymes, J. McNulty, Dispersion polymerization of styrene in
polar solvents. 6. Influence of reaction parameters on particle size and mo-
lecular weight in poly(N-vinylpyrrolidone)-stabilized reactions, Macromole-
cules 23 (1990) 3104e3109.
[2] A.J. Paine, Dispersion polymerization of styrene in polar solvents. 7. A simple
mechanistic model to predict particle size, Macromolecules 23 (1990)
3109e3117.
[3] C.K. Ober, Dispersion copolymerization in non-aqueous media, Makromol.
Chem. Macromol. Symp. 35e36 (1990) 87e104.
[4] K.P. Lok, C.K. Ober, Particle size control in dispersion polymerization of
polystyrene, Can. J. Chem. 63 (1984) 209e216.
[5] T. Corner, Polyelectrolyte stabilised latices part 1, preparation, Colloids Surf. 3
(1981) 119e129.
[
6] H. Kitano, J. Kawabata, Temperature-responsive polymer microspheres pre-
pared with a macroinitiator, Macromol. Chem. Phys. 197 (1996) 1721e1729.
7] S. Fujii, K. Aono, M. Suzaki, S. Hamasaki, S. Yusa, Y. Nakamura, pH-responsive
[
hairy particles synthesized by dispersion polymerization with
a macro-
initiator as an inistab and their use as a gas-sensitive liquid marble stabilizer,
Macromolecules 45 (2012) 2863e2873.
4
. Concluding remarks
[8] C. Lepilleur, E.J. Beckman, Dispersion polymerization of methyl methacrylate
2
in supercritical CO , Macromolecules 30 (1997) 745e756.
We carried out dispersion polymerization of styrene in the
presence of a novel macromonomer, mPLGA , as a functional sta-
n
bilizer to give narrowly-distributed polystyrene particles. The
resultant particles showed a pH-responsive colloidal behavior
n
reflecting the helixecoil transformation of the mPLGA , indicating
successful preparation of surface-functional polystyrene particles.
We represent power law dependences on macromonomer con-
centration, macromonomer length, medium composition, styrene
concentration, and AIBN concentration for surface structure as well
as size of the particles. As a result, it is clarified that the partici-
[
9] M. Muranaka, H. Yoshizawa, T. Ono, Design of polylactide-grafted copolymeric
stabilizer for dispersion polymerization of D,L-lactide, Colloid Polym. Sci. 287
(2009) 525e532.
[
10] M.T. Elsesser, A.D. Hollingsworth, Revisiting the synthesis of a well-known
comb-graft copolymer stabilizer and its application to the dispersion poly-
merization of poly(methyl methacrylate) in organic media, Langmuir 26
(2010) 17989e17996.
[11] D. Alaimo, A. Beigbeder, P. Dubois, G. Broze, C. J
eꢀ r o^ me, B. Grignard, Block,
random and palm-tree amphiphilic fluorinated copolymers: controlled syn-
thesis, surface activity and use as dispersion polymerization stabilizers,
Polym. Chem. 5 (2014) 5273e5282.
[12] G. Riess, C. Labbe, Block copolymers in emulsion and dispersion polymeriza-
tion, Macromol. Rapid Commun. 25 (2004) 401e435.
13] F.L. Baines, S. Dionisio, N.C. Billingham, S.P. Armes, Use of block copolymer
stabilizers for the dispersion polymerization of styrene in alcoholic media,
Macromolecules 29 (1996) 3096e3102.
14] T. Itoh, K. Fukutani, M. Hino, E. Ihara, K. Inoue, Effects of polystyrene-b-
poly(aminomethyl styrene)s as stabilizers on dispersion polymerization of
styrene in alcoholic media, J. Colloid Interface Sci. 330 (2009) 292e297.
[15] J. Jacobs, A. Byrne, N. Gathergood, T.E. Keyes, J.P.A. Heuts, A. Heise, Facile
synthesis of fluorescent latex nanoparticles with selective binding properties
using amphiphilic glycosylated polypeptide surfactants, Macromolecules 47
n
pation of the mPLGA was promoted as the dispersion polymeri-
[
zation proceeded. To our knowledge, this is the first experimental
study to obtain the exponent values for the surface structure in a
macromonomer-stabilizing dispersion polymerization, which
enabled us to make comparison with the theoretical model and to
control particle size, surface density, and length of surface-grafted
PLGA. A GPC curve of the resultant particles was separated in
three components which were tentatively identified as polystyrene
prepared by a) bulk polymerization in large stable particles, b) bulk
polymerization in small unstable particles which capture oligo-
meric radicals more frequently than the stable ones, and c) bimo-
lecular termination in dispersion medium, in decreasing order of
their molecular weight. Particle size as well as medium solvency,
monomer concentration, and radical concentration had a strong
effect on the three components. We propose a possibility that the
unstable particles were significantly involved in the particle-
growth mechanism and the particle-size distribution. While func-
tional polymer particles have been empirically controlled in a
macromonomer-stabilizing dispersion polymerization, the aspects
of the formation mechanism will contribute precise control of size,
surface structure, and composition of the functional particles.
[
(
2014) 7303e7310.
[16] T. Itoh, S. Komada, E. Ihara, K. Inoue, Dispersion polymerization of styrene
using a polystyrene/poly( -glutamic acid) block copolymer as a stabilizer,
L
J. Colloid Interface Sci. 388 (2012) 112e117.
17] T. Itoh, I. Abe, T. Tamamitsu, H. Shimomoto, K. Inoue, E. Ihara, Surface struc-
ture of stimuli-responsive polystyrene particles prepared by dispersion
polymerization with a polystyrene/poly(L-lysine) block copolymer as a sta-
bilizer, Polymer 55 (2014) 3961e3969.
18] K. Tomita, T. Ono, Synthesis of polyaspartate macromonomer having a vinyl
end group and application to dispersion copolymerization of styrene, Colloid
Polym. Sci. 287 (2009) 109e113.
[
[
[
[
19] S. Kawaguchi, K. Ito, Dispersion polymerization, Adv. Polym. Sci. 175 (2005)
299e328.
20] S. Kawaguchi, M.A. Winnik, K. Ito, 1H NMR study of dispersion copolymeri-
zation of n-butyl methacrylate with poly(ethylene oxide) macromonomer in
deuterated methanol-water, Macromolecules 29 (1996) 4465e4472.
21] M.B. Nugroho, S. Kawaguchi, K. Ito, M.A. Winnik, Control of particle size in
dispersion polymerization using poly(ethylene oxide) macromonomers,
J. Macromol. Sci. Part A 32 (1995) 593e601.
22] C. Wu, M. Akashi, M.-Q. Chen, A simple structural model for the polymer
microsphere stabilized by the poly(ethylene oxide) macromonomers grafted
on its surface, Macromolecules 30 (1997) 2187e2189.
[
[
[
Acknowledgment
23] M.-Q. Chen, A. Kishida, M. Akashi, Graft copolymers having hydrophobic
backbone and hydrophilic branches. XI. Preparation and thermosensitive
This work was supported by JSPS KAKENHI Grant Number
6810115.
2

T. Itoh et al. / Polymer 70 (2015) 183e193
193
properties of polystyrene microspheres having poly(N-isopropylacrylamide)
branches on their surfaces, J. Polym. Sci. Part A Polym. Chem. 34 (1996)
[33] R.J.I. Knoop, M. de Geus, G.J.M. Habraken, C.E. Koning, H. Menzel, A. Heise,
Stimuli responsive peptide conjugated polymer nanoparticles, Macromole-
cules 43 (2010) 4126e4132.
2213e2220.
[
24] J.R. McKee, V. Ladmiral, J. Niskanen, H. Tenhu, S.P. Armes, Synthesis of
sterically-stabilized polystyrene latexes using well-defined thermoresponsive
poly(N-isopropylacrylamide) macromonomers, Macromolecules 44 (2011)
[34] F. Audouin, R.J.I. Knoop, J. Huang, A. Heise, Star polymers by cross-linking of
linear poly(benzyl-
L
-glutamate) macromonomers via free-radical and RAFT
polymerization.
A
simple route toward peptide-stabilized nanoparticles,
7
692e7703.
J. Polym. Sci. Part A Polym. Chem. 48 (2010) 4602e4610.
[35] A. Bousquet, R. Perrier-Cornet, E. Ibarboure, E. Papon, C. Labrug eꢁ re,
V. H eꢀ roguez, J. Rodríguez-Hern aꢀ ndez, Design of polypeptide-functionalized
polystyrene microspheres, Biomacromolecules 9 (2008) 1811e1817.
[36] H.R. Kricheldorf, Polypeptides and 100 years of chemistry of a-amino acid N-
carboxyanhydrides, Angew. Chem. Int. Ed. 45 (2006) 5752e5784.
[37] N. Hadjichristidis, H. Iatrou, M. Pitsikalis, G. Sakellariou, Synthesis of well-
defined polypeptide-based materials via the ring-opening polymerization of
[
25] S.F. Lascelles, F. Malet, R. Mayada, N.C. Billingham, S.P. Armes, Latex syntheses
using novel tertiary amine methacrylate-based macromonomers prepared by
oxyanionic polymerization, Macromolecules 32 (1999) 2462e2471.
26] A.P. Richez, L. Farrand, M. Goulding, J.H. Wilson, S. Lawson, S. Biggs, O.J. Cayre,
Poly(dimethylsiloxane)-stabilized polymer particles from radical dispersion
polymerization in nonpolar solvent: Influence of stabilizer properties and
monomer type, Langmuir 30 (2014) 1220e1228.
[
[
[
27] L.A. Fielding, J.A. Lane, M.J. Derry, O.O. Mykhaylyk, S.P. Armes, Thermo-
responsive diblock copolymer worm gels in non-polar solvents, J. Am. Chem.
Soc. 136 (2014) 5790e5798.
28] N.J. Warren, O.O. Mykhaylyk, D. Mahmood, A.J. Ryan, S.P. Armes, RAFT
aqueous dispersion polymerization yields poly(ethylene glycol)-based diblock
copolymer nano-objects with predictable single phase morphologies, J. Am.
Chem. Soc. 136 (2014) 1023e1033.
a
-amino acid N-carboxyanhydrides, Chem. Rev. 109 (2009) 5528e5578.
[38] C. Deng, J. Wu, R. Cheng, F. Meng, H.-A. Klok, Z. Zhong, Functional polypeptide
and hybrid materials: precision synthesis via -amino acid N-carboxyanhy-
a
dride polymerization and emerging biomedical applications, Prog. Polym. Sci.
39 (2014) 330e364.
[39] E.M. Bradbury, C. Crane-Robinson, H. Goldman, H.W.E. Rattle, Proton magnetic
resonance and the helixecoil transition, Nature 217 (1968) 812e816.
[40] G. Holzwarth, P. Doty, The ultraviolet circular dichroism of polypeptides,
J. Am. Chem. Soc. 87 (1965) 218e228.
[
29] P. Shi, Q. Li, X. He, S. Li, P. Sun, W. Zhang, A new strategy to synthesize
temperature- and pH-sensitive multicompartment block copolymer nano-
particles by two macro-RAFT agents comediated dispersion polymerization,
Macromolecules 47 (2014) 7442e7452.
[41] W.H. Daly, D. Poch eꢀ , The preparation of N-carboxyanhydrides of
a-amino
acids using bis(trichloromethyl)carbonate, Tetrahedron Lett. 29 (1988)
5859e5862.
[42] M.L. Tiffany, S. Krimm, New chain conformations of poly(glutamic acid) and
polylysine, Biopolymers 6 (1968) 1379e1382.
[
[
[
30] S. Jiang, E.D. Sudol, V.L. Dimonie, M.S. El-Aasser, Kinetics of dispersion poly-
merization of methyl methacrylate and n-butyl acrylate: effect of initiator
concentration, Macromolecules 40 (2007) 4910e4916.
31] S. Jiang, E.D. Sudol, V.L. Dimonie, M.S. El-Aasser, Kinetics of dispersion poly-
merization: effect of medium composition, J. Polym. Sci. Part A Polym. Chem.
[43] A.V. Mikhonin, N.S. Myshakina, S.V. Bykov, S.A. Asher, UV resonance Raman
determination of polyproline II, extended 2.51-helix, and
energy landscape in poly-L-lysine and poly-L-glutamic acid, J. Am. Chem. Soc.
127 (2005) 7712e7720.
b-sheet
J angle
46 (2008) 3638e3647.
32] S. Jiang, E. David Sudol, V.L. Dimonie, M.S. El-Aasser, Dispersion polymeriza-
tion of methyl methacrylate: Effect of stabilizer concentration, J. Appl. Polym.
Sci. 107 (2008) 2453e2458.
[44] J. Brandrup, E.H. Immergut, E.A. Grulke (Eds.), Polymer Handbook, fourth ed.,
John Wiley and Sons, New York, 1999.