Catalytic chain transfer mediated autopolymerization of divinylbenzene: Toward facile synthesis of high alkene functional group density hyperbranched materials

Article abstract of DOI:10.1021/ma3018684

A facile and highly reproducible autopolymerization method, mediated by catalytic chain transfer, for the synthesis of hyperbranched materials with high alkene functional group density is reported. The rapid autopolymerization of divinylbenzene at 150 °C in the absence of any catalytic chain transfer agent was demonstrated to result in the formation of highly cross-linked networks/gels in less than 10 min. Exploitation of the extremely high chain transfer coefficient of bis[(difluoroboryl)diphenylglyoximato]cobalt(II) delayed gelation and produced good yields of high molecular weight hyperbranched divinylbenzene in under 1 h on a multigram scale. Gas chromatography was employed to monitor the levels of conversion over the course of the reaction. The materials produced were characterized by GPC, MALLS, and viscometry.

Full text of DOI:10.1021/ma3018684

Article
pubs.acs.org/Macromolecules
Catalytic Chain Transfer Mediated Autopolymerization of
Divinylbenzene: Toward Facile Synthesis of High Alkene Functional
Group Density Hyperbranched Materials
Ian A. Barker,†,‡ Jaouad El Harfi,†,‡ Kevin Adlington,†,‡ Steven M. Howdle,† and Derek J. Irvine*,†,‡
‡
†School of Chemistry and Department of Chemical and Environmental Engineering, Faculty of Engineering, University of
Nottingham, University Park, Nottingham NG7 2RD, United Kingdom
S
* Supporting Information
ABSTRACT: A facile and highly reproducible autopolymeri-
zation method, mediated by catalytic chain transfer, for the
synthesis of hyperbranched materials with high alkene
functional group density is reported. The rapid autopolyme-
rization of divinylbenzene at 150 °C in the absence of any
catalytic chain transfer agent was demonstrated to result in the
formation of highly cross-linked networks/gels in less than 10
min. Exploitation of the extremely high chain transfer
coefficient of bis[(difluoroboryl)diphenylglyoximato]cobalt-
(II) delayed gelation and produced good yields of high molecular weight hyperbranched divinylbenzene in under 1 h on a
multigram scale. Gas chromatography was employed to monitor the levels of conversion over the course of the reaction. The
materials produced were characterized by GPC, MALLS, and viscometry.
dilute monomer concentrations (≪10% monomer).19 In 2002,
INTRODUCTION
■
Sherrington et al.19 and Guan20 independently reported the use
Dendritic materials are three-dimensional (3D), globular
macromolecules which contain regular, well-defined, and
extremely high levels of branching. They are typically accessed
of catalytic chain transfer21−25 (CCT), an industrially viable
polymerization control strategy, to control the free radical
polymerization of multifunctional monomers to form hyper-
branched materials. In this work it was demonstrated that
soluble, highly branched methacrylic copolymers could be
made in a simple, one-step, high conversion batch polymer-
ization, where a cobalt porphyrin was successfully used to
inhibit gelation provided the branching agent (a multifunctional
monomer such as tripropylene glycol diacrylate) concentration
was kept low (5 wt %).19 Additionally, because methacrylates
are a monomer class that are particularly applicable to CCT due
to the presence of an α-methyl group on the monomer from
which facile hydrogen abstraction can occur to terminate
growth, it was demonstrated that this control could be achieved
through the use of very small amounts of CCT agent, typically
between 2.5 and 10.0 ppm.19 Furthermore, the CCT hydrogen
abstraction termination mechanism was demonstrated to
produce macromolecules with increased concentrations of
alkene chain-ends compared to those produced by the use of
more conventional chain transfer agents, e.g. thiols, leaving
them susceptible to postmodification.19 This is because the
chain transfer mechanism employed ensures that any CCT-
terminated polymer has an unsaturated carbon−carbon double
bond as the terminal group.26,27
by the application of highly controlled, very efficient, and
multistage synthetic strategies.1−5 Furthermore, the material
properties associated with these highly branched macro-
molecular structures have generated a great deal of interest
related to their potential to deliver significantly differentiated
application performance in such areas as altering polymer melt
flow characteristics,6−8 improving drug/gene delivery,9−12 and
maximizing dye transport in light harvesting and light-emitting
diodes.13 However, research into fully defining the end use
performance of such dendritic materials has been hampered by
the typically labor intensive and expensive routes required for
their production. Consequently, this has significantly con-
tributed to the lack of uptake of these materials into mass
markets. In comparison, hyperbranched materials can be
produced using relatively inexpensive, one-pot synthetic
approaches and have structures similar to that of a dendrimer,
only less well-defined.14−17 This has led to the most recent
research on branched polymeric materials being dominated by
the development of new hyperbranched polymers and
synthetic/processing systems specifically designed to achieve
such structures.18
During the formation of hyperbranched polymers, care must
be taken to avoid cross-linking, which leads to the formation of
an insoluble gel instead of the desired, solvent-soluble
hyperbranched macromolecules. Indeed, it has been demon-
strated that gelation in uncontrolled free radical systems
generally occurs below 20% monomer conversion, even at very
Received: September 10, 2012
Revised: November 13, 2012
Published: November 26, 2012
© 2012 American Chemical Society
9258
dx.doi.org/10.1021/ma3018684 | Macromolecules 2012, 45, 9258−9266

Macromolecules
Article
divinylbenzene 80% (DVB-80, technical grade, 80% difunctional
monomer (m- and p-DVB), 20% monofuntional monomer (3- and 4-
ethylstyrene)), hydroquinone (99%+), cyclohexanone (99%+),
potassium tert-butoxide (98%+), methyltriphenylphosphonium bro-
mide (98%), isophthalaldehyde (97%), terephthalaldehyde (98%),
hexane (99%+), deuterated chloroform (99.8%+), and tetrahydrofuran
(THF, anhydrous, 99%+) were all purchased from Sigma-Aldrich.
Tetrahydrofuran was passed through two alumina towers and collected
under nitrogen to ensure exclusion of water. High-purity nitrogen was
purchased from BOC gases. Chromatographic grade neutral aluminum
oxide was purchased from Acros Organics. 2,2′-Azobis-
(isobutyronitrile) (AIBN, 98%+) purchased from Wako Chemicals
was recrystallized three times from methanol before use. Bis-
[(difluoroboryl)diphenylglyoximato]cobalt(II) (CoPhBF) was ob-
tained from DuPont.
Interestingly, the work conducted by Guan did not involve
the inclusion of a monofunctional monomer; the reactions were
homopolymerizations of a difunctional monomer.20 To the best
of our knowledge, this was the first report of such a monomer
being successfully homopolymerized in the presence of a CCT
agent and was an important step toward creating hyper-
branched systems with high alkene functional group density
through the application of addition polymerization. Guan also
showed that ethylene glycol dimethacrylate (EGDMA) could
be homopolymerized in either the bulk or a solution (1,2-
dichloroethane) system. In the presence of an azo-initiator and
∼500 ppm of a CoBF derivative, high molecular weight, solvent
soluble polymers were successfully isolated up to 70%
monomer conversion.20 By comparison, when no CCT agent
was introduced, the solution formed a gel almost instantly.
Despite the robust and facile nature of the methods
developed by Sherrington and Guan, little work has been
undertaken to extend this CCT-based methodology to different
divinyl species beyond methacrylate-based monomers, despite
several being readily commercially available. One such
monomer (divinylbenzene (DVB)) has been the subject of a
number of recent developments which have enabled the
homopolymerization of this monomer to yield highly branched,
alkene functionality dense hyperbranched poly(divinylbenzene)
(HBDVB). Baskaran demonstrated the homopolymerization of
commercially obtained DVB could be achieved via anionic self-
condensing vinyl polymerization (SCVP)28 with n-BuLi.29 In
this synthesis, DVB containing 35% monofunctional impurities
was polymerized in 10 min to produce yields in excess of 90%
polymeric material. However, the materials produced in this
method typically contained anywhere from 1 to 15% insoluble
cross-linked material, and the molecular weights achieved were
moderate (Mw = 155 000 g mol−1).29 Wang et al. described the
application of the deactivation enhanced atom transfer radical
polymerization (ATRP)30−32 of DVB to produce HBDVB.33
Manipulating the polymerization in this way allowed the
controlled homopolymerization of DVB to produce very high
molecular weight HBDVB (Mw = 5 400 000 g mol−1) after 28 h
in good yields (60%).33 More recently, work by Perrier et al.34
reported that commercially available DVB, containing a mixture
of 80% m-DVB and p-DVB and 20% ethylstyrene (ES), could
be homopolymerized via reversible addition−fragmentation
chain transfer (RAFT) polymerization to realize moderately
high molecular weight (Mw = 233 000 g mol−1) highly
branched materials after ∼18 h polymerization reaction time.34
In this study, we report the first application of CCT to the
homopolymerization of commercially available DVB, aimed at
developing a facile methodology for the large scale production
of HBDVB. In doing so, a new route to the production of
hyperbranched structures with high alkene content is defined
which results in the production of tens of grams of HBDVB in
less than 1 h. Furthermore, we report the comparison of
applying this method of creating highly branched polymer
structures to both the individual homopolymerization and
copolymerization of m-DVB and p-DVB.
Preparation of m-Divinylbenzene. m-DVB was synthesized
according to a literature procedure via a Wittig reaction.35 Potassium
tert-butoxide (53.0 g, 0.47 mol) was dissolved in dry THF (750 mL) in
a 2 L, three-necked flask. Methyltriphenylphosphonium bromide
(160.8 g, 0.45 mol) was then added to the resulting solution in a single
addition, and the resulting bright yellow solution of the resultant ylide
was stirred for 20 min at room temperature. Isophthalaldehyde (25.0 g,
0.18 mol) dissolved in dry THF (250 mL) was then added dropwise to
the stirring mixture, maintaining the temperature below 40 °C. The
stirring was continued for 1 h at room temperature postaddition, after
which ice water (400 g ice in 800 mL of water) was introduced, and
the resulting organic phase was separated from the water phase in a
separating funnel. The water phase was repeatedly shaken with n-
hexane (3 × 500 mL), and the precipitate formed was removed by
filtration. Thereafter, the organic fractions were combined and dried
over anhydrous magnesium sulfate. Hydroquinone (0.03 g, 2.7 × 10−4
mol) was added to inhibit polymerization, and the hexane was
removed under vacuum. The crude product was vacuum distilled twice
at 45 °C (0.2 mbar) to yield m-DVB as a clear liquid. 9.2 g (39%
1
yield); purity >99% as determined by GC. H NMR (CDCl3, 400
MHz) at 25 °C: δ = 7.4 (s, 1H, aromatic), 7.3 (m, 3H, aromatic), 6.7−
6.8 (dd, 2H, −CHCH2), 5.7−5.8 (dd, 2H, −CHC(H)H cis),
5.2−5.3 (dd, 2H, −CHC(H)H trans). Mass spectrum m/z (relative
intensities) 130 (M+, 100.0), 129 (22.7), 128 (26.5), 127 (12.3), 115
(23.5), 77 (7.2), 63 (3.9), 51 (1.7).
Preparation of p-Divinylbenzene. p-DVB was synthesized by
the same procedure as for m-DVB, with the exception that dialdehyde
terephthalaldehyde was added to the ylide solution in this case. The
crude product was vacuum-distilled twice at 45 °C (0.2 mbar) to yield
p-DVB as a clear liquid which rapidly solidified on standing to become
a sweet smelling, opaque solid. 10.5 g (45% yield); purity >99% as
determined by GC. 1H NMR (CDCl3, 400 MHz) at 25 °C: δ = 7.4 (s,
4H, aromatic), 6.7−6.8 (dd, 2H, −CHCH2), 5.7−5.8 (dd, 2H,
−CHC(H)H cis), 5.2−5.3 (dd, 2H, −CHC(H)H trans). Mass
spectrum m/z (relative intensities) 130 (M+, 100.0), 129 (20.3), 128
(26.1), 127 (12.5), 115 (26.5), 77 (21.1), 63 (12.3), 51 (5.0).
General Polymerization Procedure. The required quantities of
degassed DVB, cyclohexanone, and CoPhBF were introduced into a
dry Schlenk flask equipped with a magnetic stirrer bar and containing
an inert nitrogen atmosphere. The reaction vessel was then immersed
in a preheated oil bath which was thermostatically controlled to remain
at 150 °C. Upon completion of the reaction, the vessel was removed
from the oil bath and immediately quenched in liquid nitrogen to
prevent further reaction and hence gelation. The mixture was then
diluted with chloroform, precipitated into cold (0 °C) methanol, and
collected via filtration. The crude polymer was dissolved and
reprecipitated a further two times before yields were determined
gravimetrically after 48 h under vacuum at 50 °C.
EXPERIMENTAL SECTION
■
Example 1: DVB-80 Polymerization. Degassed DVB-80 (30 mL,
27.6 g, 0.21 mol) and cyclohexanone (30 mL, 28.4 g) were transferred
to a Schlenk flask containing CoPhBF (0.391 g, 6.18 × 10−4 mol), a
magnetic stirrer bar, and an inert nitrogen atmosphere. The resulting
solution was then rapidly heated to 150 °C by immersion in a
preheated oil bath and held at that temperature for 45 min, after which
it was immediately quenched in liquid nitrogen to prevent gelation.
Materials. Unless otherwise stated, all reagents were used as
received and without further purification, and all procedures were
conducted under an inert nitrogen atmosphere using standard Schlenk
line techniques. m-Divinylbenzene and p-divinylbenzene (m- and p-
DVB) were synthesized individually via a Wittig reaction and purified
via bulb-to-bulb distillation just before use. Styrene (99%+),
9259
dx.doi.org/10.1021/ma3018684 | Macromolecules 2012, 45, 9258−9266

Macromolecules
Article
Once cooled, the solution was dissolved in 30 mL of chloroform, and
the subsequent solution was added dropwise to cold (0 °C) methanol.
The resulting precipitate was collected via filtration to provide a fine
off-white/brown powder (coloration due to CoPhBF) which was then
dried to constant mass. Yield = 40%. 1H NMR (CDCl3, 400 MHz) at
25 °C: δ = 7.44 −5.95 (broad m, aromatic and −CHCH2), 5.79−
5.50 (broad s, 1H, −CHCH cis), 5.25−5.05 (broad s, 1H, −CH
CH trans), 2.65−0.64 (broad m, CH2 (backbone) and −Ar−CH2−
CH3). GPC: Mn = 4800, Mw = 111 200, Đ = 23.2. MALLS: Mn = 66
000, Mw = 742 700, Đ = 11.3, Rn = 50.4 nm.
RESULTS AND DISCUSSION
■
Development of the Polymerization Process and
Determination of the Polymerization Kinetics. A CCT
control agent was employed to control the molecular weight
within the polymerizations conducted in this study to prevent
gelation. The catalyst chosen in this study was bis-
[(difluoroboryl)diphenylglyoximato]cobalt(II) (CoPhBF, Fig-
ure 1).
Example 2: Synthesis of p-DVB-co-m-DVB. p-DVB (1.5 mL,
1.38 g, 1.05 × 10−2 mol), m-DVB (1.5 mL, 1.38 g, 1.05 × 10−2 mol),
cyclohexanone (3 mL), and CoPhBF (0.0391 g, 6.18 × 10−5 mol)
were added to a Schlenk flask equipped as described in example 1. The
resulting solution was then rapidly heated to 150 °C and held at that
temperature for 20 min, after which it was taken off and immediately
quenched in liquid nitrogen to prevent gelation. The product was
1
recovered as described in example 1 with a yield of 51%. H NMR
(CDCl3, 400 MHz) at 25 °C: δ = 7.61−5.95 (broad m, aromatic and
−CHCH2), 5.88−5.50 (broad s, 1H, −CHCH cis), 5.35−5.01
(broad s, 1H, −0CHCH trans), 2.83−0.59 (broad m, CH2
(backbone)). GPC: Mn = 6000, Mw = 150 900, Đ = 25.2. MALLS:
Mn = 445 000, Mw = 1 725 000, Đ = 3.9, Rn = 62.3 nm.
Figure 1. Molecular structures of bis[(difluoroboryl)-
diphenylglyoximato]cobalt(II) (CoPhBF) (left), m-DVB (top right),
and p-DVB (bottom right).
Note on Gelation. Gelation was defined as the point at which the
solution ceased to be a free-flowing, easily stirred liquid and became a
rubbery/solidified gel. This point was a readily observable change in
physical form apparent through visual inspection. However, gelation
was confirmed by withdrawing a small sample and attempting to dilute
it in chloroform. Where gelation had occurred, material presented as
nonredissolvable translucent particles in the resulting solution. The
time at which this occurred was defined as the gelation time and
represented a change from an oligomeric/hyperbranched system to an
extended cross-linked system.
Refractive Index Measurements. HBDVB dn/dc values were
obtained using the Cirrus Multi Software and verified using a Mettler
Toledo 30GS refractometer which was calibrated against deionized
water. Several concentrations of HBDVB in THF (at both high and
low levels of monomer conversion) were prepared and warmed to 40
°C to replicate the conditions in the GPC system. The refractive index
of each solution was measured and plotted against sample
concentration, yielding a straight line with a gradient equal to the
dn/dc of the specific polymer sample. The dn/dc values of HBDVB-80
and HBDVB-100 were determined to be 0.215 and 0.228, respectively.
Branching and molecular weight were not observed to affect the
recorded dn/dc.
Gel Permeation Chromatography. Gel permeation chromatog-
raphy (GPC) was performed using a refractive index (RI) detector
with HPLC THF as the eluent. Analysis was performed at 40 °C with
a flow rate of 1 mL min−1 through two PolarGel-M columns with a
calibration range of 580−377 400 Da calibrated with 10 poly(styrene)
narrow molecular weight distribution standards. All GPC equipment
and standards were supplied by Polymer Laboratories (Varian). GPC
data were analyzed using the Cirrus GPC offline software package.
Nuclear Magnetic Resonance. 1H and 13C nuclear magnetic
resonance (NMR) spectra were obtained in CDCl3 on a Bruker
AV3500 (500 MHz) or Bruker AV3400 (400 MHz) spectrometer.
Chemical shifts are referenced against residual solvent signal (1H =
7.26 ppm, 13C = 77.16 ppm) and processed using the MestReC
software package.
It has been shown that the purity of CoPhBF is not
necessarily a good indicator of the expected activity.36 Instead,
the catalytic activity of the batch of CoPhBF which was used
throughout this study was determined to ensure that the results
could be rigorously compared with the prior literature.
Determination of the chain transfer constant (Cs) of this
batch of CCT agent was obtained through a combination of
both Mayo plots37 and chain length distribution (CLD) plots.38
In both methods, several low conversion polymerizations are
conducted containing different concentrations of CTA
following the methodology reported by Suddaby et al.39 The
resulting molecular weight distribution was subsequently
analyzed by GPC and the data manipulated to yield Cs (see
Supporting Information, Figures S1−S4). Styrene was chosen
as a suitable monomer for the determination of Cs for the
CoPhBF used in this study because it is structurally very similar
to DVB and has been extensively studied, and as such direct
comparison could be made to already determined literature
values. At the same time it removes any potential to produce
cross-linked microgels which may complicate the determination
of Cs. The results of the Cs calculations conducted as part of
this study are shown in Table 1.
Table 1. Values of Cs for CoPhBF/Styrene Obtained from
Both Mayo and CLD Plots
Cs
run
Mayo (1/Mn)
Mayo (2/Mw)
CLD
1
507
467
381
359
382
367
2
3
463
357
351
average
479 20
366 13
367 16
Gas Chromatography−Mass Spectroscopy. GC-MS was
performed on a VG/Micromass/Waters AutoSpec in EI+ mode. A
15 m capillary column of type BP-1 with 0.25 mm thickness was
employed, with helium at 80 kPa employed as the carrier gas. The
parameters used for the GC analysis of compounds taken up in
dichloromethane were the following: column temperature, from room
temperature to 50 °C in 3 min, raised up to 280 °C by 4 °C min−1;
injection temperature: 250 °C; sample: 1 mL. Typically, 50 mg of
sample was dissolved in 1 mL of dichloromethane for analysis.
The Cs was determined to be in good agreement with the
reported results of Heuts et al., who reported a Cs value of 400
for CoPhBF in styrene when using the number-average
molecular weight distribution (Mn) as the data for the Mayo
method36 and 657 and 478 for the same system when using the
weight-average molecular weight (Mw) in Mayo and the CLD
method analysis, respectively.40 It is worthy of noting that they
concluded greater accuracy was obtained from using the
9260
dx.doi.org/10.1021/ma3018684 | Macromolecules 2012, 45, 9258−9266

Macromolecules
Article
weight-average molecular weight distributions. This is sup-
ported by the data from this study, as smaller error ranges are
obtained from the application of both the Mayo and CLD
methods when using Mw to analyze the GPC data (Table 1).
The Mw-based Cs value obtained in this study was 366 15,
indicating that the catalyst had an acceptable level of activity for
the intended application and comparable with the literature.
Polymerization of Divinylbenzene. The initial polymer-
ization of DVB-80 was attempted using 2,2′-azobis-
(isobutyronitrile) (AIBN) as the source of initiating radicals
(Table 2, entries A1−A3).
DVB molecule means that a greater proportion of random
molecular collisions will have the correct geometry and
associated energy to result in the formation of an initiating
species/chain propagation, while the weakly electron-donating
nature of the vinyl groups in p-DVB may enhance the reactivity
of the monomer toward radical addition. If this latter factor is
the dominant effect, then this increased affinity for radical
addition may not be seen in m-DVB since electron donating
groups tend to be ortho/para directing in aromatic benzene-
type systems.
CoPhBF was reintroduced into the system, and a series of
reproducibility studies were undertaken to assess the duration
of polymerization required to achieve significant recoverable
yields of hyperbranched materials. A typical yield of 22% was
achieved in 30 min when 391 mg of CCT agent was introduced
(Table 2, entries B2−B4) which is a reduction of 18 h 30 min
in the reaction time compared to the previous AIBN-initiated
system (Table 2, entry A3). When the reaction time was
increased to 45 min, the average recoverable yield rose to 40%
(Table 2, entries B5−B7). The isolated material was
determined to be completely soluble in a range of organic
solvents (tetrahydrofuran, ethyl acetate, dichloromethane, and
chloroform). However, extending the reaction time beyond 45
min resulted in the onset of gelation (Table 2, entry B8).
Repeat experiments defined that this system typically became
insoluble/cross-linked in a reaction time range of 46−48 min.
Therefore, where these reaction/concentration conditions were
utilized, all reactions were quenched at 45 min to avoid
insoluble material being present in the final product.
Table 2. Conditions and Obtained Yields for the AIBN
a
Initiated or Autopolymerization of DVB-80
ID
AIBN (mg) temp (°C) CoPhBF (mg) time (min) yield (%)
A1
A2
A3
B1
B2
B3
B4
B5
B6
B7
B8
39.4
39.4
65.7
90
90
782
391
391
2880
1440
1140
10
4.1
18
gel
22
27
17
38
42
41
Gel
90
150
150
150
150
150
150
150
150
391
391
391
391
391
391
391
30
30
30
45
45
45
48
a
In each case 30 mL of DVB and 30 mL of cyclohexanone were used.
Determination of the Composition of DVB-80 by GC.
Analysis of the composition of the DVB-80 solution was
undertaken via gas chromatography (GC) to assess precisely
what monomeric species were present in the polymerization
mixture. Four unique compounds were identified in the GC
trace of DVB-80; 3- and 4-ethylstyrene which eluted at 17.85
and 18.15 min, and m-DVB and p-DVB which eluted at 19.75
and 20.25 min, respectively (Supporting Information, Figure
S5). The area under the GC peaks is a quantitative
representation of the amount of compound present in the
sample. However, to correctly ascertain the relative sensitivities
of detection of each compound by the UV detector present in
the GC instrument, it is necessary to determine the relative GC
response factors (fi) of each material. These can be calculated
using eq 1.
It was observed that under these conditions lengthy
polymerization times were required to obtain any polymeric
product. Initial CoPhBF concentrations were observed to be
too high, yielding only oligomeric products that were too low in
molecular weight to be isolated by filtration from the
supernatant after 48 h. Reducing the catalyst concentration
by 50% led to minimal improvements in yields (Table 2, A2)
while increasing the AIBN concentration by 66% did improve
the yields (Table 2, A3). However, this strategy was
unsustainable with the level of AIBN increasing to scales
large enough to render the process unsafe.
Consequently, the autopolymerization of DVB-80 was
attempted at an elevated temperature of 150 °C with the
target of significantly reducing reaction times. Under these
conditions polymerization proceeds through either a diradical-
type mechanism as suggested by Flory or a molecule-assisted
homolysis (MAH) mechanism as proposed by Mayo.41,42
Recent literature has favored the latter mechanism in which a
Diels−Alder cycloaddition of two vinyl groups occurs. This
leads to the formation of a cyclic intermediate; subsequent loss
of a hydrogen atom from this intermediate results in the
formation of the initiating radicals.43 Use of autopolymerization
over conventional azo-thermal radical initiators that are
typically used for CCT polymerization would allow a
significantly higher reaction temperature to be employed by
removing the dependence on half-life temperatures. The results
of the initial autopolymerization reactions are detailed in Table
2 along with lower temperature AIBN-initiated polymerizations
for comparison.
Ai
Ast
fi =
fst
(1)
Where Ai is the unknown sample area, Ast is the standard peak
area, and fst is the standard response factor for the machine, in
this case 1.00 and as such can be ignored. A dichloromethane
solution containing 1:1 molar ratios of m-DVB and p-DVB was
prepared and subjected to repeated GC analysis, and the data
obtained are shown in Table 3.
Thus, an average response factor of fi = 1.11 was defined for
m-DVB where fi for p-DVB was set to 1; i.e., the detector is
∼10% more sensitive to m-DVB than to p-DVB. It was not
possible to obtain pure samples of 3- and 4-ethylstyrene to
determine their individual response factors. It was noted that
the relative average fi corrected peak area for m-DVB calculated
to be 456 200 compared to 265 200 for p-DVB. From this, it
was concluded that the difunctional components of the DVB-
80 solution contains a mixture of 63% m-DVB and 37% p-DVB.
Initial autopolymerizations of DVB-80 conducted in the
absence of CoPhBF (Table 2, entry B1) was observed to be
rapid and resulted in gelation after 10 min of reaction time. The
rapid autopolymerization observed was attributed to a number
of potential reasons; the presence of two vinyl units on each
9261
dx.doi.org/10.1021/ma3018684 | Macromolecules 2012, 45, 9258−9266

Macromolecules
Article
more, the GC analysis showed that the rate of consumption of
m-DVB was observed to be significantly slower than the
consumption of p-DVB. Analysis of the standardized, response
factor corrected data set shows that the rate of consumption of
m-DVB is only 2% after 5 min, 4% after 10 min, and a total of
34% conversion after the full 22 min reaction period. In
contrast, the consumption figures for p-DVB were 12%, 44%,
and 70%, respectively, at the same time intervals. Plotting these
conversions as a ln(1/1 − c) (where c is the decimal monomer
conversion) vs time graph yielded two straight lines with
gradients equal to kapp as shown in Figure 2.
Table 3. Relative Peak Areas of a Dichloromethane Solution
Containing 1:1 Molar Ratio of p-DVB:m-DVB As
Determined by GC
peak area
monomer
run 1
run 2
run 3
m-DVB
417 600
372 800
1:1.12
365 700
316 800
1:1.15
451 400
424 400
1:1.06
p-DVB
ratio (p:m)
The total combined area of the monofunctional peaks was
found to typically account for 22% of the total peak area,
suggesting the commercially quoted 80% DVB content was
valid.
Monitoring Monomer Conversion via GC. With the
composition of the solution established and relative response
factors determined, GC analysis of samples taken from the
reaction solutions were used to follow the conversion of all the
monomers present in the system with time. First, the apparent
rates of polymerization (kapp) for m- and p-DVB in a system
free of 3- and 4-ethylstyrene were determined by following the
progress of an autoinitiated polymerization. The reaction
medium comprised of a 1:1 ratio of pure m- and p-DVB,
prepared by mixing together samples of the pure monomers
synthesized following the methods detailed in the Experimental
Section. This polymerization was repeated three times so that
average peak areas could be used to gain an estimation of the
error associated with this methodology. The polymerization
reaction results are shown in Table 4, and the response factor
corrected comparison of the average monomer peak areas with
time is detailed in Table 5.
Figure 2. Kinetics plot detailing the relative polymerization rates of m-
DVB and p-DVB. The gradient of each line is equal to kapp: p-DVB
(red) 10.62 × 10−4 0.75 × 10−4; m-DVB (black) 3.97 × 10−4 0.2
× 10−4.
As shown in Figure 2, the kapp for p-DVB is over 2.5 times
greater than that of m-DVB. Further confirmation of this trend
was provided by observing the gelation times for a subsequent
series of homoautopolymerizations conducted using 100% m-
and p-DVB as their respective starting materials following the
same standard polymerization method (Table 6).
Table 4. Reaction Times and Polymer Yields from the Mixed
a
m- and p-DVB Kinetics Measurements
ID
monomers
ratio
time (min)
yield (%)
C1
C2
C3
m-DVB/p-DVB
m-DVB/p-DVB
m-DVB/p-DVB
1:1
1:1
1:1
23.0
24.0
22.5
49
53
51
Table 6. Conditions for the Homoautopolymerization of m-
or p-DVB To Determine Gelation Times
a
a
The reactions were quenched when the viscosity appeared to
ID
monomer
time (min)
approach the point of gelation.
D1
D2
D3
D4
p-DVB
p-DVB
m-DVB
m-DVB
16
17
33
32
Table 5. Average GC Peak Area Values for Three
Polymerizations of 1:1 Molar Ratios of m-DVB and p-DVB
after Standardization of the Peak Areas against Residual
Solvent and Being Response Factor Corrected
a
3 mL of DVB and 3 mL of cyclohexanone were used in each case in
the presence of 39.1 mg of CoPhBF. The reaction temperature was
held at a constant 150 °C until the point of gelation had occurred.
standardized average peak areas
time (min)
m-DVB
p-DVB
0
5
98 692
96 873
94 625
76 433
71 246
64 702
96 446
84 580
54 447
32 642
30 471
28 509
p-DVB was found to gel in half the time required for m-DVB,
further confirming that p-DVB is much more susceptible to
autopolymerization than m-DVB. This was believed to be due
to a combination of differences in steric and electronic effects
between the two monomers. Furthermore, the magnitude of
this reduction in the gelation time is of the order that would be
predicted by taking into account the difference in the kapp values
for these monomers generated from the gas chromatogram data
from the 1:1 mixed monomer kinetic experiments. These
empirical data sets support the earlier proposal made on the
relative reactivity of DVB and styrene by demonstrating that
the weakly electron-donating effects of the vinyl groups into the
electron system in p-DVB do indeed serve to make them more
susceptible to radical attack, and since these electron-donating
10
18
20
22
Gelation times were found to vary between 22.5 and 24 min
over the three repeats, significantly shorter than the DVB-80
reactions. This suggests that the presence of the monofunc-
tional monomer may be retarding the overall reaction rate
observed with DVB-80. Thus, 22 min was identified as the last
acceptable data point when following the reaction kinetics of
the 1:1 m-/p-DVB mixture polymerizations via GC. Further-
9262
dx.doi.org/10.1021/ma3018684 | Macromolecules 2012, 45, 9258−9266

Macromolecules
Article
effects tend to be ortho/para directing, this effect is not
observed in m-DVB.
any chain ends that contain ethylstyrene radicals. However, we
note that it is difficult to be definitive on the exact nature of the
reduction in kapp in such mixed feed systems due to the
significant role that penultimate unit effects may have in such
copolymerizations.44 Further evidence that ethylstyrene led to
reduced reactivity of the hyperbranched system came from
comparing the polymerization of DVB-80 against a second
commercially available feedstock which contained only 55%
DVB, with the rest of the solution consisting of 3- and 4-
ethylstyrene. As expected, a higher content of monofunctional
ethylstyrene monomer led to a dramatic decrease in the
poly(DVB) yield from ∼40% in 45 min in DVB-80 solutions to
only 12% after 60 min reaction time in DVB-55 solutions.
Gel Permeation Chromatography and Multiangle Laser
Light Scattering Analysis. Both GPC and MALLS were
utilized to follow the changes in molecular weight of the
HBDVB materials produced via the polymerization of both the
1:1 m- and p-DVB mixture and DVB-80 over the course of the
reaction, where time-related sampling was conducted as earlier
described. Both the GPC (Figure 4) and the MALLS
(Supporting Information, Figure S6) of the polymers isolated
from the 1:1 m- and p-DVB polymerizations clearly show that
as the reaction progresses, both the molecular weight increases
and the molecular weight profile broadens significantly, taking
on the characteristic multimodal peak shape attributed to
branched structures, providing evidence of the formation of
hyperbranched structures.33
Since a retardation of the observed polymerization rate had
been observed with mixed monomer systems, a kinetics study
was undertaken to assess the impact of the presence of both the
monofunctional 3- and 4-ethylstyrene and the uneven p-
DVB:m-DVB ratio found in DVB-80. The results of this
kinetics study are shown in Figure 3.
Figure 3. GC generated kinetics plot detailing the polymerization rates
of m-DVB, p-DVB, 3-ethylstyene, and 4-ethylstyrene in the
autopolymerization of DVB-80. The gradient of each line is equal to
kapp: p-DVB (red) 7.55 × 10−4 0.84 × 10−4; m-DVB (black) 2.89 ×
10−4 0.17 × 10−4; 4-ES (blue) 1.34 × 10−4 0.02 × 10−4; 3-ES
(green) 8.66 × 10−5 0.08 × 10−5.
Furthermore, both of these data sets suggest that these
reactions comply with the classical statistics of the formation of
hyperbranched materials, as suggested by Flory. This states that
linear polymeric structures with little or no branching will
initially be formed and that branching will only occur as the
polymerization progresses.45 Both the GPC and MALLS data
for the first, lowest conversion kinetic sample, taken after only 5
min, demonstrated that the polymer formed up to this point
had a single distribution with a relatively narrow molecular
weight profile (Đ(MALLS) = 1.74, Đ(GPC) = 2.15), which is
characteristic of a linear polymeric material. At this stage, the
Mw of the material was also relatively low (MALLS = 8900 g
mol−1, GPC = 5700 g mol−1). As the reaction progressed, the
The data demonstrate that the kapp of p-DVB was still higher
than that of m-DVB, even when [p-DVB] ≪ [m-DVB]. This
again suggests that the para isomer is more reactive toward
radical addition than the meta isomer. However, while the
difference between the m-DVB and p-DVB reaction ratios
remains constant at 1:2.6, the kapp values of both difunctional
monomers are reduced compared to the homopolymerization
systems. This reduction in the kapp values was attributed to the
presence of the monofunctional monomer (the ES contami-
nant) both further diluting the concentrations of the more
reactive difunctional monomers and reducing the reactivity of
Figure 4. RI chromatographs of HBDVB produced from the polymerization of m- and p-DVB that has been isolated at different reaction times: 5
(black), 10 (red), 15 (blue), 18 (green), 20 (brown), and 22 min (magenta).
9263
dx.doi.org/10.1021/ma3018684 | Macromolecules 2012, 45, 9258−9266

Macromolecules
Article
molecular weight and distribution profiles of the polymers
started to increase dramatically, and a multimodal distribution
was generated, indicating that the polymerization was
producing nonlinear structures. It is proposed that gelation in
this system is delayed because there are significant amounts of
monomer still in the system, combined with the fact that the
CCT agent ensures that the branch lengths are very short,
keeping the reactive chain ends within a relatively high sterically
hindered environment. These two factors combine to promote
monomer-hyperbranch/oligomer-hyperbranch interaction and
discourage hyperbranch−hyperbranch coupling. However,
further propagation will eventually deplete the monomer to
the point that the reaction balance will shift to being dominated
by the combination of hyperbranched macromolecules,
resulting in gelation. This highlights the reason why maximum
yields for a successful hyperbranching polymerization are
between 40 and 50% which typically equates to a monomer
conversion of ∼70%.34
Further evidence for the formation of a branched structure is
provided by the fact that the molecular weight reported by the
MALLS instrument is consistently higher than the molecular
weight reported by GPC. This is because refractive index (RI)
detectors on the GPC system used in this study are only
sensitive to the relative concentration of the sample being
analyzed, whereas MALLS analysis is proportional to both
concentration and molecular weight. Furthermore, MALLS is
typically more representative of the true Mw and dispersity
values (Đ) values for a sample of material which contains 3D
macromolecules.33,46 This is confirmed by the comparison of
the GPC and MALLS data for polymers manufactured using
1:1 m- and p-DVB mixture (sample C3) and DVB-80 (sample
B6) feedstocks (Figure 5).
Table 7. Comparison of Data obtained via MALLS and GPC
for HBDVB Samples C3 and B6 Produced from the
Polymerization of a 1:1 Mix of m- and p-DVB and DVB-80,
Respectively
MALLS data
GPC RI data
sample ID
Mn
Mw
Đ
Mn
Mw
Đ
B6
C3
76 000
602 600
7.92
4.48
4800
6000
111 200
150 900
23.16
25.22
431 000
1 929 000
mixture of pure m- and p-DVB as opposed to use of the DVB-
80 feedstock. There are several reasons why the exclusion of
monofunctional monomer may lead to higher molecular
weights. First, the introduction of monofunctional monomer
will result in less vinyl functionality on the backbone of
branched molecules, leading to fewer reactive centers that can
propagate the structure to produce very high molecular weights.
Second, as already discussed, the DVB-80 polymerization
mixture contains approximately 48% m-DVB and 32% p-DVB.
However, because of the differing reactivity of these two
monomers, it is thought that the initial polymeric material will
consist primarily of p-DVB, with m-DVB being incorporated
more at the later stages of the polymerization. This will result in
more m-DVB being incorporated toward the “surface” of the
3D macromolecule, dramatically slowing the molecular weight
built from these points in the structure attributed to the
reduced reactivity of the chain end.
Intrinsic Viscosity Determination and Comparison to
Poly(styrene). Hyperbranched/dendritic materials display
markedly lower melt viscosities than corresponding linear
polymers of the same molecular weight.47−49 As a direct result
of the increased crowding in the molecular structure,
entanglements both inside branched structures and between
individual macromolecules are significantly reduced. This
means that entanglement dominated physical properties, such
as melt viscosity, should be severely affected.50 Thus, a classical
Mark−Houwink−Sakurada (MHS) plot of log(Mw) vs log(η)
of linear polymer standards and the corresponding hyper-
branched structure provides further evidence of the occurrence
of branching in the macromolecule, since the linear polymer
should have a significantly higher MHS exponent (gradient)
than the hyperbranched system. In this study, poly(styrene)
standards have been used as a direct comparison to HBDVB
because of the unavailability of linear poly(DVB). The intrinsic
viscosity of THF solutions of each molecular weight of the
material was determined using a BV-400 (Polymer Laborato-
ries) bridge viscometer attached in-line to the GPC system. In
all cases, the value of Mw was determined by MALLS to provide
a consistent and standardized set of results.
Figure 5. Overlay of the MALLS and GPC chromatographs of
materials C3 and B6 produced from the polymerization of a 1:1 mix of
m- and p-DVB and DVB-80, respectively.
From the data contained in Figure 6, the MHS exponent of
poly(styrene) was calculated to be ∼0.74 at 40 °C, which
compares well with the widely accepted literature value for the
MHS exponent of 0.70 at 40 °C.51 Additionally, it was
calculated through subsequent repeats that the MHS exponent
determined in this study for poly(styrene) had an associated
error of ∼6% (0.74 0.04). This level of error (both in the
viscometer and in the determination of Mw) was considered
acceptable for the intended purpose of demonstrating the
significant difference in the MHS exponent for linear
poly(styrene) and HBDVB. The MHS exponent of the
HBDVB produced in this study was 0.19 0.01, significantly
below that of poly(styrene), and was in agreement with
previously reported values.33 This much reduced MHS
While the Mn, Mw, and Đ values obtained from the GPC RI
detectors for the products from both feedstocks were found to
be comparable (Table 7) and exhibit distributions with very
similar profiles (Figure 5), the data obtained via MALLS clearly
show that the sample from the DVB-80 consists of lower
molecular weight material. This is reflected in Table 7, with the
MALLS generated Mw of material C3 being over 3 times higher
than material B6. However, Đ values acquired via MALLS are
significantly reduced due to the MALLS under-representing the
lower molar mass data.
In all cases, both MALLS and GPC analysis of the materials
demonstrated that higher molecular weight hyperbranched
materials were obtained through the polymerization of a
9264
dx.doi.org/10.1021/ma3018684 | Macromolecules 2012, 45, 9258−9266

Macromolecules
Article
quently, the homoautopolymerization of p-DVB was found to
reach the point of gelation in half the time of m-DVB.
Additionally, kinetic analysis of an autoinitiated polymerization
of a 1:1 mixture of p-DVB:m-DVB showed that this level of
differential in reactivity was maintained when in they were used
as a monomer mixture, as the ratio of their calculated kapp was
determined to be 2.6:1. The inclusion of the monofunctional
materials in DVB-80 and the disparity in the difunctional
monomer concentrations was observed to result in the
polymerizations taking longer to reach the gel point in the
autopolymerization of DVB-80 and also in DVB-55.
ASSOCIATED CONTENT
* Supporting Information
Additional GPC data detailing the determination of chain
transfer constants (Cs), a GC trace of the DVB-80 solution, and
MALLS traces for the copolymerization of m-/p-DVB. This
material is available free of charge via the Internet at http://
pubs.acs.org.
Figure 6. Plot of log(η) (intrinsic viscosity) vs log(Mw) for
poly(styrene) standards (×) and HBDVB ( ). The MHS exponent
(α) was shown to be approximately 0.74 for poly(styrene) and 0.19 for
■
□
S
HBDVB.
exponent for HBDVB compared to poly(styrene) is a direct
indication of a lower intrinsic viscosity and provides further
evidence of high levels of branching within these macro-
molecules.
AUTHOR INFORMATION
Corresponding Author
*E-mail: Derek.Irvine@nottingham.ac.uk.
Notes
■
During the formation of highly branched polymers it is
widely believed that intramolecular cyclization occurs, hence
leading to a deviation from the Flory−Stockmayer theory which
predicts that insoluble polymer networks would be formed at
monomer conversion levels much lower than have been
reported in the current literature.33,52−57 Direct quantification
of the levels of cyclization in these species is typically difficult.
To date, the only report establishing a method for
quantification of cyclics in branched structures by Rosselgong
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors acknowledge the support of the DICE initiative
(EP/D501229/1) (D.J.I.) and the EPSRC (I.A.B.) for funding
and Dr. Mick Cooper for his help with GC-MS data presented
in this work.
58
1
and Armes relies of H NMR spectroscopy. Analysis of the
1
HBDVB by H NMR is complicated by signal broadening and
poor peak resolution because of a decrease in molecular
mobility. By virtue of the large broadening of the peaks,
accurate integration of individual peaks for structural assign-
ment was not possible with a sufficient degree of confidence,
and work to improve on this is the subject of ongoing
investigation.
REFERENCES
■
(1) Buhleier, E.; Wehner, W.; Vogtle, F. Synthesis 1978, 2, 155−158.
(2) Frechet, J. M. J. Science 1994, 263 (5154), 1710−1715.
(3) Newkome, G. R.; Yao, Z. Q.; Baker, G. R.; Gupta, V. K. J. Org.
Chem. 1985, 50 (11), 2003−2004.
(4) Tomalia, D. A.; Baker, H.; Dewald, J.; Hall, M.; Kallos, G.;
Martin, S.; Roeck, J.; Ryder, J.; Smith, P. Polym. J. (Tokyo, Jpn.) 1985,
17 (1), 117−132.
(5) Tomalia, D. A.; Baker, H.; Dewald, J.; Hall, M.; Kallos, G.;
Martin, S.; Roeck, J.; Ryder, J.; Smith, P. Macromolecules 1986, 19 (9),
2466−2468.
(6) Dusek, K.; Somvarsky, J.; Smrckova, M.; Simonsick, W. J.;
Wilczek, L. Polym. Bull. (Berlin) 1999, 42 (4), 489−496.
(7) Nunez, C. M.; Chiou, B.-S.; Andrady, A. L.; Khan, S. A.
Macromolecules 2000, 33 (5), 1720−1726.
(8) Uppuluri, S.; Keinath, S. E.; Tomalia, D. A.; Dvornic, P. R.
Macromolecules 1998, 31 (14), 4498−4510.
CONCLUSIONS
■
This work has clearly shown that CCT polymerization can
successfully be utilized to synthesize hyperbranched polymers
of difunctional styrenyl monomers, in both the presence and
absence of monofunctional monomers. Furthermore, this can
be achieved in less than 1 h with good batch to batch
reproducibility by using high reaction temperatures (150 °C)
and autoinitiation as the means of radical generation. The
proposed branched molecular structures of the resulting
polymers was confirmed by comparison of the GPC-RI and
MALLS data of the products obtained and further confirmed by
comparison of intrinsic viscosity values to those of linear
poly(styrene) standards and the DVB products. The former
demonstrated that there is a significant difference between the
data collected by these different detectors, which generally
indicates 3D structure with the latter demonstrating that the
MHS exponent was 0.74 for poly(styrene) and 0.19 for
HBDVB, a difference which again indicated a significant level of
branching in materials produced by the autopolymerization of
DVB. It has also been demonstrated that the order of reactivity
of the monomeric species investigated in this study is p-DVB >
m-DVB > 4ES > 3ES under autopolymerization conditions as
determined by monitoring changes in the composition of the
polymerization system as the reaction proceeded. Conse-
(9) Kukowska-Latallo, J. F.; Candido, K. A.; Cao, Z. Y.; Nigavekar, S.
S.; Majoros, I. J.; Thomas, T. P.; Balogh, L. P.; Khan, M. K.; Baker, J.
R. Cancer Res. 2005, 65 (12), 5317−5324.
(10) Malik, N.; Evagorou, E. G.; Duncan, R. Anti-Cancer Drugs 1999,
10 (8), 767−776.
(11) Patri, A. K.; Kukowska-Latallo, J. F.; Baker, J. R. Adv. Drug
Delivery Rev. 2005, 57 (15), 2203−2214.
(12) Thomas, T. P.; Shukla, R.; Kotlyar, A.; Kukowska-Latallo, J.;
Baker, J. R. Bioorg. Med. Chem. Lett. 2010, 20 (2), 700−703.
(13) Froehling, P. E. Dyes Pigm. 2001, 48 (3), 187−195.
(14) Jikei, M.; Kakimoto, M. Prog. Polym. Sci. 2001, 26 (8), 1233−
1285.
(15) Kim, Y. H.; Webster, O. W. J. Am. Chem. Soc. 1990, 112 (11),
4592−4593.
(16) Kim, Y. H.; Webster, O. W. Macromolecules 1992, 25 (21),
5561−5572.
9265
dx.doi.org/10.1021/ma3018684 | Macromolecules 2012, 45, 9258−9266

Macromolecules
Article
(17) Liu, B.; Kazlauciunas, A.; Guthrie, J. T.; Perrier, S. Macro-
molecules 2005, 38 (6), 2131−2136.
(53) Rosselgong, J.; Armes, S. P.; Barton, W. R. S.; Price, D.
Macromolecules 2010, 43 (5), 2145−2156.
(54) Li, Y.; Armes, S. P. Macromolecules 2005, 38 (12), 5002−5009.
(55) Flory, P. J. J. Am. Chem. Soc. 1941, 63 (11), 3096−3100.
(18) Inoue, K. Prog. Polym. Sci. 2000, 25 (4), 453−571.
(19) Costello, P. A.; Martin, I. K.; Slark, A. T.; Sherrington, D. C.;
Titterton, A. Polymer 2002, 43 (2), 245−254.
(56) Butun, V.; Bannister, I.; Billingham, N. C.; Sherrington, D. C.;
̈
̈
Armes, S. P. Macromolecules 2005, 38 (12), 4977−4982.
(57) Stockmayer, W. H. J. Chem. Phys. 1944, 12 (4), 125−131.
(58) Rosselgong, J.; Armes, S. P. Macromolecules 2012, 45 (6), 2731−
2737.
(20) Guan, Z. J. Am. Chem. Soc. 2002, 124 (20), 5616−5617.
(21) Burczyk, A. F.; Odriscoll, K. F.; Rempel, G. L. J. Polym. Sci., Part
A: Polym. Chem. 1984, 22 (11), 3255−3262.
(22) Enikolopyan, N. S.; Smirnov, B. R.; Ponomarev, G. V.;
Belgovskii, I. M. J. Polym. Sci., Part A: Polym. Chem. 1981, 19 (4),
879−889.
(23) Smirnov, B. R.; Belgovskii, I. M.; Ponomarev, G. V.; Marchenko,
A. P.; Enikolopian, N. S. Dokl. Akad. Nauk SSSR 1980, 254 (1), 127−
130.
(24) Smirnov, B. R.; Morozova, I. S.; Pushchaeva, L. M.; Marchenko,
A. P.; Enikolopian, N. S. Dokl. Akad. Nauk SSSR 1980, 255 (3), 609−
612.
(25) Smirnov, B. R.; Morozova, I. S.; Marchenko, A. P.; Markevich,
M. A.; Pushchaeva, L. M.; Enikolopian, N. S. Dokl. Akad. Nauk SSSR
1980, 253 (4), 891−895.
(26) Wang, W. X.; Stenson, P. A.; Marin-Becerra, A.; McMaster, J.;
Schroder, M.; Irvine, D. J.; Freeman, D.; Howdle, S. M. Macromolecules
2004, 37 (18), 6667−6669.
(27) Haddleton, D. M.; Topping, C.; Kukulj, D.; Irvine, D. Polymer
1998, 39 (14), 3119−3128.
(28) Frechet, J. M. J.; Henmi, M.; Gitsov, I.; Aoshima, S.; Leduc, M.
R.; Grubbs, R. B. Science 1995, 269 (5227), 1080−1083.
(29) Baskaran, D. Polymer 2003, 44 (8), 2213−2220.
(30) Kato, M.; Kamigaito, M.; Sawamoto, M.; Higashimura, T.
Macromolecules 1995, 28 (5), 1721−1723.
(31) Wang, J. S.; Matyjaszewski, K. Macromolecules 1995, 28 (23),
7901−7910.
(32) Percec, V.; Barboiu, B. Macromolecules 1995, 28 (23), 7970−
7972.
(33) Wang, W. X.; Zheng, Y.; Roberts, E.; Duxbury, C. J.; Ding, L. F.;
Irvine, D. J.; Howdle, S. M. Macromolecules 2007, 40 (20), 7184−7194.
(34) Koh, M. L.; Konkolewicz, D.; Perrier, S. Macromolecules 2011,
44 (8), 2715−2724.
(35) Nyhus, A. K.; Hagen, S.; Berge, A. J. Polym. Sci., Part A: Polym.
Chem. 1999, 37 (16), 3345−3359.
(36) Heuts, J. P. A.; Kukulj, D.; Forster, D. J.; Davis, T. P.
Macromolecules 1998, 31 (9), 2894−2905.
(37) Mayo, F. R. J. Am. Chem. Soc. 1943, 65, 2324−2329.
(38) Clay, P. A.; Gilbert, R. G. Macromolecules 1995, 28 (2), 552−
569.
(39) Suddaby, K. G.; Maloney, D. R.; Haddleton, D. M.
Macromolecules 1997, 30 (4), 702−713.
(40) Heuts, J. P. A.; Forster, D. J.; Davis, T. P.; Yamada, B.; Yamazoe,
H.; Azukizawa, M. Macromolecules 1999, 32 (8), 2511−2519.
(41) Mayo, F. R. J. Am. Chem. Soc. 1968, 90 (5), 1289−1295.
(42) Flory, P. J. J. Am. Chem. Soc. 1937, 59, 241−253.
(43) Chong, Y. K.; Rizzardo, E.; Solomon, D. H. J. Am. Chem. Soc.
1983, 105 (26), 7761−7762.
(44) Coote, M. L.; Davis, T. P. Prog. Polym. Sci. 1999, 24 (9), 1217−
1251.
(45) Flory, P. J. J. Am. Chem. Soc. 1952, 74 (11), 2718−2723.
(46) Podzimek, S. J. Appl. Polym. Sci. 1994, 54 (1), 91−103.
(47) Malmstrom, E.; Hult, A.; Gedde, U. W.; Liu, F.; Boyd, R. H.
Polymer 1997, 38 (19), 4873−4879.
(48) Degennes, P. G.; Hervet, H. J. Phys., Lett. 1983, 44 (9), L351−
L360.
(49) Maciejewski, M. J. Macromol. Sci., Chem. 1982, A17 (4), 689−
703.
(50) Hawker, C. J.; Farrington, P. J.; Mackay, M. E.; Wooley, K. L.;
Frechet, J. M. J. J. Am. Chem. Soc. 1995, 117 (15), 4409−4410.
(51) Strazielle, C.; Benoit, H.; Vogl, O. Eur. Polym. J. 1978, 14 (5),
331−334.
(52) Gao, H.; Li, W.; Matyjaszewski, K. Macromolecules 2008, 41 (7),
2335−2340.
9266
dx.doi.org/10.1021/ma3018684 | Macromolecules 2012, 45, 9258−9266