Synthesis of branched methacrylic Copolymers: Comparison between RAFT and ATRP and effect of varying the monomer concentration

Article abstract of DOI:10.1021/ma902757z

The statistical copolymerization of methyl methacrylate (MMA) with varying amounts of a disulfide-based dimethacrylate (DSDMA) branching comonomer in toluene at 90 °C can lead to highly branched soluble methacrylic copolymers under appropriate conditions.

Full text of DOI:10.1021/ma902757z

Macromolecules 2010, 43, 2145–2156 2145
DOI: 10.1021/ma902757z
Synthesis of Branched Methacrylic Copolymers: Comparison between
RAFT and ATRP and Effect of Varying the Monomer Concentration
Julien Rosselgong and Steven P. Armes*
Dainton Building, Department of Chemistry, The University of Sheffield, Sheffield, South Yorkshire,
S3 7HF, U.K.
William R. S. Barton and David Price
The Lubrizol Corporation, Hazelwood Research Centre, Derbyshire, DE5 61QN, U.K.
Received December 14, 2009; Revised Manuscript Received January 28, 2010
ABSTRACT: The statistical copolymerization of methyl methacrylate (MMA) with varying amounts of a
disulfide-based dimethacrylate (DSDMA) branching comonomer in toluene at 90 °C can lead to highly
branched soluble methacrylic copolymers under appropriate conditions. This model system is utilized in
order to examine the following points: (i) the relative merits of using reversible addition-fragmentation chain
transfer (RAFT) polymerization and atom transfer radical polymerization (ATRP) in such syntheses; (ii) the
effect of varying the number of DSDMA units per primary chain; (iii) the effect of varying the initial
monomer concentration. Kinetic studies of the linear RAFT and ATRP homopolymerizations conducted in
the absence of any DSDMA confirmed their living character at 10, 30, and 50 wt % [MMA]0, where the
former monomer concentration approximately corresponds to the critical overlap concentration, c*, for
linear poly(methyl methacrylate) (PMMA) chains with a mean degree of polymerization of 50. HPLC
analysis of the monovinyl and divinyl comonomers confirms that there is statistical incorporation of the
DSDMA brancher into the growing poly(methyl methacrylate) chains. Cleavage of both RAFT- and ATRP-
synthesized branched copolymers prepared at 50 wt % [MMA]0 using tributylphosphine affords linear
primary chains with narrow molecular weight distributions; thus these retro-syntheses demonstrate the
retention of living character under branching conditions and suggest little or no chain transfer to polymer. In
principle, macroscopic gelation can be avoided provided that the number of fully reacted divinyl branching
comonomers per primary chain is less than unity. Taking into account the respective efficiencies of the RAFT
chain transfer agent and the ATRP initiator, this hypothesis holds for both ATRP and RAFT branching
copolymerizations conducted in the presence of DSDMA at 50 wt % [MMA]0 but fails at 10 wt % [MMA]0.
Thus, soluble branched copolymers can be prepared at 10 wt % [MMA]0 containing up to five fully reacted
DSDMA units per primary chain using RAFT chemistry and up to three fully reacted DSDMA units per
primary chain with the ATRP formulation; no gelation is observed even when the overall conversion of vinyl
groups exceeds 96%. These observations strongly suggest that intramolecular cyclization is prevalent at this
lower monomer concentration, regardless of the precise nature of the polymerization chemistry. In contrast,
intermolecular branching between primary chains is evidently favored at 50 wt % [MMA]0, since this
concentration substantially exceeds c*. In summary, although there are no doubt some subtle
differences between branched copolymers synthesized via RAFT and ATRP chemistry, physical factors
are arguably much more important than the precise nature of the living radical polymerization chemistry; in
particular, systematic variation of the monomer concentration clearly leads to fundamentally different
behavior.
Introduction
reacted divinyl branching comonomers per primary chain is less
than unity. This concept follows from the following simple
Free radical copolymerization of a monovinyl monomer with a
small amount of divinyl monomer usually leads to gelation: an
insoluble gel network is obtained when there is an average of two
or more branch points per chain. Assuming equal comonomer
reactivities, no intramolecular cyclization and perfectly mono-
disperse primary chains, classical Flory-Stockmayer theory
predicts that the condition for the onset of gelation should be
0.50 units of fully reacted divinyl branching comonomer per
primary chain.1-3 However, in a recent series of papers,4-18
Sherrington and co-workers have empirically shown that macro-
scopic gelation can be avoided provided that the number of fully
argument: cross-linking N chains together to form a single giant
molecule requires at least N - 1 fully reacted divinyl comono-
mers. For macroscopic gelation, N is large in the high molecular
weight limit so (N - 1)/N, or the minimum number of divinyl
comonomers per primary chain, x, tends to unity. Using this
hypothesis, the Strathclyde group developed a facile, robust one-
pot synthetic route for the synthesis of various soluble branched
copolymers using conventional free radical polymerization
(FRP) chemistry in solution,4-15 emulsion,16 or suspension.17,18
In their original paper, a thiol chain transfer agent is used to
reduce the mean primarychain lengthand hence suppressgelation
in the branching copolymerization of methyl methacrylate with
but-2-ene-1,4-diacrylate.4 This FRP approach was later extended
by both Sherrington’s group5-10 and other research groups to
*To whom correspondence should be addressed. E-mail: S.P.Armes@
sheffield.ac.uk.
r 2010 American Chemical Society
Published on Web 02/10/2010
pubs.acs.org/Macromolecules

2146 Macromolecules, Vol. 43, No. 5, 2010
Rosselgong et al.
Scheme 1. Synthesis of Linear Poly(methyl methacrylate) by RAFT Polymerization (Left) Using Cumyl Dithiobenzoate (CDB) as Chain Transfer
Agent and 1,10-Azobiscyclohexanecarbonitrile (ACCN) Initiator, and ATRP (Right) Using 3-Methylphenyl Bromoisobutyrate Initiator,
Copper(I) Chloride Catalyst and N-(n-Propyl)-2-pyridylmethanimine Ligand at 90°C in Toluenea
a The initial monomer concentration [MMA]0 was varied from 10 to 50 wt. % for both RAFT and ATRP syntheses.
include various other monovinyl and divinyl comonomers,19-22
as well as anionic polymerization,23 cationic polymerization,24
group transfer polymerization,25,26 or controlled/“living” radical
polymerizations (CLRP) such as atom transfer radical poly-
merization26-34 (ATRP) and reversible addition-fragmenta-
tion chain transfer (RAFT) polymerization.35-46 According to a
number of reports, branching copolymerizations conducted under
RAFT conditions usually exhibit strongly non-ideal behavior in
that the number of fully reacted divinyl branching comonomers
per primary chain can significantly exceed unity.35,38,39,42 On the
other hand, at least three research groups have reported that
ATRP copolymerizations conform to the Sherrington hypothesis
that gelation can be avoided if x is less than unity.28,30,31,47 In
addition, Gao and co-workers studied the ATRP copolymeriza-
tion of methyl acrylate with a divinyl cross-linker and found that
the nature of the divinyl cross-linker, the cross-linker/initiator
molar ratio, the reactant concentrations, the ATRP initiator
efficiency and the molecular weight distribution of the primary
chains all influenced the gel point.47-51 This latter set of studies has
been recently reviewed.52,53
approximation, it is reasonable to equate the effect of homopolymer
concentration in such model quaternization reactions to the effect
of monomer concentration in branching copolymerizations. This
insight allows rationalization of much of the otherwise rather con-
fusing literature data on branching copolymerizations.35,38,55
Thus, syntheses of branched copolymers conducted under rela-
tively dilute monomer concentrations (i.e., below or close to c*)
are necessarily much more wasteful of the divinyl branching
comonomer, because neighboring polymer coils do not interpe-
netrate each other in solution and are effectively isolated.46
Moreover, such low monomer concentrations are generally more
typical of RAFT syntheses35,38,55 than ATRPsyntheses.28-31,47,48
Recently, Zhu and co-workers56 compared RAFT, ATRP and
conventional FRP for the statistical copolymerization of oligo-
(ethylene glycol) monomethacrylate with oligo(ethylene glycol)
dimethacrylates in the bulk. The reaction kinetics and gelation
behavior were studied in detail using strain-controlled rheometry:
it was found that microgel formation occurred for the FRP
formulation well below the gel point, whereas macroscopic gela-
tion was obtained for the RAFT and ATRP formulations without
any evidence for prior microgel formation. This difference was
attributed to the suppression of intramolecular cyclization in the
RAFT and ATRP syntheses due to their much slower copolymer
chain growth, which allows sufficient relaxation and diffusion to
ensure evolution of a more ordered copolymer network.
In the present work, we have studied the homopolymerization
of methyl methacrylate (MMA) in toluene at 90 °C by both
ATRP and RAFT, respectively (see Scheme 1). In particular, the
effect of varying the initial monomer concentration ([MMA]0) on
the kinetics and controlled/living character of such homopolymer-
izations was examined in detail. This information was then used to
study the branching copolymerization of methyl methacrylate
(MMA) with a disulfide-based dimethacrylate comonomer
(DSDMA) using both ATRP and RAFT at [MMA]0 = 50, 30,
and 10 wt % (see Scheme 2). The latter monomer concentration
approximately corresponds to the c* calculated for linear (primary)
PMMA chains with a mean target degree of polymerization, DP,
of 50 (see Supporting Information for this calculation).
In a recent model study, Li and co-workers54 quaternized a
near-monodisperse poly[(2-dimethylamino)ethyl methacrylate]
(PDMA) homopolymer using a bifunctional reagent, bis(2-iodo-
ethoxy)ethane (BIEE) to produce highly branched copolymers.
Branching behavior was compared both above and below the coil
overlap concentration, c*, in order to assess the effect of the
concentration of the linear primary PDMA chains in determining
the relative probabilities of intramolecular cyclization and inter-
1
molecular cross-linking. According to H NMR studies, many
more BIEE molecules could be reacted per primary chain without
causing gelation when the [PDMA]o was below c*. This observa-
tion strongly suggests that intramolecular cyclization is prevalent
under these conditions, as expected. Moreover, Bannister et al.
demonstrated that linear primary chains are mainly formed
initially during the ATRP copolymerization of a monovinyl
methacrylic monomer with a dimethacrylate branching comono-
mer, with extensive branching being essentially confined to the
latter stages of this copolymerization.31 Thus, to a zeroth-order

Article
Macromolecules, Vol. 43, No. 5, 2010 2147
Scheme 2. Synthesis of Branched Copolymers by Statistical Copolymerization of Methyl Methacrylate (MMA) with a Disulfide-Based
Dimethacrylate (DSDMA) Comonomer at 90°C in Toluene by (a) RAFT Chemistry Using Cumyl Dithiobenzoate as a Chain Transfer Agent and ACCN
Initiator and (b) ATRP Using 3-Methylphenyl Bromoisobutyrate Initiator, Copper(I) Chloride Catalyst, and N-(n-Propyl)-2-pyridylmethanimine Ligand
Experimental Section
Synthesis of Linear Poly(methyl methacrylate) Homopolymer
by RAFT. The protocol used for the RAFT synthesis of a linear
poly(methyl methacrylate) homopolymer with a target DP of 50
was as follows. Cumyl dithiobenzoate (CDB, 0.654 g, 2.4 mmol)
and methyl methacrylate (MMA, 12.01 g, 120.0 mmol) mono-
mer were weighed into a 50 mL Schlenk flask, degassed using
three freeze-pump-thaw cycles and refilled with nitrogen.
Anhydrous toluene (13.8 mL, 50 wt %) was added via a N2-purged
glass syringe and the mixture was purged with nitrogen for 10 min.
1,10-Azobiscyclohexanecarbonitrile (ACCN, 0.117 g, 0.48 mmol,
CDB/ACCN molar ratio =5:1) initiator was added last under
a positive pressure of nitrogen before the flask was immersed in
a preheated oil bath at 90 °C. Aliquots (typically 0.20 mL) were
periodically extracted for GPC analysis and 1H NMR studies of
the monomer conversion. In the latter case, spectra were recor-
ded in CDCl3 and the signals due to the residual vinyl protons of
MMA monomer at δ 5.63 and 6.19 were compared with the
methyl proton signals due to both the homopolymer and MMA
monomer (δ 3.40-4.00). After 30 h, the polymerization was
terminated by exposure to air and cooling the reaction flask
temperature with liquid nitrogen. Toluene was removed under
reduced pressure and the crude homopolymer was dissolved in
minimal THF before precipitating (twice) into excess n-hexane
to remove any unreacted monomer. Finally, the homopolymer
was dried for 24 h in a vacuum oven at 50 °C to produce a light
pink powder. For the linear polymerizations conducted at lower
MMA concentrations, additional ACCN initiator (0.117 g,
0.48 mmol) was injected after 24 h when [MMA]0 = 30 wt %
and after both 24 and 48 h when [MMA]0 = 10 wt %.
Materials. All monomers, solvents, and other reagents were
purchased from Aldrich at the highest purity available and used
as received, unless otherwise stated. Methyl methacrylate
(MMA, 99%) was passed through an activated basic alumina
column (Brockmann I) to remove inhibitor before use.
1,10-Azobiscyclohexanecarbonitrile (ACCN, 99%) was recrys-
tallized from methanol. Cumyl dithiobenzoate (CDB),57 the
disulfide-based dimethacrylate comonomer (DSDMA; see
Scheme 2)46 and N-(n-propyl)-2-pyridylmethanimine (n-Pr)58
were synthesized following previously reported methods.
Synthesis of 3-Methylphenyl Bromoisobutyrate (MP-Br) Ini-
tiator. 3-Methylphenol (13.52 g, 0.125 mol, 1.0 equiv), distil-
led triethylamine (52.2 mL, 0.375 mol, 3.0 equiv), and
4-(dimethylamino)pyridine (1.53 g, 12.5 mmol, 0.1 equiv) were
dissolved in 250 mL of dry dichloromethane in a 1 L two-neck
round-bottomed flask under a nitrogen atmosphere. This flask
was immersed in an ice bath for 15 min, and then 2-bromoiso-
butyryl bromide (34.49 g, 0.15 mol, 1.2 equiv) was added
dropwise to the stirred solution using an addition funnel over
1 h. The resulting heterogeneous mixture was stirred at 20 °C for
24 h and then filtered to remove the triethylamine hydrobromide
byproduct. This solution was washed three times with a satu-
rated aqueous solution of sodium hydrogen carbonate (500 mL)
and three times with deionized water (500 mL). The purified
organic solution was dried using anhydrous MgSO4, and the
dichloromethane was removed under reduced pressure. The
crude brown product was then purified by column chromatog-
raphy using silica gel as stationary phase and a 1:15 ethyl
acetate/petroleum ether mixed eluent as mobile phase. The final
ATRP initiator (3-methylphenyl bromoisobutyrate; MP-Br)
was obtained as a slightly yellow liquid (32.1 g, 87%) and was
stored in a freezer under nitrogen prior to use. 1H NMR
(CD2Cl2): δ 7.37 (t, 1H), 7.17 (d, 1H), 7.01 (d, 2H), 2.45 (s,
3H), 2.14 (s, 6H); an assigned 1H NMR spectrum of this
compound is shown in the Supporting Information. MS
(EIþ): m/z = 256 Da. Anal. Calcd for C11H13BrO2: C, 51.38;
H, 5.10; Br, 31.08. Found: C, 51.61; H, 5.25; Br, 31.09.
Synthesis of Linear Poly(methyl methacrylate) Homopolymer
by ATRP. The protocol used for the ATRP synthesis of a linear
poly(methyl methacrylate) homopolymer with a target DP of 50
with [MMA]0 = 50 wt % was as follows. The 3-methylphenyl
bromoisobutyrate initiator (MP-Br, 0.617 g, 2.4 mmol), Schiff
base ligand N-(n-propyl)-2-pyridylmethanimine (n-Pr, 0.712 g,
4.8 mmol), and methyl methacrylate (MMA, 12.01 g, 120.0 mmol)
were weighed into a 50 mL Schlenk flask, degassed using three
freeze-pump-thaw cycles, and refilled with nitrogen. Anhydrous

2148 Macromolecules, Vol. 43, No. 5, 2010
Rosselgong et al.
toluene (13.8 mL, 50 wt %) was added via a N2-purged glass
syringe and the mixture was purged with nitrogen for 10 min.
Copper(I) chloride (0.238 g, 2.4 mmol) catalyst was added last
under a positive pressure of nitrogen before the flask was
immersed in a preheated oil bath at 90 °C. Aliquots (typically
0.20 mL) were periodically extracted for GPC analysis and
1H NMR studies of the monomer conversion. In the latter case,
spectra were recorded in CDCl3 and the signals due to the
residual vinyl protons of MMA monomer at δ 5.63 and 6.19
were compared with the methyl proton signals due to both the
polymer and MMA monomer (δ 3.40 to 4.00). After 20 h, the
polymerization was terminated by exposure to air and cooling
the reaction flask temperature with liquid nitrogen. The reaction
mixture was passed through a silica column with THF as eluent
to remove the spent catalyst. The solution was concentrated
under reduced pressure before precipitating (twice) into excess
n-hexane to remove any unreacted monomer. Finally, the
purified homopolymer was dried for 24 h in a vacuum oven at
50 °C to produce an off-white powder.
Synthesis of Poly(methyl methacrylate)-Based Branched Co-
polymers by RAFT or ATRP. The experimental protocol for the
synthesis of branched copolymers was very similar to that used
for the linear polymerizations, except that various amounts of
DSDMA branching comonomer were also added to the for-
mulations. Copolymerizations were conducted at 90 °C and
aliquots were periodically withdrawn for 1H NMR, HPLC, and
GPC characterization of monomer conversions and polymer
molecular weights. For [MMA]0 = 10 wt %, such copolymer-
izations required up to 96 h (RAFT) or 168 h (ATRP) at 90 °C in
order to ensure that high conversions (>95%) were achieved.
Reductive Cleavage of Branched Copolymers Using Tributyl-
phosphine. The PMMA50-DSDMA0.90 branched copolymers
prepared by either RAFT or ATRP (0.200 g, 0.038 mmol
disulfide bonds) were dissolved in 3.0 mL THF containing an
approximately 20-fold excess of tributylphosphine (Bu3P; 15.3 mg,
0.757 mmol) and deionized water (680 μL, 0.038 mmol). These
reaction solutions were stirred at room temperature under a
nitrogen atmosphere for 24 h, concentrated under reduced
pressure, precipitated in n-hexane, filtered, and finally dried
for 24 h in a vacuum oven at 50 °C. The isolated colorless
polymers were then analyzed by GPC (see below). Branched
copolymers containing higher proportions of DSDMA como-
nomer were cleaved using a correspondingly higher amount of
Bu3P such that the Bu3P/disulfide molar ratio was always fixed
at 20.
GPC analysis. All such copolymers were analyzed in THF at a
concentration of 5.0 g/L.
Kinetic Studies of Comonomer Depletion by HPLC. The
following HPLC protocol has been used to monitor the deple-
tion of the MMA monomer and DSDMA branching agent from
both RAFT and ATRP copolymerizing solutions. The Waters
2695 Separations Module HPLC setup comprised Waters Spheri-
sorb S5 ODS2 analytical HPLC column (125 mm ꢀ 4.6 mm), a
Waters UV detector set at 220 nm, and a Waters HPLC pump
operating at a flow rate of 1.0 mL min-1. The HPLC eluent was
a gradient mobile phase initially comprising a mixture of 40%
THF and 60% deionized water. The THF content was increased
from 40% to 75% over 15 min and maintained at 75% THF for
a further 10 min. This protocol not only allowed good discri-
mination between MMA (which elutes at 2.86 min) and
DSDMA (which elutes at 7.19 min) but also ensured good
separation from the other species that were present in the
copolymerizing solution (i.e., ACCN, toluene and copolymer
for the RAFT formulations; transition metal complex, toluene
and copolymer for the ATRP formulations). Thus, this UV
HPLC protocol allowed convenient monitoring of the progres-
sive depletion of both MMA and DSDMA branching comono-
mers from copolymerizing solutions.
Results and Discussion
In order to make a meaningful comparison between
RAFT and ATRP under branching conditions, we sought to
hold constant as many experimental parameters as possible.
MMA was chosen as the model methacrylic monomer since
there are many papers describing MMA-based branched co-
polymers.4-8,10,14,16,18,26,32,35,44,60,61 Moreover, PMMA homo-
polymer is very convenient for GPC analyses, since a wide range
of near-monodisperse PMMAcalibrationstandards are available
at relatively low cost. The DSDMA divinyl comonomer is an
extremely useful model branching agent as it allows facile retro-
synthesis of the primary chains via selective cleavage of its
disulfide bonds under mild reductive conditions.28,42,46,60 A rela-
tively high polymerization temperature of 90 °C was selected to
ensure a fairly rapid rate of propagation under RAFT condi-
tions.62 This temperature was also known to afford good control
over the solution polymerization of MMA in toluene according
to a well-documented ATRP formulation.58,63 Numerous pre-
vious RAFT39 and ATRP31 studies had indicated that a target
degree of polymerization (DP) of 50 for the primary chains
should be sufficient to obtain a high degree of branching within
a reasonable time scale. For the RAFT formulation, CDB was
selected for its well-known excellent control over the polymeri-
zation of methacrylic monomers,64,65 both in terms of its high
chain transfer efficiency and also the narrow molecular weight
distributions of the final polymers. ACCN was chosen as the
radical initiator for its relatively long half-life (approximately
10 h in toluene at 90 °C).66 A subset of our earlier published
RAFT data (previously reported in ref 54) has been included in
this manuscript in order to facilitate close comparison with that
obtained using the ATRP formulation. A phenolic-based ATRP
initiator63 was selected for its relatively high initiator efficiency
in the polymerization of methacrylic monomers63 and the N-(n-
propyl)-2-pyridylmethanimine ligand58 was utilized since it was
known to solubilize (and stabilize) the Cu-based catalyst in
toluene at 90 °C.
Characterization of Branched Copolymers. 1H NMR spectra
were recorded in either CDCl3 or CD2Cl2 using a Bruker AC 400
MHz spectrometer. The molecular weight distributions of the
branched copolymers were examined using a PL-GPC50 inte-
grated GPC system from Polymer Laboratories. Both linear and
branched copolymers were characterized at 30 °C using the
following GPC setup: THF eluent containing 2% v/v triethyla-
mine at a flow rate of 1.0 mL min-1; two 5 μm (30 cm) “Mixed
C” columns from Polymer Laboratories; a WellChrom K-2301
refractive index detector operating at 950 ( 30 nm, a Precision
detector PD 2020 light scattering detector (with scattering
angles of 90° and 15°), and a BV400RT viscosity detector.
Molecular weights of the branched copolymers were determined
by the triple detection method using PL Cirrus Multi online
software (version 2.0) supplied by Polymer Laboratories. A
series of 10 near-monodisperse linear poly(methyl methacrylate)
calibration standards (Mp from 1280 to 330 000 g mol-1) were
purchased from Polymer Laboratories and employed with the
above refractive index detector for the analysis of the linear
PMMA50 homopolymer and the Bu3P-degraded branched co-
polymers. A mean refractive index increment (dn/dc) of 0.079
was taken from the literature for these branched copolymers.59
For kinetic studies, aliquots (typically 0.20 mL) were extracted
from the (co)polymerizing solutions, diluted with THF as
required and ultrafiltered using 0.22 μm Teflon filters prior to
Linear Homopolymerizations Conducted at Various [MMA]0.
The kinetics of MMA homopolymerization at 90 °C using
either ATRP or RAFT was studied at [MMA]0=10, 30, and
50 wt % for a fixed DP = 50. Monomer conversions were
determined by 1H NMR spectroscopy and molecular
weights were obtained by THF GPC, see Table 1. As expec-
ted, a linear relationship was obtained for the evolution of
molecular weight with monomer conversion for both ATRP

Article
Macromolecules, Vol. 43, No. 5, 2010 2149
Table 1. Summary of Final Monomer Conversions, GPC Molecular Weights and Polydispersities of Linear Poly(methyl methacrylate)
Homopolymers Prepared in Toluene at 90°C at Three Different [MMA]0 by Either RAFT Using a Cumyl Dithiobenzoate Chain Transfer Agent
(at Relative Molar Ratios of [MMA]0:[CDB]:[ACCN] = 50:1.0:0.2) or by ATRP Using 3-Methylphenyl Bromoisobutyrate Initiator, Copper(I)
Chloride and N-(n-Propyl)-2-pyridylmethanimine (at Relative Molar Ratios of [MMA]:[MP-Br]:[CuCl]:[n-Pr] = 50:1:1:2)a
conversionb (%)
time (h)
Mn
Mw/Mn
1H NMR efficiencyd (%)
GPC efficiencyd (%)
c
formulation
[MMA]0 (wt %)
ATRP
ATRP
ATRP
RAFT
RAFT
RAFT
50
30
10
50
30
10
99.5
99.9
94.6
96.1
97.0
95.5
20
48
168
30
48
96
5800
5900
5800
5600
5900
6000
1.19
1.27
1.19
1.23
1.25
1.29
90.2
88.9
86.5
89.6
86.9
84.2
90.4
89.2
86.0
90.4
86.6
84.8
a RAFT CTA and ATRP initiator efficiencies are calculated at each [MMA]0. b Monomer conversions were determined by 1H NMR spectroscopy.
c Molecular weight data were obtained by THF GPC using a series of poly(methyl methacrylate) calibration standards. d “Efficiency” refers to either
initiator efficiency (ATRP) or chain transfer agent efficiency (RAFT).
Figure 1. Evolution of number-average molecular weight and polydispersity with monomer conversion for the linear homopolymerization of methyl
methacrylate: (a) via RAFT using cumyl dithiobenzoate as a chain transfer agent at a relative molar ratio of [MMA]:[CDB]:[ACCN] 50:1.0:0.2, and (b)
via ATRP using 3-methylphenyl bromoisobutyrate as initiator, copper(I) chloride and N-(n-propyl)-2-pyridylmethanimine as catalyst at a relative
molar ratio of [MMA]:[MP-Br]:[CuCl]:[n-Pr] 50:1:1:2 using various [MMA]0 in toluene at 90 °C (the straight lines represent the calculated theoretical
Mn values, assuming 100% CTA or initiator efficiency).
and RAFT formulations, see Figure 1. This suggests that
reasonably good living character is achieved, regardless of
the initial monomer concentration. The ATRP initiator and
the RAFT chain transfer agent (CTA) efficiencies were
calculated to be 90% by end group analysis of their respec-
tive aromatic 1H NMR signals. Almost identical efficiencies
were estimated from THF GPC (calibrated using a series of
near-monodisperse PMMA standards).
Calculation of CDB Chain Transfer Agent and MP-Br
Initiator Efficiencies. The CDB chain transfer agent and
MP-Br initiator efficiencies are crucial parameters that dictate
the actual primary chain length of the branched copolymers.
1
These efficiencies were calculated using H NMR and GPC
from linear RAFT and ATRP homopolymerizations (where
the target DP was 50 in both cases). 1H NMR spectra of the
resulting purified linear homopolymers were recorded in
CD2Cl2 in order to compare the three methyl ester protons of
the PMMA (at δ 3.6) with either the ten aromatic protons due
to CDB (RAFT formulation) or the four aromatic protons due
to the MP-Br (ATRP formulation). For example, for the
linear ATRP homopolymerization conducted at [MMA]0 =
50 wt %, an experimental methyl ester/MP-Br molar ratio of
165.5: 4 (theoretical molar ratio is 150: 4) was determined at a
final monomer conversion of 99.5%, which indicates an ini-
tiator efficiency of 90.2%. The MP-Br efficiency can also been
calculated by comparing the GPC number-average molecular
weight obtained for the linear homopolymer (vs PMMA
standards) with its theoretical value. This alternative approach
yielded an initiator efficiency of 90.4%, which is in very good
agreement with the NMR value. Calculation of the CTA
efficiency for this specific RAFT formulation has been des-
cribed previously.46 Table 1 summarizes the RAFT CTA and
ATRP initiator efficiencies calculated for the three different
monomer concentrations examined in this study. Reduced
efficiencies were obtained at lower monomer concentrations.
However, even at [MMA]0 = 10 wt %, these efficiencies were
around 85%.
At high monomer concentration (i.e., [MMA]0=50 wt %),
a linear semilogarithmic plot indicating first-order kinetics with
respect to monomer was obtained for both RAFT and ATRP
formulations (see Figure 2). Using RAFT, a relatively constant
radical flux during the polymerization was achieved by addition
of further ACCN initiator after approximately every two half-
lives in order to ensure that high monomer conversions were
achieved at [MMA]0 = 10 and 30%. For the ATRP syntheses
conducted at the two lowest [MMA]0, negative deviations from
first-order kinetics were observed, suggesting that termination
reactions were more prevalent under these conditions. Never-
theless, very high monomer conversions were still obtained
after 48 h for [MMA]0=30 wt % and after 168 h for [MMA]0=
10 wt %. Homopolymers with narrow polydispersities and
molecular weights close to those targeted were obtained at each
of the three monomer concentrations using both RAFT and
ATRP formulations (see Table 1). Regardless of the monomer
concentration, the final conversions achieved by RAFT were
typically 96-97% whereas monomer conversions obtained by
ATRP were slightly higher at ∼99%, particularly at [MMA]0=
50 wt %.

2150 Macromolecules, Vol. 43, No. 5, 2010
Rosselgong et al.
Figure 2. Semilogarithmic plot for the homopolymerization of methyl methacrylate in toluene at 90 °C, with [MMA]0=10, 30, and 50 wt %: (a) via
RAFT using cumyl dithiobenzoate as a chain transfer agent at a relative molar ratio of [MMA]:[CDB]:[ACCN] 50:1.0:0.2, and (b) via ATRP using
3-methylphenyl bromoisobutyrate as initiator, copper(I) chloride and N-(n-propyl)-2-pyridylmethanimine as catalytic system at a relative molar ratio
of [MMA]:[MP-Br]:[CuCl]:[n-Pr] = 50:1:1:2.
Figure 3. Conversion vs time plot for the consumption of disulfide-based dimethacrylate branching agent (DSDMA, triangles) and methyl
methacrylate monomer (MMA, circles) as determined using HPLC, and overall vinyl bond conversion determined by 1H NMR (squares) during
the synthesis of a copolymer with targeted composition PMMA50-DSDMA0.90. (a) RAFT copolymerization mediated by cumyl dithiobenzoate
(CDB) in toluene at 90 °C and (b) ATRP copolymerization using MPBr/CuCl/n-Pr at a relative molar ratio of 1/1/2 as initiator/catalytic system in
toluene at 90 °C.
HPLC Studies of the Depletion of the Monovinyl and
Divinyl Comonomers during Branching Copolymerizations
by RAFT or ATRP Conducted at [MMA]0 = 50 wt % for a
Targeted Composition of PMMA50-DSDMA0.90. One of the
inherent assumptions in Flory’s mean field theory is equal
monomer reactivities in the statistical copolymerization of
the monovinyl monomer with the divinyl comonomer. If this
assumption is valid, then microgels cannot be formed
at low conversions by preferential consumption of the
divinyl monomer. Moreover, if the reactivity of the divinyl
monomer is comparable to that of the monovinyl monomer,
then there is only a very low probability that the former
species remains unreacted at the end of the copolymeriza-
tion. In order to assess whether the copolymerization of
conversion with reaction time (calculated by 1H NMR)
and also the individual consumptions of MMA and
DSDMA comonomers (determined by UV HPLC) obtained
with both RAFT46 and ATRP formulations at [MMA]0 =
50 wt % for a target copolymer composition of PMMA50-
DSDMA0.90
.
In each case the DSDMA branching agent is consumed
significantly faster than the MMA. On the other hand, the
probability of this divinyl monomer becoming incorporated
into the copolymer chains is twice that of the monovinyl
monomer. As soon as one of the vinyl bonds of the DSDMA
reacts, it is removed from the reaction solution and can no
longer be detected by UV HPLC. According to a recent study
in our group,31 the remaining fraction of unreacted divinyl
monomer, p, in a statistical conversion of vinyl bonds is
connected to the overall fractional conversion of all
double bonds, F, by the following simple relation: p=(1 - F)2.
This theoretical curve is plotted in Figure 4, along with the
experimental data shown in Figure 3 for DSDMA.
The relatively good fit indicates that the DSDMA copoly-
merizes statistically with MMA using either RAFT46
or ATRP chemistry. Thus equal comonomer reactivity,
which is an important assumption in Flory-Stockmayer
theory, appears to be valid for this particular model
system.3
1
MMA with DSDMA was truly statistical, H NMR was
used to evaluate the overall vinyl monomer conversion and
UV HPLC was utilized to determine the depletion of MMA
and DSDMA during both RAFT and ATRP syntheses
(the vinyl signals of the two comonomers overlap in the 1H
NMR spectra, whereas using a gradient eluent allows
very good discrimination between these two species and
hence a robust HPLC protocol). Linear calibration curves
were used to quantify the concentration of each com-
onomer after sampling the copolymerizing solution at
various times. Figure 3 shows the change in total vinyl group

Article
Table 2 summarizes the overall vinyl conversions as
Macromolecules, Vol. 43, No. 5, 2010 2151
in Table 2 were repeated at least twice, with generally good
experimental reproducibility being observed.
judged by 1H NMR and the weight-average molecular
weights and polydispersities obtained using a light scattering
detector for branched copolymerizations conducted using
either ATRP or RAFT at 10 wt %, [MMA]0 = 30 wt %
and 50 wt %. Following the work of Bannister
et al.,31 we calculate (see the last two columns) both nb
Branched Copolymerization at [MMA]0 = 10 wt % Using
DSDMA. A series of branched PMMA50-DSDMAx with
DSDMA/CDB or DSDMA/MP-Br molar ratios varying
from 1.0 to 5.0 were conducted at an initial monomer
concentration of 10 wt %, which corresponds to the critical
overlap concentration, c*, estimated for PMMA50 (see ent-
ries 1-6 in Table 2). Close inspection of these entries con-
firms that both RAFT and ATRP formulations tolerate high
proportions of DSDMA brancher per primary chain without
causing gelation: up to three DSDMA branchers can be
present under ATRP conditions and up to five DSDMA
branchers using RAFT chemistry. Given the relatively high
monomer conversions achieved in both cases, most of the
DSDMA brancher must therefore be wasted in intramole-
cular cyclization side-reactions, which are assumed to be
negligible in Flory-Stockmayer theory. The ATRP formu-
lation seems to be somewhat less non-ideal: macroscopic
gelation was finally observed at a DSDMA/MP-Br molar
ratio of 5.0, whereas the RAFT branched copolymer pre-
pared under the equivalent conditions remained soluble. In
principle, this suppression of gelation may be related to the
reinitiation required for the RAFT copolymerization, be-
cause this could increase the number of primary chains. In
general, copolymerizations conducted at around c* (i.e.,
[MMA]0 = 10 wt %) produced relatively low molecular
weight branched copolymers compared to those conducted
at higher monomer concentrations (see later). This observa-
tion can be attributed to increased participation of the
DSDMA in wasteful intramolecular cyclizations, rather
than intermolecular branching. However, the lower overall
vinyl conversions that can be achieved under these condi-
tions most likely also play a role, since the extent of branch-
ing is very sensitive to the overall conversion.31
l
b
(using the formula nb = (Mn /Mn ) - 1), a parameter that
corresponds to the average number of DSDMA comonomer
l
units per copolymer molecule and also the ratio (Mnb - Mn )/
Mn , which equals unity when an infinite gel network is
l
l
produced. Here Mn corresponds to the number-average
b
molecular weight of the linear homopolymer and Mn cor-
responds to the number-average molecular weight of the
branched copolymer at the equivalent overall vinyl double
bond conversion. We emphasize that all of the entries shown
Figure 4. Fraction of unreacted DSDMA, p, vs fraction conversion of
double bonds, F, obtained for the DSDMA data shown in Figures 3a
and 3b. The dash line is the reasonable fit obtained for these data,
assuming that p=(1 - F)2, the triangles represent the DSDMA deple-
tion for the RAFT technique and the squares represent the DSDMA
depletion for the ATRP technique.
Branched Copolymerization at [MMA]0 = 30 wt % Using
DSDMA. Although relatively few experiments were con-
ducted under these conditions, the data shown in Table 2
(see entries 7 to 10) indicate that both ATRP and RAFT
Table 2. Summary of Final Monomer Conversions, GPC Molecular Weights (Determined Using the Triple Detection Method) and Polydispersities
Obtained for PMMA50-DSDMAx Branched Copolymers Prepared by Either ATRP or RAFT in Toluene at 90°C at Various [MMA]0
target copolymer
composition
conversiona
(%)
b
Mn
b
b
c
nb
l d
l
formulation
[MMA]0 wt %
Mw
Mw/Mn
(Mnb - Mn ) /Mn
ATRP
ATRP
ATRP
RAFT
RAFT
RAFT
ATRP
ATRP
RAFT
RAFT
ATRP
ATRP
ATRP
ATRP
ATRP
ATRP
RAFT
RAFT
RAFT
RAFT
RAFT
RAFT
PMMA50-DSDMA1.0
PMMA50-DSDMA3.0
PMMA50-DSDMA5.0
PMMA50-DSDMA1.00
PMMA50-DSDMA3.00
PMMA50-DSDMA5.00
PMMA50-DSDMA1.25
PMMA50-DSDMA1.50
PMMA50-DSDMA1.50
PMMA50-DSDMA1.60
PMMA50-DSDMA0.60
PMMA50-DSDMA0.70
PMMA50-DSDMA0.80
PMMA50-DSDMA0.85
PMMA50-DSDMA0.90
PMMA50-DSDMA0.95
PMMA50-DSDMA0.60
PMMA50-DSDMA0.70
PMMA50-DSDMA0.80
PMMA50-DSDMA0.85
PMMA50-DSDMA0.90
PMMA50-DSDMA0.95
10
10
10
10
10
10
30
30
30
30
50
50
50
50
50
50
50
50
50
50
50
50
97.6
96.7
96.9
96.7
96.1
96.4
99.1
98.9
98.7
98.5
99.1
98.9
98.8
98.9
99.0
98.5
96.2
96.5
97.3
96.1
96.6
96.1
39 600
59 000
54 800
126 300
1.38
2.14
5.88
9.34
0.85
0.90
macroscopic gelation
10 200
18 600
43 800
43 900
17 200
69 700
608 700
92 300
1.69
0.79
2.28
6.70
6.51
0.42
0.68
0.87
0.87
3.74
13.89
2.10
macroscopic gelation
209 200
4 508 800 21.56
macroscopic gelation
34.92
0.97
17 100
19 600
32 200
45 000
71 800
58 300
3.41
3.69
3.77
4.42
6.09
1.93
2.37
4.54
6.71
0.65
0.70
0.82
0.87
0.92
72 400
121 200
199 000
437 700
11.38
macroscopic gelation
8800
12 000
13 900
28 700
92 000
41 400
68 300
331 500
1 126 700
3 324 900
4.70
0.55
1.11
1.42
4.07
15.2
0.33
0.51
0.58
0.79
0.94
5.67
23.92
39.29
36.14
macroscopic gelation
a Monomer conversions were determined by 1H NMR spectroscopy. b Molecular weight data were obtained by THF GPC using a triple detection
system comprising refractive index, viscosity and light scattering (15° and 90°) detectors. c nb is the average number of fully reacted DSDMA branching
comonomers per copolymer molecule. d (Mnb - Mn )/Mnl = 1 is the condition for the formation of an infinite gel network. Mnb and Mnl are the number-
l
average molecular weight of, respectively, weight of the branched copolymer and the linear homopolymer, respectively.

2152 Macromolecules, Vol. 43, No. 5, 2010
Rosselgong et al.
formulations behave less non-ideally at this higher mono-
mer concentration. Thus, ATRP can now tolerate only
∼1.25 DSDMA units per primary chain, with gelation
occurring if 1.50 DSDMA units are utilized. Similarly, the
RAFT formulation remains soluble (while very close to
gelation) at 1.50 DSDMA units per chain, but gelation is
observed if 1.60 DSDMA units are utilized. It is perhaps also
noteworthy that the ATRP syntheses are marginally less
non-ideal than those conducted under RAFT conditions.
Branched Copolymerizations Conducted at [MMA]0
=
50 wt % Using DSDMA. At high monomer concentration,
i.e., well above c*, RAFT or ATRP copolymerization of
MMA with DSDMA enables a range of soluble highly
branched copolymers to be obtained simply by varying the
initial DSDMA/CDB or DSDMA/MP-Br molar ratio,
respectively. When such molar ratios are increased from
0.60 to 0.90 (see entries 11-16 and 17-22 in Table 2 for
ATRP and RAFT formulations, respectively), the weight-
average molecular weight and polydispersity both increase
dramatically. Moreover, the average number of fully reacted
Figure 5. Evolution of weight-average molecular weight (Mw) with
total vinyl group conversion for the synthesis of branched PMMA50-
DSDMA0.90 copolymers via RAFT ([) and via ATRP (b) using
an initial monomer concentration [MMA]0 = 50 wt % (light
scattering detector). The Mw of the linear homopolymers prepared in
the absence of DSDMA branching agent are also shown for the RAFT
formulation (2), and the ATRP formulation (9) (refractive index
detector).
branching comonomer units per copolymer molecule, nb,
b
increases with the amount of DSDMA introduced and (Mn
Mn )/Mn approaches unity (indicating the formation of an
-
l
l
infinite gel network) in both cases for the targeted composi-
b
distributions are observed even at relatively low comonomer
conversions (26.6%) and these features persist up to very
high monomer conversions. In contrast, the same detector
indicates unimodal distributions over a wide range of con-
versions for the ATRP formulation.
The refractive index detector (see Figures 6b and 7b) is
much more sensitive to the presence of linear primary chains,
which can still be detected even at high comonomer conver-
sions (the characteristic peak retention time for linear
PMMA50 homopolymer prepared in the absence of any
DSDMA is approximately 16 min in the refractive
index chromatograms shown in Figures 6b and 7b). This
observation is consistent with a recent Monte Carlo simulation
study, which predicts that a significant proportion of linear
chains remain at the end of these copolymerizations due to the
purely statistical nature of the branching process.67
The last three chromatograms shown in Figure 7a suggest
continued evolution of the molecular weight with comono-
mer conversion. However, there may also be some evidence
for an artificial cut-off at such low retention times, which
suggests possible formation of an unanalyzable microgel
fraction. This hypothesis is supported by inspection of the
GPC traces recorded using the refractive index detector: the
curves at 95.0%, 97.5%, and 99.0% comonomer conversion
are almost identical (see Figure 7b), whereas a high mole-
cular weight shoulder might be expected (as observed in
Figure 6a for the corresponding RAFT-synthesized bran-
ched copolymer). The presence of a microgel fraction is also
consistent with the observation that the branched copolymers
obtained at high conversions are very difficult to filter prior to
GPC analysis. However, further studies are required to exam-
ine this hypothesis. It is perhaps worth emphasizing that the
rate of polymerization obtained with the ATRP formulation
conducted at [MMA]0=50 wt % is approximately three times
faster than that obtained with the RAFT formulation under
the same conditions. In principle, such kinetic differences could
be an important consideration when rationalizing our experi-
mental observation of subtle differences between the RAFT
and ATRP formulations. In the case of RAFT, the slower
chain growth should allow greater diffusion of the branched
copolymer chains, whereas for ATRP significantly less relaxa-
tion of the growing copolymer chains is possible. Further
studies are required to examine this possibility.
tion of PMMA50-DSDMA0.90. More specifically, (Mn
-
l
l
Mn )/Mn =0.92 or 0.94 for the ATRP and RAFT formula-
tions respectively, see entries 15 and 21 in Table 2. These
values are comparable to those calculated by Bannister et
al.31 for the ATRP branched copolymerization of 2-hydro-
xypropyl methacrylate with ethylene glycol dimethacrylate
just below the gel point. Macroscopic gelation is observed at
high comonomer conversions for both RAFT and ATRP
when the proportion of DSDMA per primary chain is
increased from 0.90 to 0.95. The targeted PMMA50-
DSDMA0.90 formulation lies very close to the gel point for
both formulations. Taking into account the CDB (or
MP-Br) efficiency of around 90%, this approximately
corresponds to one fully reacted DSDMA branching agent
per primary chain, which is consistent with near-ideal beha-
vior according to Sherrington’s hypothesis, if not Flor-
y-Stockmayer theory. In the rest of this manuscript, the
terms “ideal” and “non-ideal” are used solely in the context
of the former postulate, rather than the latter theory.
Figure 5 compares the evolution of weight-average mole-
cular weight (Mw) with comonomer conversion for the linear
PMMA50 and the PMMA50-DSDMA0.90 branched copoly-
mer obtained using the RAFT and ATRP (see Figure 5b)
formulations at [MMA]0 =50 wt %. In both cases the onset
of branching, which corresponds to the initial deviation from
linearity, is observed at around 70% comonomer conver-
sion. Beyond this conversion, the branched copolymer
Mw increases rapidly for the RAFT formulation but a less
dramatic change is observed under ATRP conditions. How-
ever, the final Mw obtained in the latter case is still compar-
able to that reported by Li et al.28 and Bannister et al.31
for branching compositions very close to gelation using the
same target degree of polymerization of 50 for the primary
chains (which in these earlier examples comprised poly-
(2-hydroxypropyl methacrylate), rather than PMMA).
Figures 6 and 7 show the evolution of the GPC curves for
the nominal targeted composition of PMMA50-DSD-
MA0.90 obtained at [MMA]o = 50 wt % using (a) the light
scattering detector and (b) the refractive index detector for
the RAFT and the ATRP formulations, respectively. The
former detector is clearly much more sensitive to the presence
of high molecular weight copolymer, as expected.28,31,39
For the RAFT formulation (see Figure 6a), bimodal

Article
Macromolecules, Vol. 43, No. 5, 2010 2153
Figure 6. GPC traces recorded at various monomer conversions for the RAFT synthesis of branched PMMA50-DSDMA0.90 copolymer in toluene at
90 °C with a [MMA]0 = 50 wt % using (a) a 90° light scattering detector and (b) a refractive index detector calibrated with PMMA standards.
Figure 7. GPC traces recorded at various monomer conversions for the ATRP synthesis of branched PMMA50-DSDMA0.90 copolymer in toluene at
90 °C with a [MMA]0 = 50 wt % using (a) a 90° light scattering detector and (b) a refractive index detector calibrated with PMMA standards.
Figure 8. Schematic representation of the retrosynthesis from branched copolymer to linear primary chains by disulfide bond cleavage using reduction
using tributylphosphine (Bu3P): (left) GPC traces recorded using the refractive index detector and the light scattering detector for the
PMMA50-DSDMA0.90 branched copolymer prepared by RAFT at [MMA]0 = 50 wt %; (right) GPC traces recorded using the refractive index
detector of the degraded polymer chains obtained after reductive cleavage of disulfide bonds by Bu3P and the PMMA50 homopolymer prepared in the
absence of any DSDMA branching agent.
In summary, there certainly appears to be some subtle
differences in the evolution of molecular weight in the ATRP
and RAFT branching copolymerizations conducted at
50 wt %. However, close inspection of Table 2 confirms that
it is essential to utilize a relatively high monomer concentra-
tion (i.e., well above c*) if near-ideal Sherrington-type
behavior is desired, regardless of whether a ATRP or RAFT
formulation is selected for the branching copolymerization.
This is primarily because intermolecular branching can only
be really efficient if individual primary chains (which exist as
random coils in solution) are able to interpenetrate each other,
otherwise intramolecular cyclization will inevitably be favored.
Reductive Cleavage of Disulfide Bonds in PMMA50-DSD-
MA0.90 Branched Copolymers Using Tributylphosphine. Selec-
tive reductive cleavage of the disulfide bonds in PMMA50-
DSDMA0.90 branched copolymers prepared at [MMA]0 =
50 wt % using tributylphosphine (Bu3P) leads to the formation
of low polydispersity primary chains, as judged by gel permeation
chromatography. The molecular weight distributions of these
degraded chains are comparable to either RAFT-synthesized or

2154 Macromolecules, Vol. 43, No. 5, 2010
Rosselgong et al.
Figure 9. Schematic representation of the retrosynthesis from branched copolymer to linear primary chains by disulfide bond cleavage using reduction
using tributylphosphine (Bu3P): (left) GPC traces recorded using the refractive index detector and the light scattering detector for the
PMMA50-DSDMA0.90 branched copolymer prepared by ATRP at [MMA]0 = 50 wt %; (right) GPC traces recorded using the refractive index
detector of the degraded polymer chains obtained after reductive cleavage of disulfide bonds by Bu3P and the PMMA50 homopolymer prepared in the
absence of any DSDMA branching agent.
Figure 10. Schematic representation of the retrosynthesis from branched copolymer to linear primary chains by disulfide bond cleavage using
reduction using tributylphosphine (Bu3P): (left) GPC traces recorded using the refractive index detector and the light scattering detector for the
PMMA50-DSDMA5.00 branched copolymer prepared by RAFT at [MMA]0 = 10 wt %; (right) GPC traces recorded using the refractive index
detector of the degraded polymer chains obtained after reductive cleavage of disulfide bonds by Bu3P and the PMMA50 homopolymer prepared in the
absence of any DSDMA branching agent.
ATRP-synthesized linear poly(methyl methacrylate) homopo-
lymers prepared in the absence of any DSDMA branching
agent, see Figures 8 and 9. This confirms that good control over
the copolymerization is achieved under branching conditions
and that the polydisperse highly branched chains simply com-
prise randomly coupled, near-monodisperse primary chains, as
expected. It also suggests that chain transfer to polymer must be
negligible in such branched copolymer syntheses, since such
side-reactions would necessarily involve the formation of new
C-C bonds that could not be cleaved by the Bu3P.
Figure 10 shows the GPC traces of the PMMA50-DSD-
MA5.00 branched copolymer synthesized under RAFT con-
ditions at [MMA]0 = 10 wt % and its linear counterpart
obtained after reductive cleavage with Bu3P. The cleaved
thiol-containing PMMA has a slightly higher molecular
weight than the corresponding PMMA homopolymer pre-
pared in the absence of any DSDMA. This is not unexpected,
since the former species contains an additional five copoly-
merized DSDMA repeat units.
monomer conversions. Our ATRP formulation also produced
homopolymers with narrow polydispersities in very high yields
but required a reaction time of 1 week at 90 °C. RAFT CTA and
ATRP initiator efficiencies were comparable for a given initial
monomer concentration and varied from ∼90% at [MMA]0 =
50 wt % to ∼85% at [MMA]0 = 10 wt %. Combined 1H NMR
and UV HPLC studies confirmed that the DSDMA divinyl
comonomer reacts statistically with MMA during branched
copolymerizations conducted with either RAFT or ATRP for-
mulations. The approximately equal reactivities of these two
comonomers reduces the possibility of microgel formation during
the early stages of the copolymerization and also minimizes the
probability of any unreacted DSDMA remaining at high como-
nomer conversion. Taking into account the RAFT CTA or
ATRP initiator efficiency, our results confirm that branching
copolymerizations conducted at high monomer concentration
(i.e., 50 wt %) cannot tolerate more than one DSDMA unit per
primary chain if gelation is to be avoided at high conversion.
Highly branched copolymers with compositions that lie close to
the gel point (such as PMMA50-DSDMA0.90) have been chemi-
cally degraded to their constituent primary chains by reductive
cleavage of the disulfide bonds in the DSDMA branching
comonomer. For both RAFT and ATRP formulations, this
affords low polydispersity primary chains with molecular weights
that are comparable to PMMA homopolymers prepared
under identical conditions in the absence of any DSDMA.
When branching copolymerizations are conducted at
Conclusions
Linear homopolymerization of MMA in toluene at 90 °C using
either RAFT or ATRP chemistry has relatively good controlled/
living character for monomer concentrations ranging from 10 to
50 wt %. At the lowest monomer concentration investigated, our
RAFT formulation required periodic addition of further initiator
to maintain an appropriate radical flux and hence achieve high

Article
Macromolecules, Vol. 43, No. 5, 2010 2155
[MMA]0 = 10 wt % (i.e., at around the critical overlap concen-
tration c* for linear PMMA chains of DP 50) a relatively high
proportion of branching agent can be tolerated without causing
gelation; up to three DSDMA units per primary chain for the
ATRP formulation and up to five DSDMA units per primary
chain for the RAFT formulation. This strongly suggests that the
majority of the DSDMA actually participates in intramolecular
(17) Durie, S.; Jerabek, K.; Mason, C.; Sherrington, D. C. Macromo-
lecules 2002, 35, 9665–9672.
(18) Chisholm, M.; Hudson, N.; Kirtley, N.; Vilela, F.; Sherrington,
D. C. Macromolecules 2009, 42, 7745–7752.
(19) Weaver, J. V. M.; Williams, R. T.; Royles, B. J. L.; Findlay, P. H.;
Cooper, A. I.; Rannard, S. P. Soft Matter 2008, 4, 985–992.
(20) He, T.; Adams, D. J.; Butler, M. F.; Cooper, A. I.; Rannard, S. P.
J. Am. Chem. Soc. 2009, 131, 1495–1501.
(21) Weaver, J. V. M.; Rannard, S. P.; Cooper, A. I. Angew. Chem.
2009, 48, 2131–2134.
1
cyclizations, rather than forming intermolecular branches. H
NMR studies (data not shown here) confirm the presence of these
“excess” DSDMA units in the branched copolymers. GPC
studies of the disulfide-cleaved copolymers prepared at 10 wt
% indicate slightly highly molecular weights for the thiol-func-
tionalized primary chains, as expected. In summary, we have
demonstrated that the initial monomer concentration is an
important parameter in determining the nature of the products
of the branched copolymerization of MMA and DSDMA under
both RAFT and ATRP conditions. The DSDMA branching
comonomer forms intermolecular branches at both high and low
monomer concentration but reacts preferentially via intramole-
cular cyclization under the latter conditions. Although there are
certainly some subtle differences between RAFT- and ATRP-
synthesized branched copolymers, the initial monomer concen-
tration seems to be much more important than the precise nature
of the polymerization chemistry in determining the microstruc-
ture of the branched copolymer products.
(22) Woodward, R. T.; Slater, R. A.; Higgins, S.; Rannard, S. P.;
Cooper, A. I.; Royles, B. J. L.; Findlay, P. H.; Weaver, J. V. M.
Chem. Commun. 2009, 24, 3554–3556.
(23) Li, Y. T.; Armes, S. P. Macromolecules 2005, 38, 5002–5009.
(24) Camerlynck, S.; Cormack, P. A. G.; Sherrington, D. C. Eur. Polym.
J. 2006, 42, 3286–3293.
€ €
(25) Butun, V.; Bannister, I.; Billingham, N. C.; Sherrington, D. C.;
Armes, S. P. Macromolecules 2005, 38, 4977–4982.
(26) Isaure, F.; Cormack, P. A. G.; Graham, S.; Sherrington, D. C.;
€ €
Armes, S. P.; Butun, V. Chem. Commun. 2004, 9, 1138–1139.
(27) Wang, A. R.; Zhu, S. Macromolecules 2002, 35, 9926–9933.
(28) Li, Y. T.; Armes, S. P. Macromolecules 2005, 38, 8155–8162.
(29) Wang, A. R.; Zhu, S. J. Polym. Sci., Part A: Polym. Chem. 2005, 43,
5710–5714.
(30) Wang, A. R.; Zhu, S. Polym. Eng. Sci. 2005, 45, 720–727.
(31) Bannister, I.; Billingham, N. C.; Armes, S. P.; Rannard, S. P.;
Findlay, P. Macromolecules 2006, 39, 7483–7492.
(32) Bouhier, M.-H.; Cormack, P. A. G.; Graham, S.; Sherrington,
D. C. J. Polym. Sci., Part A: Polym. Chem. 2007, 45, 2375–2386.
(33) Yu, Q.; Qin, Z.; Li, J.; Zhu, S. Polym. Eng. Sci. 2008, 48, 1254–
1260.
Acknowledgment. Dr. Peter Wright and Jay Syrett (Warwick
University) are thanked for their advice on the choice of ATRP
initiator and appropriate ATRP homopolymerization condi-
tions. We thank Lubrizol (Hazelwood, UK) for funding a PhD
studentship for JR and for permission to publish this research.
SPA is the recipient of a five-year Royal Society/Wolfson Re-
search Merit Award.
(34) Yang, H.-J.; Jiang, B.-B.; Huang, W.-Y.; Zhang, D.-L.; Kong,
L.-Z.; Chen, J.-H.; Liu, C.-L.; Gong, F.-H.; Yu, Q.; Yang, Y.
Macromolecules 2009, 42, 5976–5982.
(35) Liu, B.; Kazlauciunas, A.; Guthrie, J. T.; Perrier, S. Macromole-
cules 2005, 38, 2131–2136.
(36) Norisuye, T.; Morinaga, T.; Tran-Cong-Miyata, Q.; Goto, A.;
Fukuda, T.; Shibayama, M. Polymer 2005, 46, 1982–1994.
(37) Lin, Y.; Liu, X.; Li, X.; Zhan, J.; Li, Y. J. Polym. Sci., Part A:
Polym. Chem. 2006, 45, 26–40.
Supporting Information Available: Calculation of c* for a
linear PMMA50 homopolymer; 1H NMR spectra for the MP-Br
initiator and DSDMA branching comonomer. This material is
available free of charge via the Internet at http://pubs.acs.org.
(38) Taton, D.; Baussard, J.-F.; Dupayage, L.; Poly, J.; Gnanou, Y.;
Ponsinet, V.; Destarac, M.; Mignaud, C.; Pitois, C. Chem. Com-
mun. 2006, 18, 1953–1955.
(39) Vo, C.-D.; Rosselgong, J.; Armes, S. P.; Billingham, N. C. Macro-
molecules 2007, 40, 7119–7125.
(40) Poly, J.; Wilson, D. J.; Destarac, M.; Taton, D. Macromol. Rapid
Commun. 2008, 29, 1965–1972.
(41) Tao, L.; Liu, J.; Tan, B. H.; Davis, T. P. Macromolecules 2009, 42,
4960–4962.
(42) Li, Y. T.; Armes, S. P. Macromolecules 2009, 42, 939–945.
(43) Mounteney, P.; Rannard, S.; Findlay, P.; Duncalf, D. J.; Perrier, S.
Polym. Prepr. 2006, 47, 636–637.
References and Notes
(1) Flory, P. J. J. Am. Chem. Soc. 1941, 63, 3083–90.
(2) Stockmayer, W. H. J. Chem. Phys. 1944, 12, 125–31.
(3) Flory, P. J. Principles of Polymer Chemistry; Cornell University
Press: Ithaca, NY, 1953.
(44) Mounteney, P.; Rannard, S.; Findlay, P.; Duncalf, D. J.; Perrier, S.
PMSE Prepr. 2007, 96, 648–649.
(4) O’Brien, N.; McKee, A.; Sherrington, D. C.; Slark, A. T.; Titterton,
A. Polymer 2000, 41, 6027–6031.
(45) Yu, Q.; Gan, Q.; Zhang, H.; Zhu, S. ACS Symp. Ser. 2009, 1024,
181–193.
(46) Rosselgong, J.; Armes, S. P.; Barton, W.; Price, D. Macromolecules
(5) Costello, P. A.; Martin, I. K.; Slark, A. T.; Sherrington, D. C.;
Titterton, A. Polymer 2001, 43, 245–254.
(6) Isaure, F.; Cormack, P. A. G.; Sherrington, D. C. J. Mater. Chem.
2009, 42, 5919–5924.
2003, 13, 2701–2710.
(47) Gao, H.; Min, K.; Matyjaszewski, K. Macromolecules 2007, 40,
7763–7770.
(48) Gao, H.; Li, W.; Matyjaszewski, K. Macromolecules 2008, 41,
2335–2340.
(49) Gao, H.; Miasnikova, A.; Matyjaszewski, K. Macromolecules
2008, 41, 7843–7849.
(50) Gao, H.; Min, K.; Matyjaszewski, K. Macromolecules 2009, 42,
8039–8043.
(7) Slark, A. T.; Sherrington, D. C.; Titterton, A.; Martin, I. K.
J. Mater. Chem. 2003, 13, 2711–2720.
(8) Isaure, F.; Cormack, P. A. G.; Sherrington, D. C. Macromolecules
2004, 37, 2096–2105.
(9) Graham, S.; Cormack, P. A. G.; Sherrington, D. C. Macromole-
cules 2005, 38, 86–90.
(10) Saunders, G.; Cormack, P. A. G.; Graham, S.; Sherrington, D. C.
Macromolecules 2005, 38, 6418–6422.
(51) Li, W.; Gao, H.; Matyjaszewski, K. Macromolecules 2009, 42, 927–
932.
(52) Gao, H.; Li, W.; Min, K.; Matyjaszewski, K. ACS Symp. Ser. 2009,
1023, 203–213.
(11) Baudry, R.; Sherrington, D. C. Macromolecules 2006, 39, 5230–
5237.
(12) Isaure, F.; Cormack, P. A. G.; Sherrington, D. C. React. Funct.
Polym. 2006, 66, 65–79.
(53) Gao, H.; Matyjaszewski, K. Prog. Polym. Sci. 2009, 34, 317–
350.
(13) Graham, S.; Rannard, S. P.; Cormack, P. A. G.; Sherrington, D. C.
J. Mater. Chem. 2007, 17, 545–552.
(54) Li, Y. T.; Ryan, A. J.; Armes, S. P. Macromolecules 2008, 41, 5577–
5581.
(55) Poly, J.; Wilson, D. J.; Destarac, M.; Taton, D. J. Polym. Sci., Part
(14) Camerlynck, S.; Cormack, P.; Sherrington, D.; Saunders, G.
J. Macromol. Sci., Part B: Phys 2005, 44, 881–895.
(15) Besenius, P.; Slavin, S.; Vilela, F.; Sherrington, D. C. React. Funct.
Polym. 2008, 68, 1524–1533.
A: Polym. Chem. 2009, 47, 5313–5327.
(56) Yu, Q.; Xu, S.; Zhang, H.; Ding, Y.; Zhu, S. Polymer 2009, 50,
3488–3494.
(16) Baudry, R.; Sherrington, D. C. Macromolecules 2006, 39, 1455–
1460.

2156 Macromolecules, Vol. 43, No. 5, 2010
Rosselgong et al.
(57) Perrier, S.; Barner-Kowollik, C.; Quinn, J. F.; Vana, P.; Davis,
T. P. Macromolecules 2002, 35, 8300–8306.
(62) Benaglia, M.; Rizzardo, E.; Alberti, A.; Guerra, M. Macromole-
cules 2005, 38, 3129–3140.
(58) Haddleton, D. M.; Crossman, M. C.; Dana, B. H.; Duncalf, D. J.;
Heming, A. M.; Kukulj, D.; Shooter, A. J. Macromolecules 1999,
32, 2110–2119.
(63) Haddleton, D. M.; Waterson, C. Macromolecules 1999, 32, 8732–
8739.
(64) Moad, G.; Rizzardo, E.; Thang, S. H. Aust. J. Chem. 2005, 58, 379–
(59) Bouhier, M.-H.; Cormack, P. A. G.; Graham, S.; Sherrington,
D. C. J. Polym. Sci., Part A: Polym. Chem. 2007, 45, 2375–
2386.
(60) Tsarevsky, N. V.; Matyjaszewski, K. Macromolecules 2005, 38,
3087–3092.
(61) Min, K.; Gao, H.; Yoon, J. A.; Wu, W.; Kowalewski, T.;
410.
(65) Moad, G.; Rizzardo, E.; Thang, S. H. Acc. Chem. Res. 2008, 41,
1133–1142.
(66) Cook, W. D.; Chen, F.; Pattison, D. W.; Hopson, P.; Beaujon, M.
Polym. Int. 2007, 56, 1572–1579.
(67) Bannister, I.; Billingham, N. C.; Armes, S. P. Soft Matter 2009, 5,
3495–3504.
Matyjaszewski, K. Macromolecules 2009, 42, 1597–1603.