Thiocarbonylthio compounds [S=C(Ph)S-R) in free radical polymerization with reversible addition-fragmentation chain transfer (RAFT polymerization). Role of the free-radical leaving group (R)

Article abstract of DOI:10.1021/ma020882h

Radical polymerization with reversible addition-fragmentation chain transfer (RAFT polymerization) can be used to synthesize a wide range of polymers of controlled architecture and narrow molecular weight distribution. The polymerization use addition-frag

Full text of DOI:10.1021/ma020882h

2
256
Macromolecules 2003, 36, 2256-2272
Thiocarbonylthio Compounds [SdC(Ph)S-R] in Free Radical
Polymerization with Reversible Addition-Fragmentation Chain Transfer
(RAFT Polymerization). Role of the Free-Radical Leaving Group (R)
(
Bill) Y. K. Ch on g, J u lia Kr stin a , Ta m P . T. Le, Gr a em e Moa d ,* Alm a r P ostm a ,
Ezio Rizza r d o,* a n d Sa n H. Th a n g*
CSIRO Molecular Science, Bag 10, Clayton South, Victoria 3169, Australia
Received J une 7, 2002; Revised Manuscript Received December 16, 2002
ABSTRACT: Radical polymerization with reversible addition-fragmentation chain transfer (RAFT
polymerization) can be used to synthesize a wide range of polymers of controlled architecture and narrow
molecular weight distribution. The polymerizations use addition-fragmentation chain transfer agents
(RAFT agents) that possess high transfer coefficients in free radical polymerization and confer living
character on the polymerization. This paper explores the effect of the substituents R of dithiobenzoate
RAFT agents [SdC(Ph)S-R] on the outcome of polymerizations of styrene, methyl methacrylate (MMA)
and butyl (BA) or methyl acrylate (MA). In MMA polymerization at 60 °C, effectiveness depends strongly
on R decreasing in the order where R is: -C(Alkyl)
C(Me) C(dO)NH(alkyl) > -C(Me) CH C(Me) g -C(Me)HPh > -C(Me)
the compounds with R ) -C(Me) Ph and -C(Me) CN provided polymers with substantially narrowed
2
CN ∼ -C(Me)
2
Ar > -C(Me)
2
C(dO)O(alkyl) >
-
2
2
2
3
3
∼ -CH Ph. Of these, only
2
2
2
polydispersities in batch polymerization and gave molecular weight control of the form expected for a
living polymerization. These compounds have high transfer coefficients in MMA polymerization (transfer
coefficient >20 for 0.003 M RAFT agent in bulk MMA). Reagents with R ) -C(Me)
C(Me) C(dO)NH(alkyl), -C(Me) CH C(Me) , -C(Me)HPh, -C(Me) , and -CH Ph appear relatively
ineffective in MMA polymerization because they have much lower transfer coefficients (ca. 1.7, 0.7, 0.4,
.15, 0.03, and 0.03 respectively). In polymerizations of styrene or acrylates at 60 °C, transfer coefficients
2
C(dO)O(alkyl),
2
2
2
3
3
2
0
of RAFT agents are several orders of magnitude higher and all compounds in the above series yield narrow
polydispersity polymers. The transfer coefficients of benzyl dithiobenzoate in styrene and methyl acrylate
polymerization at 60 °C are ca. 50 (0.0093 M RAFT agent in bulk monomer) and 105 (with 0.003-0.006
M RAFT agent in 4.43 M monomer), respectively. The rate constant for addition to the thiocarbonyl
group does not depend strongly on R. Differences in activity depend on the properties of R and the
•
propagating chain as free radical leaving groups and on the partitioning of R between adding to monomer
•
(
to reinitiate) and adding to polymeric RAFT agent. When the reaction of R with the polymeric RAFT
agent to reform the initial RAFT agent is significant, the transfer coefficient may reduce with concentration
of RAFT agent and with conversion. In these circumstances, the activity of RAFT agents is best
characterized by two transfer constants associated with the forward and reverse reactions. Steric factors,
•
polar factors, and the stability of R all appear to play an important role in determining the transfer
coefficient and the effectiveness of RAFT agents.
In tr od u ction
mers obtained by the RAFT process can be chain
extended or used as precursors to block copolymers by
In recent years, considerable effort1-7 has been ex-
pended to develop free radical polymerizations that
display the essential characteristics of living polymer-
1
5,16,18,19,21-23
the addition of further monomer(s).
RAFT
polymerization also provides a route to stars and more
1
6,19,24-26
complex architectures.
8
izations. One driving force behind this work has been
We and others have also reported that the effective-
ness of RAFT agents (1) strongly depends on the nature
of the groups Z and R, the monomer and the polymer-
to achieve the level of control associated with living
polymerization while retaining as much as possible of
the scope and versatility of the free radical process.
Recently, we described a new method for achieving
1
5,18-20,27-34
ization conditions.
In this paper, we examine
in greater detail the effect of varying the free radical
leaving group, R, of dithiobenzoate derivatives (2,
Scheme 1) on polymerizations of several monomers:
styrene, methyl methacrylate (MMA), and methyl or
9
,10
living free radical polymerization
(RAFT polymeri-
zation) that involves, as the key step, reversible addi-
tion-fragmentation chain transfer.1
1-20
The more ef-
fective RAFT agents are certain thiocarbonylthio com-
3
5
butyl acrylate (MA or BA). In the accompanying
pounds (1) where Z is a group that modifies the
3
6
paper, we report on the influence of the reactivity-
modifying group, Z, in a similar manner.
reactivity of the thiocarbonyl group toward free radical
addition and R is a free radical leaving group.1
4-16
With
appropriate choice of the RAFT agent and reaction
conditions, RAFT polymerization can be successfully
used to produce narrow polydispersity polymers with
Exp er im en ta l Section
Gen er a l Da ta . Tetrahydrofuran (THF) used in synthesis
was freshly distilled from sodium benzophenone. Reagent
chemicals were obtained from Aldrich and used without
purification unless otherwise indicated. Solvents used for
column chromatography were of AR grade and were distilled.
Petroleum spirits refers to the fraction with bp 40-60 °C. The
silica was Kieselgel-60 (Merck), 70-230 mesh. Nuclear mag-
netic resonance (NMR) spectra were obtained with a Bruker
molecular weights predetermined by the conversion and
RAFT agent concentration.1
4,16-20
Moreover, the poly-
*
To whom correspondence should be addressed. E-mail:
graeme.moad@csiro.au,
csiro.au.
ezio.rizzardo@csiro.au,
san.thang@
1
0.1021/ma020882h CCC: $25.00 © 2003 American Chemical Society
Published on Web 03/15/2003

Macromolecules, Vol. 36, No. 7, 2003
RAFT Polymerization with Thiocarbonylthio Compounds 2257
Sch em e 1. Over a ll Rea ction a
removed in vacuo to afford 2.5 g of crude product which was
chromatographed on silica with hexane:diethyl ether (9:1) as
the eluent. The isolated yield of pure 2-cyanoprop-2-yl dithioben-
1
zoate (2e) was 1.6 g (92%). The H NMR spectrum was
identical with that of the sample prepared from the reaction
bis(thiobenzoyl) disulfide with AIBN.19
4
-Cya n o-1-h yd r oxyp en t -4-yl Dit h iob en zoa t e (2f). A
solution of di(thiobenzoyl) disulfide (0.73 g, 2.4 mmol) and 4,4′-
azobis(4-cyano-1-pentanol)39 (0.50 g, 2 mmol) in ethyl acetate
(10 mL) was prepared. The mixture was degassed and heated
to 70 °C in a sealed vessel for 24 h. The volatiles were removed
under reduced pressure and the residue chromatographed on
silica with ethyl acetate/petroleum spirits/acetonitrile (6:14:
a
Key: 2a , R ) C(CH
)Ph; 2d , R ) CH
)(CN)CH CH CH OH; 2g, R ) C(CH
R ) C(CH C(dO)NHCH CH OH; 2i, R ) C(CH
j, R ) C(CH ; 2k, R ) SC(CH ; 2l, R ) [C(CH
R′; 2m , R ) [CH(Ph)] -R′; 2n , R ) [CH(CO
3
)
2
Ph; 2b, R ) C(CH
3
)
2
C
3
6
H
2
4
Cl; 2c, R )
CN; 2f, R )
; 2h ,
C(CH
CH )]
-R′.
1
CH(CH
C(CH
3
2
Ph; 2e, R ) C(CH
)
1) as eluent to give the title compound (2f) (0.49 g, 46%). H
3
2
2
2
3
)
2
C(dO)OC
CH
)(CO
)]
2
H
5
NMR, δ 1.90 (3H, CH
CH CH CH OH), 3.75 (2H, CH
(m, 5H, ArH).
-(Eth oxycar bon yl)pr op-2-yl Dith ioben zoate (2g). Phen-
3
), 2.10 (2H, CH
2
CH
2 2
CH OH), 2.30 (2H,
3
)
2
2
2
3
)
2
3
2
3
)
3
;
2
2
2
2
CH CH
2
2
OH), 7.40, 7.50, 7.90
2
3
)
3
3
)
3
2
3
n
-
n
2
C
4
H
9
n
2
ylmagnesium bromide was prepared from bromobenzene (6.28
g, 0.04 mol) and magnesium turnings (1 g, 0.041 mol) in dry
THF (30 mL). The solution was warmed to 40 °C and carbon
disulfide (3.05 g, 0.04 mol) was added over 15 min while
maintaining the reaction temperature at 40 °C. Ethyl R-bro-
moisobutyrate (7 g, 0.036 mol) was added to the resultant dark
brown solution and the reaction temperature then raised to
80 °C and maintained for 60 h. Ice-cold water (50 mL) was
added, and the organic products were extracted with diethyl
ether (3 × 50 mL). The combined organic extracts were washed
with water and brine and dried over anhydrous magnesium
sulfate. After removal of solvent and purification by column
chromatography on silica with n-hexane/diethyl ether (9:1) as
eluent. 2-(ethoxycarbonyl)prop-2-yl dithiobenzoate (2g) was
AC200 spectrometer on samples dissolved in deuteriochloro-
form. Chemical shifts are reported in ppm from tetramethyl-
silane. High-resolution mass spectra were obtained with either
a J EOL J MS DX303 mass spectrometer (chemical ionization,
CI) or a Bruker Bio Apex Fourier transform mass spectrometer
(electron impact, EI). Methane was used as reagent gas for
the CI mass spectra.
Cu m yl Dith ioben zoa te (2a ). The synthesis of cumyl (or
2
-phenylprop-2-yl) dithiobenzoate is described elsewhere.19
2
-(4-Ch lor oph en yl)pr op-2-yl Dith ioben zoate (2b). Dithio-
1
9
benzoic acid (6.3 g, 0.041 mol) and 4-chloro-R-methylstyrene
6 g, 0.039 mol) were combined, and the mixture was heated
(
at 70 °C for 16 h. The residue was subjected to column
chromatography on silica with n-hexane as eluent to give the
1
obtained as a red oil (4.52 g, 42.2% yield). H NMR, δ: 1.25 (t,
title compound (2b) as a purple solid (4.3 g, 34.2% yield), mp
3H, CH
2
CH
3
), 1.77 (s, 6H, 2 × CH
3 2 3
), 4.17 (q, 2H, OCH CH ),
1
7
7
5
1
1
7-78 °C. H NMR, δ: 1.97 (s, 6H, 2 × CH
3
), 7.20-7.52 (m,
),
Cl), 126.5 (phenyl-C3 and 4-chlrophenyl-C3),
7.35 (dd, 2H, m-ArH), 7.52 (dd, 1H, p-ArH), 7.95 (d, 2H, o-ArH).
1
3
H, ArH) and 7.86 (d, 2H, o-ArH). C NMR, δ: 28.4 (2 × CH
3
16 2 2
CI mass spectrum: found, 269.0675 (M + 1); C13H S O
requires 269.0670.
5.7 (C(CH
3
)
2
C
6
H
4
28.1 (phenyl-C2 and 4-chlrophenyl-C2), 131.9 (phenyl-C4),
32.4 (4-chlorophenyl-C4), 142.8 (4-chlorophenyl-C1), 146.0
2-[N-(2-Hyd r oxyeth yl)ca r ba m oyl]p r op -2-yl Dith ioben -
zoa te (2 h ). A solution of di(thiobenzoyl) disulfide (0.61 g, 2
mmol) and 4,4′-azobis(2-methyl-N-(2-hydroxyethyl)propion-
amide) (0.87 g, 3 mmol, Wako) in ethyl acetate (20 mL) and
ethanol (10 mL) was degassed at reflux for 24 h. The volatiles
were removed under reduced pressure and the residue chro-
matographed on silica with ethyl acetate as eluent to give the
(
phenyl-C1). The signal due to CdS was not observed. EI mass
+
spectrum: found, 306.0304 (M ); C16
-P h en yleth yl Dith ioben zoa te (2c). Dithiobenzoic acid
9.9 g, 0.064 mol), styrene (10 mL 0.087 mol) and carbon
15 2
H S Cl requires 306.0298.
1
9
1
(
tetrachloride (30 mL) were combined and the mixture heated
at 70 °C for 4 h. The resultant mixture was concentrated on a
rotary evaporator and the residue subjected to column chro-
matography on alumina (activity III) with, petroleum spirits
1
title compound (2h ) (0.1 g, 18%). H NMR, δ: 1.74 (6H, 2 ×
CH
O-CH
3
2
), 1.89 (br s, 1H, OH), 3.34 (m, 2H, N-CH ), 3.64 (m, 2H,
2
), 6.74 (br s, 1H, NH), 7.35, 7.51, 7.92 (m, 5H, ArH).
+
as eluent to give 1-phenylethyl dithiobenzoate (2c) (7.2 g,
17 2 2
EI mass spectrum: found, 283.0695 (M ); C13H S O N re-
1
4
CH
3.4%). H NMR, δ: 1.92 (d, 3H, CH-CH
3
3
); 5.39 (q, 1H, CH-
quires 283.0701.
); 7.34-7.62 (m, 8H, ArH), 8.08 (d, 2H, o-ArH of dithioben-
zoate). EI mass spectrum: EI found, 258.0531 (M ); C15
2,4,4-Tr im eth ylp en t-2-yl Dith ioben zoa te (2i). A mixture
of dithiobenzoic acid (5 g, 0.032 mol), 2,4,4-trimethylpentene
(7.3 g, 0.065 mol), and carbon tetrachloride (25 mL) was heated
at 70 °C for 2 days. The volatiles were removed under reduced
pressure, and the residue was chromatographed on silica with
petroleum spirits as eluent to give 2,4,4-trimethylpent-2-yl
+
14 2
H S
requires 258.0537.
Ben zyl Dith ioben zoa te (2d ). The one-pot procedure de-
3
7
scribed by Meijer and Vermeer was modified as follows. A
solution of phenylmagnesium bromide (from bromobenzene
(
62.8 g, 0.40 mol) and magnesium turnings (10 g, 0.42 mol) in
dithiobenzoate (2i) (2.74 g, 31.7% yield) as a dark purple oil.
1
dry tetrahydrofuran (300 mL) was warmed to 40 °C, and
carbon disulfide (30.4 g, 0.41 mol) was added over 15 min at
such a rate that the reaction temperature was maintained at
H NMR, δ: 1.08 (s, 9H, C(CH
C(CH ), 7.35 (dd, 2H, m-ArH), 7.49 (dd, 1H,
p-ArH), 7.85 (d, 2H, o-ArH). C NMR, δ: 28.3 (SC(CH
31.5 (CH C(CH ), 32.8 (CH C(CH ), 50.5 (SC(CH
57.7 (SC(CH CH
ArC), 147.9 (Ar-C1). The signal due to CdS was not observed.
3
)
3
), 1.77 (s, 6H, 2 × CH
3
), 2.20
(s, 2H, CCH
2
3 3
)
1
3
3
3
)
2
CH
CH
2
2
),
),
4
0 °C. Benzyl bromide (77 g, 0.45 mol) was then added to the
2
3
)
3
2
3
)
3
)
2
resultant dark brown mixture over 15 min, after which the
reaction temperature was raised to 50 °C and maintained at
that temperature for a further 45 min. Ice-cold water (1.5 L)
was added and the mixture extracted with diethyl ether (total
3
)
2
2
), 126.6 (o-ArC), 128.1 (m-ArC), 131.5 (p-
+
EI mass spectrum: found, 266.1157 (M ); C15
22 2
H S requires
266.1163.
2
L). The combined ethereal extracts were washed with water
ter t-Bu tyl Dith ioben zoa te (2j). tert-Butyl mercaptan (6.15
g, 0.068 mol) was added dropwise to a solution of benzoyl
chloride (10.5 g, 0.075 mol) in pyridine (6 g). The resulting
mixture was allowed to stir for 2 h at room temperature then
poured onto ice-water and the mixture extracted with diethyl
ether. The organic extract was washed with dilute HCl, water,
and brine and finally dried over anhydrous sodium sulfate.
After removal of solvent and vacuum distillation, S-tert-butyl
thiobenzoate was obtained (6.64 g, 50.1% yield), bp 86 °C (0.8
(1 L) and brine (500 mL) and dried over anhydrous magnesium
sulfate. After removal of solvent, vacuum distillation of the
residue gave benzyl dithiobenzoate (2d ) as a red oil (60.2 g,
6
3
8
1.7%), bp 152 °C (0.02 mmHg) [lit. bp 179-180 °C (3
1
mmHg)]. H NMR, δ: 4.60 (s, 2H, CH
H, ArH), 8.02 (m, 2H, ArH).
-Cya n op r op -2-yl Dith ioben zoa te (2e). A solution of
2
-Ph), 7.30-7.60 (m,
8
2
1
9
cumyl dithiobenzoate (2a ) (2.0 g, 0.0074 mol) and an excess
of AIBN (2.4 g, 0.0146 mol) in benzene (20 mL) was heated
under reflux for 20.5 h. The solution was cooled and the solvent
mmHg). 1H NMR, δ: 1.60 (s, 9H, C(CH
m-ArH), 7.54 (m, 1H, p-ArH), 7.94 (d, 2H, o-ArH). C NMR,
3 3
) ), 7.41 (m, 2H,
1
3

2
258 Chong et al.
Macromolecules, Vol. 36, No. 7, 2003
Ta ble 1. Molecu la r Weigh t a n d Con ver sion Da ta for P olym er iza tion a of MMA a t 60 °C in th e P r esen ce of Dith ioben zoa te
Der iva tives
a
Reaction medium: 75% (v/v) MMA in benzene ([MMA]o 7.02 M, [AIBN]0 6.10 × 10 M). b Numbers in parentheses are calculated
-3
c
conversions based on the molecular weights shown. Transfer coefficients based on single point Mayo analysis are shown in parentheses
(see text).
δ: 29.8 (SC(CH
3
)
3
), 48.0 (SC(CH
3
)
3
), 126.8 (o-ArC), 128.3 (m-
(s, 9H, C(CH
7.88 (d, 2H, o-ArH). C NMR, δ: 28.2 (SC(CH
(SC(CH ), 126.6 (o-ArC), 128.1 (m-ArC), 131.7 (p-ArC), 147.0
(Ar-C1). The signal due to CdS was not observed.
3 3
) ), 7.36 (m, 2H, m-ArH), 7.50 (m, 1H, p-ArH),
1
3
ArC), 132.7 (p-ArC), 138.6 (Ar-C1), 192.9 (C(dO)S). A mixture
of S-tert-butyl thiobenzoate (1.94 g, 0.01 mol) and Lawesson’s
reagent (2,4-bis(4-methoxyphenyl)-1,3-dithia-2,4-diphosphe-
tane-2,4-disufide) (2.43 g, 0.006 mol) in anhydrous toluene (10
mL) was refluxed for 25 h. After cooling to room temperature,
the reaction mixture was concentrated and the residue sub-
jected to column chromatography on silica with petroleum
spirits/diethyl ether 19:1 as eluent. The title compound (2j)
3 3
) ), 52.2
3 3
)
ter t-Bu tyl Tr ith iop er ben zoa te (2k ). The procedure de-
scribed by Aycock and J urch40 was applied. The crude reaction
product was subjected to column chromatography on silica
with n-hexane as eluent to give tert-butyl trithioperbenzoate
1
(2k ) as a dark purple oil in 60% yield. H NMR, δ: 1.32 (s,
1
was obtained as a purple oil, 1.37 g (65.5%). H NMR, δ: 1.69
3 3
9H, C(CH ) ), 7.45 (m, 3H, m, p-ArH), 8.00 (m, 2H, o-ArH).

Macromolecules, Vol. 36, No. 7, 2003
RAFT Polymerization with Thiocarbonylthio Compounds 2259
Ta ble 2. Molecu la r Weigh t a n d Con ver sion Da ta for Bu lk P olym er iza tion of MMA a t 60 °C in th e P r esen ce of
Dith ioben zoa te Der iva tivesa
a
Reaction time 1 h, [AIBN] ) 0.012 M.
Ta ble 3. Molecu la r Weigh ts a n d Con ver sion s Obser ved
in th e P olym er iza tion of MMA a t 60 °C in th e P r esen ce of
Ta ble 4. Molecu la r Weigh ts a n d Con ver sion s Obser ved
in th e P olym er iza tion of MMA a t 60 °C in th e P r esen ce of
2-Cya n op r op -2-yl Dith ioben zoa te (2e)
Cu m yl Dith ioben zoa te (2a )
convn
convn
convn (CTA)
b
[
2a]o
[AIBN]o reaction
convn (CTA)a
(%)
[2e]o
[AIBN]o reaction
2
2
2
2
(
M × 10 ) (M × 10 ) time (h)
Mh n
Mh w/ Mh n
(%)
(M × 10 ) (M × 10 ) time (h)
2.901
2.901
2.901
Mh n
Mh w/ Mh n
(%)
(%)
4.405a
1
4
16
1
4
16
1
6710 1.26
20
95
2
2
2
2
2
2
2
2
2
2
2
2
2
0
0
0
.900
.900
.900
.900
.910
.910
.910
.893
.893
.893
.883
.883
.883
.586
.586
.586
5.67
0.5
5610
8570
14 700
34 400
4280
6940
11 900
3500
9110
30 300
1790
3230
8110
1.32
1.28
1.24
1.25
1.35
1.29
1.23
1.38
1.26
1.22
1.35
1.36
1.26
1.43
1.33
1.20
11
22
88
99
8.6
19
42
5.4
29
>99
1.1
5.2
26
6.1
12
22
58
75
a
5.67
1
4.405
29 100 1.27 >99
31 000 1.29 >99
3380 1.41
8500 1.20
31 900 1.23 >99
3510 1.39
9000 1.19
28 100 1.29 >99
2550 1.50
3040 1.46
5200 1.32
2660 1.40
4.405a
5.67
2
2
2
2
2
2
2
2
2
2
2
2
2
0
0
.937
.937
.937
.941
.941
.941
.901
.901
.901
.931
.931
.931
.576
.576
0.462a
6.5
28
65
68
5.67
4
a
0.462
0.556
0.556
0.556
0.572
0.572
0.572
0.0549
0.0549
0.0549
0.0587
0.0587
0.0587
1.5
3
9
1
4
16
1
4
16
2
4
60
85
0.462a
0.589
7.1
32
0.589
4
45
0.589
16
0.570
0.333
1.5
4.1
14
1
7
36
7.2
15
31
97
20
46
86
13
61
0.570
0.667
5
47
0.570
2
1
4
16
1
0.0546
0.0546
0.0546
0.577
3850 1.38
12 800
19 100
33 700
75
96
10 100 1.18
13 900 1.39
26 600 1.59
49 000 1.55
152 300 3.32
8500 1.74
82
8
0.577
2
a
0.576
0.577
4
Conversion of transfer agent evaluated using eq 6 with
0
0
0
0
0
.576
.578
.578
.578
.578
0.577
8
Mn(calcd) from eq 7.
0.0585
0.0585
0.0585
0.0585
0.0585
0.0585
0.0585
0.0585
0.75
1.5
3
4
1
2
4
8
1.4
28
56
85
92
41
71
75
9600 1.73
3.2
7.1
9.8
2.4
5.4
6.6
P olym er iza tion s. Monomers (Styrene Aldrich 99%, MMA
Aldrich 99%, BA Aldrich 99+%, MA Aldrich 99%) were purified
by filtration through neutral alumina (to remove inhibitors),
fractionated under reduced pressure and flash distilled im-
mediately prior to use. Azobis(isobutyronitrile) (AIBN) was
obtained from Tokyo Kasei and recrystallized twice from
chloroform-methanol. All concentrations are relate to an
ambient temperature of 22 ( 2 °C. Unless stated otherwise,
conversions were determined gravimetrically following drying
to constant weight under vacuum and relate the total mass of
product to the mass of monomer used. In applying these
numbers in calculations these numbers were reduced by the
initial amount of RAFT agent used. Gel permeation chroma-
tography (GPC) was performed on a Waters Associates liquid
chromatograph equipped with differential refractometer and
13 600 1.51
17 400 1.37
9700 1.54
12 400 1.42
14 200 1.40
30 000 1.29
0.581
0.581
0.581
0.581
18
a
Initiator: AIBMe, azobis(methyl isobutyrate). b Conversion of
transfer agent evaluated using eq 6 with Mn(calcd) from eq 7.
and solvent removed by evaporation at ambient temperature
under vacuum and the residues analyzed directly by GPC.
The conditions used and results of MMA polymerizations
are shown in Tables 1-4.
Styr en e P olym er iza tion s. The following procedure is
typical. A stock solution comprising styrene (60 mL) and AIBN
(16.9 mg) was prepared. Aliquots (5 mL) were removed and
transferred to ampules containing benzyl dithiobenzoate (2d )
(11.4 mg). The ampules were degassed with three freeze-
evacuate-thaw cycles, sealed and heated at 60 ( 1 °C for the
periods of time indicated in Table 5. The excess monomer and
solvent removed by evaporation at ambient temperature under
vacuum and the residues analyzed directly by GPC.
6
5
4
3
a set of six Ultrastyragel columns (10 , 10 , 10 , 10 , 500, and
00 Å). Tetrahydrofuran (flow rate of 1.0 mL/min) was used
1
as eluent at 22 ( 2 °C. The columns were calibrated with
narrow polydispersity polystyrene or poly(methyl methacry-
late) standards (data in Tables 3 and 4) obtained from Polymer
Laboratories. Poly(butyl acrylate) molecular weights are re-
ported in polystyrene equivalents. No precipitation or other
form of fractionation was performed prior to analysis.
MMA P olym er iza tion s. The following procedure is typical.
A stock solution comprising MMA (15 mL), AIBN (20 mg),
cumyl dithiobenzoate (2a ) (60.7 mg), and benzene (5 mL) was
prepared. Aliquots (4 mL) were transferred to four ampules,
degassed with three freeze-evacuate-thaw cycles and sealed.
The ampules were heated at 60(1 °C in a thermostated oil
bath for the times indicated in Table 1. The excess monomer
The conditions used and results of styrene polymerizations
are summarized in Table 5.
P olym er iza tion s of Acr yla te Ester s. The following pro-
cedure is typical. Two stock solutions were prepared (a)
comprising AIBN (13.4 mg) in benzene (50 mL) and (b)
comprising benzyl dithiobenzoate (2d ) (9.62 mg) in benzene
(10 mL). Aliquots of these solutions (4 mL of solution a and 2

2
260 Chong et al.
Macromolecules, Vol. 36, No. 7, 2003
Ta ble 5. Molecu la r Weigh t a n d Con ver sion Da ta for Bu lk a P olym er iza tion of Styr en e a t 60 °C in th e P r esen ce of
Dith ioben zoa te Der iva tives
a
Bulk [styrene] ) 8.728 M. b Bimodal distribution observed.
mL of solution b) were transferred to ampules containing BA
2 mL). The ampules were degassed with three freeze-
Sch em e 2. Mech a n ism of RAF T P olym er iza tion
(
evacuate-thaw cycles, sealed, and heated at 60 ( 1 °C for the
periods of time indicated. The excess monomer and solvent
removed by evaporation under vacuum and the residues
analyzed directly by GPC. The results of this and related
polymerizations are summarized in Table 6.
ω-(Th iob en zoylt h io)p olyst yr en e. The ω-(thiobenzoyl-
thio)polystyrene was prepared as follows: A solution compris-
ing AIBN (28 mg) and cumyl dithiobenzoate (2a ) (228 mg)
styrene (100 mL) was prepared, transferred to an ampule,
degassed, and sealed. The solution was heated at 60 °C for
3
0.5 h. The ω-(thiobenzoylthio)polystyrene isolated by evapo-
ration of the residual monomer under reduced pressure (18.8
g, 21% conversion) had Mh 20 300, Mh / Mh 1.15.
n
w
n
P olym er iza tion of Meth yl Meth a cr yla te in th e P r es-
en ce of ω-(Th ioben zoylth io)polystyr en e. The above ω-(thio-
benzoylthio)polystyrene (0.5 g) was dissolved in tetrahydro-
furan and the solution transferred to an ampule. The tetrahy-
drofuran was removed under reduced pressure, and AIBN
(
3.88 mg) and MMA (2 mL) were added. The solution was
P olystyr en e-block-p oly(N,N-d im eth yla cr yla m id e). ω-
degassed, sealed in an ampule, and heated at 60 °C for 3.8 h.
After removal of the unreacted styrene, the product was
analyzed by GPC, and the mixture comprised apparently
unchanged ω-(thiobenzoylthio)polystyrene and poly(methyl
methacrylate): Mh
P olystyr en e-block-p oly(4-m eth ylstyr en e). ω-(Thioben-
zoylthio)polystyrene (0.5 g, M 20 300, M /M 1.15, prepared
as described above), 4-methylstyrene (2 mL), AIBN (2.5 mg),
and benzene (0.5 mL) were transferred to an ampule. The
resulting solution was degassed, sealed, and heated at 60 °C
for 3 h. Volatiles were removed under reduced pressure to give
polystyrene-block-poly(4-methylstyrene) (0.81 g, 17.1% conver-
n w n
(Thiobenzoylthio)polystyrene (0.2 g, M 20 300, M /M 1.15,
prepared as described above), N,N-dimethylacrylamide (2 mL),
AIBN (0.5 mg), and benzene (8 mL) was transferred to an
ampule. The resulting mixture was degassed, sealed and
heated at 60 °C for 1 h. The volatiles were removed in vacuo
to give polystyrene-block-poly(N,N-dimethylacrylamide) at 0.4
n
290 000, Mh
w
/Mh
n
2.63 (conversion ca. 70%).
n
w
n
g, 10.4% conversion, with Mh
n
43 000 and Mh
w
n
/ Mh 1.24 (GPC with
N,N-dimethylformamide eluent at 1 mL/min).
P olym er iza tion of Styr en e in th e P r esen ce of ω-(Th io-
ben zoylth io)p oly(m eth yl m eth a cr yla te). The ω-(thioben-
zoylthio)(methyl methacrylate) was prepared as follows: A
solution comprising AIBN (53 mg), cumyl dithiobenzoate (2a )
(160 mg) MMA (40 mL), benzene (13 mL) was prepared and
sion, Mh
n
25 400, and Mh
w
n
/ Mh 1.19).

Macromolecules, Vol. 36, No. 7, 2003
RAFT Polymerization with Thiocarbonylthio Compounds 2261
Ta ble 6. Molecu la r Weigh t a n d Con ver sion Da ta for P olym er iza tion of Acr yla te Ester sa a t 60 °C in th e P r esen ce of
Dith ioben zoa te Der iva tives
a
b
Reaction medium: 40% (v/v) BA in benzene ([BA]0 ) 2.79 M, [AIBN]0 ) 0.00032 M, and [2]0 ) 0.0016 M. [MA]0 ) 4.45 M, [AIBN]
c
)
0.00033 M, and [2]0 ) 0.0037 M. [MA]0 ) 4.43 M, [AIBN] ) 0.00033 M, and [2]0 ) 0.006 M. d Values in parentheses from universal
7
6
e
calibration with Mark Houwink constants for poly(methyl acrylate): K ) 0.000261 and R ) 0.659.
Part of the molecular weight
distribution exceeds upper exclusion limit of column set. f Bimodal distribution observed.
transferred to an ampule, which was degassed and sealed. The
solution was heated at 60 °C for 4 h. The ω-(thiobenzoylthio)-
poly(methyl methacrylate) was isolated by evaporation of the
solvent and residual monomer under reduced pressure (10.3
mechanism as defined in Scheme 2 and using the method of
4
1
moments as described in an earlier publication. In the
present work, numerical integration was performed using the
Mathematica 4.1 (Wolfram Research Inc.) function NDSolve.
g, 25% conversion) had Mh
n
14 700, Mh
w
n
/ Mh 1.21).
For the kinetic simulation of styrene polymerization, it was
-
1
The reaction time was chosen to give an approximate
doubling of the molecular weight under conditions similar to
those used in styrene homopolymerization. The above poly-
assumed that k
add p
does not depend on R and is 12000k M
-1. This value is based on the value of C of ∼6000 estimated
s
tr
by Goto et al.30 for 2m ( Mh n ∼ 2000) at 60 °C. The rate constant
(methyl methacrylate) (1.7 g) was dissolved in ethyl acetate
(k-â ) for benzyl or cumyl adding to the polymeric RAFT agent
and the solution transferred to an ampule. The ethyl acetate
was removed under reduced pressure, and AIBN (2.82 mg) and
styrene (10 mL) were added. The solution was degassed, sealed
in an ampule, and heated at 60 °C for 20 h. After removal of
volatiles, poly(methyl methacrylate)-block-polystyrene was
obtained (3.9 g, 23.5% conversion) with Mh
No residual poly(methyl methacrylate) was apparent in the
GPC chromatogram.
-1 -1 42
has been taken as 12000k
i
M
s . Differences in the transfer
coefficient for the various RAFT agents are then due to
differences in the partition coefficient (1:1 for the polystyryl
propagating radical, 1:29 for benzyl, and 1:5 for cumyl). The
-1
â
polymerization kinetics are insensitive to k >10 s . A lower
n
35 000, Mh
w
n
/ Mh 1.24.
k
causes a broadening of the molecular weight distribution
â
for low conversions and much lower values lead to retardation.
-1
â
A value of k ) 1000 s was used; the actual value may be
Kin etic Sim u la tion . Kinetic simulation was performed by
directly solving the differential equations associated with the
higher or lower than this. Other rate constants used in the
-
1
-1 43
8
-1 -1 44
simulation are k
p
) 340 M
s
,
k
t
) 1.08 × 10 M
s ,

2
262 Chong et al.
Macromolecules, Vol. 36, No. 7, 2003
) 9.65 × 10 M-1
-6
s
-1,4 and f ) 0.6, and
k
k
tc/(ktc + ktd) ) 1, k
i
d
(for benzyl, cumyl, and 2-cyanoprop-2-yl) was as shown in
Table 10.
Molecu la r Or bita l Ca lcu la tion s. Semiempirical molec-
ular orbital calculations were performed with MOPAC 6.0
using the ChemBats3D Ultra package on an Apple Macintosh
computer as the graphical user interface. For each compound
a complete energy minimization was carried out using the
keyword “PRECISE”. For radical species the keyword “UHF”
was specified. Multiple conformations were used a starting
points for geometry optimization to ensure a global minimum
was achieved.
The AM1 Hamiltonian was used for all calculations reported
in this paper. Use of the PM3 Hamiltonian was also briefly
explored. The trend in LUMO energies was the same. The AM1
was finally selected over the PM3 Hamiltonian as it provides
better heat of formation data for free radical species.45
F igu r e 1. Relationship between transfer constant and the
amount of residual RAFT agent for various monomer conver-
sions.
Resu lts a n d Discu ssion
Tr a n sfer Coefficien t Estim a tion . The Mayo equa-
4
6
tion may be used to estimate transfer coefficients of
dithiobenzoates in circumstances where the consump-
tion of dithiobenzoate and monomer are negligible. For
the case of the less active dithiobenzoates, these condi-
tions are easily met in low conversion polymerizations.
We have used this method to estimate transfer coef-
ficients of the less active dithiobenzoates in MMA
polymerization (Table 2). For the case of the more active
dithiobenzoates, which provide very narrow polydisper-
sities, and where the molecular weight is seen to
increase with conversion even at very low conversion,
direct application of the Mayo method will underesti-
mate the transfer coefficient and a different approach
is required.
In this case, the slope of a plot of ln[M] vs ln[2] yields
the transfer coefficient. If the rate of the reaction of R•
with the polymeric RAFT agent (4) is not negligible, the
transfer coefficient obtained from such a plot will be
lower than the actual transfer constant. The fit of
experimental data to eq 5 or the Mayo equation does
•
not prove that the reaction of R with the polymeric
RAFT agent (4) is negligible. For this reason the term
transfer coefficient has been adopted rather than trans-
fer constant to refer to numbers obtained by either the
Mayo method or by application of eq 5.
A simple method of evaluating the conversion of
transfer agent is based on a comparison of found and
calculated molecular weights (eq 6).
In the case of reversible chain transfer, the rate of
consumption of the transfer agent depends on two
transfer coefficients, Ctr ()ktr/kp) and C-tr ()k-tr/ki),
which describe the reactivity of the propagating radical
[2] - [2]t
[M] - [M]t
o
o
conversion of 2 )
)
{
}
[2]o
[2]o
/
•
•
[M] - [M]
Xh (calcd)
(
1
Pn ), and the expelled radical (R ) respectively (see eq
).
o
t
n
)
(6)
4
7
{
}
[2] - [2]t
X (found)
o
n
d[2]
[2]
where Xh n(calcd) is the expected number-average degree
of polymerization assuming complete consumption of
transfer agent and Xh n(found) is the measured number-
average degree of polymerization. In some cases, a small
correction is needed to allow for initiator-derived chains
(see below).
≈
Ctr
(1)
d[M]
[M] + C [2] + C-tr[4]
tr
In chain transfer by addition-fragmentation (Scheme
2
), the rate constant for chain transfer (ktr) is given by
4
8
the following expression (eq 2).
It is useful to at least place a lower limit on the value
of the transfer constant in circumstances where the
found and calculated molecular weights agree (or are
within experimental error) at the first conversion time
point. If the found and calculated degrees of polymeri-
zation (or molecular weights) agree, it suggests that the
transfer agent is completely consumed (no residual
RAFT agent). With eqs 5 (or 1), it is possible to predict
the dependence of the amount of residual transfer agent
on the transfer coefficient for a given monomer conver-
sion. Figure 1 shows this dependence for three different
monomer conversions (1%, 2%, and 5%) estimated using
eq 5.
kâ
ktr ) kadd
×
(2)
(3)
k-add + kâ
Similarly (eq 3)
k-add
ktr ) k-â×
k-add + kâ
These relationships assume that adduct 3 undergoes no
reactions other than fragmentation.
If the rate of the reverse reaction between R and the
polymeric RAFT agent (4) is negligible, and chains are
•
In principle, transfer coefficients of RAFT agents may
also be estimated from the rate at which the polydis-
persity or weight-average molecular weight reduces
with conversion/time. Kinetic simulation shows that
these parameters are more sensitive to variations in the
transfer coefficient than is the number-average molec-
ular weight. Polydispersities depend on the transfer
coefficients of both the initial RAFT agent (2) and those
of the polymeric species (4) formed during polymeriza-
long, this expression (eq 1) simplifies to eq 4 that also
describes conventional chain transfer.
4
9
d[2]
[2]
≈
Ctr
(4)
(5)
d[M]
[M]
ktr [M] d[2]
d(ln[2])
Ctr
)
≈
)
kp
[2] d[M] d(ln[M])

Macromolecules, Vol. 36, No. 7, 2003
RAFT Polymerization with Thiocarbonylthio Compounds 2263
Sch em e 3
Sch em e 4
Ctr(2) polydispersities are more sensitive to the value
of the transfer coefficient than are the number-average
molecular weights. The simulations also show that, with
Ctr(4) . Ctr(2) and C-tr(2) negligible, the predicted
polydispersities are independent of the precise value of
Ctr(4) and are indicative of the value of Ctr(2). Polydis-
persities predicted with these equations can therefore
be useful for placing a lower limit on the transfer
coefficient and for initial parameter estimation. A
finding that polydispersities reduce with conversion at
a rate greater than suggested by these equations is
evidence that C-tr(2) is not negligible.
Ca lcu la ted Molecu la r Weigh ts. Molecular weights
in RAFT polymerizations were calculated using eq 7.
[
M] - [M]t
o
F igu r e 2. Predicted dependence of (a) polydispersity and (b)
degree of polymerization on conversion in polymerizations
involving reversible chain transfer as a function of the chain
transfer constant. Predictions are based on equations proposed
Mh (calcd) )
m
(7)
n
[2] + df([I] - [I] )
o
o
t
where m is the monomer molecular weight.
df([I]o - [I]t) is the number of initiator-derived chains
5
0,51
-7
by Mueller et al.
05. Experimental data points are for methyl methacrylate
polymerization in the presence of dithiobenzoate esters PhC-
dS)SR (0.0116 M) where R ) -C(Me) CO Et (2g) (O) and R
-C(Me) Ph (2a ) (0). For reaction conditions, see Table 1.
with R ) 10 , â as indicated, and γ )
6
produced and
(
2
2
-
kdt
)
2
[
I] - [I] ) [I] (1 - e
)
(8)
o
t
o
tion. Recent work of Goto et al.30 suggests that ω-(thio-
benzoylthio)poly(methyl methacrylate) (2l) and ω-(thio-
benzoylthio)polystyrene (2m ) have transfer coefficients
of >6000 and ca. 140 at 60 °C in styrene and MMA
polymerizations, respectively. The transfer coefficients
of other polymeric dithiobenzoates are unknown. How-
ever, for the systems included in the present study, the
transfer coefficient of the polymeric RAFT agents (in
the monomer from which they are formed) do not appear
to be a limiting factor. We have used polydispersities
predicted by kinetic simulation to estimate Ctr and C-tr
for benzyl and cumyl dithiobenzoates in styrene polym-
erization (see below).
where k is the rate constant for initiator decomposition,
f is the initiator efficiency (for AIBN at 60 °C, k ) 9.65
d
4
d
-
6
-1 -1
× 10
M
s , and f ) 0.6) and d is the number of
chains produced from radical-radical reaction (d ) 1.67
for MMA and d ) 1.0 for styrene). These parameters
should generally hold for low conversions <50%. For
higher conversions use of these parameters will over-
estimate the number of initiator-derived chains and it
is necessary to allow for the decrease in the initiator
efficiency with conversion. In many experiments (those
with short reaction times or low initiator concentrations)
the term df([I] (1 - e
and can be neglected.
5
4
-
k t
d
o
) is small with respect to [2]o
There has been some use of equations derived by
Mueller et al.50,51 to describe polymerization with re-
versible chain transfer to predict molecular weights and
Syn th esis of Dith ioben zoa te Der iva tives. Dithio-
benzoate derivatives are available in moderate to excel-
lent yields by a variety of methods. For a summary of
5
2,53
55
polydispersities in RAFT polymerization (Figure 2).
the literature on dithioester synthesis see review.
5
6
Use of these equations requires an assumption that the
rate of consumption of the initial RAFT agent (2) is not
kinetically significant and that there is no radical-
radical termination (or other side reactions). Full kinetic
simulations carried out based on numerical integration
of the appropriate differential equations (i.e. including
initiation and radical-radical termination) using meth-
ods described previously41 demonstrate the validity of
these equations and indicate they provide a reasonable
estimate of the molecular weight and polydispersity for
low monomer conversions (<20% for MMA polymeriza-
tion, <5% for styrene polymerization under the reaction
conditions used in this work) as long as C-tr(2) is
negligible. It is also clear from Figure 2 that for high
Sudalai et al. have described a method based on the
use of phosphorus pentasulfide which reportedly pro-
vides dithiobenzoates in high yields. The four methods
have been used in this work are summarized below.
(a) The first is sequential treatment of a phenyl
Grignard reagent with carbon disulfide and an alkyla-
ting agent (typically an alkyl bromide) in a “one pot”
37
procedure (Scheme 3). This process was used to
prepare benzyl dithiobenzoate (2d) (62% yield), 2-(ethoxy-
carbonyl)prop-2-yl dithiobenzoate (2g) (42% yield) and
1
9
2-cyanoprop-2-yl dithiobenzoate (2e) (43% yield).
(b) Second is addition of dithiobenzoic acid across an
5
7
olefinic double bond (Scheme 4). Dithiobenzoic acid is
readily prepared by oxidation of benzyl chloride with

2
264 Chong et al.
Macromolecules, Vol. 36, No. 7, 2003
Sch em e 5
Sch em e 6
F igu r e 3. Evolution of molecular weight with conversion for
polymerization of methyl methacrylate (7.02 M in benzene)
at 60 °C with AIBN initiator (0.0061 M) in the presence of
dithiobenzoate esters PhC(dS)SR (0.0116 M) where R )
-
-
-
C(Me)
2
CO
2
Et (2g) (b), R ) -C(Me)
Cl (2b) (O), R ) -C(Me) CN (2e) (4), or R )
CH CH OH (2f) (]). For experimental details,
2
Ph (2a ) (0), R )
C(Me)
2
C
6
H
4
2
C(Me)(CN)CH
2
2
2
see Table 1. Key: calculated molecular weights assuming no
initiator-derived chains (- -); calculated molecular weight
including initiator-derived chains (- - -) based on df ) 1.0.
In flu en ce R Su b st it u en t on t h e Act ivit y of
Dith ioben zoa te Der iva tives. To examine the effect
of the R substituent on RAFT polymerization in greater
detail, polymerizations of MMA, styrene, and MA or BA
have been carried out in the presence of a series of
dithiobenzoate derivatives (2). Polymerizations were
conducted at 60 °C either in bulk or in benzene solution
for various reaction times and conversions. For MMA
and styrene polymerization, the concentration of RAFT
agent was chosen to be sufficiently low to avoid the
substantial retardation that may be observed when high
concentrations of RAFT agents are used (e.g. cumyl
dithiobenzoatessee further discussion below).19 The rate
of polymerization of butyl acrylate, and other acrylates,
is significantly retarded by dithiobenzoate derivatives.19
The results (values of molecular weight and polydisper-
sity vs conversion) obtained are summarized in Tables
elemental sulfur.58 Dithiobenzoic acid is unstable and
should be used immediately or stored at low tempera-
ture (<-20 °C). The process was used to prepare cumyl
1
9
dithiobenzoate (2a ) from R-methylstyrene (33% yield),
1
-phenylethyl dithiobenzoate (2c) from styrene (43%
yield), 2-(4-chlorophenyl)prop-2-yl dithiobenzoate (2b)
from 4-chloro-R-methylstyrene (34% yield), and 2,4,4-
trimethylpent-2-yl dithiobenzoate (2i) (32% yield). It
(
Scheme 4) cannot be used to prepare tertiary dithioben-
zoates from electrophilic monomers (e.g. MMA, MAN)
where the primary dithiobenzoate is formed preferen-
tially.57
(
c) Third is radical-induced decomposition of bis-
(
thiobenzoyl) disulfide (Scheme 5). The procedure was
used to prepare 4-cyano-1-hydroxypent-4-yl dithioben-
zoate (2j) (46% yield) and 2-[N-(2-hydroxyethyl)carbam-
oyl]prop-2-yl dithiobenzoate (2h ) (18% yield) with the
corresponding azo-compound as the source of free
radicals. This route is described in a recent communica-
tions and appears to be the method of choice for
1
-6.
MMA P olym er iza tion . Data for MMA polymeriza-
tions carried out in the presence of a series of dithioben-
zoates (2) under a standard set of reaction conditions
(60 °C, 7.02 M in benzene solvent) are shown in Table
2
1,59,60
preparing functional tertiary dithiobenzoates.
d) Finally there is radical-induced ester exchange.
1
and Figures 3 and 4. Under these conditions, the
(
ability of dithiobenzoate derivatives [2, SdC(Ph)S-R]
to give narrow polydispersity and molecular
weight control decreases in the order where R is
Treatment of a dithiobenzoate ester with a source of free
radicals can result in ester exchange as shown in
Scheme 6. The reaction was used to prepare 2-cyano-
prop-2-yl dithiobenzoate (2e) from cumyl dithiobenzoate
-
C(CH3)(CH2CH2CH2OH)CN
C(Me)2Ar > -C(Me)2C(dO)OEt > -C(Me)2C(dO)NH-
>
-C(Me)2CN
∼
-
(
2a ) (>90% yield) with AIBN as the source of free
(alkyl) > -C(Me)2CH2C(Me)3, g -C(Me)HPh > -C(Me)3
radicals. The R group of the precursor dithiobenzoate
should be a good free radical leaving group with respect
to desired R group. For example, benzyl dithiobenzoate
appears not to react with the 2-cyanoprop-2-yl radical
∼
-CH2Ph.
Of the compounds listed, only the tertiary cyanoalkyl
(2e,f) and cumyl dithiobenzoates (2a ,b) give substan-
tially narrowed polydispersities under the present reac-
tion conditions. The close correspondence between found
and calculated molecular weights (see Figure 3) indi-
cates that the RAFT agent is largely consumed at the
first time point (<15% conversion of monomer) and
therefore these data do not permit a direct estimate of
the transfer coefficient. Relatively high transfer coef-
ficients (>10) are also indicated by the narrow polydis-
persity polymers (<1.3) that are obtained even at the
first conversion point (Figure 2). The transfer coef-
ficients of compounds 2a and 2e were evaluated in
separate experiments where the polymerization was
(
though it may react reversibly). It appears essentially
unchanged (NMR) after heating with AIBN in benzene
at 80 °C.
All RAFT agents were purified by column chroma-
tography, crystallization and/or distillation as appropri-
ate and their purity established by NMR analysis and
thin-layer chromatography prior to use. The yields
quoted are for the isolated and purified products,
reaction conditions have not been optimized. Confirma-
tion of identity in the case of novel compounds also came
from high-resolution chemical ionization mass spec-
trometry.

Macromolecules, Vol. 36, No. 7, 2003
RAFT Polymerization with Thiocarbonylthio Compounds 2265
F igu r e 6. Double log plot of transfer agent concentration vs
F igu r e 4. Evolution of polydispersity with conversion for
polymerization of methyl methacrylate (7.02 M in benzene)
at 60 °C with AIBN initiator (0.0061 M) in the presence of
dithiobenzoate esters PhC(dS)SR (0.0116 M) where R )
monomer concentration for polymerization of methyl meth-
4
7
2
acrylate in the presence of PhC(dS)SC(Me) Ph (2a ). Data
are for bulk MMA polymerization at 60 °C with AIBN initiator
in the range 0.0005-0.04 and 0.0058 M (b) or 0.029 M (O) 2a
as RAFT agent. For additional details, see Table 3. Lines are
those predicted by eq 1 with Ctr ) 56 and C-tr ) 2500.
-
-
-
C(Me)
2
CO
2
Et (2g) (b), R ) -C(Me)
Cl (2b) (O), R ) -C(Me) CN (2e) (4), or R )
CH CH OH (2f) (]). For experimental details,
2
Ph (2a ) (0), R )
C(Me)
2
C
6
H
4
2
C(Me)(CN)CH
2
2
2
see Table 1.
F igu r e 7. Double log plot of transfer agent concentration vs
F igu r e 5. Plot of molecular weight vs conversion for polym-
erization of methyl methacrylate in the presence of PhC(dS)-
SC(Me) CN (2e). Data are for bulk MMA polymerization at
2
monomer concentration for polymerization of methyl meth-
4
7
2
acrylate in the presence of PhC(dS)SC(Me) CN (2e). Data
4
7
are for bulk MMA polymerization at 60 °C with AIBN initiator
in the range 0.0005-0.04 and 0.0058 M (b) or 0.029 M (O) 2e
as RAFT agent. For additional details, see Table 4. Lines are
those predicted by eq 1 with Ctr ) 25 and C-tr ) 450.
6
0 °C with AIBN initiator in the range 0.0005-0.04 and 0.0058
M (b) or 0.029 M (O) cumyl dithiobenzoate as RAFT agent.
For additional details, see Table 4.
carried out in bulk and some experiments were carried
to relatively low conversions of monomer and RAFT
agent (Tables 3 and 4, respectively). A plot of molecular
weight vs conversion shows that the transfer agent is
not fully consumed to at least 20% conversion (for 2e,
see Figure 5). For these compounds the transfer coef-
ficients (obtained by direct use of eq 5) depend on RAFT
agent concentration (see Figures 6 and 7; apparent Ctr
)
9 with 0.029 M 2a , 25 with 0.0058 M 2a ; apparent
Ctr ) 13 with 0.029 M 2e, 21 with 0.0058 M 2e). In these
circumstances, the rate of consumption of RAFT agent
is determined by eq 1. This equation can be solved
6
1
numerically and Ctr ) 56, C-tr ) 2500 for 2a and Ctr
)
25, C-tr ) 450 for 2e (bulk MMA, 60 °C) provides a
good fit to the experimental data (see Figures 6 and 7).
This means that 2-cyanoprop-2-yl dithiobenzoate (2e),
even though it has a lower Ctr, is more effective than
cumyl dithiobenzoate (2a ) for high concentrations of
RAFT agent. Both reagents are highly effective when
using low RAFT agent concentrations (e0.03 M) to
make higher molecular weight polymers.
F igu r e 8. Double log plots of transfer agent concentration
vs monomer concentration for polymerization of methyl meth-
acrylate (7.02 M in benzene) at 60 °C with AIBN initiator
(
2 2
0.0061 M) in the presence of PhC(dS)SC(Me) CO Et (0.0116
M 2g). For additional experimental details, see Table 1. Line
of best fit is -7.792 +1.6981 ln[M]; correlation coefficient )
0.99877.
When R ) -C(Me)2CO2Et (2g) the polydispersity
reduces with conversion to a final value of Mh w/ Mh n ) 1.5
Figure 8). The difference between found and calculated
molecular weights enables the conversion of transfer
agent to be estimated (from eq 5) and a plot of ln[2g] vs
ln[M] suggests a transfer coefficient of ca. 1.7 at 60 °C
(Figure 8).
(
Figure 4). The molecular weights predicted for low
conversions are significantly higher than those expected
based on complete consumption of the RAFT agent (see

2
266 Chong et al.
Macromolecules, Vol. 36, No. 7, 2003
F igu r e 9. Kinetic plot for polymerization of methyl meth-
acrylate (7.02 M in benzene) at 60 °C with AIBN initiator
p
F igu r e 10. Plot of log (initial R ) vs log (initial initiator
concentration). Data are for bulk MMA polymerization at 60
°C with AIBN initiator (0.0005-0.04 M) and either cumyl (9)
or 2-cyanoprop-2-yl dithiobenzoate (O) as RAFT agent 0.006-
0.03 M. A least-squares fit provides slope 0.507, R ) 0.98655.
(
0.0061 M) in the presence of dithiobenzoate esters PhC(dS)-
SR, where R ) -C(Me) CO Et (2g) (b), R ) -C(Me) Ph (2a )
0), R ) -C(Me) Cl (2b) (O), R ) -C(Me) CN (2e) (4), or
CH CH OH (2f) (]). For experimental
2
2
2
(
2
C
6
H
4
2
R ) -C(Me)(CN)CH
2
2
2
details see Table 1. Key: (- -) a line of best fit; (- - -) result
of kinetic simulation.
The dithiobenzoates with R ) -C(Me)2CH2C(Me)3
2i), -C(Me)HPh (2c), -C(Me)3 (2j), and -CH2Ph (2d )
(
appear ineffective as RAFT agents in MMA polymeri-
zation. No significant narrowing of the molecular weight
distribution was observed (see Tables 1 and 2). No
molecular weight control of the form expected for a
living polymerization was obtained. There is only a
small (e2-fold) decrease in molecular weight with
respect to the control. These poor results are attributed
to these compounds having relatively low transfer
coefficients in MMA polymerization. The transfer coef-
ficients of 2i, 2c, 2j, and 2d at 60 °C were determined
either using the conventional Mayo method or by
analysis of the rate of RAFT agent consumption and are
estimated to be ca. 0.4, 0.15, 0.03, and 0.03, respectively.
The compound with R ) -StBu (2k ) provides a
narrow polydispersity product suggesting a high trans-
fer coefficient. However, molecular weights are signifi-
cantly higher than expected based on the concentration
of RAFT agent. This is attributed to the instability of
the trithioperbenzoate under the reaction conditions
causing the concentration of the reagent to be lower
than expected.
F igu r e 11. Evolution of molecular weight with conversion for
bulk polymerization of styrene 60 °C with AIBN initiator
(
0.0017 M) in the presence of dithiobenzoate esters PhC(dS)-
SR, where R ) -C(Me)
or -C(CH CH C(CH
Table 3. Key: calculated molecular weights assuming no
initiator-derived chains (- -); calculated molecular weight
including initiator-derived chains (- - -) based on df ) 1.4.
2
2 2 2 5
Ph (2a ) (]), -C(Me) CO C H (2g) (4),
3
)
2
2
3
)
3
(2i) (0). For experimental details, see
Styr en e P olym er iza tion s. In styrene polymeriza-
tion, all of the dithiobenzoates (2) tested were found to
give both narrowed polydispersities and good molecular
weight control (see Table 5 and Figures 11 and 12). In
the case of dithiobenzoates with R ) -C(Me)2Ph, (2a ),
-C(Me)2CO2Et (2g), or -C(Me)2CH2C(Me)3 (2i), narrow
polydispersities were evident at the first time/conversion
point (e1.13 at e2% conversion, Figure 12), and the
correspondence between found and calculated molecular
weight indicates complete consumption of the initial
RAFT agent. Complete (>95%) utilization of the initial
RAFT agent at <2% conversion suggests Ctr > 150
(Figure 1). Furthermore, to attain a polydispersity of
1.13 at <2% conversion, a substantially higher transfer
coefficient (Ctr > 500) is required (Figure 2a). The
transfer coefficient of the ω-(thiobenzoylthio)polystyrene
(2m ) in bulk styrene polymerization has been deter-
The effectiveness of dithiobenzoate derivatives in
promoting living polymerization of MMA thus depends
strongly on the nature of R.
For the polymerizations reported in Table 1, the rate
of polymerization shows no marked dependence of the
particular dithiobenzoate derivative used (Figure 9), and
while the rate of polymerization is slow with respect to
the rate of polymerization expected in the absence of
RAFT agent, it appears slightly accelerated when
compared to the rate expected in the absence of a gel
effect (that predicted by kinetic simulation with low
conversion propagation and termination rate constants).
For the polymerizations reported in Tables 3 and 4, a
plot of log[initial rate of polymerization] vs log[AIBN]
shows the (initial) rate of polymerization to be 0.5 order
in initiator (as expected for conventional radical polym-
erization initiated by AIBN) and over a 5-fold range of
RAFT agent concentrations (0.0006-0.003 M of 2a or
3
0
mined to be g6000. Kinetic simulation with Ctr ) 2000
and a C-tr ) 10 000 with kâ g 10 provides an acceptable
fit to the observed variation of molecular weight and
polydispersity with conversion for RAFT agent 2a
(Figure 12). Additional experiments are required to
refine these values.
2
e) and the rate of polymerization seems independent
1
9
of the RAFT agent concentration (Figure 10) with
slight retardation apparent only for the highest con-
centration of 2a . For higher concentrations of 2a ,
polymerization may be strongly retarded.61
The rate of polydispersity narrowing with conversion/
time was lower with benzyl dithiobenzoate (2d ) than
with the other dithiobenzoates examined (see Figure
12). When using this reagent (2d ), the molecular

Macromolecules, Vol. 36, No. 7, 2003
RAFT Polymerization with Thiocarbonylthio Compounds 2267
Sch em e 7. P oten tia l Sid e Rea ction s
was marked retardation (manifest as a slow initial rate
of polymerization) for very high concentrations of cumyl
dithiobenzoate (0.018 M 2a , at 110 °C 50% w/w styrene
in toluene). Much less retardation was observed19 with
2
-cyanoprop-2-yl dithiobenzoate (2e) under these condi-
tions or (in accord with the present work) with 2a at
lower concentrations (0.003 M 2a ). Monteiro and de
Brouwer use relatively high RAFT agent concentra-
6
2
F igu r e 12. Evolution of polydispersity with conversion for
bulk polymerization of styrene with AIBN initiator (0.0017 M)
in the presence of dithiobenzoate esters PhC(dS)SR, where R
64
tions (0.04-0.06 M) as do Kwak et al. (0.1 M). Thus,
we have no reason to believe there is any inconsistency
between our and their results. Our current findings
appear in contrast with recent work of Barner-Kowollik
)
-CH
]), -C(Me)
CH (0.0084 M 2i) (0). For experimental details, see Table
. Lines come from the results of kinetic simulation with Ctr
2
Ph (0.0093 M 2d ) (b), R ) -C(Me)
2
Ph (0.0083 M 2a )
(
(
2
CO (0.0084 M 2g) (4), or -C(CH
2
C
2
H
5
3
)
2
CH C-
2
6
3
)
3
et al. who found that bulk styrene polymerization with
cumyl dithiobenzoate at 60 °C is strongly retarded.
3
3
)
50 and C-tr ) 0 (- -), Ctr ) 400 and C-tr ) 11 600 (s), or
63
Conditions differ, however: their concentration range
C
tr ) 2000 and C-tr ) 10 000 (- - -). Other kinetic parameters
for 2a (0.0025-0.033 M) spans that used in the present
work (0.0084 M), and their initiator concentration
are provided in the Experimental Section.
(
0
[AIBN] 0.0035 M) is 2-fold higher than ours ([AIBN]
.0017 M). It is not clear why they should find strong
retardation with their lower RAFT agent concentra-
tions.
The rate of polymerization observed with 2d appears
consistent with a ca. 4-fold higher termination rate
constant. Attempts to model the conversion time profile
by including side reaction of the intermediates 3 or 5
(
as shown in Scheme 7) did not give the observed
conversion time profile. It can also be noted that, the
development of a bimodal molecular weight distribution
was observed for high conversions (>50%) with either
2
a or 2d with the appearance of a high molecular weight
shoulder with peak Mh n ca. twice that of the principal
peak (see discussion below).
F igu r e 13. Kinetic plot for polymerization of styrene in the
presence of dithiobenzoate esters PhC(dS)SR, where R )
Acr yla te P olym er iza tion . Molecular weight, poly-
dispersity, and conversion data for polymerizations of
BA or MA in the presence of several dithiobenzoates (2)
are shown in Table 6. All reagents were effective in
controlling the molecular weight and provided narrow
polydispersity. These reagents include those that were
ineffective in controlling MMA polymerization (R )
-
CH
]), -C(Me)
CH (0.0084 M 2i) (0). For experimental details, see Table
. The line comes from the results of kinetic simulation with
2
Ph (0.0093 M 2d ) (b), R ) -C(Me)
2
Ph (0.0083 M 2a )
(
(
3
2
CO
2
C
2
H
5
(0.0084 M 2g) (4), or -C(CH CH C-
3
)
2
2
3 3
)
kinetic parameters provided in the Experimental Section.
weights observed for short reaction times are higher
than those expected on the basis of the monomer
conversion assuming complete utilization of the RAFT
agent (Table 5). Analysis of the latter result suggests a
Ctr for 2d of ca. 50 (if C-tr ) 0) under the reaction
conditions (by application of eq 5). However, the rate of
narrowing of the polydispersity is consistent with a
transfer coefficient that is significantly higher (Figure
-C(Me)HPh (2c), -CH2Ph (2d ). The correspondence
between found and calculated molecular weights and
the fact that very narrow polydispersities are obtained
even at low conversion (e1%) suggests that the transfer
coefficients of the RAFT agents are very high. Where
2
d is used, molecular weights are higher than calculated
for low conversions (residual 2d was also detected in
the NMR spectrum of the reaction mixtures - signal
at 4.66 ppm for CH2Ph). The slope of the plot of ln [2d ]
vs ln [MA] suggests a Ctr of ca. 105 for 2d in MA
polymerization (Figure 14) assuming C-tr is negligible.
The kinetic plot shows that the rate of polymerization
was strongly retarded compared to that observed in the
absence of RAFT agent or that observed in the presence
of other RAFT agents (e.g. benzyl dithioacetate ) but
does not depend markedly on R (Figure 15). This is
consistent with our previous findings. The rate of
polymerization reduces when the RAFT agent concen-
tration is increased (Table 5). Despite this, polymeriza-
tions do proceed to high conversions after long reaction
times (Table 6). The results suggest that retardation
under these conditions are mainly due to a property of
the polymeric RAFT agent. We have previously shown
1
0). Kinetic simulation shows that Ctr ) 400 and a C-tr
)
11 600 provides an acceptable fit to the variation of
both molecular weight and polydispersity with conver-
sion (Figure 12).
The rate of styrene polymerization with cumyl
dithiobenzoate (2a ) as RAFT agent is, under these
conditions, similar to that expected for styrene polym-
erization in the absence of a gel effectsFigure 13). In
contrast, styrene polymerization with benzyl dithioben-
zoate (2d ) appears slightly retarded (conversions are
lower for the same reaction time).
65
19
We19 along with Monteiro and de Brouwer, Barner-
62
6
3
64
Kowollik et al., and Kwak et al. have recently
reported on retardation in styrene polymerization caused
by dithiobenzoate derivatives. We reported19 that there

2
268 Chong et al.
Macromolecules, Vol. 36, No. 7, 2003
Sch em e 8
F igu r e 14. Double log plot of transfer agent concentration
vs monomer concentration for polymerization of methyl acry-
acrylamide (DMA) polymerization.32 DMA was polym-
erized in the presence of a series of four dithiobenzoate
RAFT agents. Retardation (manifest as an inhibition
period) was observed with 2a and 2d (most severe with
late ([MA]
o
) 4.45 M) in benzene at 60 °C with AIBN initiator
(0.00033 M) in the presence of PhC(dS)SCH
2
Ph [2d ] ) 0.006
0
M (0) and 0.0037 M (O). For additional experimental details
see Table 5. Lines of best fit are -162.16 + 105.13 ln[M],
correlation coefficient 0.99665 (- -), and -162.08 + 105.43,
correlation coefficient 0.99877 (- - -).
2
a ). Bimodal distributions were observed and were
pronounced in the case of benzyl dithiobenzoate (2d ).
Inefficient transfer due to poor reinitiation efficiencies
•
32
by the expelled benzyl radicals (R ) was suggested as
a possible explanation. This should not apply in MA
polymerization. The reported value of ki for benzyl
radical is high and approximately 2.5 times that for
6
7
cyanoisopropyl radical (Table 10). Even though 2d has
a lower transfer coefficient (than 2e), it is consumed
rapidly and there is little difference in the rate of MA
polymerization with 2d and 2e (Figure 15).
The possible origins of bimodal molecular weight
distributions in high conversion RAFT polymerization
6
5
are discussed elsewhere.
Or igin of Su bstitu en t Effects. The sequence of
events for a polymerization carried out in the presence
of dithiobenzoates (2) can be envisioned as follows
F igu r e 15. Kinetic plot for polymerization of methyl acrylate
(
(
[MA]
0.00033 M) in the absence (O) or presence of dithiobenzoate
Ph (0.0037 M 2d ) (0) or
CN (0.0037 M 2e) ((). For experimental details,
o
) 4.45 M) in benzene at 60 °C with AIBN initiator
1
4
(Scheme 2). Propagating radicals are generated as in
a conventional free radical polymerization. A propagat-
esters PhC(dS)SR, where R ) -CH
2
R ) -C(Me)
2
•
ing radical (Pn ) adds to the dithiobenzoate derivative
see Table 6. The lines are lines of best fit through the
experimental points.
(
2) to form an adduct (3) which can fragment to form a
•
polymeric dithiobenzoate (4) and a radical (R ). The
liberated radical (R ) then reacts with monomer to form
a new propagating radical (Pm ).
•
that in polymerization of acrylate esters, the adduct
•
radicals (3 and 5) are present in high concentration
6
6
relative the propagating radical.
Chain extension of the polymeric dithiobenzoate (4)
involves the same process. The reversible addition-
fragmentation sequence in which the SdC(Ph)S- moi-
ety is transferred between dormant and active chains,
thus maintaining the living character of the polymeri-
zation. The formation of intermediates (3 and 5) during
polymerizations of BA or styrene in the presence of
cumyl dithiobenzoate (2a ) has recently been confirmed
by electron spin resonance spectroscopy under condi-
As in the case of styrene polymerization, the appear-
ance of a high molecular weight shoulder on the mo-
lecular weight distribution with peak molecular weight
approximately twice that of the principal peak was
observed for high conversions. The high molecular
weight shoulder might, in part, be attributed to the
product formed by bimolecular termination between
polystyrene propagating radicals. Another process that
may contribute to the formation of this peak is the
reaction of adduct (5) with a small radical or a propa-
6
6
tions of high radical flux.
This mechanism of RAFT polymerization (Scheme 2)
should not be confused with the “iniferter” mechanism
•
gating radical (Q ) by combination or disproportionation
2
0,62
3
(Scheme 7).
However, the molar yield of product
first described by Otsu et al. That process relies on
from any of these processes cannot exceed the molar
concentration of radicals generated. We have found in
related polymerizations, by GPC with UV analysis, that
the high molecular weight peak retains the thiocarbo-
nylthio chromophore.65 This observation, and size of the
peak, which is substantially greater than that antici-
pated on the basis of the above considerations, suggests
that some other mechanism(s) is responsible for the
formation of the major part the high molecular weight
shoulder. An observation of bimodal molecular weight
distributions was recently reported for N,N-dimethyl-
reversible photodissociation of the C-S bond of a
dithiocarbamate derivative to provide a mechanism for
chain activation and deactivation (Scheme 8). Although
it was originally considered that reversible chain trans-
fer might occur during these polymerizations, subse-
quent work has shown that the transfer coefficient of
3
N,N-dialkyldithiocarbamates in polymerizations of meth-
1
7,36,68-70
acrylates and styrenes are very low (,1).
The
thermal dissociation of the dithiobenzoates (2) analo-
gous to Scheme 8 is not detectable under our reaction
conditions as demonstrated by control experiments that

Macromolecules, Vol. 36, No. 7, 2003
RAFT Polymerization with Thiocarbonylthio Compounds 2269
Ta ble 7. Ra te Con sta n ts for F or m a tion for Ra d ica ls
Ta ble 8. Bon d Dissocia tion En er gies for Com p ou n d s
CTUV‚
CTUV-H
reaction
radical
substituents
T,U,V
DC-H80
DC-H81
(kJ mol-1)
(kJ mol-1)
Me• tBuO
•
Me•
substituents
T,U,V
rel Ctr
in MMA abs
azo
a
b
b
c
decd
abs
addn
Me,Me,Ph
Me,H,Ph
H,H,Ph
Me,Me,CN
Me,Me,CO2Me(Et)
Me,Me,Me3CCH2
Me,Me,Me
365
378
Me,Me,Ph
Me,Me,CN
Me,Me,CO2Me(Et)
Me,H,Ph
370
170
11
3.2
5.0
1.0
0.8
2.1
1.4
1.0
48
2.4
2.1
375
400
385
388
1.0
1.0
1.0
Me,Me,Me3CCH2
H,H,Ph
Me,Me,Me
3
0.2
0.2
0.006
0.0009
0.0000009
401
0.2
0.9
0.3
1.2
Sch em e 9
a
Relative transfer coefficient of dithiobenzoate in MMA polym-
b
erization. Relative rate constant for hydrogen atom abstraction
7
7
c
from C(TUV)-H at 60 °C.
radical addition to CH2dCUV at 65 °C.
for decomposition of azo compound C(TUV)-NdN-C(TUV) at 60
Relative rate constant for methyl
7
8 d
Relative rate constant
7
9
transfer coefficient of the dithiobenzoate derivatives in
MMA polymerization are compared with the relative
rate constants of other reactions that generate radicals
°
C.
showed no polymer is formed in the absence of added
initiator.
•
CTUV. In all reactions the rate constants decrease in
On the basis of the addition-fragmentation chain
transfer mechanism (Scheme 2), at least four factors can
be seen to influence the effectiveness of the dithioben-
zoates (2) as RAFT agents: (a) the rate constant of
reaction of 2 with the propagating (or initiator-derived
radicals) (kadd); (b) the partitioning of the adduct (3)
between starting materials and products (determined
by the relative magnitude of (k-add and kâ); (c) the values
of the rate constants for fragmentation of the intermedi-
ate radicals (3 and 5); (d) the ability of the expelled
the series where the radical is C(Me)2Ar > -C(Me)HPh
> -CH2Ph and -C(Me)2CH2C(Me)3 > -C(Me)3. How-
ever, there is no general correlation between the rate
constants for the various reactions and there is also no
•
general correlation with the stability of R . This could
reflect the different steric demands of the transition
states for the various processes, a greater or lesser
influence of polar terms, and perhaps other factors.
The C-H bond dissociation energies provide one
measure of radical stability. Values for C-H bond
dissociation energies for compounds H-CTUV ()H-R)
are summarized in Table 8. For tertiary dithiobenzoates
Ph(CdS)S-R even though there is no direct correlation
there is nonetheless a trend for the transfer coefficients
to increase according to the stability of R. If the transfer
coefficients were to reflect the stability of the free
•
radicals (R ) to reinitiate polymerization.
Factors a and b should be directly reflected in the
magnitude of the transfer coefficient of 2. Factor c
should not directly affect the transfer coefficient. How-
ever, if fragmentation is particularly slow (i.e., both
k-add and kâ are small) or reinitiation of polymerization
is slow with respect to propagation, then polymerization
may be retarded and the likelihood of the radicals 3 and/
•
radicals formed (R ) it would be anticipated, for example,
that 2d (R ) -CH2Ph) or 2c (R ) -CH(CH3)Ph) would
have higher transfer coefficients than, for example,
either 2i (R ) -C(CH3)2CCH2(CH3)3) or 2g (R )
-C(CH3)2CO2Et). The experimental data indicate that,
in polymerizations of styrene and MMA, benzyl dithio-
benzoate (2d ) has a substantially lower transfer coef-
ficient.
•
or R undergoing side reactions leading to some degree
of inhibition is increased. Retardation may be observed
in RAFT polymerization with certain dithiobenzoate
derivatives when relatively high concentrations of re-
1
9
agent are used.
Since chain transfer is reversible, the rate of con-
sumption of the RAFT agent (2) and the transfer
coefficient depend on the reinitiation rate constant ki
Steric effects come into play because of the B-strain
3
2
engendered by the change from sp to sp hybridization
when the adduct (7) fragments to form the radical (8)
(Scheme 9). Steric interactions between T, U, and V
groups will be greater for a tetrahedral configuration
(in 7) than for a planar configuration (in 8) due to the
closer proximity of those groups. The strain relieved on
rehybridization should increase in the series tertiary >
secondary > primary. We also expect on this basis, that
oligomeric or polymeric radicals should be better leaving
groups than monomeric species. This offers an explana-
tion for the finding that 2g (R ) -C(CH3)2CO2Et) is a
poor RAFT agent in MMA polymerization, with Ctr ∼
2, even though it resembles the RAFT agent (2l) derived
from the MMA propagating radical for which Ctr is
approximately 2 orders of magnitude higher (∼140 at
•
and the partitioning of the radical R between adding
to monomer and reacting with the polymeric RAFT
•
agent (4). The expelled radical R may also react with
the initial RAFT agent (2). However, this process is
degenerate, involving no overall change.
We have previously reported that chain transfer
activity of reagents 1 is a function of the substituents
1
4-16,20
R and Z and the particular propagating radical.
The rate of addition to the thiocarbonyl group (CdS) of
should not be dramatically affected by the nature of
1
the group R since that group is remote from the CdS
double bond and not directly conjugated with it. Thus,
the magnitude of the transfer coefficient should reflect
the partitioning of the intermediate between starting
materials and products and the relative leaving group
ability of R and the propagating radical. We therefore
sought to establish if there was a correlation between
the transfer coefficient (in MMA polymerization) and
3
0
60 °C). Such steric factors have been previously
invoked to rationalize the transfer coefficients of meth-
acrylate macromonomers in RAFT polymerization and
provide one explanation for the very low transfer
coefficient of “MMA dimer” (Ctr ∼ 0.005 at 60 °C) in
relation to those of higher macromonomers (Ctr ∼ 0.04
for n g 3); the ratio Ctr(isobutyrate):Ctr(oligo/polymeric
•
the properties of the radical R
It has been stated that rates of free radical processes
are determined by a “complex interplay of steric, polar,
and bond strength terms”.67,71,72 In Table 7, the relative
48
species) is of similar magnitude for these two systems.

2
270 Chong et al.
Macromolecules, Vol. 36, No. 7, 2003
calculated for different Z groups).36 This is consistent
with a hypothesis that the R group does not significantly
affect the rate of addition.
Ta ble 9. HUMO a n d LUMO En er gies, P a r tia l Ch a r ges on
Su lfu r for Dith ioben zoa tes P h C(dS)S-R (In Or d er of
Decr ea sin g LUMO En er gy) a n d Rela tive Hea ts of
Rea ction for Ad d ition of Meth yl Ra d ica l
Another factor to consider is the ability of the radicals
R to reinitiate polymerization. As in conventional chain
c
∆
∆H
kJ
dSa -Sa dSb -Sb mol-1)
r
•
energy (eV)
atomic charge
(
R
HOMO LUMO
transfer, the rate of reinitiation should be equal to or
greater than the rate of propagation to avoid retarda-
tion. In RAFT polymerization, it is also necessary to
consider the rate of reinitiation in relation to the rate
Ph(Me)
Me)
Ph(Me)CH
Me
2
C
-8.45 -1.07 -0.06 0.25 0.01 0.29 -4.68
-8.46 -1.06 -0.05 0.24 0.01 0.28 -7.99
(
2
C
-8.67 -1.14 -0.05 0.26
-2.54
-8.71 -1.17 -0.03 0.27 0.03 0.32
-8.70 -1.18 -0.04 0.27 0.02 0.32
-8.71 -1.28 -0.02 0.28 0.04 0.32
-8.81 -1.38 -0.02 0.27 0.04 0.31
0
0
4.06
3.70
•
of reaction of R with the polymeric RAFT agent. If the
PhCH
CO Me)(CH
2
rate of reinitiation is small, the rate of consumption of
the initial RAFT agent will be reduced and this, in turn,
may cause the transfer coefficient to be dependent on
RAFT agent concentration and conversion. The transfer
coefficients of cumyl dithiobenzoate (2a ) and 2-cyano-
prop-2-yl dithiobenzoate (2e) in MMA polymerization
depend on the RAFT agent concentration (see above).
The data show that while cumyl dithiobenzoate (2a ) has
a higher transfer coefficient, for high concentrations of
RAFT agent (when making low molecular weight poly-
mers) the 2-cyanoprop-2-yl derivative (2e) is the more
effective RAFT agent. This is attributable to the very
high reactivity of cumyl radicals toward the RAFT
agent. For high concentrations of 2a (greater than the
concentrations used in the present experiments) sub-
(
2
3
)
2
C
CN(Me)2C
a
Milliken charge. b Atomic charge. c Relative heat of reaction
)
∆Hf (2) - ∆Hf (3) (referenced to R ) Me).
Such steric factors also help us to understand why tert-
butyl dithiobenzoate (2j) and benzyl dithiobenzoate (2d )
have significantly lower transfer coefficients than iso-
octyl dithiobenzoate (2i). We can also note that 2f (R )
-
C(CH3)(CH2CH2CH2OH)CN is more effective than 2e
R ) -C(CH3)2CN). Of the compounds studied, 2f is the
(
most effective RAFT agent in MMA polymerization
under our experimental conditions (based on the poly-
dispersity at a given conversion).
We indicated that for a series of dithiobenzoates 2,
variation in the reactivity of the CdS toward free radical
addition should be small and the magnitude of the
transfer coefficient should be dependent primarily on
6
1,73
stantial retardation may be observed in each MMA,
styrene,
1
9,62,63
19
and BA, polymerization. For MMA and
styrene polymerization with 2-cyanoprop-2-yl dithioben-
•
zoate (2e) as the RAFT agent, this retardation is absent
the leaving group ability of the radical R . To provide a
1
9,61,73
or much diminished.
qualitative assessment of the effect of the group R on
the rate constant for addition of radicals to dithioben-
zoates 2, we calculated the LUMO and HOMO energies
of a series of dithiobenzoates using semiempirical mo-
lecular orbital calculations. These data are provided in
Table 9. The rate constant for addition of a given radical
should be highest for the compound with the lowest
Literature data suggest no marked difference in ki in
•
74 •
75
MMA polymerization with C(CH3)2CN, C(CH3)2Ph,
•
75
and CH2Ph as the initiating species (see Table 10).
The rate constants ki for C(CH3)3 and other alkyl
•
radicals are several orders of magnitude higher. The
rate constants for the reaction of these radicals with the
polymeric RAFT agents are unknown. However, the
similarity in ki values suggests that the different
behaviors of 2-cyanoprop-2-yl and cumyl dithiobenzoates
3
6
LUMO energy. For RAFT agents with different Z
substituents and common R, there is a reasonable
correlation between ln(Ctr) and the LUMO energies
estimated using semiempirical calculations or higher
level ab initio calculations.36 The calculations suggest
that the rate of addition to the CdS is favored by alkyl
or phenyl substitution of the R-carbon but disfavored
by polar groups (nitrile or ester). The LUMO energy for
the 2-cyanoprop-2-yl dithiobenzoate (2e) is lower than
that of cumyl dithiobenzoate (2a ) and on this basis
should be less reactive toward addition. The lower Ctr
for 2e vs. 2a may in part be due to a lower kadd.
However, the differences in LUMO (and HOMO) ener-
gies are small with respect to the error inherent in the
calculations (and when considered in relation to those
•
are due to the rate constant (k-â) for C(CH3)2CN being
•
lower than that for C(CH ) Ph. Our molecular orbital
3 2
calculations provide some support for this hypothesis.
Releva n ce in Block Cop olym er Syn th esis. In a
1
5
communication, we reported that it is critical when
making styrene-MMA block copolymers to prepare the
poly(methyl methacrylate) block first. Attempted block
copolymer syntheses involving polymerization of MMA
in the presence of ω-(thiobenzoylthio)polystyrene (2m )
in a batch polymerization were unsuccessful. Thus, an
experiment using conditions similar to those used for
the other MMA polymerizations mentioned abovessee
Ta ble 10. Absolu te Ra te Con sta n ts for Ra d ica l Ad d ition to Va r iou s Mon om er s
rate constant (M- s-1
1
)
•C(CH3)375,82
•C(CH3)2CO2Me83
•C(CH3)2CN74
(315 K)
•C(CH3)2Ph75
(296 K)
•CH2Ph75
(296 K)
monomer
MMA
(300 K)
(294 K)
a
b
c
a
b
c
a
b
c
660 000
3170
-
1590
-
3291
367
-
865
2410
4041
4896
2700
10 225
7829
800
2100
-
8348
430
2134
2042
1200
4557
4685
952 970
110 000
9722
1150
3129
3412
5500
12 508
13 455
MA
-
1523800
132 000
226 622
224 850
2650
1100
3803
3802
styrene
a
Value provided in original paper for temperature indicated. b Value at 60 °C calculated using Arrhenius parameters provided in the
c
reference indicated (- indicates no Arrhenius parameters given). Value at 60 °C calculated using log A 7.5 (tertiary radicals) or log A 8.5
benzyl radical) and activation energy as reported.
6
7
(

Macromolecules, Vol. 36, No. 7, 2003
RAFT Polymerization with Thiocarbonylthio Compounds 2271
•
Sch em e 10
the relative leaving group ability of R and the propa-
gating radical. Steric factors, radical stability, and polar
factors all appear important in determining the leaving
•
group ability of R (more stable, more electrophilic, more
bulky radicals are better leaving groups). The partition-
•
ing of R between adding to monomer (to reinitiate) and
adding to polymeric RAFT agent can also have a
significant effect on the rate of consumption of RAFT
agent. The propensity of benzylic radicals to add RAFT
agent rather than monomer may explain why certain
reagents (e.g., cumyl dithiobenzoate 2a ) are less effec-
tive RAFT agents when used in high concentrations to
make lower molecular weight polymers.
Note Ad d ed in P r oof. Two papers on retardation
in acrylate polymerization using dithiobenzoate esters
8
4
Experimental Sectionsprovided poly(methyl methacry-
late) with molecular weight and polydispersity values
similar to those obtained in the absence of added 2m .
Other block copolymers where the B-block is com-
posed of a monosubstituted acrylic or styrenic monomer
have recently appeared.
Ack n ow led gm en t. We are grateful to Drs. C. Berge,
M. Fryd, and R. Matheson of DuPont Performance
Coatings for their support of this work and for valuable
discussion.
(e.g., N,N-dimethylacrylamide or 4-methylstyrene) were
successfully prepared from this same ω-(thiobenzoyl-
thio)polystyrene in a batch polymerization process
Refer en ces a n d Notes
(
Scheme 10).15 For details, see the Experimental Sec-
(
(
1) Matyjaszewski, K. ACS Symp. Ser. 2000, 768, 2-26.
2) Matyjaszewski, K. ACS Symp. Ser. 1998, 685, 2-30.
tion.
Polymerization of styrene in the presence of ω-(thioben-
zoylthio)poly(methyl methacrylate) (2l) ( Mh n 17408, Mh / Mh n
(3) Otsu, T.; Matsumoto, A. Adv. Polym. Sci. 1998, 136, 75-137.
(4) Moad, G.; Solomon, D. H. The Chemistry of Free Radical
Polymerization Pergamon: Oxford, England, 1995.
w
1
.20) was successful in giving a narrow polydispersity
(
5) Hawker, C. J .; Bosman, A. W.; Harth, E. Chem. Rev. 2001,
block copolymer Mh n 35 000, Mh w/ Mh n 1.24) with complete
1
01, 3661-88.
(
>95%) incorporation of the precursor polymer.
It appears that the polystyrene propagating species
-Pn ) -[CH2-CHPh]n in Scheme 2), like the 1-phen-
(
6) Matyjaszewski, K.; Xia, J . Chem. Rev. 2001, 101, 2921-90.
(7) Kamigaito, M.; Ando, T.; Sawamoto, M. Chem. Rev. 2001, 101,
3689-745.
(
(
8) Quirk, R. P.; Lee, B. Polym. Int. 1992, 27, 359-67.
9) Darling, T. R.; Davis, T. P.; Fryd, M.; Gridnev, A. A.;
Haddleton, D. M.; Ittel, S. D.; Matheson, R. R., J r.; Moad,
G.; Rizzardo, E. J . Polym. Sci., Part A: Polym. Chem. 2000,
38, 1709.
ylethyl radical, is a very poor free radical leaving group
with respect to the poly(methyl methacrylate) propagat-
ing species (-Pn ) -[CH2-C(Me)(CO2Me)]n in Scheme
(
2
). Thus, for synthesis of poly(methyl methacrylate)-
block-polystyrene under batch polymerization condi-
tions, it is necessary to make the poly(methyl meth-
acrylate) block first in order to make a narrow polydis-
persity block copolymer.
(10) The use of the adjective “living” to describe a radical polym-
erization which displays the attributes of a living polymer-
ization is preferred over other terms that have been advo-
cated in this context (e.g. “controlled”, “living/controlled”,
“pseudoliving”, “living polymerization with reversible deac-
tivation”) even though some side reactions (radical-radical
termination) must inevitably complicate the process.
We anticipate that since the transfer coefficient of
ω-(thiobenzoylthio)polystyrene (2m ) should be g0.15
(
11) Le, T. P.; Moad, G.; Rizzardo, E.; Thang, S. H. PCT Int. Appl.
(transfer coefficient of phenylethyl dithiobenzoate 2c),
WO 98/01478 (Chem. Abstr. 1998, 128, 115390).
it should be possible to prepare block copolymers from
this starting material using an appropriate feed addition
protocol. We have successfully prepared poly(methyl
methacrylate)-block-polystyrene from a ω-(thioacetyl-
thio)polystyrene by a feed emulsion polymerization
where the concentration of RAFT agent with respect to
that of monomer is maintained at a low (approximately
(
12) Krstina, J .; Moad, G.; Rizzardo, E.; Winzor, C. L.; Berge, C.
T.; Fryd, M. Macromolecules 1995, 28, 5381-5.
(13) Krstina, J .; Moad, C. L.; Moad, G.; Rizzardo, E.; Berge, C.
T.; Fryd, M. Macromol. Symp. 1996, 111, 13-23.
(
14) Chiefari, J .; Chong, Y. K.; Ercole, F.; Krstina, J .; J effery, J .;
Le, T. P. T.; Mayadunne, R. T. A.; Meijs, G. F.; Moad, C. L.;
Moad, G.; Rizzardo, E.; Thang, S. H. Macromolecules 1998,
31, 5559-62.
1
9
(15) Chong, B. Y. K.; Le, T. P. T.; Moad, G.; Rizzardo, E.; Thang,
constant) level. Feed emulsion polymerization has also
been successful in other circumstances when the trans-
fer coefficient of the RAFT agent is low. For example,
narrow polydispersity methacrylate block copolymer can
be prepared from methacrylate macromonomers (the
transfer coefficients of methacrylate macromonomers in
S. H. Macromolecules 1999, 32, 2071-4.
(
16) Rizzardo, E.; Chiefari, J .; Chong, Y. K.; Ercole, F.; Krstina,
J .; J effery, J .; Le, T. P. T.; Mayadunne, R. T. A.; Meijs, G. F.;
Moad, C. L.; Moad, G.; Thang, S. H. Macromol. Symp. 1999,
143, 291-307.
(
17) Mayadunne, R. T. A.; Rizzardo, E.; Chiefari, J .; Moad, G.;
Thang, S. H. Macromolecules 1999, 32, 6977-80.
18) Mayadunne, R. T. A.; Rizzardo, E.; Chiefari, J .; Krstina, J .;
Moad, G.; Postma, A.; Thang, S. H. Macromolecules 2000,
1
2,13
MMA polymerization are <0.5).
(
Con clu sion s
3
3, 243-5.
(19) Moad, G.; Chiefari, J .; Krstina, J .; Postma, A.; Mayadunne,
In MMA polymerization, the ability of the RAFT
agents [SdC(Ph)S-R] to promote living radical polym-
erization depends strongly on the nature of R. In
polymerizations of styrene and BA, a much wider range
of R groups is available. A major factor determining the
transfer coefficient of dithiobenzoate derivatives is the
way the intermediate 3 partitions between starting
materials and products which is in turn determined by
R. T. A.; Rizzardo, E.; Thang, S. H. Polym. Int. 2000, 49, 933-
1
001.
(20) Rizzardo, E.; Chiefari, J .; Mayadunne, R. T. A.; Moad, G.;
Thang, S. H. ACS Symp. Ser. 2000, 768, 278-96.
(
21) Mitsukami, Y.; Donovan, M. S.; Lowe, A. B.; McCormick, C.
L. Macromolecules 2001, 34, 2248-56.
(
22) De Brouwer, H.; Schellekens, M. A. J .; Klumperman, B.;
Monteiro, M. J .; German, A. L. J . Polym. Sci., Part A: Polym.
Chem. 2000, 38, 3596-603.

2
272 Chong et al.
Macromolecules, Vol. 36, No. 7, 2003
(
(
(
23) Monteiro, M. J .; Sjoberg, M.; van der Vlist, J .; Gottgens, C.
M. J . Polym. Sci., Part A: Polym. Chem. 2000, 38, 4206-17.
24) Stenzel-Rosenbaum, M.; Davis, T. P.; Chen, V.; Fane, A. G.
J . Polym. Sci., Part A: Polym. Chem., 2001, 39, 2777-83.
25) Tsujii, T.; Ejaz, M.; Sato, K.; Goto, A.; Fukuda, T. Macro-
molecules 2001, 34, 8872-8.
(54) Moad, G.; Rizzardo, E.; Solomon, D. H.; J ohns, S. R.; Willing,
R. I. Makromol. Chem., Rapid Commun. 1984, 5, 793-8.
(55) Mayer, R.; Scheithauer, S. In Houben-Weyl Methods of
Organic Chemistry; Buechel, K. H., Falbe, J ., Hagemann, H.,
Hanack, M. Eds.; Thieme: Stuttgart, Germany, 1985; Vol.
E, pp 891-930.
(
(
(
(
(
(
56) Sudalai, A.; Kanagasabapathy, S.; Benicewicz, B. C. Org. Lett.
000, 2, 3213-6.
57) Oae, S.; Yagihara, T.; Okabe, T. Tetrahedron 1972, 28, 3203-
6.
58) Becke, F.; Hagen, H. DE 1,274,121, 1968 (Chem. Abstr. 1969,
0, 3573v).
(
(
26) Baum, M.; Brittain, W. J . Macromolecules 2002, 35, 610-5.
27) Barner-Kowollik, C.; Quinn, J . F.; Nguyen, T. L. U.; Heuts,
J . P. A.; Davis, T. P. Macromolecules 2001, 34, 7849-57.
28) Destarac, M.; Charmot, D.; Franck, X.; Zard, S. Z. Macromol.
Rapid Commun. 2000, 21, 1035-9.
2
1
(
(
(
(
(
(
(
7
29) Ladaviere, C.; Doerr, N.; Claverie, J . P. Macromolecules,
59) Thang, S. H.; Chong, Y. K.; Mayadunne, R. T. A.; Moad, G.;
Rizzardo, E. Tetrahedron Lett. 1999, 40, 2435-8.
2
001, 34, 5370-2.
30) Goto, A.; Sato, K.; Tsujii, Y.; Fukuda, T.; Moad, G.; Rizzardo,
E.; Thang, S. H. Macromolecules 2001, 34, 402-8.
31) Severac, R.; Lacroix-Desmazes, P.; Boutevin, B. Polym. Int.
60) Bouhadir, G.; Legrand, N.; Quiclet-Sire, B.; Zard, S. Z.
Tetrahedron Lett. 1999, 40, 277-80.
61) Moad, G.; Chiefari, J .; Mayadunne, R. T. A.; Moad, C. L.;
Postma, A.; Rizzardo, E.; Thang, S. H. Macromol. Symp.
2
002, 51, 1117-22.
32) Donovan, M. S.; Lowe, A. B.; Sumerlin, B. S.; McCormick, C.
L. Macromolecules 2002, 35, 4123-32.
33) Donovan, M. S.; Lowe, A. B.; Sumerlin, B. S.; McCormick, C.
L. Macromolecules 2001, 34, 6561-4.
34) Donovan, M. S.; Sanford, T. A.; Lowe, A. B.; Sumerlin, B. S.;
Mitsukami, Y.; McCormick, C. L. Macromolecules 2002, 35,
2
002, 182, 65-80.
(
62) Monteiro, M. J .; de Brouwer, H. Macromolecules 2001, 34,
349-52.
(63) Barner-Kowollik, C.; Quinn, J . F.; Morsley, D. R.; Davis, T.
P. J . Polym. Sci., Part A: Polym. Chem. 2001, 39, 1353-65.
(64) Kwak, Y.; Goto, A.; Tsujii, Y.; Murata, Y.; Komatsu, K.;
Fukuda, T. Macromolecules 2002, 38, 3026-9.
4
570-2.
(
(
(
35) Some results mentioned in this paper have been published
previously as part of conference papers (refs 19, 47) and in a
patent (ref 11).
(65) Moad, G.; Mayadunne, R. T. A.; Rizzardo, E.; Skidmore, M.
A.; Thang, S. H. ACS Symp. Ser. 2003, in press.
(66) Hawthorne, D. G.; Moad, G.; Rizzardo, E.; Thang, S. H.
Macromolecules 1999, 5457-9.
36) Chiefari, J .; Mayadunne, R. T. A.; Moad, C. L.; Moad, G.;
Postma, A.; Rizzardo, E.; Postma, A., Skidmore, M. A.; Thang,
S. H. Macromolecules 2003, 36, 2273.
(
(
(
(
67) Fischer, H.; L, R. Angew. Chem., Int. Ed. 2001, 40, 1340-
7
1.
68) Manga, J . D.; Polton, A.; Tardi, M.; Sigwalt, P. Polym. Int.
37) Meijer, J .; Vermeer, P. J . R. Chem. Soc. Pays-Bas 1973, 92,
1
998, 45, 243-54.
6
01.
69) Otsu, T.; Matsunaga, T.; Doi, T.; Matumoto, A. Eur. Polym.
J . 1995, 31, 67-78.
(
(
(
38) Wood; Bost J . Am. Chem. Soc. 1937, 59, 1011.
39) Reed, S. F. J . Polym. Sci., Part A-1 1971, 9, 2029-38.
40) Aycock, D. F.; J urch, G. R., J r. J . Org. Chem. 1979, 44, 569-
70) Zaremskii, M. Y.; Olenin, A. V.; Garina, Y. S.; Kuchanov, S.
I.; Golubev, V. B.; Kabanov, V. A. Polym. Sci. 1991, 33, 2167-
7
2.
7
5.
(
(
(
41) Moad, G.; Ercole, F.; J ohnson, C. H.; Krstina, J .; Moad, C.
L.; Rizzardo, E.; Spurling, T. H.; Thang, S. H.; Anderson, A.
G. ACS Symp. Ser 1998, 685, 332.
42) The simulation results are not sensitive to the absolute value
of Ctr or C-tr and their magnitudes may differ significantly
from that shown.
(71) Tedder, J . M. Angew. Chem., Int. Ed. Engl. 1982, 21, 401.
(72) Tedder, J . M.; Walton, J . C. Tetrahedron 1980, 36, 701.
(73) Chiefari, J .; Mayadunne, R.; Moad, G.; Postma, A.; Rizzardo,
E.; Thang, S. 24th Aust. Polym. Symp. 2001, C6/3.
(
(
(
74) Heberger, K.; Fischer, H. Int. J . Chem. Kinet. 1993, 25, 249-
6
3.
43) Buback, M.; Gilbert, R. G.; Hutchinson, R. A.; Klumperman,
B.; Kuchta, F.-D.; Manders, B. G.; O’Driscoll, K. F.; Russell,
G. T.; Schweer, J . Macromol. Chem. Phys. 1995, 196, 3267-
75) Walbiner, M.; Wu, J . Q.; Fischer, H. Helv. Chim. Acta 1995,
7
8, 910-24.
76) Hutchinson, R. A.; Paquet, D. A., J r.; McMinn, J . H.
8
0.
DECHEMA Monogr. 1995, 131, 467-92.
(
44) Buback, M.; Kowollik, C.; Kurz, C.; Wahl, A. Macromol.
(77) Hendry, D. G.; Mill, T.; Piszkiewicz, L.; Howard, J . A.;
Eigenmann, H. K. J . Phys. Chem. Ref. Data 1974, 3, 937-
78.
Chem. Phys. 2000, 201, 464-9.
(
(
(
45) Moad, G.; Rizzardo, E. Macromolecules 1995, 28, 8722-8.
46) Mayo, F. R. J . Am. Chem. Soc. 1943, 65, 2324.
47) Moad, C. L.; Moad, G.; Postma, A.; Mayadunne, R. T. A.;
Rizzardo, E.; Thang, S. H. Macromolecules 2003, to be
submitted.
(78) Herk, L.; Stefani, A.; Szwarc, M. J . Am. Chem. Soc. 1961,
8
3, 3008-11.
(
(
79) Engel, P. S. Chem. Rev. 1980, 80, 99-150.
80) Brocks, J . J .; Beckhaus, H.-D.; Beckwith, A. L. J .; R u¨ chardt,
C. J . Org. Chem. 1998, 63, 1935-43.
(
48) Moad, G.; Moad, C. L.; Rizzardo, E.; Thang, S. H. Macromol-
ecules 1996, 29, 7717-26.
(
81) Tsang, W. In Energetics of Organic Free Radicals; Simoes,
J . A. M., Greenberg, A., Liebman, J . F., Eds.; Blackie:
London, 1996; pp 22-58.
(
49) Bamford, C. H.; Barb, W. G.; J enkins, A. D.; Onyon, P. F.
The Kinetics of Vinyl Polymerization by Radical Mechanisms
Butterworth: London, 1958.
(
(
(
82) M u¨ nger, K.; Fischer, H. Int. J . Chem. Kinet. 1985, 17, 809-
2
9.
(
(
(
(
50) M u¨ ller, A. H. E.; Zhuang, R.; Yan, D.; Litvenko, G. Macro-
83) Zytowski, T.; Knuehl, B.; Fischer, H. Helv. Chim. Acta 2000,
3, 658-75.
molecules 1995, 28, 4326-33.
8
51) M u¨ ller, A. H. E.; Litvenko, G. Macromolecules 1997, 30,
84) See: Vana, P.; Davis, T. P.; Barner-Kowollik, C. Macromol.
Theory Simul. 2002, 11, 823-835. Perrier, S.; Barner-
Kowollik, C.; Quinn, J . F.; Vana, P.; Davis, T. P. Macromol-
ecules 2002, 35, 8300-8306.
1
253-66.
52) Tsavalas, J . G.; Schork, F. J .; de Brouwer, H.; Monteiro, M.
J . Macromolecules, 2001, 34, 3938-46.
53) Monteiro, M. J .; de Barbeyrac, J . Macromolecules, 2001, 34,
4
416-23.
MA020882H