Thiocarbonylthio compounds (S=C(Z)S-R) in free radical polymerization with reversible addition-fragmentation chain transfer (RAFT polymerization). Effect of the activating Group Z

Article abstract of DOI:10.1021/ma020883+

Free-radical polymerization in the presence of suitable addition-fragmentation chain transfer agents [S=C(Z)S-R] (RAFT agents) possess the characteristics of a living polymerization (i.e., polymer products can be reactivated for chain extension and/or block synthesis, molecular weights are predetermined by RAFT agent concentration and conversion, narrow polydispersities are possible). Styrene polymerizations (110°C, thermal initiation) were performed for two series of RAFT agents [S=C(Z)S-CH₂Ph and S=C(Z)S-C(Me)₂CN]. The chain transfer coefficients decrease in the series where Z is Ph > SCH₂Ph ～ SMe ～ Me ～ N-pyrrolo ? OC₆F₅ > N-lactam > OC₆H₅ > O(alkyl) ? N(alkyl)₂ (only the first five in this series provide narrow polydispersity polystyrene (< 1.2) in batch polymerization). More generally, chain transfer coefficients decrease in the series dithiobenzoates > trithiocarbonates ～ dithioalkanoates > dithiocarbonates (xanthates) > dithiocarbamates. However, electron-withdrawing substituents on Z can enhance the activity of RAFT agents to modify the above order. Thus, substituents that render the oxygen or nitrogen lone pair less available for delocalization with the C=S can substantially enhance the effectiveness of xanthates or dithiocarbamates, respectively. The trend in relative effectiveness of the RAFT agents is rationalized in terms of interaction of Z with the C=S double bond to activate or deactivate that group toward free radical addition. Molecular orbital calculations and the estimated LUMO energies of the RAFT agents can be used in a qualitative manner to predict the effect of the Z substituent on the activity of RAFT agents.

Full text of DOI:10.1021/ma020883+

Macromolecules 2003, 36, 2273-2283
2273
Thiocarbonylthio Compounds (SdC(Z)S-R) in Free Radical
Polymerization with Reversible Addition-Fragmentation Chain Transfer
(RAFT Polymerization). Effect of the Activating Group Z
J oh n Ch iefa r i, Rosh a n T. A. Ma ya d u n n e, Ca th er in e L. Moa d , Gr a em e Moa d ,*
Ezio Rizza r d o,* Alm a r P ostm a , Melissa A. Sk id m or e, 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: Free-radical polymerization in the presence of suitable addition-fragmentation chain
transfer agents [SdC(Z)S-R] (RAFT agents) possess the characteristics of a living polymerization (i.e.,
polymer products can be reactivated for chain extension and/or block synthesis, molecular weights are
predetermined by RAFT agent concentration and conversion, narrow polydispersities are possible). Styrene
polymerizations (110 °C, thermal initiation) were performed for two series of RAFT agents [SdC(Z)S-
CH2Ph and SdC(Z)S-C(Me)2CN]. The chain transfer coefficients decrease in the series where Z is Ph >
SCH2Ph ∼ SMe ∼ Me ∼ N-pyrrolo . OC6F5 > N-lactam > OC6H5 > O(alkyl) . N(alkyl)2 (only the first
five in this series provide narrow polydispersity polystyrene (< 1.2) in batch polymerization). More
generally, chain transfer coefficients decrease in the series dithiobenzoates > trithiocarbonates ∼
dithioalkanoates > dithiocarbonates (xanthates) > dithiocarbamates. However, electron-withdrawing
substituents on Z can enhance the activity of RAFT agents to modify the above order. Thus, substituents
that render the oxygen or nitrogen lone pair less available for delocalization with the CdS can substantially
enhance the effectiveness of xanthates or dithiocarbamates, respectively. The trend in relative effectiveness
of the RAFT agents is rationalized in terms of interaction of Z with the CdS double bond to activate or
deactivate that group toward free radical addition. Molecular orbital calculations and the estimated LUMO
energies of the RAFT agents can be used in a qualitative manner to predict the effect of the Z substituent
on the activity of RAFT agents.
Sch em e 1
In tr od u ction
In recent communications, we have demonstrated
that free radical polymerization in the presence of
reagents that give reversible addition-fragmentation
chain transfer (Scheme 1) can be used to produce narrow
polydispersity polymers. These polymers can be chain
extended to form block, star, or other polymers of
complex architecture.1-9 The polymerizations (desig-
nated RAFT polymerizations) have the characteristics
usually associated with living polymerization. However,
they appear considerably more versatile than other
processes described as “living” or “controlled” free radi-
cal polymerization in that they are compatible with a
wider range of monomers and reaction conditions. The
most effective RAFT agents are certain thiocarbonylthio
compounds (1).3-9
We3-10 and others11-14 have reported that, for polym-
erization with thiocarbonylthio compounds (SdC(Z)S-
R), the polydispersity and the degree of molecular
weight control obtained under a particular set of reac-
tion conditions depend on the nature of the groups Z
and R. R is a homolytic leaving group and the radical
R• must efficiently reinitiate polymerization to give
chain transfer. The influence of the R substituent on
the effectiveness of dithiobenzoate derivatives (1, Z )
Ph) as RAFT agents is detailed in a companion publica-
tion.15 Z is a group that modifies the reactivity of the
thiocarbonylthio compound and of the derived adduct
radical. In this paper, we examine the effect of the Z
group on the activity of thiocarbonylthio compounds (1)
in promoting living polymerization of styrene and draw
some general conclusions of the relative effectiveness
of various RAFT agents in promoting living radical
polymerization.16
Exp er im en ta l Section
Gen er a l Da ta . Styrene (Aldrich, 99+%) was purified by
filtration through alumina (to remove inhibitors) fractionated
under reduced pressure and flash distilled immediately prior
to use. Azobis(isobutyronitrile) (AIBN) was obtained from
Tokyo Kasei and recrystallized twice from chloroform-
methanol. Tetrahydrofuran (THF) used in synthesis was
freshly distilled from sodium benzophenone. Solvents used for
column chromatography were of AR grade and were distilled.
* To whom correspondence should be addressed. E-mail
addresses:
graeme.moad@csiro.au;
ezio.rizzardo@csiro.au;
san.thang@csiro.au.
10.1021/ma020883+ CCC: $25.00 © 2003 American Chemical Society
Published on Web 03/15/2003

2274 Chiefari et al.
Macromolecules, Vol. 36, No. 7, 2003
the solvent removed. The crude product was chromatographed
on silica eluting with 5% ethyl acetate in petroleum spirits to
provide benzyl 1-pyrrolecarbodithioate as a yellow oil (2.34 g,
50% yield). 1H NMR, δ: 4.60 (s, 2H, CH2Ph), 6.30 (m, 2H,
pyrrole-H3), 7.40 (m, 5H, CH2Ph), 7.70 (m, 2H, pyrrole-H2).
13C NMR, δ: 41.7 (CH2Ph), 114.2 (pyrrole-C3), 120.6 (pyrrole-
C2), 128.0 (phenyl-C2), 128.8 (phenyl-C4), 129.4 (phenyl-C3),
135.0 (phenyl-C1), 189.0 (CdS).
O-P en taflu or oph en yl S-Ben zyl Xan th ate (2e, Z ) C6F5).
Thiophosgene (1.93 g, 0.017 mol) in CHCl3 (10 mL) at 0 °C
was treated dropwise with pentafluorophenol (3.13 g, 0.017
mol) in 5% NaOH (15 mL) cooled to 0-10 °C. The solution
was stirred for 1 h at the same temperature, the CHCl3 layer
separated and washed with 5% NaOH (10 mL), 5% HCl (10
mL), and H2O (10 mL). The organic layer was dried with
MgSO4 and filtered and the solvent removed to give the
perfluorophenyl chlorothioformate (3.76 g).
Benzyl mercaptan (1.24 g, 0.01 mol) was added to 0.8 g of
NaOH (0.02 mol) dissolved in 20 mL of H2O and allowed to
stir for 10 min. The crude chlorothioformate (2.63 g, 0.01 mol)
was added to the solution and stirred for 2 h. The aqueous
solution was extracted with diethyl ether (3 × 30 mL), organic
portions were combined, dried with Na2SO4, and filtered, and
the solvent wasremoved. The residue was purified by chro-
matography on silica with 2% ethyl acetate in petroleum
spirits as the eluent to afford O-pentafluorophenyl S-benzyl
Petroleum spirits refers to the fraction with bp 40-60 °C. The
silica was Kieselgel-60 (Merck), 70-230 mesh. Reagent chemi-
cals were obtained from Aldrich and were used without further
purification unless indicated otherwise. Nuclear magnetic
resonance (NMR) spectra were obtained with a Bruker Avance
DRX500 or a Bruker AC200 spectrometer on samples dissolved
in deuteriochloroform (CDCl3). Chemical shifts are reported
in ppm from tetramethylsilane. High-resolution chemical
ionization mass spectra were obtained with a J EOL J MS
DX303 spectrometer with methane as reagent gas. Gel per-
meation chromatography (GPC) was performed on a Waters
Associates liquid chromatograph equipped with differential
refractometer and a set of Ultrastyragel columns (106, 105, 104,
103, 500, and 100 Å) at 22 °C. Tetrahydrofuran (flow rate of
1.0 mL/min) was used as eluent. The columns were calibrated
with narrow polydispersity polystyrene standards (Polymer
Laboratories). Conversions reported in Tables 1-3 were
determined gravimetrically.
1
xanthate (0.89 g, 25% yield). H NMR, δ: 4.5 (s, 2H, CH2Ph),
7.3 (m, 5H, ArH). 13C NMR, δ: 42.9 (CH2Ph), 128.3 (phenyl-
C4), 128.9 (phenyl-C2*), 129.3 (phenyl-C3*), 129.3 (C6F5-C1),
133.9 (phenyl-C1), 138.2 (C6F5-C3, J 1
252 Hz), 140.1 (C6F5-
247 Hz), 210.9 (CdS)
C-F
C4, J 1
252 Hz), 141.2 (C6F5-C2, J 1
C-F
C-F
(* assignments tentative, may be reversed). 19F NMR (CDCl3),
δ: -162.5 (m, 2F, ortho-F), -156.9 (t J ) 22 Hz, 1F, para-F),
-151.5 (m, 2F, meta-F). Mass spectrum: found 350.9917 (M
+ 1); C14H7S2OF5 requires 350.9937.
Ben zyl 2-P yr r olid in on e-1-ca r bod ith ioa te (2f). Benzyl
chloride (0.8 g, 0.0064 mol) was added to a suspension solution
of 2-pyrrolidone-1-carbodithioic acid19 (0.97 g, 0.006 mol) and
potassium carbonate (0.84 g, 0.0067 mol) in absolute ethanol
(10 mL) at room temperature, and the resulting mixture was
stirred at room temperature for 3 h. Water (25 mL) was then
added, and the mixture was extracted with ethyl acetate (3 ×
20 mL). The combined organic extracts were dried over
anhydrous sodium sulfate. The solvent was evaporated under
reduced pressure and the residue subjected to column chro-
matography on silica eluting sequentially with n-hexane (to
give unreacted benzyl chloride) and with ethyl acetate/n-
hexane (3:7) to give benzyl 2-pyrrolidinone-1-carbodithioate
(1.1 g, 73% yield) as a bright yellow solid, mp 57-58 °C. 1H
NMR, δ: 2.11 (ddt, 2H, -CH2CH2CH2-), 2.73 (t, 2H, -CH2C-
(dO)-), 4.25 (dd, 2H, -CH2-N<), 4.40 (s, 2H, CH2Ph), 7.20-
7.40 (m, 5H, ArH). Mass spectrum: found 252.0510 (M + 1);
Ben zyl Dith ioben zoa te (2a , Z ) P h ). The preparation of
this compound is described elsewhere.15
Diben zyl Tr ith ioca r bon a te (2b, Z ) SCH2P h ). The
compound was prepared according to the procedure of Leung
et al.17
Ben zyl Dith ioa ceta te (2c, Z ) Me). Methylmagnesium
chloride (10 mL, 0.03 mol, 3M solution in THF, Aldrich) was
diluted with THF (10 mL) and the resulting solution warmed
to 40 °C. Carbon disulfide (2.28 g, 0.03 mol) was added over
10 min while maintaining the reaction temperature at 40 °C.
The reaction was cooled to room temperature before adding
benzyl bromide (5.1 g, 0.03 mol) over 15 min. The reaction
temperature was increased to 50 °C and maintained for a
further 45 min. Water (100 mL) was added and the organic
products extracted with n-hexane (3 × 60 mL). The combined
organic extracts were washed with water and saturated brine
and dried over anhydrous magnesium sulfate. After removal
of solvent and column chromatography on silica with 5%
diethyl ether in n-hexane as eluent, pure benzyl dithioacetate
was obtained as a golden oil (3 g, 55% yield). 1H NMR, δ: 2.90
(s, 3H, CH3), 4.46 (s, 2H, CH2Ph), 7.31 (m, 5H, ArH). Mass
spectrum: found 183.0291 (M + 1); C9H10S2 requires 183.0302.
Ben zyl 1-P yr r oleca r bod ith ioa te (2d , Z ) N-P yr r olo).18
Pyrrole (1.34 g, 0.02 mol) was added dropwise to a stirred
suspension of sodium hydride (0.48 g, 0.02 mol) in dimethyl
sulfoxide (20 mL). On completion of addition, the resulting
brown solution was stirred at room temperature for 30 min
before the addition of carbon disulfide (1.52 g, 0.02 mol). The
solution was allowed to stir at room temperature for a further
half-hour and benzyl chloride (2.53 g, 0.02 mol) added. Water
(20 mL) was added after 1 h, followed by diethyl ether (20 mL).
The organic layer was separated and the aqueous layer
extracted with diethyl ether (2 × 20 mL). The combined
extracts were dried with magnesium sulfate and filtered and
C
12H13S2ON requires 252.0517.
O-P h en yl S-Ben zyl Xa n th a te (2 g, Z ) P h ).20 Benzyl
mercaptan (1.24 g, 0.01 mol) was added to an aqueous (20 mL)
solution of NaOH (0.8 g) at room temperature and stirred for
15 min. Phenyl thionochloroformate (2.07 g, 0.012 mol) was
next added dropwise to this solution at the same temperature
and stirred for a further 2 h. Diethyl ether (20 mL) and water
(50 mL) were added and the organic layer separated. The
aqueous layer was extracted with diethyl ether (3 × 20 mL).
The combined organic fractions were dried with Na2SO4 and
filtered, the solvent was removed under vacuum, and the crude
product was chromatographed on silica with 2% ethyl acetate
in petroleum spirits as eluent to afford O-phenyl S-benzyl
1
xanthate (1.95 g, 75% yield) as a yellow oil. H NMR, δ: 4.43
(s, 2H, CH2Ph), 7.10-7.50 (m, 10H, ArH). 13C NMR, δ: 41.7
(CH2Ph), 122.1 (phenoxy-C2), 126.7 (phenoxy-C4), 127.8 (phenyl-
C2), 128.8 (phenyl-C4), 129.3 (phenoxy-C3), 129.6 (phenyl-C3),
135.1 (phenyl-C1), 154.0 (phenoxy-C1), 213.0 (CdS).
N,N-Dieth yl S-Ben zyl Dith ioca r ba m a te (2i, Z dNMe2).
A 100 mL, three-neck round-bottom flask was equipped with
magnetic stirrer, condenser, and dropping funnel. Sodium N,N-
diethyldithiocarbamate trihydrate (8.20 g, 0.036 mol) was
dissolved in ethanol (40 mL, 99%) and added to the reaction

Macromolecules, Vol. 36, No. 7, 2003
Thiocarbonylthio RAFT Polymerization 2275
Ta ble 1. Molecu la r Weigh ts a n d P olyd isp er sities Obta in ed in Th er m a l P olym er iza tion s of Styr en e w ith Ben zyl RAF T
Agen ts (2) a t 110 °C
a
b
AIBN initiator, 80 °C. Di-tert-butyl peroxide initiator, 120 °C.
flask under a nitrogen atmosphere at 0 °C. A solution of benzyl
chloride (5.10 g, 0.041 mol) in ethanol (10 mL) was added
dropwise over 30 min. The reaction was gradually warmed to
room temperature, and stirring was continued for a further
65 h. The precipitate (sodium chloride) was filtered off and
the filtrate concentrated under vacuum. Vacuum distillation
gave 6.18 g (64.5% yield) of a pale yellow liquid, bp 154-156
NCH2-CH3), 4.57 (s, 2H, SCH2Ph), 7.2-7.5 (m, 5H, ArH).
2-Cya n op r op -2-yl Dith ioben zoa te (3a , Z ) P h ). The
preparation of this compound is described elsewhere.15
S-Cya n op r op -2-yl S′-Meth yl Tr ith ioca r bon a te (3b, Z )
SMe). A solution of bis((methylthio)thiocarbonyl)disulfide22
(1.23 g, 0.005 mol) and AIBN (1.25 g, 0.0076 mol) in benzene
(10 mL) was degassed with a stream of nitrogen and heated
under reflux for 24 h. The solvent was removed under vacuum
and the residue chromatographed on silica with 20% ethyl
1
°C (0.5 mmHg) (lit.21 bp 154-155 °C (1 mmHg)). H NMR, δ:
1.3 (t, 6H, NCH2-CH3), 3.7 (q 2H, NCH2-CH3), 4.05 (q, 2H,

2276 Chiefari et al.
Macromolecules, Vol. 36, No. 7, 2003
Ta ble 2. Molecu la r Weigh ts a n d P olyd isp er sities
Obta in ed in Th er m a l P olym er iza tion s of Styr en e w ith
Va r iou s Cya n oisop r op yl RAF T Agen ts (3) a t 110 °C
acetate in petroleum spirits as eluent to afford S-cyanoprop-
1
2-yl S′-methyl trithiocarbonate (0.09 g, 47% yield). H NMR,
δ: 1.85 (s, 6H, 2 × CH3), 2.75 (s, 3H, SCH3). 13C NMR (CDCl3),
δ: 19.1 (SCH3), 27.0 (2 × CH3), 43.0 (C(CH3)2CN), 120.0 (CN),
199.0 (CdS). Mass spectrum: found 191.9989 (M + 1);
C6H9S3N requires 191.9976.
2-Cya n op r op -2-yl Dith ioa ceta te (3c, Z ) CH3). Methyl-
magnesium chloride (20 mL, 0.06 mol, 3M solution in THF,
Aldrich) was diluted with THF (10 mL) and warmed to 40 °C,
and carbon disulfide (4.56 g, 0.06 mol) was added dropwise
over 15 min while maintaining the reaction temperature at
40 °C. After 1 h, the mixture was allowed to cool to room
temperature and then poured slowly into ice water (100 mL)
and extracted with diethyl ether (50 mL). The aqueous layer
was acidified to pH 2 with cold hydrochloric acid (10% aqueous
solution) and extracted with diethyl ether (2 × 50 mL). The
organic extracts were combined, dried over anhydrous mag-
nesium sulfate, and filtered, and the solvent was evaporated
to leave dithioacetic acid (5.2 g, 94% yield) as a dark yellow
oil. Dithioacetic acid is unstable and was used immediately
as follows.
A mixture of dithioacetic acid (5.2 g, 0.056 mol) and
R-methylstyrene (10 mL) were heated under nitrogen at 70
°C for 72 h. The excess R-methylstyrene was removed under
vacuum and the residue chromatographed on silica with 3%
ethyl acetate in petroleum spirits as eluent to afford 2-phen-
ylprop-2-yl dithioacetate23 (4.5 g, 45% yield) as a yellow oil.
1H NMR, δ: 1.90 (s, 6H, 2 × CH3), 2.70 (s, 3H, CH3), 7.20-
7.25 (m, 3H, meta, para-ArH), 7.45 (d, 2H, ortho-ArH). 13C
NMR, δ: 27.9 (2 × CH3), 40.6 (CH3), 56.1 (C(CH3)2Ph), 126.6
(phenyl-C2), 126.8 (phenyl-C4), 128.1 (phenyl-C3), 144.5 (phenyl-
C1).
A solution of 2-phenylprop-2-yl dithioacetate (1 g) and AIBN
(1.17 g) in benzene (10 mL) was degassed and heated in vacuo
at 80 °C for 18 h. The solvent was removed under vacuum to
provide a residue (1.99 g) which was chromatographed on silica
with 6% ethyl acetate in petroleum spirits as eluent to afford
2-cyanoprop-2-yl dithioacetate (380 mg, 50%) as a yellow oil.
1H NMR (CDCl3), δ: 1.82 (s, 6H, 2 × CH3), 2.77 (s, 3H, CH3).
Cyanoisopropyl dithioacetate is unstable and decomposes
over several days at -20 °C.
2-Cya n op r op -2-yl 1-P yr r oleca r b od it h ioa t e (3d , Z )
P yr r ole). A solution of pyrrole N-thiocarbonyl disulfide (0.15
g)24 and AIBN (0.16 g) in ethyl acetate (5 mL) was degassed
and heated in vacuo at 70 °C for 24 h. The solvent was removed
under vacuum and the residue chromatographed on silica with
10% ethyl acetate in petroleum spirits as eluent to afford
2-cyanoprop-2-yl 1-pyrrolecarbodithioate (135 mg, 61% yield).
1H NMR, δ: 1.99 (s, 6H, 2 × CH3), 6.38 (m, 2H, pyrrole-H3),
7.61 (m, 2H, pyrrole-H2). 13C NMR, δ: 27.0 (2 × CH3), 44.0
(C(CH3)2CN), 114.7 (pyrrole-C3), 120.7 [(pyrrole-C2) and CN],
193.2 (CdS). Mass spectrum: found 211.0358 (M + 1); C9H10
N2S2 requires 211.0364.
Styr en e P olym er iza tion s. Solutions comprising styrene
and the thiocarbonylthio compound were heated at 110 ( 1
°C (polymerization times and exact concentrations are shown
in Table 1 or Table 2). The following procedure is typical.
Benzyl dithiobenzoate (35.6 mg) was dissolved in styrene
(4.55 g). Aliquots (1 mL; 1 h sample) or (2 mL; 4, 16 h samples)
were transferred to ampules which were degassed by four
freeze-evacuate-thaw cycles (<10-4 mmHg), sealed, and
heated at 110 ( 1 °C for the requisite time in a thermostated
oil bath. The ampules were then removed and cooled rapidly.
An aliquot (ca. 50 mg) of the reaction mixture was taken for
NMR analysis. The excess styrene was removed at ambient
temperature under vacuum to leave a residue that was
weighed to determine conversion and analyzed by GPC (Table
1).
The results of these and similar polymerizations are sum-
marized in Tables 1-3.
Tr a n sfer Coefficien t of N,N-Dieth yl S-Ben zyl Dith io-
ca r ba m a te (2i). Stock solutions of AIBN (9.7 mg) in styrene
(10 mL) and di-tert-butyl peroxide (12.0 mg) in styrene (100
mL) were prepared. N,N-Diethyl S-benzyl dithiocarbamate
(details of amounts used provided in Table 1) was weighed into
ampules to which 1 mL of the initiator solution was added.
The solutions were degassed by three freeze-evacuate-thaw
cycles (<10-4 mmHg), sealed, and heated at 80 ( 1 °C for 45
min (AIBN) or 120 ( 1 °C for 150 min (di-tert-butyl peroxide).
Conversions were estimated by 1H NMR on the reaction
mixtures by integration signals of the olefinic protons. Samples
were analyzed by GPC to provide the molecular weight data
in Table 1.
Molecu la r Or bita l Ca lcu la tion s. Semiempirical molec-
ular orbital calculations were performed with MOPAC 6.0
using the Chem3D 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. Trends in LUMO energies for dithiobenzoates were
the same. The AM1 was finally selected over the PM3
Hamiltonian as it provides better heat of formation data for
free radical species.25
All ab initio calculations were performed using the GAUSS-
IAN9826 program on a NEC SX-5 computer. Geometry opti-
mizations were performed using standard gradient techniques
at SCF, MP2 and B3LYP levels of theory using RHF for closed
shell systems with the basis sets indicated. Vibrational
frequencies were calculated for each optimized structure, all
structures were confirmed to be ground states. LUMO energies
are reported in Figure 8. The Gaussian archive entries for the
optimized geometries are provided as Supporting Information.
Syn th esis of Th ioca r bon ylth io Com p ou n d s. Thiocar-
bonylthio compounds are available in moderate to excellent
yields by a variety of methods. For a summary of the literature,
see recent reviews.27 In the present work, benzyl esters (2)
were synthesized by reaction of the appropriate carbodithioate
salt with benzyl bromide (Scheme 2).
For NMR the aliquot of reaction mixture (ca. 50 mg) was
transferred to an NMR tube and diluted with CDCl3 to ca. 500
µL. The residual styrene and benzyl dithiobenzoate was
determined by integration of the doublets at δ 5.3 and 5.8 ppm
(styrene PhCHdCH2) and the singlet at δ 4.65 ppm (2a
PhCH2).
Cyanoisopropyl esters (3b and 3d ) were prepared by the
reaction of the corresponding disulfide with cyanoisopropyl
radicals generated thermally from AIBN (Scheme 3).28 Cy-
anoisopropyl dithioacetate (3c) was prepared by the reaction
of cumyl dithioacetate23 with cyanoisopropyl radicals. We have

Macromolecules, Vol. 36, No. 7, 2003
Thiocarbonylthio RAFT Polymerization 2277
Ta ble 3. Tr a n sfer Coefficien ts for Ben zyl RAF T Agen ts (2) in Th er m a l P olym er iza tion s of Styr en e a t 110 °C
a
b
Evaluated by determination of residual CTA by 1H NMR (eq 7). Evaluated from the discrepancy between found and calculated
molecular weights (eq 8). c Transfer coefficient of RAFT agent evaluated from NMR conversions Value in parentheses comes from molecular
weight data by application of eq 8. Literature value at 70 °C,35 determined by Mayo method. e Value at 80 °C, determined by Mayo
d
method with data from Table 1. Literature value36 at 60 °C is 0.0044. f Value at 120 °C, determined by Mayo method with data from
Table 1.
shown elsewhere15 that cyanoisopropyl dithiobenzoate (3a ) can
also be prepared in high yield by a similar method.
agent Z substituent on RAFT polymerization, styrene poly-
merizations for RAFT agents 2 (R ) benzyl) and 3 (R )
2-cyanoprop-2-yl) were carried out in bulk at 110 °C for various
reaction times. All experiments were conducted with ∼0.03
M RAFT agent. The results (concentrations, molecular weights,
polydispersities, conversions) of these experiments are sum-
marized in Tables 1 and 2, respectively.
The evolution of molecular weight and polydispersity vs
conversion for two of the more active RAFT agents, the
dithiobenzoate and trithiocarbonate derivatives, are also sum-
All RAFT agents were purified by column chromatography,
crystallization, and/or distillation as appropriate, and the
purity was established by NMR analysis and thin-layer
chromatography prior to use. Confirmation of identity in the
case of novel compounds also came from high-resolution
chemical ionization mass spectrometry.
Styr en e P olym er iza tion in th e P r esen ce of Th ioca r -
bon ylth io Com p ou n d s. To examine the effect of the RAFT

2278 Chiefari et al.
Macromolecules, Vol. 36, No. 7, 2003
Sch em e 2
Sch em e 3
F igu r e 1. Evolution of polydispersity (9) and molecular
weight (b) with conversion for thermal styrene polymerization
at 110 °C with benzyl dithiobenzoate (2a ) (filled symbols) or
cyanoisopropyl dithiobenzoate (open symbols). Solid line is
calculated molecular weight. Dotted and dashed lines are lines
of best fit.
marized in Figures 1 and 2. The polydispersity at high
conversion is narrow for both benzyl and cyanoisopropyl
derivatives.
The RAFT agent concentration (∼0.03 M) was chosen to be
sufficiently high that the proportion of chain ends formed by
radical-radical termination would be small and sufficiently
low that retardation, which may be pronounced for high
concentrations of certain RAFT agents,9,11,29-32 would also be
small. Yields of polymer obtained in the various polymeriza-
tions are generally reduced from those obtained in the absence
of the RAFT agent but, in general, show no marked depen-
dence on the particular RAFT agent used. The small reduction
in yield (e.g. from 72% in the absence of the RAFT agent to
ca. 60% with RAFT agent for 16 h reaction time) may be
attributed to a lesser gel effect. It is well-established that
polymerizations giving lower molecular weight polymers show
a reduced gel effect.33 We have observed that a similar
conversion (ca. 60% after 16 h at 110 °C) is obtained in
nitroxide-mediated polymerizations initiated by alkoxyamines
under these reaction conditions.34 Slightly lower conversions
were observed with the dithiobenzoates 2a and 3a (ca. 53%
after 16 h) this may be indicative of some retardation (see later
discussion); however, the extent is small and not expected to
significantly affect the conclusions drawn in this paper.
Somewhat higher than anticipated conversions were obtained
with 3e (1, R ) 2-cyanoprop-2-yl, Z ) N-pyrrolo). This RAFT
agent appears to be thermally unstable and may decompose
to form radicals under the polymerization conditions.
Tr an sfer Coefficien ts of Th iocar bon ylth io Com pou n ds
in Styr en e P olym er iza tion . There are few values for the
transfer coefficients of thiocarbonylthio compounds (1) in
styrene polymerization (or other polymerizations) in the
literature. Niwa et al.35 have determined the transfer coef-
ficient of S-benzyl-O-ethyl dithiocarbonate as 0.105 at 70 °C
(this is much lower than that of the corresponding bis-
(xanthogen disulfide) 4.4435). The transfer coefficient of S-
benzyl-N,N-diethyl dithiocarbamate is reported by Otsu et al.36
as 0.0044 at 60 °C (that for the corresponding dithiuram
disulfide is 0.2936). We obtain a similar value at 110 °C (see
Table 3).
F igu r e 2. Evolution of polydispersity (9) and molecular
weight (b) with conversion for thermal styrene polymerization
at 110 °C with S,S′-dibenzyl trithiocarbonate (2b) (filled
symbols) or S-methyl S′-cyanoisopropyl trithiocarbonate (open
symbols). The solid line is calculated molecular weight. Dotted
and dashed lines are lines of best fit to the experimental data.
Elimination of the radical concentrations by a steady-state
approximation provides eq 2.
d[1]
[1]
≈ Ctr
(2)
d[M]
[M] + Ctr[1] + C-tr[5]
This equation can, in principle, be solved numerically to
provide estimates of Ctr and C-tr. If the rate of the reverse
reaction between R• and the polymeric RAFT agent (5) is
negligible (low C-tr and/or low [5]), this expression simplifies
to an expression (eq 3) that describes conventional chain
transfer and the transfer constant can be evaluated from the
slope of a plot of ln[1] vs ln[M].40
d[1]
[1]
≈ Ctr
(3)
(4)
d[M]
[M]
d ln[1]
Ctr
≈
d ln[M]
There are also few previous reports on the kinetics of free
radical addition to thiocarbonyl compounds37,38 but little on
the substituent effects on the rate constant for addition. It is,
nonetheless apparent from this work that the thiocarbonyl
double bond (CdS) can be very reactive toward free radicals.
In the case of reversible chain transfer, we have shown that
the rate of consumption of the transfer agent depends on two
transfer constants, Ctr ()ktr/kp) and C-tr ()k-tr/ki) which
describe the reactivity of the propagating radical (Pn•), and the
expelled radical (R•) respectively (see eq 1).39
Values of Ctr reported in this paper have been calculated using
eq 4 and should therefore be considered as transfer coefficients
for the given reaction conditions rather than transfer constants
and be taken as minimum values pending further investigation
over a wider range of RAFT agent concentrations.
In chain transfer by addition-fragmentation (refer to
Scheme 1), the rate constant for chain transfer (ktr) is given
by the following expression (eq 5).41
ktr[1][P•] - k-tr[5][R•]
d[1]
kâ
≈
(1)
ktr ) kadd
×
(5)
d[M] kp[M][Pn•] + ki[M][R•]
k-add + kâ

Macromolecules, Vol. 36, No. 7, 2003
Thiocarbonylthio RAFT Polymerization 2279
Similarly, the reverse transfer constant (k-tr) is given by eq 6.
k-add
k-tr ) k-â
(6)
k-add + kâ
In this work, the concentration of residual RAFT agent in the
reaction mixtures has, where possible, been determined di-
rectly by NMR analysis. The 1H NMR resonance associated
with the benzylic methylene of RAFT agents 2 appears at 4.4-
4.6 ppm where there is no interference from other signals. The
cyanoisopropyl methyls of 3 appear in the region 1.8-2.0 ppm
and may be obscured by signals due to polystyrene.
The signals due to 2 were compared with those due to the
residual monomer to give ([2]/[M])t. Then, since the conversion
(and [M]t) is known independently
F igu r e 3. Double log plot of [RAFT agent] vs [styrene] for
thermal polymerizations at 110 °C. Residual RAFT agent
determined from molecular weight data by application of eq 4
(filled symbols) or from NMR (open symbols). RAFT agent: 2a
(b, s); 2b (1, - -); 2c (9, ‚‚‚); 2d (2, - - -). Lines are lines of
best fit to molecular weight derived data that provide transfer
constants shown in Table 3.
conversion of 2) (([2]/[M])t [M]t)/[2]o
(7)
The concentration of residual RAFT agent can also be calcu-
lated from a comparison of found and calculated molecular
weights using the following relationship39
[2]0 - [2]t
conversion of 2 )
)
[2]o
[M]0 - [M]t
[2]o
[M]0 - [M]t
Xhn(calcd)
Xhn(found)
≈
(8)
{
} {
}
[2]o - [2]t
/
where Xhn(calcd) is the expected degree of polymerization
assuming complete consumption of transfer agent and Xhn-
(found) is the measured degree of polymerization. Xhn(calcd)
should be corrected for initiator derived chains as discussed
elsewhere.15 For the present polymerization conditions at
reaction times e4 h, this correction is small and can be
neglected as being within experimental error.
The transfer constants might also, in principle, be estimated
from the rate at which the polydispersity or the molecular
weight distribution narrows with conversion.15,31 The polydis-
persity is more sensitive to changes in Ctr and C-tr than the
number-average molecular weight (see Figures 1 and 2). In
other work, we have estimated the value of Ctr and C-tr for
the dithiobenzoate (2a ) at 60 °C to be ∼400 and 11600
respectively based on kinetic simulation of the rate of nar-
rowing of the molecular weight distribution (the Ctr (60 °C),
calculated from the molecular weight data by assuming C-tr
) 0 was 50).15 This method was not applied in the current
work due to insufficient experimental data. It is, nonetheless,
evident that C-tr is significant at 110 °C in that the polydis-
persities obtained at low conversion are substantially lower
than those expected on the basis of the transfer constants
shown which assume C-tr ) 0.15
Plots of ln[monomer] vs ln[RAFT agent] for the benzyl RAFT
agents are shown in Figures 3 and 4. The transfer coefficients
of the RAFT agents (2, R ) benzyl), based on the assumption
that eq 4 applies, are reported in Table 3. The value for 2d
seems to reduce with conversion. This may indicate a high C-tr
but may also be due to experimental error. The low conversion
Ctr is quoted in Table 3. There is excellent agreement between
values obtained using the two methods of estimating residual
RAFT agent (i.e., via eqs 7 and 8). Thus, even though the
transfer coefficients are based on limited data (two to three
data points), they are supported by two independent methods
of measurement. The transfer coefficients vary over 4 orders
of magnitude (0.01-30), depending on the structure of Z.
The errors in transfer coefficients are difficult to estimate
since there is insufficient data for a statistical analysis. Errors
are likely to be higher for those RAFT agents having large
transfer coefficients. On the basis of the estimated error in
determining conversions, in NMR integration and the errors
in molecular weight measurement, we estimate the error in
the transfer coefficient of 2a to be <10%. However, since many
of the errors are systematic, relative values should be subject
to a lesser error.
F igu r e 4. Double log plot of [RAFT agent] vs [styrene] for
thermal polymerizations at 110 °C. Residual RAFT agent
determined from molecular weight data by application of eq 4
(filled symbols) or from NMR (open symbols). RAFT agent: 2e
(9, s), 2f (b, - - -), 2g (2, -- --). Lines are lines of best fit
to molecular weight derived data that provide transfer con-
stants shown in Table 3.
For RAFT agents 3 (1, R ) 2-cyanoprop-2-yl), the transfer
coefficients were too large to be readily measured by the above-
mentioned method. The calculated and found molecular weights
agree and residual dithioester was not detected by NMR even
after short reaction times (at the lowest monomer conversion).
A very high transfer constant for 2-cyanoprop-2-yl dithioben-
zoate in styrene polymerization is also indicated by the fact
that very narrow polydispersities are obtained even at the
lowest conversion (see Figure 1 and Table 2). Complete (>95%)
utilization of the initial RAFT agent at 5% conversion suggests
Ctr > 50.15 The actual Ctr may be much higher. With C-tr ) 0
it is possible to show that attaining a polydispersity of <1.1
at 5% conversion requires a Ctr of >500.15
Mech a n ism of RAF T P olym er iza tion . To understand the
effect of the Z group on the effectiveness of the RAFT agent
the mechanism of RAFT process must be known. The proposed
mechanism for styrene polymerization in the presence of a
RAFT agent is shown in Scheme 1.3 Propagating radicals are
generated as in a conventional radical polymerization under
the reaction conditions (i.e., thermally from styrene.42,43 The
RAFT agent (1) is transformed into a polymeric RAFT agent
(5) through reaction with a propagating radical (Pn•) by an
addition-fragmentation process. The radical liberated (R•)
then reacts with monomer to form a new propagating radical
(Pm•). Chain extension of the polymeric RAFT agent (5)
involves essentially the same process. The existence of the
radical adducts 4 (and 6) as intermediates in the addition-
fragmentation process has been confirmed by ESR spectrom-
etry.44 The reversible addition-fragmentation steps transfer
the SdC(Z)S- moiety between active and dormant chains and

2280 Chiefari et al.
Macromolecules, Vol. 36, No. 7, 2003
Ta ble 4. C-H Bon d Dissocia tion En er gies (DC-H) for
Com p ou n d s ZXYC-H
provide a mechanism for all chains to grow with similar rate
and uniformity. The efficiency of this process determines the
living character on the polymerization.
DC-H (kJ mol-1)a for X, Y )
On the basis of the addition-fragmentation mechanism, at
least four factors are expected to influence the effectiveness
of thiocarbonylthio compounds (1): (a) the rate constant of
reaction of 1 with the propagating (or initiating) radicals (kadd);
(b) the partitioning of the adduct (4) between starting materi-
als and products (determined by the relative magnitude of
(k-add and kâ); (c) the absolute rate constant for fragmentation
of the intermediate radicals (4) (kâ); (d) the rate and efficiency
at which the expelled radicals (R•) reinitiate polymerization.
45
Z
CH3, CH3
400
CH3, H45
H, H49
H, H50
H
439
439
423
CH3
(CH3)3C
Cl
413
423
422
419
417
415
402
389d
F
CH3S
(CH3)3Si
(CH3)2P
RO
Factors a and b should be directly reflected in the magnitude
of the transfer coefficient of 1.
393b
394
406c
406
402d
391
H2N
If fragmentation is slow (i.e., both k-add and kâ are small) or
reinitiation of polymerization is slow with respect to propaga-
tion, then polymerization may be retarded, and the likelihood
of the radicals 4 and/or R• undergoing side reactions leading
to some degree of inhibition is increased. If readdition to reform
the adduct radical (4) becomes a significant pathway, the
situation may arise where the transfer coefficient for chain
transfer is dependent on the concentration of the RAFT agent
(see above and eq 2).
(CH3)2N
CO2C2H5
CN
Ph
CH2dCH
352
387
385
365
349
400
397
378
360
397
375
367
a
All numbers rounded to nearest kJ . b Z ) C3H7O. c Z ) C4H9O.
d
Z ) CH3O.
Sch em e 4
Discu ssion
Within each series of RAFT agent (benzyl (2) or
cyanoisopropyl (3)) the differences in transfer coefficient
(Table 3) should mainly reflect the influence of the Z
group on the rate of free radical addition to the CdS
double bond. The partitioning of the adduct radical (4)
between starting materials and products should then
be determined by the relative leaving group ability of
R and the polystyryl radical, and R is constant within
each series. The dependence, if any, of the slope of the
plots of ln[1] vs ln[M] on RAFT agent concentration has
not been examined and the extent of the back reaction
under these conditions is unknown but based on other
data is substantial when R is benzyl (see above).15 The
transfer coefficients given in Table 3 should therefore
be considered as minimum values. However, the trend
in activity within a series (constant R) is not expected
to vary from that shown. The transfer coefficients of
benzyl dithiobenzoate (2a ) (ca. 26) and acetate (2c) (ca.
10, see Table 3) under our conditions are several orders
of magnitude less than the recently reported10 transfer
constant of the corresponding polystyryl derivatives (ca.
6000 and 180 respectively at 60 °C). It is clear that
benzyl is a very poor leaving group with respect to the
polystyryl radical (i.e., k-add . kâ). The better leaving
group ability of polystyryl vs benzyl is attributed to the
influence of steric factors.15
Dibenzyl trithiocarbonate (2b) possesses two identical
leaving groups. Other work8 shows that both benzyl
groups act as leaving groups such that the trithiocar-
bonate moiety is in the center of the chain. This
compound should have a higher transfer coefficient than
a monobenzyl compound because there are two path-
ways for fragmentation from the intermediate adduct.
Several factors can be considered to be of importance
in determining the rate of free radical addition to
thiocarbonylthio compounds. It is clear that product
radical stabilities or, rather, the ability of the group Z
to stabilize an adjacent radical center is not, by itself, a
predictive tool in estimating transfer coefficients. Lit-
erature data on C-H bond dissociation energies for
compounds ZXYC-H are summarized in Table 4 from
which one can infer the relative stabilities of ZXYC•
radicals. In attempting to correlate the transfer coef-
ficients of RAFT agents (2) with these data, one feature
which stands out is the relatively low activity of dithio-
carbamates (Z ) N(alkyl)2) and xanthate (Z ) O-alkyl)
derivatives. These Z substituents might be anticipated
to favor addition because of their ability stabilize an
adjacent radical center more than (for example) an
adjacent aliphatic (Z ) alkyl) or sulfur (Z ) S-alkyl)
substituent (Table 4).43
The effect of the Z substituent on the double bond
character of CdS should be considered. The relatively
low activity of O-alkyl dithiocarbamate and N,N-dialkyl
xanthates derivatives can be qualitatively understood
in terms of the importance of the zwitterionic canonical
forms (see Scheme 4) which arise through interaction
between the O or N lone pairs and the CdS double bond.
These reduce the CdS double bond character and raise
the energy of the LUMO (and HOMO) (see Table 5).7
These factors should also reduce the reactivity of the
CdS double bond of dithiocarbamates and xanthate
derivatives toward free radical addition. These same
considerations lead to the expectation that substituents
that are electron withdrawing or which are able to
delocalize the lone pair should enhance the activity of
these compounds. Thus, by changing the substituent on
nitrogen their activity can be substantially modified
such that they become very effective RAFT agents.6,7,12,46
The reactions of xanthates with free radicals to give
reversible addition-fragmentation and their use in this
context as synthetic reagents in organic chemistry has
known been for some time.47 However, prior to the
present work,46 there are little data on what factors
affect the rate of radical addition (or of fragmentation
from the species formed). We can also note that alkyl
xanthates, while comparatively poor RAFT agents in
styrene polymerization, are very effective as RAFT
agents in vinyl acetate polymerization.6,46,48 This indi-
cates that the more reactive vinyl acetate propagating
radical is able to give facile addition to the xanthate but,
more importantly, that fragmentation is also facile. As
with dithiocarbamates, the activity of xanthates can be
tuned by changing the substituent on oxygen. In styrene
polymerization, we find the transfer coefficient increases

Macromolecules, Vol. 36, No. 7, 2003
Thiocarbonylthio RAFT Polymerization 2281
Ta ble 5. HUMO a n d LUMO En er gies, P a r tia l Ch a r ges on Su lfu r for Th ioca r bon ylth io Com p ou n d s ZC(dS)S-CH 3 (in
Or d er of Decr ea sin g LUMO En er gy), a n d Rela tive Hea ts of Rea ction for F r ee Ra d ica l Ad d ition of Meth yl Ra d ica l
energy (eV)
atomic charge
c
∆∆Hr
CdS
(Å)
Z
HOMO
LUMO
dSa
-Sa
dSb
-Sb
(kJ mol-1
)
(CH3)2N-
lactam-
PhO-
CH3O-
(CH3)3Si-
CH3-
CH3S-
Ph-
-8.59275
-8.63381
-8.9495
-9.0183
-8.37473
-8.75059
-8.7506
-8.71945
-8.89483
-9.44907
-9.1006
-9.3363
-0.29084
-0.71265
-0.72940
-0.76045
-0.87821
-0.94518
-1.0759
-0.171
0.029
-0.043
-0.021
0.072
0.039
0.089
0.032
0.033
0.197
0.158
0.337
0.312
0.363
0.305
0.283
0.312
0.353
0.347
0.403
0.426
-0.222
-0.065
-0.103
-0.087
0.012
-0.030
0.002
-0.037
0.035
0.000
0.168
0.132
0.292
0.317
0.329
0.262
0.235
0.279
0.292
0.398
0.349
0.368
41.5
72.6
64.1
50.9
55.5
64.5
66.5
70.5
74.5
78.7
87.5
88.3
1.59
1.56
1.55
1.57
1.53
1.54
1.55
1.55
1.57
1.55
1.53
1.52
-1.17647
-1.33411
-1.41857
-1.6338
N-pyrrolo
C6F5O-
CCl3
-0.067
0.162
0.882
0.077
0.0970
CF3
-1.7739
a
b
Milliken charge. Atomic charge. c Relative heat of reaction ) ∆Hf(1) - ∆Hf(4).
F igu r e 5. Plot of logarithm of transfer coefficient (of ZC(d
F igu r e 6. Correlation of LUMO energy (0, s) and relative
heat of reaction for methyl radical addition to methyl RAFT
agents (ZC(dS)S-CH3). Data come form AM1 calculations.
Details are provided in Table 5. The point removed from the
line is the system with Z ) lactam.
S)S-CH2Ph) vs calculated LUMO energy for methyl (0, ZC-
(dS)S-CH3, from Table 5) and benzyl (2, ZC(dS)S-CH2Ph)
RAFT agents.
20-fold over the series EtO < C6H5O < C6F5O (see Table
2) as the oxygen lone pair becomes less available for
interaction with the CdS double bond.
We have calculated heats of reaction for addition of
various radicals to the CdS double bond of thiocarbo-
nylthio compounds using semiempirical molecular or-
bital methods. Energies of the LUMO and HOMO and
atomic charge densities for various thiocarbonylthio
compounds have also been calculated. Full geometry
optimizations for both the RAFT agent (1) and the
corresponding adduct (4) were performed using the
program MOPAC with the AM1 Hamiltonian. Data for
various RAFT agents (1) with R ) methyl are sum-
marized in Table 5. Some calculations were also per-
formed for RAFT agents (1) with R ) benzyl. While the
parameters are dependent on R, the trend in values
appears essentially independent of R (Figure 5). The
effect of substituents on R on the activity of RAFT
agents is discussed elsewhere.15 There is a good cor-
relation between relative LUMO energies and overall
heats of reaction (Figure 6).
F igu r e 7. Correlation of LUMO energy (0, s) and charge on
dS of methyl RAFT agents (ZC(dS)S-CH3). Data come from
AM1 calculations. Details are provided in Table 5.
with the LUMO energy is poorer with these examples
included. However, this may reflect the uncertainty in
the values of Ctr (see above). There is a reasonable
correlation of the LUMO energy with the charge local-
ized on the thiocarbonyl sulfur (Figure 7), which is
consistent with a hypothesis that electron-withdrawing
Z groups facilitate addition. There appears to be no
correlation with CdS bond length or other geometric
parameters (Table 5).
Ab initio methods have also been applied to calculate
the properties of the ground states of a series of RAFT
agents (1), R ) methyl. The results validate the trends
with LUMO energy seen with the semiempirical calcu-
lations. The data are shown in Figure 8. The trends are
similar although absolute values differ.
We observe that for dithiocarbamates and xanthates
there is a significant charge localized on the thiocarbo-
nyl sulfur (consistent with the above-mentioned hypoth-
esissScheme 4) and that this is reduced when the lone
pair is able to interact directly with a π-system or an
electron-withdrawing group. The correlation of ln(Ctr)
One interesting feature of the geometry apparent
from both AM1 and ab initio calculations is that the
phenyl ring of the dithiobenzoate prefers not to lie
coplanar with the CdS double bond. The dihedral angle
is 38.4 (AM1) or 35.6° (ab initio) for methyl dithioben-
zoate. However, the activation barrier for rotation

2282 Chiefari et al.
Macromolecules, Vol. 36, No. 7, 2003
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. We thank Ngoc Le for assistance in carrying
out the polymerizations with N,N-diethyl S-benzyl
dithiocarbamate.
Su p p or tin g In for m a tion Ava ila ble: A figure showing
the atom numbering for a table giving a summary of the
geometry and atomic charges for the thiocarbonylthio group
of methyl RAFT agents (ZC(dS)S-CH3) from Gaussian 98
calculations with MP2/D95 basis set and text giving Gaussian
archive entries for the optimized geometries with 3/21G*,
6/31G*, MP2/D95, and B3LYP/6-31G* basis sets. This material
is available free of charge via the Internet at http://
pubs.acs.org.
F igu r e 8. Plot of logarithm of transfer coefficient (of ZC(d
S)S-CH2Ph) vs calculated LUMO energy for methyl RAFT
agents (ZC(dS)S-CH3). Values from AM1 calculations (0), ab
initio calculations with Gaussian 98 and 3/21G* (O), 6/31G*
(2) or MP2/D95 (1) basis sets, or density functional calcula-
tions with basis set B3LYP/6-31G* (9).
Refer en ces a n d Notes
(1) Krstina, J .; Moad, G.; Rizzardo, E.; Winzor, C. L.; Berge, C.
T.; Fryd, M. Macromolecules 1995, 28, 5381-5.
(2) Krstina, J .; Moad, C. L.; Moad, G.; Rizzardo, E.; Berge, C.
T.; Fryd, M. Macromol. Symp. 1996, 111, 13-23.
(3) 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.
(4) Chong, B. Y. K.; Le, T. P. T.; Moad, G.; Rizzardo, E.; Thang,
S. H. Macromolecules 1999, 32, 2071-4.
(5) 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.
(6) Rizzardo, E.; Chiefari, J .; Mayadunne, R. T. A.; Moad, G.;
Thang, S. H. ACS Symp. Ser. 2000, 768, 278-96.
(7) Mayadunne, R. T. A.; Rizzardo, E.; Chiefari, J .; Moad, G.;
Thang, S. H. Macromolecules 1999, 32, 6977-80.
(8) Mayadunne, R. T. A.; Rizzardo, E.; Chiefari, J .; Krstina, J .;
Moad, G.; Postma, A.; Thang, S. H. Macromolecules 2000,
33, 243-5.
(9) Moad, G.; Chiefari, J .; Krstina, J .; Postma, A.; Mayadunne,
R. T. A.; Rizzardo, E.; Thang, S. H. Polym. Int. 2000, 49, 933-
1001.
(10) Goto, A.; Sato, K.; Tsujii, Y.; Fukuda, T.; Moad, G.; Rizzardo,
E.; Thang, S. H. Macromolecules 2001, 34, 402-8.
(11) Barner-Kowollik, C.; Quinn, J . F.; Nguyen, T. L. U.; Heuts,
J . P. A.; Davis, T. P. Macromolecules 2001, 34, 7849-57.
(12) Destarac, M.; Charmot, D.; Franck, X.; Zard, S. Z. Macromol.
Rapid Commun. 2000, 21, 1035-9.
through the plane of the CdS bond is very small (∼0.2
kJ mol-1).
Factors which enhance the rate of free radical addi-
tion to the CdS double bond of 1 will, in general, also
reduce the rate of fragmentation of the adduct 4
(Scheme 1). For efficient chain transfer both addition
and fragmentation are required to be facile. A slow
overall rate of fragmentation is one possible cause of
retardation.9,31 Rates of polymerization in the present
study show no marked dependence on Z and are only
slightly reduced for the dithiobenzoates (RdPh) over
those seen with other Z substituents. The retardation
with these RAFT agents appears independent of R (the
yield after 16 h with 2a or 3a is ca. 53% vs ca. 60-65%
with most other RAFT agentssTables 1 and 2). The
more severe retardation seen with high concentrations
of dithiobenzoate RAFT agents is dependent on R and
has therefore been associated with slow fragmentation
of the initial adduct (4).9,29 The smaller retardation seen
here may be associated with slow fragmentation of the
adduct 6. Other work suggests that slow fragmentation,
by itself, is unlikely to be responsible for retardation.32,44
If, however, fragmentation is slower there is a greater
likelihood that side reactions involving 4 and/or 6 such
as reversible or irreversible coupling with other radicals
will assume greater importance.11,29,30,32
(13) Ladaviere, C.; Doerr, N.; Claverie, J . P. Macromolecules 2001,
34, 5370-2.
(14) Donovan, M. S.; Lowe, A. B.; Sumerlin, B. S.; McCormick, C.
L. Macromolecules 2002, 35, 4123-32.
(15) Chong, Y. K.; Krstina, J .; Le, T. P. T.; Moad, G.; Rizzardo,
E.; Thang, S. H. Macromolecules 2003, 36, 2256.
(16) Some results mentioned in this paper have been published
previously as part of a conference paper.9
Con clu sion s
The effect of varying the substituent Z of RAFT agents
1 on the course of RAFT polymerization has been
examined. In general, the transfer coefficients of the
RAFT agents decreases in the order dithiobenzoates >
trithiocarbonates ∼ dithioalkanoates > dithiocarbonates
(xanthates) > dithiocarbamates. RAFT agents with
electrophilic Z substituents with lone pairs directly
conjugated to the CdS double bond (O-, N<) have low
transfer coefficients. However, electron-withdrawing
groups on O or N (in particular, groups able to delocalize
the nitrogen lone pair in the case of dithiocarbamates)
can significantly enhance the activity of RAFT agents
to modify the above order. The relative effectiveness of
the RAFT agents is rationalized in terms of interaction
of the Z substituent with the CdS double bond to
activate or deactivate that group toward free radical
addition. Semiempirical molecular orbital calculations
and the estimated LUMO energies or heats of reaction
can be used in a qualitative manner to predict relative
activity of RAFT agents.
(17) Leung, M.-K.; Hsieh, D.-T.; Lee, K.-H.; Liou, J .-C. J . Chem.
Res. Synop. 1995, 478-9.
(18) Katritzky, A. R.; Lewis, J .; Musumarra, G.; Ogretir, G. Chim.
Acta Turc. 1976, 4, 71-80.
(19) Takeshima, T.; Ikeda, M.; Yokoyama, M.; Fukada, N.; Mu-
raoka, M. J . Chem. Soc., Perkin Trans. I 1979, 692-5.
(20) Wakamori, S.; Yoshida, Y.; Tsuchidate, T.; Ishii, Y. Agric.
Biol. Chem. 1969, 33, 1691-9.
(21) Otsu, T.; Kuriyama, A. Polym. Bull. (Berlin) 1984, 11, 135-
42.
(22) Fuchigami, T.; Chen, C. S.; Nomaka, T.; Yen, M. Y.; Tien, H.
J . Bull. Chem. Soc. J pn. 1986, 59, 487-91.
(23) Oae, S.; Yagihara, T.; Okabe, T. Tetrahedron 1972, 28, 3203-
16.
(24) Oddo, B.; Alberti, C. Gazz. Chim. Ital. 1938, 68, 204-14.
(25) Moad, G.; Rizzardo, E. Macromolecules 1995, 28, 8722-8.
(26) Frisch, M. J .; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J . R.; Zakrzewski, V. G.; J . A.
Montgomery, J .; Stratmann, R. E.; Burant, J . C.; Dapprich,
S.; Millam, J . M.; Daniels, A. D.; Kudin, K. N.; Strain, M. C.;
Farkas, O.; Tomasi, J .; Barone, V.; Cossi, M.; Cammi, R.;
Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski,
J .; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.;
Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman,

Macromolecules, Vol. 36, No. 7, 2003
Thiocarbonylthio RAFT Polymerization 2283
J . B.; Cioslowski, J .; Ortiz, J . V.; Baboul, A. G.; Stefanov, B.
B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Gom-
perts, R.; Martin, R. L.; Fox, D. J .; Keith, T.; Al-Laham, M.
A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe,
M.; Gill, P. M. W.; J ohnson, B. G.; Chen, W.; Wong, M. W.;
Andres, J . L.; Head-Gordon, M.; Replogle, E. S.; J . A. Pople
Gaussian 98 (Revision A.9); Gaussian, Inc.: Pittsburgh, PA,
1998.
(35) Niwa, M.; Matsumoto, T.; Izumi, H. Makromol. Chem. 1987,
189, 2187-99.
(36) Otsu, T.; Matsunaga, T.; Doi, T.; Matumoto, A. Eur. Polym.
J . 1995, 31, 67-78.
(37) Alberti, A.; Benaglia, M.; Bonini, B. F.; Pedulli, G. F. J . Chem.
Soc., Faraday Trans. 1 1988, 84, 3347-58.
(38) Scaiano, J . C.; Ingold, K. U. J . Am. Chem. Soc. 1976, 98,
4727-32.
(39) Moad, C. L.; Moad, G.; Postma, A.; Mayadunne, R. T. A.;
Rizzardo, E.; Thang, S. H. Macromolecules 2002, to be
submitted.
(40) Bamford, C. H.; Barb, W. G.; J enkins, A. D.; Onyon, P. F.
The Kinetics of Vinyl Polymerization by Radical Mechanisms;
Butterworth: London, 1958.
(27) 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.
(28) Thang, S. H.; Chong, Y. K.; Mayadunne, R. T. A.; Moad, G.;
Rizzardo, E. Tetrahedron Lett. 1999, 40, 2435-8.
(29) Moad, G.; Chiefari, J .; Mayadunne, R. T. A.; Moad, C. L.;
Postma, A.; Rizzardo, E.; Thang, S. H. Macromol. Symp.
2002, 182, 65-80.
(30) Monteiro, M. J .; de Brouwer, H. Macromolecules 2001, 34,
349-52.
(41) Moad, G.; Moad, C. L.; Rizzardo, E.; Thang, S. H. Macromol-
ecules 1996, 29, 7717-26.
(42) Mayo, F. R. J . Am. Chem. Soc. 1968, 90, 1289.
(43) Chong, Y. K.; Rizzardo, E.; Solomon, D. H. J . Am. Chem. Soc.
1983, 103, 7761-2.
(31) Barner-Kowollik, C.; Quinn, J . F.; Morsley, D. R.; Davis, T.
P. J . Polym. Sci., Part A: Polym. Chem. 2001, 39, 1353-65.
(32) Kwak, Y.; Goto, A.; Tsujii, Y.; Murata, Y.; Komatsu, K.;
Fukuda, T. Macromolecules 2002, 38, 3026-9.
(44) Hawthorne, D. G.; Moad, G.; Rizzardo, E.; Thang, S. H.
Macromolecules 1999, 5457-9.
(45) Brocks, J . J .; Beckhaus, H.-D.; Beckwith, A. L. J .; Ru¨chardt,
C. J . Org. Chem. 1998, 63, 1935-43.
(46) Chiefari, J .; Mayadunne, R. T. A.; Moad, G.; Rizzardo, E.;
Thang, S. H. PCT Int. Appl. WO 99/31144 (Chem. Abstr.
1999, 131, 45250).
(47) Zard, S. Z. Angew. Chem., Int. Ed. Engl. 1997, 36, 672-85.
(48) Charmot, D.; Corpart, P.; Adam, H.; Zard, S. Z.; Biadatti, T.;
Bouhadir, G. Macromol. Symp. 2000, 150, 23-32.
(49) Ro¨mer, B.; Gatev, G.; Zhong, M.; Brauman, J . L. J . Am. Chem.
Soc. 1998, 120, 2919-24.
(50) J ursic, B. S.; Timberlake, J . W.; Engel, P. S. Tetrahedron Lett.
1996, 37, 6473-4.
(33) O’Neil, G. A.; Wisnudel, M. B.; Torkelson, J . M. Macromol-
ecules 1998, 31, 4537-45.
(34) Postma, A.; Moad, G. Unpublished data. For example:
A
solution of 1-(2-tert-butoxy-1-phenylethoxy)-2,5-diethyl-2,5-
dimethylimidazolidin-4-one (Chong, Y. K.; Ercole, F.; Moad,
G.; Rizzardo, E.; Thang, S. H.; Anderson, A. G. Macromol-
ecules 1999, 32, 6895-903) (12.65 mg) in styrene (0.91 g) was
placed in an ampule and degassed through three freeze-
evacuate-thaw cycles, and the ampule was sealed and heated
at 110 °C for 16 h. The ampule was cooled and opened. The
styrene was evaporated under reduced pressure and the
residue dried to contant weight. The polystyrene (0.59 g, 65%
conversion) was analyzed by GPC and had Mn 16 900, Mw/
Mn 1.48.
MA020883+