Dependence of thermal stability on molecular structure of RAFT/MADIX agents: A kinetic and mechanistic study

Article abstract of DOI:10.1021/ma201570f

The thermal decomposition of different classes of RAFT/MADIX agents, namely dithioesters, trithiocarbonates, xanthates, and dithiocarbamates, were investigated through heating in solution. It was found that the decomposition behavior is complicated interplay of the effects of stabilizing Z-group and leaving R-group. The mechanism of the decomposition is mainly through three pathways, i.e., β-elimination, α-elimination, and homolysis of dithiocarbamate (particularly for universal RAFT agent). The most important pathway is the β-elimination of thiocarbonylthio compounds possessing β-hydrogen, leading to the formation unsaturated species. For the leaving group containing solely α-hydrogen, such as benzyl, α-elimination takes place, resulting in the formation of (E)-stilbene through a carbene intermediate. Homolysis occurs specifically in the case of a universal RAFT agent, in which a thiocarbonyl radical and an alkylthio radical are generated, finally forming thiolactone through a radical process. The stabilities of the RAFT/MADIX agents are investigated by measuring the apparent kinetics and activation energy of the thermal decomposition reactions. Both Z-group and R-group influence the stability of the agents through electronic and steric effects. Lone pair electron donating heteroatoms of Z-group show a remarkable stabilizing effect while electron withdrawing substituents, either in Z- or R-group, tends to destabilize the agent. In addition, bulkier or more β-hydrogens result in faster decomposition rate or lower decomposition temperature. Thus, the stability of the RAFT/MAIDX agents decreases in the order where R is (with identical Z = phenyl) -CH₂Ph (5) > -PS (PS-RAFT 15) > -C(Me)HPh (2) > -C(Me)₂C(=O)OC₂H₅ (7) > -C(Me)₂Ph(1) > -PMMA (PMMA-RAFT 16) > -C(Me) ₂CN (6). For those possessing identical leaving group such as 1-phenylethyl, the stability decreases in the order of O-ethyl (11) > -N(CH₂CH₃)₂ (13) > -SCH(CH₃)Ph (8) > -Ph (2) > -CH₂Ph (4) > -PhNO₂ (3). These results consort with the chain transfer acitivities measured by the CSIRO group and agree well with the ab initio theoretical results by Coote. In addition, the difference between thermal stabilities of the universal RAFT agents at neutral and protonated states has also been demonstrated.

Full text of DOI:10.1021/ma201570f

ARTICLE
pubs.acs.org/Macromolecules
Dependence of Thermal Stability on Molecular Structure of RAFT/
MADIX Agents: A Kinetic and Mechanistic Study
Yanwu Zhou, Junpo He,* Changxi Li, Linxiang Hong, and Yuliang Yang
The State Key Laboratory of Molecular Engineering of Polymers, Department of Macromolecular Science, Fudan University,
Shanghai 200433, China
S Supporting Information
b
ABSTRACT: The thermal decomposition of different classes of
RAFT/MADIX agents, namely dithioesters, trithiocarbonates,
xanthates, and dithiocarbamates, were investigated through
heating in solution. It was found that the decomposition
behavior is complicated interplay of the effects of stabilizing
Z-group and leaving R-group. The mechanism of the decom-
position is mainly through three pathways, i.e., β-elimination,
α-elimination, and homolysis of dithiocarbamate (particularly for
universal RAFT agent). The most important pathway is the β-elimination of thiocarbonylthio compounds possessing β-hydrogen,
leading to the formation unsaturated species. For the leaving group containing solely α-hydrogen, such as benzyl, α-elimination
takes place, resulting in the formation of (E)-stilbene through a carbene intermediate. Homolysis occurs specifically in the case of a
universal RAFT agent, in which a thiocarbonyl radical and an alkylthio radical are generated, finally forming thiolactone through a
radical process. The stabilities of the RAFT/MADIX agents are investigated by measuring the apparent kinetics and activation
energy of the thermal decomposition reactions. Both Z-group and R-group influence the stability of the agents through electronic
and steric effects. Lone pair electron donating heteroatoms of Z-group show a remarkable stabilizing effect while electron
withdrawing substituents, either in Z- or R-group, tends to destabilize the agent. In addition, bulkier or more β-hydrogens result in
faster decomposition rate or lower decomposition temperature. Thus, the stability of the RAFT/MAIDX agents decreases in the
order where R is (with identical Z = phenyl) ꢀCH2Ph (5) > ꢀPS (PS-RAFT 15) > ꢀC(Me)HPh (2) > ꢀC(Me)2C(dO)OC2H5
(7) > ꢀC(Me)2Ph(1) > ꢀPMMA (PMMA-RAFT 16) > ꢀC(Me)2CN (6). For those possessing identical leaving group such as
1-phenylethyl, the stability decreases in the order of O-ethyl (11) > ꢀN(CH2CH3)2 (13) > ꢀSCH(CH3)Ph (8) > ꢀPh (2) >
ꢀCH2Ph (4) > ꢀPhNO2 (3). These results consort with the chain transfer acitivities measured by the CSIRO group and agree well
with the ab initio theoretical results by Coote. In addition, the difference between thermal stabilities of the universal RAFT agents at
neutral and protonated states has also been demonstrated.
’ INTRODUCTION
of RAFT polymerization products has been highlighted in com-
prehensive reviews very recently.46ꢀ48
Thiocarbonylthio compounds play the central role in rever-
sible additionꢀfragmentation chain transfer process (RAFT)1ꢀ7
and macromolecular architecture design by interchange of
xanthates (MADIX)8ꢀ10 based radical polymerizations. After poly-
merization these moieties are preserved at the chain end of the
product,11,12 as indicated by nuclear magnetic resonance spec-
troscopy (NMR)13ꢀ15 and matrix-assisted laser desorption ioniza-
tion (MALDI) or electrospray ionization (ESI) mass spectroscopy
(MS) results,16ꢀ25 providing further possibility of functionaliza-
tion and coupling. One of the most important functionalization is
the conversion of thiocarbonylthio moiety into thiol group by
aminolysis,26ꢀ35 hydrolysis,36,37 and reduction.38ꢀ40 Direct cou-
pling of electron-deficient dithioester moieties and appropriate
dienes through hetero-DielsꢀAlder reaction has been developed
to synthesize block and star (co)polymers.41ꢀ43 The oxidation of
dithioester termini into thioesters, either by peroxides16 or by air
in the presence of radical,44 has also been observed. The oxida-
tion has been used to convert RAFT polymerization products
into polymers with hydroxyl termini.45 The end functionalization
Thermal decomposition of RAFT agent has been observed
some years ago and has caused attention with the aim of end
group removal to stabilize the RAFT polymerization products.34,49ꢀ55
The decomposition also closely relates to the fidelity of the
thiocarbonythio chain ends and therefore to the efficiencies of
end functionlization and block copolymerization. If decomposi-
tion occurs during the polymerization, it will cause retardation to
polymerization rate56,57 through the loss of the living chain end
and subsequent radical quenching by the resulting dithiocarboxy-
lic acid, in resemblance to the retardation caused by the hydro-
lysis of,58 or impurities in,59,60 the RAFT/MADIX agents. The
primary pathway of thermal decomposition is through Chugaev-
like elimination,61ꢀ65 in which thiocarbonylthio moiety and a
Received: July 10, 2011
Revised:
September 14, 2011
Published: October 04, 2011
r
2011 American Chemical Society
8446
dx.doi.org/10.1021/ma201570f Macromolecules 2011, 44, 8446–8457
|

Macromolecules
ARTICLE
β-hydrogen are expelled, resulting in unsaturated species34,52ꢀ57,66
and the corresponding acids.52ꢀ54,56,66
impact ionization. Analytes were separated by a HP-5MS capillary column
of 30 m (0.25 i.d. and 0.25 μm film thickness), which was inserted
directly into the ion source of the MS system. Helium (99.999%) was the
carrier gas maintained at a flow rate of 1.0 mL/min. The injector was
kept at 280 °C, and samples were split injected (1 μL, split ratio 100:1).
The column oven temperature was programmed to start at 80 °C for
2 min, ramp at 10 °C/min to 280 °C, and held at 280 °C for 5 min. EI
spectra were scanned between 50 and 550 Da in the full-scan acquisition
mode. The transfer line and EI source temperature were both 230 °C.
The electron multiplier voltage was set on 70 eV. Matrix-assisted laser
desorption ionization time-of-flight mass spectroscopy (MALDI-TOF
MS) measurements were performed on a Voyager-DE STR (Applied
Biosystems, Framingham, MA) equipped with a nitrogen laser emitting
at 337 nm with a 3 ns pulse duration. The instrument was calibrated with
angiotensin and insulin and operated in the reflector mode. The accel-
erating potential is 20 kV. For each spectrum, 1000 laser shots were
accumulated. The matrix used was (R)-cyano-4-hydroxycinnamic acid
(CHCA). Samples were prepared by mixing the solution of polymer
(2 μL, 10 mg/mL in THF), the matrix (10 μL, 20 mg/mL in THF), and
silver trifluoroacetate (TFA, cationizing agent, 1 μL, 10 mg/mL in THF).
An aliquot of 1 μL of the resulting mixture was spotted on the target plate
and air-dried. Element analysis was carried on a Heraeus 1106 instrument.
Synthesis of the RAFT/MADIX Agents. For the synthetic
procedures of agents which have been reported, the readers refer to
the corresponding literature indicated in the following. The character-
ization results of all agents are presented.
Cumyl Dithiobenzoate (1). The synthesis follows that in ref 76. The
product is purple crystals. 1H NMR (CDCl3): δ (ppm) = 2.00 (s, 6H,
CH3ꢀCꢀCH3), 7.85 (d, 2H, o-ArH of dithiobenzoate), 7.46 (dd, 1H, p-
ArH of dithiobenzoate), 7.55 (d, 2H, o-ArH), 7.22(dd, 1H, p-ArH),
7.31ꢀ7.39 (m, 4H, ArH). FT-IR: ν (cmꢀ1) = 1218 and 1042 (CdS).
UVꢀvis max (cyclohexane): 299 and 444 nm. GC-MS (EI): m/e = 272.
Purity by HPLC = 99.0%.
1-Phenylethyl Dithiobenzoate (2). The synthesis follows that in ref 76.
The product is red crystals. 1H NMR (CDCl3): δ (ppm) = 1.81 (d, 3H,
CHꢀCH3), 5.26(q, 1H, CHꢀCH3), 7.95 (d, 2H, o-ArHof dithiobenzoate),
7.50 (dd, 1H, p-ArH of dithiobenzoate), 7.25ꢀ7.45 (m, 7H, ArH), FT-IR: ν
(cmꢀ1) = 1225 and 1042 (CdS). UVꢀvis max (cyclohexane): 298 and
500 nm. GC-MS (EI): m/e = 258. Purity by HPLC = 99.2%.
1-Phenylethyl 4-Nitrobenzodithioate (3). 1-(Bromomethyl)-4-nitro-
benzene (21.60 g, 0.10 mol), sulfur (7.05 g, 0.22 mol) and triethylamine
(40 mL) were added to a round bottom flask. Then, 1-bromoethylben-
zene (20.36 g, 0.11 mol) and DMF (130 mL) were added dropwise over
one hour to the flask. The mixture was then warmed at 60 C for 12 h with
magnetic stirring. The mixture was filtered. The obtained solution was
condensed by evaporation of DMF. The crude product was purified by
column chromatography on silica using petroleum ether as eluent to obtain
a purple oil. 1H NMR (CDCl3): δ (ppm) = 1.83 (d, 3H, CHꢀCH3), 5.21
(q, 1H, CHꢀCH3), 7.31 (t, 1H, p-ArH), 7.36 (t, 2H, m-ArH), 7.43 (d, 2H,
o-ArH), 8.02 (d, 2H, o-ArH of dithiobenzoate), 8.20 (d, 2H, m-ArH of
dithiobenzoate). FT-IR: ν(cmꢀ1) = 1062 and 1220 cmꢀ1 (CdS). UVꢀvis
max (cyclohexane): 297 and 519 nm. Purity by HPLC = 99.0%.
1-Phenylethyl Phenyldithioacetate (4). The synthesis follows that in
ref 77. The product is fine yellow crystals. 1H NMR (CDCl3): δ (ppm) =
1.68 (d, 3H, CHꢀCH3), 5.05 (q, 1H, CHꢀCH3), 4.26 (s, 2H,
CH2ꢀPh), 7.17ꢀ7.35 (m, 10H, ArH). FT-IR: ν (cmꢀ1) = 1219 and
1037 (CdS). UVꢀvis max (cyclohexane): 308 nm. GC-MS (EI): m/e =
272. Purity by HPLC = 99.0%.
Homolytic decomposition generating a thiocarbonylthio radical
and a leaving group-derived radical has also been proposed for
polystyrene,55 poly(methyl methacrylate) (PMMA), and poly-
(butyl acrylate) (PBA)52,54 prepared by trithiocarbonate-mediated
polymerization. In that case, the leaving propagating radical may
undergo chain end backbiting followed by β-scission, resulting in
short chain species. In the case of PMMA, depolymerization from
the leaving radical, releasing MMA monomer, has been observed. In
another study of thermolysis by GC-MS and TGA,67 it has been
found that the carbonꢀsulfur single bond is the most labile and is
affected by the structure of Z- and R-substituents of the agents.
In the present work, a series of RAFT/MADIX agents are
subject to thermal decomposition in solution. The experimental
conditions are more close to large-scale post-treatment than
thermolysis in TGA instrument to remove the sulfur-containing
chain end. Where necessary, the decomposition products are
isolated and characterized. Therefore, it is possible to discuss the
reaction mechanism in the light of final products instead of tran-
sient species as those in thermolysis. The relationship between
the thermal stabilitiesandmolecular structures of thiocarbonythio
compounds are investigated with respect to the reaction mechan-
ism and discussed in correlation with the chain transfer activity
measured by the CSIRO group68,69 and with the stability of RAFT
agents calculated by Coote using ab initio calculation.70ꢀ74 In
addition, the thermal stability of a universal RAFT agent75 at the
neutral and protonated states is also investigated.
’ EXPERIMENTAL SECTION
Materials. α,α0-Azobis(isobutyronitrile) (AIBN, Shanghai 4th Fac-
tory of Chemicals, 99%) was recrystallized from methanol. Methanol
(Lingfeng Chemicals, 99.5%), petroleum ether (60ꢀ90 °C range,
Shenxiang Chemicals, 99%), and dichloromethane (Lingfeng Chemi-
cals, 99%) were dried over CaCl2 and distilled before use. Carbon
disulfide (CS2, Shanghai 4th Factory of Chemicals, 99%) was purified by
vigorously shaking with KMnO4 (0.4 wt % based on CS2) followed by
filtration and distillation to collect a colorless fraction. Cyclohexane
(Feida, 99.5%), benzene (Shenxiang Chemicals, 99%), tetrahydrofuran
(Shanghai Qiangsheng Chemicals, 99%), and toluene (Lingfeng Chemi-
cals, 99%) were dried over sodium and distilled before use. tert-
Butylbenzene (Aldrich, 99%), diphenyl ether (Acros, 99%), ethyl
α-bromoisobutyrate (Aldrich, 98%), DMSO (Aladdin, HPLC grade),
and p-toluenesulfonic acid monohydrate (Aldrich, 98.5%) were used as
received. Methyl methacrylate (MMA, Shanghai 4th Factory of Chemi-
cals, 98%) and styrene (Yonghua Special Chemicals, 99%) were distilled
after being stirred over CaH2 for 24 h.
Measurements. High-pressure liquid chromatography (HPLC)
was performed on an instrument composed of a Waters 515 pump, a
C-18 column, and a UV detector with the wavelength set at 254 and
300 nm. Acetonitrile/water (v/v = 83/17) mixture was used as eluent
(1.0 mL/min) at 40 °C. 1H and 13C NMR and HSQC measurements
were carried out on a Bruker (500 MHz) NMR instrument, using tetra-
methylsilane (TMS) as the interior reference. Gel permeation chroma-
tography (GPC) was performed on a Waters 410 system equipped with
a guard column and three TSK columns (TSK Gel H-type, pore size 15,
30, and 200 Å) in series, a Waters 410 RI, and a Waters 486 UV detector,
using THF as the eluent at a flow rate of 1 mL/min at 40 °C. The colu-
mns were calibrated by narrow polystyrene standard with molecular
weight ranging from 2.2 ꢁ 103 to 5.2 ꢁ 105 g/mol. Gas chromatographyꢀ
mass spectrometry (GC-MS) was performed on an Agilent Technologies
6890N system, equipped with a mass-selective detector using electron
Benzyl Dithiobenzoate (5). The synthesis follows that in ref 76. The
product is red oil. 1H NMR (CDCl3): δ (ppm) = 4.6 (s, 2H, CH2ꢀPh),
8.00 (d, 2H, o-ArH of dithiobenzoate), 7.53 (dd, 1H, p-ArH of
dithiobenzoate), 7.25ꢀ7.45 (m, 7H, ArH). FT-IR: ν (cmꢀ1) = 1226
and 1044 (CdS). UVꢀvis max (cyclohexane): 302 and 504 nm. GC-MS
(EI): m/e = 244. Purity by HPLC = 99.4%.
8447
dx.doi.org/10.1021/ma201570f |Macromolecules 2011, 44, 8446–8457

Macromolecules
ARTICLE
Table 1. Results of Thermal Decomposition Reaction of RAFT/MADIX Agents
a kdec,333 is the calculated apparent rate coefficient of thermal decomposition at 333 K. b The results are from ref 66, in which kdec at 333 K was
erroneously calculated as 4.64 ꢁ 10ꢀ5 sꢀ1. The numbers in parentheses are average values from three parallel experiments. c Expected product according
to ref 61. d Expected product according to refs 53 and 56. e 1H NMR spectrum of the reaction mixture is given in the next section.
2-Cyanoprop-2-yl Dithiobenzoate (6). The synthesis follows that in
ref 78. The product is pink oil. 1H NMR (CDCl3): δ (ppm) = 1.95 (s,
6H, CH3ꢀCꢀCH3), 7.92 (d, 2H, o-ArH of dithiobenzoate), 7.56(dd,
1H, p-ArH of dithiobenzoate), 7.40 (dd, 2H, m-ArH of dithiobenzoate).
FT-IR: ν (cmꢀ1) = 1229 and 1047 (CdS); ν (cmꢀ1) = 2231(CN).
UVꢀvis max (cyclohexane): 296 and 529 nm. GC-MS (EI): m/e = 221.
Purity by HPLC = 99.0%.
(m, 20H, ArH). FT-IR: ν (cmꢀ1) 1208 and 1069 (CdS). UVꢀvis max
(cyclohexane): 313 nm. Purity by HPLC = 99.0%.
Dibenzyl Trithiocarbonate (9). The synthesis follows that in ref 80.
The product is yellow oil. 1H NMR: δ (ppm) = 4.62 (s, 4H, CH2ꢀPh),
7.26ꢀ7.36 (m, 20H, ArH). FT-IR: ν (cmꢀ1) = 1223 and 1063 (CdS).
UVꢀvis max (cyclohexane): 307 nm. GC-MS (EI): m/e = 290. Purity by
HPLC = 99.5%.
2-(Ethoxycarbonyl)prop-2-yl Dithiobenzoate (7). The synthesis
follows that in ref 31. The product is red oil. H NMR (CDCl3): δ
O-Ethyl S-(2-Phenylpropan-2-yl)xanthate (10). The synthesis fol-
lows that in ref 81. The product is white crystals. 1H NMR (CDCl3): δ
(ppm) = 1.82 (s, 6H, CH3-C-CH3), 4.38 (q, 2H, CH3ꢀCH2), 1.06 (t,
3H, CH3ꢀCH2), 7.52 (d, 1H, o-ArH), 7.31 (dd, 2H, m-ArH), 7.22 (dd,
2H, p-ArH). FT-IR: ν (cmꢀ1) = 1238 and 1042 (CdS). UVꢀvis max
(cyclohexane): 286 nm. Purity by HPLC = 98.6%.
O-Ethyl S-(1-phenylethyl)xanthate (11). The synthesis follows that
in ref 81. The product is pale yellow oil. 1H NMR (CDCl3): δ (ppm) =
1.71 (d, 3H, CHꢀCH3), 4.89 (q, H, CHꢀCH3), 4.61 (q, 2H, CH3ꢀ
CH2), 1.38 (t, 3H, CH3ꢀCH2), 7.37 (d, 2H, o-ArH), 7.31 (dd, 2H,
m-ArH), 7.25 (dd, H, p-ArH). FT-IR: ν (cmꢀ1) = 1213 and 1041 (CdS).
1
(ppm) = 1.24 (t, 3H, CH2ꢀCH3), 1.76 (s, 6H, CH3ꢀCꢀCH3), 4.16
(q, 2H, OꢀCH2CH3), 7.35 (dd, 2H, m-ArH of dithiobenzoate), 7.52
(dd, 1H, p-ArH of dithiobenzoate), 7.95 (d, 2H, o-ArH of dithiobenzoate).
FT-IR: ν (cmꢀ1) = 1260 and 1042 (CdS); ν (cmꢀ1) = 1732 (CdO).
UVꢀvis max (cyclohexane): 296 and 513 nm. GC-MS (EI): m/e = 268.
Purity by HPLC = 99.8%.
Bis(1-phenylethyl) Trithiocarbonate (8). The synthesis follows that
1
in ref 79. The product is yellow oil. H NMR (CDCl3): δ (ppm) =
1.72ꢀ1.75 (dd, 6H, CH3ꢀCH), 5.31 (q, 4H, CH2ꢀPh), 7.26ꢀ7.36
8448
dx.doi.org/10.1021/ma201570f |Macromolecules 2011, 44, 8446–8457

Macromolecules
ARTICLE
Figure 1. (a) UVꢀvis spectra following the thermal decomposition reaction of protonated 1-methoxycarbonylethyl N-methyl-N-pyridin-4-
yldithiocarbamate (18), (0.0135 g, 5.00 ꢁ 10ꢀ2 mmol) in DMSO (5 mL) at 180 °C (inset: photographs showing the color change during heating)
and (b) kinetic plots of heating 18 at different temperatures: 9, 393 K; b, 413 K; 2, 433 K; 1, 453 K.
GC-MS (EI): m/e = 226. UVꢀvis max (cyclohexane): 283 nm. Purity by
HPLC = 99.0%.
O-Ethyl S-Benzylxanthate (12). The synthesis follows that in ref 79.
1
The product is pale yellow oil. H NMR (CDCl3): δ (ppm) = 4.37
(s, 2H, CH2ꢀAr), 4.65 (q, 2H, CH3ꢀCH2), 1.42 (t, 3H, CH3ꢀCH2),
7.25ꢀ7.35 (m, 5H, ArH). FT-IR: ν (cmꢀ1) = 1216 and 1047 (CdS).
GC-MS (EI): m/e = 212. UVꢀvis max (cyclohexane): 279 nm. Purity by
HPLC = 99.8%.
1-Phenylethyl-N,N0-diethyldithiocarbamate (13). The synthesis fol-
lows that in ref 82. The product is pale yellow oil. 1H NMR (CDCl3): δ
(ppm) = 1.24 and 1.27 (t, 6H, CH2ꢀCH3), 1.78 (d, 3H, CHꢀCH3),
3.70 and 4.02 (q, 4H, CH2ꢀCH3), 5.27 (q, H, CHꢀCH3), 7.25ꢀ7.43
(m, 5H, ArH). FT-IR: ν (cmꢀ1) = 1208 and 1069 (CdS). GC-MS
(EI): m/e = 253. UVꢀvis max (cyclohexane): 283 nm. Purity by
Figure 2. 1H NMR spectra of the reaction mixture of heating 1-pheny-
HPLC = 99.5%.
Benzyl-N,N0-diethyldithicarbamate (14). The synthesis follows that
lethyldithiobenzoate (2) at 160°C in DMSO-d6 (upper, 18 h;lower, 0 h).
in ref 82. The product is pale yellow oil. 1H NMR (CDCl3): δ (ppm) =
1.26 and 1.29 (t, 6H, CH2ꢀCH3), 3.72 and 4.04 (q, 4H, CH2ꢀCH3),
Protonation of 17 with TsOH (18). The synthesis follows that in ref
4.54 (s, 2H, CH2ꢀAr), 7.25ꢀ7.38 (m, 5H, ArH). FT-IR: ν (cmꢀ1) =
75. The product is yellow solid. 1H NMR (DMSO-d6): δ (ppm) = 1.53
1208 and 1069 (CdS). GC-MS (EI): m/e = 239. UVꢀvis max
(d, 3H, CHCH3); 2.33 (s, 3H, p-toluenesulfonate, CH3); 3.70, 3.75
(cyclohexane): 280 nm. Purity by HPLC = 99.9%.
(s, 6H,COOCH3 and NꢀCH3); 4.57 (q, 1H, CHCH3); 7.14 (d, 2H,
PS with Dithiobenzoate Terminus (15). Styrene (8.87 g, 0.09 mol)
bulk polymerization, initiated by AIBN (0.03 g, 0.20 mmol), was carried
out in the presence of CDB (0.11 g, 6.41 mmol) in a 50 mL flask under
argon at 60 °C for 10 h. The resulting PS-RAFT was obtained as a pink
powder after three times of dissolving in THF and precipitating from
methanol. The product was analyzed by GPC. Number-average molec-
ular weight: Mn,GPC = 2600 g/mol, PDI = 1.10.
PMMA with Dithiobenzoate Terminus (16). Methyl methacrylate
(MMA) (6.20 g, 0.06 mol) polymerization, initiated by AIBN (0.16 g,
0.97 mmol), was carried out in the presence of CDB (0.57 g, 2.10 mmol)
in benzene (16 mL) under argon at 60 °C for 8 h. PMMA-RAFT was
obtained as a pink powder after three times of dissolving in THF and
precipitating from petroleum ether. The product was analyzed by GPC.
Number-average molecular weight: Mn,GPC = 2800 g/mol, PDI = 1.14.
1-Methoxycarbonylethyl-N-methyl-N-pyridin-4-yldithiocarbamate
(17). The synthesis follows that in ref 75. The product is off-white solid.
1H NMR (CDCl3): δ (ppm) = 1.53 (d, 3H, CHCH3); 3.73 (s, 6H,
COOCH3 and NꢀCH3); 4.65 (q, 1H, CHCH3); 7.30 (d, 2H, m-ArH);
8.75 (br s, 2H, o-ArH). FT-IR: ν (cmꢀ1) = 1226 and 1094 (CdS). GC-
MS (EI): m/e = 270. UVꢀvis max (cyclohexane): 296 nm. Purity by
HPLC = 99.0%.
p-toluenesulfonate m-ArH); 7.73 (d, 2H, p-toluenesulfonate o-ArH);
7.90 (d, 2H, m-ArH); 8.94 (d, 2H, o-ArH).
Thermal Decomposition of RAFT/MADIX Agents. RAFT/
MADIX agent (listed in Table 1) was dissolved in tert-butylbenzene,
diphenyl ether, or DMSO in a 5 mL flask, keeping the initial concentra-
tion ca. 0.01 mol Lꢀ1. Then the reaction mixture was transferred to
individual ampules (∼1 mL) and deaerated through three cycles of
freezingꢀpumpingꢀthawing. The sealed ampules were then placed in
an oil bath. Each ampule was removed after a predetermined time
interval. The reactions were stopped by quenching into liquid nitrogen.
The concentration of the starting material was measured by HPLC
detected at 254 or 300 nm.
’ RESULTS AND DISCUSSION
Mechanism of Thermal Decomposition. Thermal decom-
position reactions are performed by heating the RAFT/MADIX
agents in well-deaerated solution of tert-butylbenzene, diphenyl
ether, or DMSO-d6. The products are characterized by at least
one method of HPLC, NMR, or MALDI-TOF MS. Appropriate
temperatures are applied for each compound due to different
8449
dx.doi.org/10.1021/ma201570f |Macromolecules 2011, 44, 8446–8457

Macromolecules
ARTICLE
Compound 6, with the leaving group derived from AIBN,
releases α-methylacrylonitrile, of which the vinylic and α-methyl
protons appear at δ = 5.94 ppm (doublet) and 1.73 ppm (Figure 3).
Compound 7, resembling the living chain in methacrylate RAFT
polymerization, decomposes into ethyl methacrylate. The vinylic
protons give signals at δ = 5.65 and 6.00 ppm as shown in
Figure 4.
On the other hand, dithioester 5 and trithiocarbonate 9
possessing benzyl as the leaving group (no β-hydrogen) are
not able to decompose through Chugaev process. Both 5 and 9
yield stilbene in pure (E)-isomer as the main product. The char-
acterizations of the isolated product using 1H and 13C NMR are
shown in Figure 5. The vinylic protons give signal at δ = 7.10
ppm, while aromatic protons give signals at δ = 7.24, 7.38, and
7.50 ppm, which are typical for (E)-stilbene.83 In 13C NMR, the
vinylic carbon is observed at δ = 126 ppm. The product is also
analyzed by GC-MS (Supporting Information), showing molec-
ular ion at m/z = 180. In addition, both of these reactions release
H2S as detected by Pb(OAC)2.
Figure 3. 1H NMR spectra of the reaction mixture of heating 2-cya-
noisopropyl dithiobenzoate (6) at 160 °C in DMSO-d6 (upper, 3 h;
lower, 0 h).
The elimination of trithiocarbonate or carbonylthiocarbonate
to form alkene is well documented.84ꢀ88 Nevertheless, those
syntheses usually started from cyclic trithiocarbonate precursors
and underwent the 1,2-elimination process. It should be noted
that the thermal decomposition of 5 and 9 in the present work
forms new CdC bond derived from benzyl groups. The process
is more likely to involve a carbene intermediate which is formed
by α-elimination on the methylene, as shown in Scheme 1. The
first step is the leaving of a proton, possibly aided by the solvent,
diphenyl ether, a weak base, and then the leaving of dithiocar-
boxylic or trithiocarbonic anion to form benzylic carbene, 5-2.
The carbene species is very active and can insert into the CꢀH
bond of the starting material, forming a coupled precursor, 5-3,
which will undergo further β-elimination to form stilbene. The
two-step process, e.g., the formation of precursor and subsequent
formation of stilbene, resembles the process in the preparation of
poly(phenylenevinylene) (PPV) derivatives from xylene-type
precursors of xanthate89,90 and dithiocarbamate.91ꢀ96 However,
it is believed that PPV synthesis involves a p-quinodimethane
intermediate93 derived by delocalization of carbanion on bifunc-
tionalized aromatic ring. Neither 5 nor 9 in the present work is
able to form such an intermediate. The stereoselectivity comes
from the second step, in which elimination takes place only
between thiocarbonylthio and cis β-hydrogen. Thus, for the three
conformers of 5-3 (Scheme 1), the most favored is 5-3-a whereas
the energy barrier of 5-3-b is high and 5-3-c does not have
cis β-hydrogen. This leads to the predominant formation of
the E-isomer of stilbene. The release of H2S may be a con-
sequence of possible anhydridization between the resulting
dithiobenzoic acids.
Trithiocarbonate 9 may undergo not only similar process as in
Scheme 1 but also another possible pathway, e.g., the extrusion of
CS2 to form a dibenzyl sulfur ether intermediate, which will
further release H2S to form (E)-stilbene (Scheme 2).97 This is
supported by the evidence of sulfur ether species in GC-MS of 9
(Supporting Information).
Xanthate 12 also possesses benzyl as the leaving group but
tends to decomposes from the O-ethyl side, releasing ethylene
and COS according to Chugaev elimination (Scheme 3).98 This
reaction should be in competition with α-elimination on the
S-benzyl side, which would otherwise result in (E)-stilbene. How-
ever, no stilbene has been detected by HPLC (Supporting
Information). Therefore, Chugaev elimination on O-ethyl side
Figure 4. 1H NMR spectra of the reaction mixture of heating 2--
(ethoxycarbonyl)prop-2-yl dithiobenzoate (7) at 160 °C in DMSO-d6
(upper, 1.5 h; lower, 0 h).
stabilities. The kinetic results are measured by HPLC, meanwhile
all reactions are also followed by UVꢀvis. An example of heating
protonated 1-methoxycarbonylethyl N-methyl-N-pyridin-4-yl-
dithiocarbamate, 18 (the universal RAFT agents in the proto-
nated form in Table 1), is shown in Figure 1, in which the
intensity of the absorption band of thiocarbonylthio moiety at
330 nm of 18 decreases along with reaction time.
Four classes of RAFT/MADIX agents, e.g., dithioesters,
trithiocarbonates, xanthates, and dithiocarbamates, are investi-
gated. The results are listed in Table 1. In most cases the main
products are olefins formed from Chugaev-like elimination,
provided that there is β-hydrogen in the molecule. For example,
dithioester 166 and xanthate 10 release α-methylstyrene upon
heating (HPLC in Supporting Information) because they both
possess cumyl substituent as the leaving group. Similarly, those
possessing 1-phenylethyl leaving group, such as dithioester 2, 3,
4, trithiocarbonate 8, xanthate 11, and dithiocarbamate 13, yields
styrene which is observed by HPLC (RT = 4.40 min, Supporting
Information). NMR spectra of the reaction mixture of 2 during
heating are shown in Figure 2, in which signals of vinylic protons
in the product, styrene, are observed at chemical shifts δ = 5.27,
5.84, and 6.75 ppm. The ratio of integrations (data underneath
the peaks in the inset) and the splitting constant agree well with
the structure of styrene.
8450
dx.doi.org/10.1021/ma201570f |Macromolecules 2011, 44, 8446–8457

Macromolecules
ARTICLE
Figure 5. (a) HPLC diagrams of the reaction mixtures of heating dibenzyl trithiocarbonate (9) (0.0145 g, 4.99 ꢁ 10ꢀ2 mmol) and benzyl
dithiobenzoate (5) (0.0122 g, 4.99 ꢁ 10ꢀ2 mmol), respectively, in diphenyl ether (5 mL). The main products indicated in the red frame are isolated and
characterized using 1H (b) and 13C (c) NMR, which are assignable to (E)-stilbene.
Scheme 1. Proposed Mechanism of Decomposition Pathway of Benzyl Dithiobenzoate (5) and the Formation of (E)-Stilbene
is predominant. Dithiocarbamate 14 is stable up to 240 °C for 30
min without any hint of decomposition, although the leaving
group is benzyl as well.
Polystyrenyl dithiobenzoate, 15, decomposes to give polystyr-
ene with an unsaturated terminus as detected by MALDI-TOF
MS of the product precipitated from methanol (Figure 6). The
main series of the peaks, P1 (m/z = 2518.3 in the enlarged
spectrum), and a minor series, P4 (m/z = 2570.3), are assignable
to polystyrene chains with ω-unsaturated terminus and two
kinds of α-termini, e.g., cumyl derived from the RAFT agents
and cyanoisopropyl from AIBN, respectively. Thus, dithioester
15 undergoes the same elimination process as trithiocarbonate
8451
dx.doi.org/10.1021/ma201570f |Macromolecules 2011, 44, 8446–8457

Macromolecules
ARTICLE
from polystyrene chain.55 The peak at m/z = 2533.3 is assignable
to coupling product bearing two cumyl termini, while the origin
for m/z = 2546.3 is unclear. It is surprising to find the coupling
product since its formation requires homolysis at the SꢀR single
bond, which is not the case in the RAFT polymerization. It
seems, however, that homolytic decomposition may take place at
elevated temperature (200 °C). Indeed, clear decrease in mo-
lecular weight is observed by GPC results for PS-RAFT after
heating, which may be ascribed to depolymerization from the
resulting propagating radical (Supporting Information). Ther-
mal elimination of PMMA prepared by RAFT polymerization
has already been studied in great detail in the literature.28,29,53,56
Recently, the CSIRO group reported N-(4-pyridinyl)-N-
methyl dithiocarbamates as a new class of universal RAFT agents
to provide excellent control on both active monomer such as
styrene and MMA and less active monomer such as vinyl acetate
and N-vinylpyrolidinone, using protonationꢀdeprotonation
switches on pyridinyl ring and the adjacent nitrogen atom.75
Dithiocarbamate 17, as well as its protonated form 18, is one
example of the switchable RAFT agents. It is interesting to find
from Figure 7 that the main decomposition product from
universal RAFT agent is a thiolactone species 19 (RT = 3.27 min,
spectrum shows strong bands for thiocarbonyl and carbonyl at
1124 and 1728 cmꢀ1, respectively (Supporting Information).
Results of element analysis shows the content of C: 49.14%, H:
4.35%, N: 10.71%, and S: 24.48%, which agree well with the
expected thiolactone structure 19 (calculated: C: 50.35%, H:
4.24%, N: 11.75%, S: 26.90%).
The formation of 19 may involve a radical process. As shown
in Scheme 4, homolysis of 17 generates a thiocarbamoyl radical
17-1 and an alkylthio radical 17-2. The latter may undergo
hydride migration and then cyclization to form radical 17-3,
which will recombine with 17-1 to form thiolactone substituted
thiocarbonyl amide 19. Since there is no vinylic protons ob-
servable in 1H NMR (Supporting Information), the decomposi-
tion of 17 is not through Chugaev elimination, although it
possess β-hydrogen. This unusual property clearly indicates the
high stability of the SꢀR single bond caused by the lone pair
donating ability of the nitrogen. Marginal stabilization of 17 may
also be imparted by the canonical structure between nitrogen and
the pyridinyl ring.
It is therefore summarized that the decomposition mechanism
of the RAFT agents is dominated by the structure of the leaving
group. When there is β-hydrogen to the thiocarbonylthio moiety,
the decomposition proceeds according to Chugaev elimination
mechanism. Molecules without β-hydrogen may involve carbene
intermediate, which is formed through α-elimination, followed
by β-elimination in the second step, although the two steps
may occur simultaneously in the whole system. Nevertheless,
it should be pointed out that the decomposition may be more
complicated, keeping in mind that only main products are
analyzed in the present work. Sulfur-containing byproduct,
possibly dithiocarboxylic acid and its derivatives, have not been
precisely characterized due to the difficulty in obtaining pure
substances.
1
P1 in HPLC diagram). The H NMR spectrum shows two
doublets at δ = 7.44 and 8.68 ppm, which are assignable to
protons on pyridine ring, while singlets at δ = 3.46 and 2.12
ppm are for methyl groups, respectively. These signals show the
expected ratio of integration, 2:2:3:3. In the 13C NMR spectrum,
the signals for thiocarbonyl and carbonyl are observed at δ = 198
and 177 ppm, respectively. The connections between carbons
and protons are confirmed by HSQC measurement. The MS
spectrum shows a molecular ion at m/z = 238. In addition, IR
Scheme 2. An Additional Pathway for the Thermal Decom-
position of Dibenzyl Trithiocarbonate (9)
Kinetics of the Thermal Decomposition. Despite the differ-
ent reaction pathways, the kinetics of the decomposition is
investigated by isothermal reaction in solution at different
temperatures. For simplification, all the thermal decomposition
reactions are regarded as unimolecular process with the ignorance
Scheme 3. Chugaev Elimination on O-Side of O-Ethyl S-Benzylxanthate (12)98
Figure 6. Matrix-assisted laser desorption ionization time-of-flight mass spectrum (MALDI-TOF MS) of thermal decomposition product of PS-RAFT
(Mn = 2600 g/mol, Mw/Mn = 1.10) in diphenyl ether at 473 K for 70 h (a: full view; b: partial enlargement).
8452
dx.doi.org/10.1021/ma201570f |Macromolecules 2011, 44, 8446–8457

Macromolecules
ARTICLE
Figure 7. (a) HPLC of the reaction mixture heating 1-methoxycarbonylethyl-N-methyl-N-pyridin-4-yl dithiocarbamate (universal RAFT agent: 17)
(0.0135 g, 4.99 ꢁ 10ꢀ2 mmol) in diphenyl ether at 240 °C for 70 min. The main product, P1, is isolated and characterized by 1H (b) and 13C (c) NMR
and HSQC (d).
Different values of kds are obtained as the slopes of the fitted lines.
The activation energy, Eact, is obtained from Arrhenius plots, by
which the rate coefficient at 333 K, kdec,333, can be calculated
(Table 1). Standard errors in Eact are estimated to be (2.9 and
(5.9 kJ/mol for 1 and 16, respectively.
Scheme 4. Proposed Mechanism of Thermal Decomposition
of 1-Methoxycarbonylethyl-N-methyl-N-pyridin-4-yl Dithio-
carbamate (Universal RAFT Agent: 17)
One must be aware, however, that eq 1 expresses only the
apparent kinetics. For a number of thiocarbonylthio compounds
there could be more decomposition pathways and, therefore,
more elementary reactions. The apparent activation energy ob-
tained from the Arrhenius plots is better regarded as the tem-
perature dependence of measured kinetic rate coefficients within
the range of performing temperature. The stability of the RAFT
agents is discussed in terms of the decomposition rate coeffi-
cients at the corresponding temperature, for example, kdec,333, at
60 °C.
The Arrhenius plots of various RAFT/MADIX agents are
shown in Figure 8, grouped according to the structure of Z- and
R-substituents. It appears that the apparent decomposition rate
and activation energy depend on both the structures of Z- and
R-group. The effect of Z-substituents is inferred by comparison
among different classes of thiocarbonylthio compounds, such
as 2, 3, 4, 8, 11, and 13, with identical leaving group R =
1-phenylethyl. In the temperature range 100ꢀ160 °C, dithioe-
sters 2, 3, and 4 readily decompose while the others seems quite
stable. The decompositions of 8, 11, and 13 are observable only
under high temperature, e.g., 200ꢀ240 °C, as shown in Figure 8a.
In addition, the values of Eact of these compounds are larger than
those of dithioesters, indicative of larger energy barrier of the
reaction intermediate. A subtle effect of the Z-group is also
of any reverse reaction. Therefore, the kinetic equation is expressed
by eq 1 in which [ZSSR]0 and [ZSSR]t represent HPLC measured
initial and instantaneous concentrations of the RAFT agents and
kdec is the apparent decomposition rate coefficient.
!
½ZSSRꢂ
½ZSSRꢂ0t
ln
¼ kdec
t
ð1Þ
The concentrations are plotted against reaction time at different
temperatures. One example of heating 18 is shown in Figure 1b.
8453
dx.doi.org/10.1021/ma201570f |Macromolecules 2011, 44, 8446–8457

Macromolecules
ARTICLE
Figure 8. Arrhenius plots of various RAFT/MADIX agents. The dotted extrapolations for systems at higher temperature are drawn only for guiding eyes
for comparison. Red lines: heated in tert-butylbenzene; blue lines: heated in diphenyl ether; green line: heated in DMSO. The associated values of pre-
exponential factor, Ad, and apparent activation energy, Eact, are given in Table 1.
Scheme 5. Canonical Structure of the Thiocarbonylthio
Compounds
(Scheme 5). The electron donating nature of the heteroatom
favors the resonance isomers by partially canceling the positive
charge and thus increases the stability of the compounds. On this
basis, electron withdrawing groups tend to destabilize the
RAFT/MADIX agents since they cannot lower the energy by
conjugation with the partially positive charged atom.
Nevertheless, for those compounds in which the leaving group
is benzyl, e.g., 5, 9, 12, and 14, the decomposition rates increase
in the order where Z is ꢀN(C2H5)2 (14) < ꢀPh (5) < ꢀSCH2Ph
(9) < O-ethyl (12) (vide infra, Figure 8c). This is very different
from those with R = 1-phenylethyl. Here, xanthate 12 becomes
the most active because it decomposes at O-ethyl side. Trithio-
carbonate 9 decomposes readily in double pathways as shown in
Schemes 1 and 2, thus displaying higher reactivity than dithioe-
ster 5 in spite of the presence of electron donating Z-group in 9.
Dithiocarbamate 14 does not show any decomposition under the
present conditions.
On the other hand, the effect of R-group is inferred by
comparison among various dithioesters with identical Z-group,
phenyl (Figure 8b). The thermal stability decreases in the order
where R is ꢀCH2Ph (5) > PSꢀ (15) > ꢀC(Me)HPh (2) >
ꢀC(Me)2C(dO)OC2H5 (7) > ꢀC(Me)2Ph (1) > ꢀPMMAꢀ
(16) > ꢀC(Me)2CN (6). It is clear that R = benzyl gives the
most stable molecule due to specific decomposition mechanism,
α-elimination. The stability of those having β-hydrogen is affected
observed among 2, 3, and 4, in which dithioesters decompose
more quickly in the order of Z = ꢀPh < ꢀCH2Ph < ꢀPhNO2,
although the values of Eact are quite close. The overall sequence
of stability is as follows: O-ethyl (11) > ꢀN(CH2CH3)2 (13) >
ꢀSCH (CH3) Ph (8) > ꢀPh (2) > ꢀCH2Ph (4) > PhNO2 (3).
The sequence correlates well with the order of apparent chain
transfer coefficient measured by Rizzardo and co-workers.68,69
These authors and Coote70ꢀ74 pointed out that lone pair
electron donating Z-groups such as alkoxy and dialkylamine
stabilize the RAFT agents through resonance between the
heteroatom in Z-group and the thiocarbonyl, exerting negative
charge on the sulfur atom in the latter. The same effect also exists
in estimating the thermal stability of the RAFT/MADIX agents
by considering further delocalization of the positive charge
between two heteroatoms connected to the thiocarbonyl
8454
dx.doi.org/10.1021/ma201570f |Macromolecules 2011, 44, 8446–8457

Macromolecules
ARTICLE
the same temperature range (not shown in the figure), possibly
due to the weak basicity of the ether which promotes the eli-
mination of β-hydrogen. However, the apparent activation
energy does not show significant difference with each other
(102.5 kJ/mol in the latter vs 106.6 kJ/mol in the former). Thus,
it is difficult to unambiguously distinguish the solvent effect from
the temperature dependence of the decomposition reaction.
Table 1 shows that the apparent activation energy for the
decomposition reaction follows roughly a decreasing sequence as
the reaction rate coefficient increases. The energy barrier for
xanthates and dithiocarbonates are generally higher than those of
trithiocarbonates and dithioesters. Striking deviations are none-
theless observed. For example, cumyl dithiobenzoate (1) shows
unusually large Eact, which could be a consequence of the reverse
coupling reaction. Trithiocarbonate 9 shows lower Eact than
dithioesters 5 because the former decomposes through two
pathways, as mentioned above. PMMA-RAFT (16) has Eact
larger than expected, possibly due to difficulty in rotation of
the CꢀS bond bearing a polymer chain. These side reactions and
factors definitely cause complexity in the macroscopic reaction
behavior of the thermal decomposition.
Figure 9. 1H NMR spectra of the reaction mixture of heating proto-
nated form of 1-methoxycarbonylethyl-N-methyl-N-pyridine-4-yl
dithiocarbamate (18) at 180 °C in DMSO-d6 (upper, 6 h; middle, 0 h;
lower, as-prepared, 1-methoxycarbonylethyl-N-methyl-N-pyridine-4-yl
dithiocarbamate (universal RAFT agent: 17).
by statistical, electronic, and steric factors. Statistical factor98
refers to the probability of conformers that provides cis hydrogen
to the thiocarbonylthio group, which finally leads to the concerted
cis elimination. Obviously, more β-hydrogen results in faster
elimination reaction. For example, 1 decomposes much faster
than 2 because cumyl has three more β-hydrogens than pheny-
lethyl. The same rule seemingly holds for polymeric RAFT
agents, PS-RAFT (15) and PMMA-RAFT (16) with two and
five β-hydrogens, respectively. In the latter, there should be two
directions for the elimination, e.g., either toward methyl to form
1-olefin or toward methylene to form 2-olefin. The former is
predominant. Nonetheless, the 1H NMR spectra in refs 52 and
56 indeed show small peaks both at δ = 5.5 ppm, which are
assignable to internal olefinic proton, in addition to those
for =CH2 at δ = 5.5 and 6.2 ppm. The calculated decomposition
rate coefficient of 16 at 60 °C is ca. 70 times larger than that of 15
(Table 1). This is the main reason why styrene RAFT polymer-
ization at elevated temperature runs well99 while MMA may not.
The electronic factor refers to that R-group with electron
withdrawing substituent tends to destabilize the RAFT agent
because it does not favor the canonical structure in Scheme 5.
This has been proved experimentally68,81 and theoretically73 by
other researchers. Keeping this in mind, it is easy to understand
that 2-cyanoisopropyl (6) is the most active dithioester. This is
because not only the large number of β-hydrogens but also the
presence of strong electronic withdrawing cyano group.
Steric factor also contributes to the destabilization effect by
weakening the SꢀR bond through distorted conformation.73
Nonetheless, it is difficult to distinguish between the steric effect
and the statistical factor because more bulky substituents such
as cumyl usually has more β-hydrogens. These two factors affect
the stability of the thiocarbonylthio compounds in the same
direction.
It is interesting to find from Figure 8c that the thermal stability
of the universal RAFT agent strongly depends on its resting state,
e.g., the neutral or protonated form. The neutral form, 17, is very
stable with Eact = 154.7 kJ/mol and kdec, 333= 4.36 ꢁ 10ꢀ13 sꢀ1
,
the second highest stability of all RAFT/MADIX agents in-
spected. The protonated form, 18, shows much higher tendency
to decompose into a number of products. The products of the
latter have not been isolated due to the low-resolution between
peaks in HPLC (Supporting Information). Also, the reaction
may have been complicated by the presence of p-toluenesulfonic
acid. In Figure 9, the 1H NMR spectrum of the reaction mixture
after heating for 6 h in DMSO-d6 shows remarkable weakening
and diversifying of aromatic proton signals, possibly due to the
coexistence of canonical structures of (4-pyridinyl)amine sub-
stituent. There is no or negligible signal assignable to vinyl
protons of methyl acrylate, the expected product of β-elimina-
tion. The decomposition rate is 104 times faster than that of the
neutral form (Table 1). This indicate that the universal RAFT
agent does show two state of activities, which can be switched by
protonation and deprotonation. The reason for larger activity in
protonated form is due to strong electron withdrawing effect of
quaternary ammonium ion.75 The switch between two states
renders the RAFT agent fascinating property to mediate the
polymerizations of monomers with very different reactivity, such
as MMA and vinyl acetate.75
’ CONCLUSION
Both the rate and mechanism of thermal decomposition
closely relate to molecular structure of RAFT/MADIX agents.
For thiocarbonylthio compounds possessing β-hydrogen, the
primary pathway is Chugaev-type elimination. For those pos-
sessing α- but not β-hydrogen, the decomposition involves a
phenylmethine carbene intermediate, resulting in the formation
of (E)-stilbene. The universal RAFT agents shows a somewhat
unexpected homolytic process, generating a thiocarbamoyl and
an alkylthio radical, which undergoes rearrangement and cou-
pling to form a new thiocarbamoyl species. In addition, thermal
decomposition of xanthate on O- and S-sides seems to compete
with each other.
The relationship between R-group and thermal stability is also
applicable for xanthates. As shown in Figure 8c, xanthate with R =
cumyl (10) gives the fastest decomposition rate, followed by
those R = 1-phenylethyl (11) and benzyl (12). The latter
decomposes from the O-ethyl side due to high stability of the
S-benzyl single bond.
From Figure 8, it seems that the reaction rate shows stronger
temperature dependence in diphenyl ether. The decomposition
of 1 is faster in diphenyl ether than in tert-butylbenzene within
8455
dx.doi.org/10.1021/ma201570f |Macromolecules 2011, 44, 8446–8457

Macromolecules
ARTICLE
The kinetic results demonstrate that the thermal stability of
RAFT/MADIX agents is influenced concurrently by the struc-
ture of Z- and R-groups and closely relates to the decomposition
pathway. For those undergo β-elimination, a remarkable differ-
ence in stability is observed among classes of thiocarbonylthio
compounds with various Z-groups. Namely, dithioesters are much
less stable than trithiocarbonates, xanthates, and dithiocarbamate.
Here the stabilizing effect of lone pair donating Z-group, such as
alkylthio (in trithiocarbonate), alkyloxy (in xanthates), and alkyl-
amine (in dithiocarbamate), consorts with that observed in chain
transfer reactivity. Within each class, those with more β-hydro-
gens and larger leaving groups give faster decomposition rate due
to easier formation of the coplanar intermediate for concerted
elimination. For those possessing only α-hydrogen, xanthates
may be more active due to elimination from the O-alkyl side,
while dithiocarbamates show highest stability without any hint of
decomposition.
The destabilizing effect by the electron withdrawing
Z-group is clearly illustrated by different stabilities of dithioe-
sters in which Z = phenyl and nitrophenyl and by the universal
RAFT agent at neutral and protonated state. This again con-
sorts with the observation of activating effect of electron
withdrawing groups on chain transfer reactivity of RAFT/
MADIX agents.
’ REFERENCES
(1) Chiefari, J. C.; Chong, Y. K.; Ercole, F.; Krstina, J.; Jeffery, 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, 5559–5562.
(2) Moad, G.; Rizzardo, E.; Thang, S. H. Polymer 2008, 49, 1079–1131.
(3) Moad, G.; Rizzardo, E.; Thang, S. H. Acc. Chem. Res. 2008, 41,
1133–1142.
(4) Boyer, C.; Bulmus, V.; Davis, T. P.; Ladmiral, V.; Liu, J.; Perrier,
S. Chem. Rev. 2009, 109, 5402–5436.
(5) Moad, G.; Rizzardo, E.; Thang, S. H. Aust. J. Chem. 2005, 58,
379–410.
(6) Moad, G.; Rizzardo, E.; Thang, S. H. Aust. J. Chem. 2006, 59,
669–692.
(7) Favier, A.; Charreyre, M. Macromol. Rapid Commun. 2006, 27,
653–692.
(8) Corpart, P. C.; Charmot, D.; Biadatti, T.; Zard, S. Z.; Michelet,
D. PCT Int. Appl. WO9858974,1998.
(9) Taton, D.; Wilczewska, A.-Z.; Destarac, M. Macromol. Rapid
Commun. 2001, 22, 1497–1503.
(10) Destarac, M.; Bzducha, W.; Taton, D.; Gauthier-Gillaizeau, I.;
Zard, S. Z. Macromol. Rapid Commun. 2002, 23, 1049–1054.
(11) Crotez-Lemus, N. A.; Salgado-Rodriguez, R.; Licea-Claverie, A.
J. Polym. Sci., Part A: Polym. Chem. 2010, 48, 3033–3051.
(12) Menteiro, M. J.; Lonsdale, D. E. J. Polym. Sci., Part A: Polym.
Chem. 2010, 48, 4496–4503.
(13) D’Agosto, F.; Hughes, R.; Charreyre, M.-T.; Pichot, C.; Gilbert,
R. G. Macromolecules 2003, 36, 621–629.
The present work not only provides fundamental knowledge
on thermal decomposition mechanism and kinetics, which is
important for the removal of thiocarbonylthio termini of the RAFT/
MADIX polymerizations products for the purpose of stabilization,
but also sets the temperature limit of applicability of specific thio-
carbonylthio compounds in RAFT polymerizations of different
monomers. While most RAFT agents are stable at normal RAFT
polymerization temperatures such as 60 °C, there are exceptions.
For example, that cyanoisopropyl dithiobenzoate (6) decomposes
at the rate coefficient of 2.57 ꢁ 10ꢀ6 sꢀ1 at 60 °C (calculated),
which is comparable to the rate of radical generation of AIBN
(14) Pound, G.; McLeary, J. B.; McKenzie, J. M.; Lange, R. F. M.;
Klumperman, B. Macromolecules 2006, 39, 7796–7797.
(15) Pound, G.; Aguesse, F.; McLeary, J. B.; Lange, R. F. M.;
Klumperman, B. Macromolecules 2007, 40, 8861–8871.
(16) Vana, P.; Albertin, L.; Barner, L.; Davis, T. P.; Barner-Kowollik,
C. J. Polym. Sci., Part A: Polym. Chem. 2002, 40, 4032–4037.
(17) Toy, A. A.; Vana, P.; Davis, T. P.; Barner-Kowollik, C. Macro-
molecules 2004, 37, 744–751.
(18) Favier, A.; Ladaviere, C.; Marie-Therese, C.; Pichot, C. Macro-
molecules 2004, 37, 2026–2034.
(19) Feldermann, A.; Toy, A. A.; Davis, T. P.; Stenzel, M. H.; Barner-
Kowollik, C. Polymer 2005, 46, 8448–8457.
(20) Hart-Smith, G.; Lovestead, T. M.; Davis, T. P.; Stenzel, M. H.;
Barner-Kowollik, C. Biomacromolecules 2007, 8, 2404–2415.
(21) Lovestead, T. M.; Hart-Smith, G.; Davis, T. P.; Stenzel, M. H.;
Barner-Kowollik, C. Macromolecules 2007, 40, 4142–4153.
(22) Hart-Smith, G.; Chaffey-Millar, H.; Barner-Kowollik, C. Macro-
molecules 2008, 41, 3023–3041.
(8.45 ꢁ 10ꢀ6
s
ꢀ1). If the conversion of 6 into corresponding
polymeric dithioester is slow, the decomposition will cause
retardation through radical quenching by the resulting dithio-
benzoic acid. Thus, the data of stability of thiocarbonylthio
compounds are very useful in optimizing the reaction conditions
for the RAFT/MADIX polymerizations.
(23) Gruendling, T.; Hart-Smith, G.; Davis, T. P.; Stenzel, M. H.;
Barner-Kowollik, C. Macromolecules 2008, 41, 1966–1971.
(24) Ladaviere, C.; Lacroix-Desmaze, P.; Delolme, F. Macromole-
cules 2009, 42, 70–84.
(25) Fandrich, N.; Falkenhagen, J.; Weidner, S. M.; Pfeifer, D.; Staal,
B.; Thunemann, A. F.; Laschewsky, A. Macromol. Chem. Phys. 2010,
211, 869–878.
(26) Mayadunne, R. T. A.; Rizzardo, E.; Chiefari, J.; Krstina, J.;
Moad, G.; Postma, A.; Thang, S. H. Macromolecules 2000, 33, 243–245.
(27) Wang, Z.; He, J.; Tao, Y.; Liu, Y.; Jiang, H.; Yang, Y. Macro-
molecules 2003, 36, 7446–7452.
(28) Patton, D. L.; Mullings, M.; Fulghum, T.; Advincula, R. C.
Macromolecules 2005, 38, 8597–8602.
’ ASSOCIATED CONTENT
Supporting Information. HPLC, GC-MS, 1H NMR,
S
b
GPC, FT-IR, and UVꢀvis of concerned products and monitor-
ing thermal decomposition reactions of certain thiocarbonylthio
compounds. This material is available free of charge via the
Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
*Phone: +86-21-65643509. E-mail: jphe@fudan.edu.cn.
(29) Lima, V.; Jiang, X.; Brokken-Zijp, J.; Schoenmakers, P. J.;
Klumperman, B.; Vanderlinde, R. J. Polym. Sci., Part A: Polym. Chem.
2005, 43, 959–973.
(30) Plummer, R.; Hill, D. J. T.; Whittaker, A. K. Macromolecules
2006, 39, 8379–8388.
(31) Xu, J.; He, J.; Fan, D.; Wang, X.; Yang, Y. Macromolecules 2006,
39, 8616–8624.
(32) Boyer, C.; Granville, A.; Davis, T. P.; Bulmus, V. J. Polym. Sci.,
Part A: Polym. Chem. 2009, 47, 3773–3794.
’ ACKNOWLEDGMENT
This work is subsidized by the National Science Foundation of
China (21174030). The authors thank Professor Ewald Daltroz-
zo for many helpful discussions.
8456
dx.doi.org/10.1021/ma201570f |Macromolecules 2011, 44, 8446–8457

Macromolecules
ARTICLE
(33) Boyer, C.; Liu, J.; Bulmus, V.; Davis, T. P. Aust. J. Chem. 2009,
62, 830–847.
(34) Moad, G.; Chong, Y. K.; Postma, A.; Rizzardo, E.; Thang, S. H.
Polymer 2005, 46, 8458–8468.
(35) Koo, S. P. S.; Stamenovic, M. M.; Prasath, R. A.; Inglis, A. J.;
Prez, E. D.; Barner-Kowollik, C.; Camp, W. V.; Junkers, T. J. Polym. Sci.,
Part A: Polym. Chem. 2010, 48, 1699–1713.
(36) Schilli, C.; Lanzendoerfer, M. G.; Mueller, A. H. E. Macro-
molecules 2002, 35, 6819–6827.
(37) Aqil, A.; Qiu, H.; Greisch, J. F.; Jerome, R.; Pauw, E. D.; Jerome,
C. Polymer 2008, 49, 1145–1153.
(38) Sumerlin, B. S.; Lowe, A. B.; Stroud, P. A.; Zhang, P.; Urban,
M. W.; McCormick, C. L. Langmuir 2003, 19, 5559–5562.
(39) Scales, C. W.; Convertine, A. J.; McCormick, C. L. Biomacro-
molecules 2006, 7, 1389–1392.
(68) Chong, Y. K.; Krstina, J.; Le, T. P. T.; Moad, G.; Postma, A.;
Rizzardo, E.; Thang, S. H. Macromolecules 2003, 36, 2256–2272.
(69) Chiefari, J.; Krstina, J.; Le, T. P. T.; Moad, G.; Postma, A.;
Rizzardo, E.; Thang, S. H. Macromolecules 2003, 36, 2273–2283.
(70) Coote, M. L.; Radom, L. J. Am. Chem. Soc. 2003, 125,
1490–1491.
(71) Coote, M. L. Macromolecules 2004, 37, 5023–5031.
(72) Coote, M. L.; Radom, L. Macromolecules 2004, 37, 590–596.
(73) Coote, M. L.; Henry, D. J. Macromolecules 2005, 38, 1415–
1433.
(74) Coote, M. L. J. Phys. Chem. A 2005, 109, 1230–1239.
(75) (a) Benaglia, M.; Chiefari, J.; Chong, Y. K.; Moad, G.; Rizzardo,
E.; Thang, S. H. J. Am. Chem. Soc. 2009, 131, 6914–6915. (b) Benaglia,
M.; Chen, M.; Chong, Y. K.; Moad, G.; Rizzardo, E.; Thang, S. H.
Macromolecules 2009, 42, 9384–9386. (c) Keddie, D. J.; Guerrero-
Sanchez, C.; Moad, G.; Rizzardo, E.; Thang, S. H. Macromolecules
2011, 44, 6738–6745.
(40) Zelikin, A. N.; Such, G. K.; Postma, A.; Caruso, F. Biomacro-
molecules 2007, 8, 2950–2953.
(41) Inglis, A. J.; Sinnwell, S.; Davis, T. P.; Barner-Kowollik, C.;
Stenzel, M. H. Macromolecules 2008, 41, 4120–4126.
(42) Inglis, A. J.; Sinnwell, S.; Martina, H.; Stenzel, M. H.; Barner-
Kowollik, C. Angew. Chem., Int. Ed. 2009, 48, 2411–2414.
(43) Sinnwell, S.; Inglis, A. J.; Davis, T. P.; Stenzel, M. H.; Barner-
Kowollik, C. Chem. Commun. 2008, 2052–2054.
(44) Li, C.; He, J.; Zhou, Y.; Gu, Y.; Yang, Y. J. Polym. Sci., Part A:
Polym. Chem. 2011, 49, 1351–1360.
(45) Dietrich, M.; Glassner, M.; Gruendling, T.; Schmid, C.;
Falkenhagen, J.; Barner-Kowollik, C. Polym. Chem. 2010, 1, 634–644.
(46) Willcock, H.; O’Reilly, R. K. Polym. Chem. 2010, 1, 149–157.
(47) Harvison, M. A.; Peter, J. R.; Davis, T. P.; Lowe, A. B. Aust. J.
Chem. 2011, 64, 992–1006.
(76) Le, T. P. ; Moad, G.; Rizzardo, E.; Thang, S. H. PCT Int. Appl.
WO 9801478.
(77) Quinn, J. F.; Rizzardo, E.; Davis, T. P. Chem. Commun. 2001,
1044–1045.
(78) Thang, S. H.; Chong, Y. K.; Mayadunne, R. T. A.; Moad, G.;
Rizzardo, E. Tetrahedron Lett. 1999, 40, 2435–2438.
(79) Ladaviere, C.; Doerr, N.; Claverie, J. Macromolecules 2001,
34, 5370–5372.
(80) Movassagh, B.; Soleiman-Beigi, M.; Nazari, M. Chem. Lett.
2008, 37, 22–23.
(81) Destarac, M.; Brochon, C.; Catala, J. M.; Wilczewska, A.; Zard,
S. Z. Macromol. Chem. Phys. 2002, 203, 2281–2289.
(82) Ishizu, K.; Khan, R. A.; Ohta, Y.; Furo, M. J. Polym. Sci., Part A:
Polym. Chem. 2004, 42, 76–82.
(83) Ohara, K.; Inokuma, Y.; Fujita, M. Angew. Chem., Int. Ed. 2010,
49, 5507–5509.
(48) Moad, G.; Rizzardo, E.; Thang, S. H. Polym. Int. 2011, 60, 9–25.
(49) Perrier, S.; Takolpuckdee, P.; Mars, C. A. Macromolecules 2005,
38, 2033–2036.
(84) Moore, J. A.; Isaacs, T. Tetrahedron Lett. 1973, 14, 5033–5036.
(85) Hatanaka, K.; Tanimoto, S.; Oida, T.; Okano, M. Tetrahedron
Lett. 1981, 22, 5195–5196.
(86) Hatanaka, K.;Tanimoto, S. Tetrahedron Lett. 1982, 23, 425–426.
(87) Corey, E. J.; Winter, E. J. Am. Chem. Soc. 1963, 85, 2677–2678.
(88) Horton, D.; Tindall, C. G., Jr. J. Org. Chem. 1970, 35,
3558–3559.
(89) Son, S.; Dodabalapur, A.; Lovinger, A. J.; Galvin, M. E. Science
1995, 269, 376–378.
(90) Kesters, E.; Gillissen, S.; Motmans, F.; Lutsen, L.; Vanderzande,
D. J. M. Macromolecules 2002, 35, 7902–7910.
(50) Chong, Y. K.; Moad, G.; Rizzardo, E.; Thang, S. H. Macro-
molecules 2007, 40, 4446–4455.
(51) Chen, M.; Moad, G.; Rizzardo, E. J. Polym. Sci., Part A: Polym.
Chem. 2009, 47, 6704–6714.
(52) Postma, A.; Davis, T. P.; Moad, G.; O’Shea, M. S. Macromole-
cules 2005, 38, 5371–5374.
(53) Chong, B.; Moad, G.; Rizzardo, E.; Skidmore, M.; Thang, S. H.
Aust. J. Chem. 2006, 59, 755–762.
(54) Postma, A.; Davis, T. P.; Li, G.; Moad, G.; O’Shea, M. S.
Macromolecules 2006, 39, 5307–5318.
(55) Postma, A.; Davis, T. P.; Evans, R. A.; Li, G.; Moad, G.; O’Shea,
(91) Banishoeib, F.; Fourier, S.; Cleij, T. J.; Lutsen, L.; Vanderzande,
D. J. M. Eur. Phys. J.: Appl. Phys. 2007, 37, 237–240.
(92) Nguyen, L. H.; Cleij, T. J.; Lutsen, L.; Vanderzande, D. J. M. Sol.
Energy Mater. Sol. Cells 2006, 90, 2815–2828.
(93) Dilien, H.; Vandenbergh, J.; Banishoeb, F.; Adriaensens, P.;
Cleij, T. J.; Lutsen, L.; Vanderzande, D. J. M. Macromolecules 2011,
44, 711–718.
M. S. Macromolecules 2006, 39, 5293–5306.
(56) Xu, J.; He, J.; Fan, D.; Tang, W.; Yang, Y. Macromolecules 2006,
39, 3753–3759.
(57) Nejad, E. H.; Castignolles, P.; Gilbert, R. G.; Guillaneuf, Y.
J. Polym. Sci., Part A: Polym. Chem. 2008, 46, 2277–2289.
(58) Albertin, L.; Stenzel, M. H.; Barner-Kowollik, C.; Davis, T. P.
Polymer 2006, 47, 1011–1019.
(94) Henckens, A.; Colladet, K.; Fourier, S.; Cleij, T. J.; Lutsen, L.;
Gelan, J. Macromolecules 2005, 38, 19–26.
(59) Plummer, R.; Goh, Y.-K.; Whittaker, A. K.; Monteiro, M. J.
Macromolecules 2005, 38, 5352–5355.
(95) Henckens, A.; Duyssens, I.; Lutsen, L.; Vanderzande, D. J. M.;
Cleij, T. J. Polymer 2006, 47, 123–131.
(60) Favier, A.; Barner-Kowollik, C.; Davis, T. P.; Stenzel, M. H.
Macromol. Chem. Phys. 2004, 205, 925–936.
(96) Palmaerts, A.;van Haren, M.; Lutsen, L.; Cleij, T. J.; Vanderzande,
D. J. M. Macromolecules 2006, 39, 2438–2440.
(97) Voronkov, M. G.; Panova, G. M.; Timokhina, L. V.; Gromkova,
R. A. Russ. J. Gen. Chem. 2004, 74, 1043–1045.
(61) Chugaev, L. Ber. Dtsch. Chem. Ges. 1899, 32, 3332.
(62) Oconnor, G. L.; Nace, H. R. J. Am. Chem. Soc. 1952, 74,
5454–5459.
(63) Alexander, E. R.; Mudrak, A. J. Am. Chem. Soc. 1950, 72,
(98) DePuy, C. H.; King, R. W. Chem. Rev. 1960, 60, 431–457.
(99) Arita, T.; Buback, M.; Vana, P. Macromolecules 2005, 38,
7935–7943.
3194–3198.
(64) Alexander, E. R.; Mudrak, A. J. Am. Chem. Soc. 1951, 73,
59–62.
(65) Ahmed, S.; Luke A. Baker, L. A.; Grainger, R. S.; Innocenti, P.;
Quevedo, C. E. J. Org. Chem. 2008, 73, 8116–8119.
(66) Liu, Y.; He, J.; Xu, J.; Fan, D.; Tang, W.; Yang, Y. Macro-
molecules 2005, 38, 10332–10335.
(67) Legge, T. M.; Slark, A. T.; Perrier, S. J. Polym. Sci., Part A: Polym.
Chem. 2006, 44, 6980–6987.
8457
dx.doi.org/10.1021/ma201570f |Macromolecules 2011, 44, 8446–8457