Controlled radical polymerization of acrylonitrile in the presence of trithiocarbonates as reversible addition-fragmentation chain transfer agents

Article abstract of DOI:10.1134/S1070427211060231

Specific features of pseudo-living radical polymerization of acrylonitrile in dimethyl sulfoxide solution in the presence of low-molecular-weight and polymeric trithiocarbonates as reversible addition-fragmentation chain transfer agents were studied.

Full text of DOI:10.1134/S1070427211060231

ISSN 1070-4272, Russian Journal of Applied Chemistry, 2011, Vol. 84, No. 6, pp. 1031–1039. © Pleiades Publishing, Ltd., 2011.
Original Russian Text © E.V. Chernikova, Z.A. Poteryaeva, S.S. Belyaev, E.V. Sivtsov, 2011, published in Zhurnal Prikladnoi Khimii, 2011, Vol. 84, No. 6,
pp. 1010–1018.
MACROMOLECULAR CHEMISTRY
AND POLYMERIC MATERIALS
Controlled Radical Polymerization of Acrylonitrile
in the Presence of Trithiocarbonates
as Reversible Addition–Fragmentation Chain Transfer Agents
E. V. Chernikovaa, Z. A. Poteryaevaa, S. S. Belyaeva, and E. V. Sivtsovb
a Lomonosov Moscow State University, Moscow, Russia
b St. Petersburg State Institute of Technology (Technical University), St. Petersburg, Russia
Received November 9, 2010
Abstract—Specific features of pseudo-living radical polymerization of acrylonitrile in dimethyl sulfoxide
solution in the presence of low-molecular-weight and polymeric trithiocarbonates as reversible addition–
fragmentation chain transfer agents were studied.
DOI: 10.1134/S1070427211060231
Polyacrylonitrile (PAN) and copolymers based on it
are important precursors for preparing carbon
materials. Interest in them is caused by their unique
chemical, mechanical, and electrical properties, which
allows these materials to be used in nanocomposites, in
gas holders, and in catalysis [1–4]. Polyacrylonitrile
intended for use as carbon fiber precursor should have
controllable molecular-weight characteristics and
architecture of macromolecules [5]. At the same time,
commercially available PAN and its copolymers are
prepared by classical radical polymerization, which
does not allow control of these parameters. In parti-
cular, the polymers formed are characterized by broad
molecular-weight distribution (MWD) (Mw/Mn > 3);
furthermore, copolymers are compositionally hetero-
geneous. These problems can be solved in principle by
using “living” anionic polymerization, but in this case
a branched polymer is formed, which appreciably
complicates preparation of fibers from PAN [6].
In the past decades, there has been much progress
in the field of controllable, or pseudo-living, radical
polymerization, which also allows preparation of poly-
mers of preset structure with controllable molecular-
weight characteristics but is simpler in implementation
than anionic polymerization [7].
The process consists in introduction into the
reaction medium of stable radicals or reversible
addition–fragmentation chain transfer (RAFT) agents
capable of reversible reaction with macroradicals [8–
10]. In the process, the macromolecules formed
constantly “revive” with the regeneration of macro-
radicals which can participate again in the chain
propagation until the next termination event (reversible
termination or transfer) occurs. The reversibility of
chain termination reactions ensures successive growth
of macromolecules in the course of polymerization,
narrowing of MWD, and formation of compositionally
homogeneous copolymers.
Recently Dong et al. prepared narrow-MWD PAN
with MW 2000–15000 [11]. It should be noted,
however, that synthesis of PAN of higher MW with
narrow MWD (Mw/Mn < 1.5) by controlled radical
polymerization has not been reported.
The goal of this study was to develop a procedure
for controlled synthesis of narrow-MWD PAN of
preset molecular weight by pseudo-living radical
polymerization occurring under RAFT conditions,
based on the results of previous studies [12–15].
EXPERIMENTAL
Acrylonitrile (AN) (99%, Fluka) was distilled at
atmospheric pressure. Dimethyl sulfoxide (DMSO)
(99%, Fluka) was used without additional purification.
Azobis(isobutyronitrile) (AIBN) was recrystallized
1031

1032
CHERNIKOVA et al.
twice from absolute ethanol and vacuum-dried to
constant weight.
liquid phase, after which the carbon disulfide and
unchanghed bromide were distilled off in a vacuum.
The product purity was judged from the absence of
these components (analysis by chromatography–mass
spectrometry). Pure METC is an amber-colored liquid
with a characteristic odor. 1H NMR spectrum (CDCl3),
δ, ppm: 1.6 d, 3.7 s, 4.8 q.
Dibenzyl trithiocarbonate (BTC) was synthesized
by the procedure described in [15]. Sodium sulfide
(5.85 g, 0.045 mol) was dissolved in 50 ml of water,
and 0.30 g of a 75% aqueous solution of methyltri-
butylammonium chloride and 3.54 g (0.046 mol) of
carbon disulfide were added. The mixture was stirred
for 1 h at 35°С. Then 10.08 g (0.08 mol) of benzyl
chloride was added over a period of 15 min, and the
mixture was stirred for 3 h at 50°С and then for 30 min
at 70°С. To bring the reaction to completion, 0.15 g of
methyltributylammonium chloride was added, and the
mixture was stirred for an additional 15 h at room
temperature. After the reaction completion, the
aqueous phase was decanted off, and the yellow mass
obtained was treated with 70 ml of ethanol. The
crystalline product was filtered off, washed with
ethanol, and dried to constant weight. BTC was thus
obtained as a yellow crystalline substance, yield
83.6%. 13C NMR spectrum (CDCl3), δ, ppm: 41.37,
127.60, 128.52, 129.08, 134.75, 222.35.
Samples for RAFT polymerization were prepared
by dissolving the calculated amounts of AIBN and
RAFT agent in a mixture of AN and DMSO, taken in a
preset ratio. The solutions were placed in ampules,
which were sealed off after degassing by repeated
freezing–pumping–thawing. The polymerization was
performed at 60, 75, or 80°С. After the polymerization
completion, the ampules were cooled with liquid
nitrogen and opened; the contents were diluted with
DMSO if necessary, and the product was precipitated
into a tenfold excess of water. Then the product was
filtered off on a glass frit, washed with water, and
vacuum-dried to constant weight; the conversion was
determined gravimetrically.
The molecular-weight characteristics of the
polymers were studied by gel permeation chro-
matography with a GPC-120 chromatograph (Polymer-
Labs). Analysis was performed at 50°С in DMF
containing 0.1 wt % LiBr at a flow rate of 1 ml min–1.
For the separation, we used two PLgel 5 μm MIXED B
columns of total length 600 mm (interval of molecular
weights 500–10000000). The polymer concentration in
LiBr-containing DMF solution was 0.7–1.0 mg ml–1.
The molecular weight was calculated using poly
(methyl methacrylate) references and the Mark–Kuhn–
Houwink coefficients known for PAN and PMMA
from the literature [17].
Bis(2-methoxycarbonyl-2-propyl) trithiocarbonate
(MPTC) R–S–C(=S)–S–R, where R = C(CH3)2·
(COOCH3), was synthesized by the procedure based
on that reported in [16]. A 30-g portion of Ambersep
900 anion-exchange resin (OH form) vacuum-dried at
110°С was added with stirring at room temperature to
150 ml of carbon disulfide acting as both the solvent
and the reactant. The resin immediately turned bright
red, and 10 min later excess (25 g) methyl 2-bromo-
isobutyrate was added dropwise. The reaction mixture
was stirred at room temperature for a week. In the
process, the color changed from red to dark amber.
The resin was separated from the aqueous phase and
washed with a small amount of carbon disulfide. The
solvent and excess bromide were removed in a vacuum
first at room temperature and then at 120°С; in so
doing, the product crystallized. The trithiocarbonate
obtained was purified by double recrystallization from
chloroform. The pure product consisted of bright
yellow needle-like crystals. 1H NMR spectrum
(CDCl3), δ, ppm: 1.6 s, 3.7 s.
The radical polymerization kinetics was studied by
the calorimetric method with a DAK-1-1а differential
automatic calorimeter in the mode of direct recording
of the heat release rate ∂Q/∂t under isothermal
conditions at 60, 75, and 80°С. The working cell was
charged with the ampule containing the reaction
mixture, and the reference cell, with an ampule con-
taining an approximately equal amount of the already
polymerized monomer. The kinetic parameters were
calculated from the calorimetric curves using a
specially developed program. The enthalpy of homo-
polymerization of acrylonitrile ∆H was taken equal to
−77.5 kJ mol–1 [17]. The calorimetric curves were
integrated by the Simpson method. The process was
monitored gravimetrically. The kinetic parameters of
Bis(1-methoxycarbonylethyl) trithiocarbonate
(METC) R–S–C(=S)–S–R, where R = CH(CH3)·
(COOCH3), was synthesized similarly using an excess
(15 g) of the corresponding bromide (methyl 2-
bromopropionate). The synthesis time at room tem-
perature was 30 h. The resin was separated from the
RUSSIAN JOURNAL OF APPLIED CHEMISTRY Vol. 84 No. 6 2011

CONTROLLED RADICAL POLYMERIZATION OF ACRYLONITRILE
(a) (b)
1033
P, %
P, %
1'
2'
t, h
t, h
Fig. 1. Conversion Р as a function of time t for the polymerization of AN in DMSO solution, initiated by AIBN, (1–3) in the
presence of BTC and (1', 2') without it. (a) [AIBN] = [BTC] = 10–2 M; Т = 75°С; AN/DMSO volume ratio: (1) 1/1 and (2) 3/7;
(b) [AIBN] = 10–3, [BTC] = 5 × 10–3 M; AN/DMSO volume ratio: (1) 4/6 and (2, 3) 3/7; Т, °С: (1, 2) 80 and (3) 60.
the polymerization were calculated using programs
developed at the laboratory of polymerization pro-
cesses, Chair of Macromolecular Compounds, Moscow
State University.
agent. In these experiments, we varied the component
ratios and temperature. The results are presented in
Fig. 1 and in the table.
Figure 1a shows that the initial rate of AN
polymerization in the presence and in the absence of
the RAFT agent are close, irrespective of the
monomeric mixture composition. Without BTC, at a
high content of AN in the mixture (curve 1'), after 30%
conversion a small gel effect is observed, which
disappears both on decreasing the initial monomer
concentration (curve 2') and on adding BTC (curve 1).
It is noteworthy that addition of BTC allows higher
conversions to be attained, compared to the classical
radical polymerization of AN. As in the classical
radical polymerization, a decrease in the temperature
and AIBN concentration leads to a decrease in the rate
of the RAFT process (Figs. 1a, 1b).
The thermal effects occurring on heating the
polymers were studied with a DSC 204 differential
scanning calorimeter (Phoenix, Netzsch Geratebau
GmbH, Germany) in a dry atmosphere (air, argon) at a
flow rate of 50 ml min–1 in the temperature range 40–
400°С (heating rate 10 deg min–1). For the measu-
rements, we took a 4–6-mg portion of the polymer,
which was placed in a standard aluminum crucible
without lid. The results were processed using Netzsch
Proteus program.
Polymerization of acrylonitrile in the presence of
low-molecular-weight trithiocarbonates. Polymeriza-
tion of AN in a DMSO solution has a characteristic
feature: PAN is insoluble in its own monomer but
readily soluble in DMSO [17]. Correspondingly, the
choice of the initial ratio AN/DMSO is critical for
performing a homophase process. Another important
factor is the polymerization rate, which is controlled
by the concentrations of the radical initiator (AIBN in
our study) and RAFT agent and by the temperature.
Therefore, it seemed necessary to find optimal
conditions for RAFT polymerization of AN.
It should be noted that the PAN prepared at the
limiting conversion at the AN/DMSO volume ratio of
1/1 showed strong opalescence, irrespective of the
presence of BTC and of the AIBN concentration. At
decreased AN concentration in the reaction mixture
(AN/DMSO volume ratio 4/6, [AIBN] = 10–3 M),
without BTC the sample became turbid with time, and
in the presence of BTC it remained transparent.
Further dilution (AN/DMSO volume ratio 3/7) led to
the formation of transparent PAN at deep conversions,
irrespective of the preparation procedure.
It is known that, in most cases, introduction of
trithiocarbonates only slightly affects the initial rate of
RAFT polymerization as compared to the rate of
classical radical polymerization at similar values of the
initiator concentration and temperature [14, 15, 18].
First we performed experiments with BTC as RAFT
The molecular-weight characteristics of PAN
prepared by classical radical polymerization and in the
presence of BTC are given in the table. Addition of
BTC to the reaction mixture leads to a drastic decrease
RUSSIAN JOURNAL OF APPLIED CHEMISTRY Vol. 84 No. 6 2011

1034
CHERNIKOVA et al.
Molecular-weight characteristics of PAN
[BTC]×103 [AIBN]×103
AN/DMSO
volume ratio
T, °C
Conversion, %
Mn×10–3
Mn/Mw
M
60
10
10
10
5
1
1
1/1
1/1
1/1
3/7
3/7
3/7
3/7
3/7
1/1
1/1
3/7
3/7
3/7
3/7
3/7
3/7
3/7
3.5
24.5
59.2
1.6
28.1
41.7
47.0
22.2
21.5
24.8
25.6
27.6
66.2
22.9
29.4
13.4
15.7
18.0
18.3
22.1
25.0
1.56
1.41
1.35
1.46
1.51
1.43
1.45
1.43
2.87
1.57
2.55
1.57
1.86
1.77
1.78
1.60
1.59
1
1
5
1
3.6
5
1
7.4
5
1
12.1
16.4
68.0
77.1
56.7
66.2
2.1
5
1
75
80
–
10
10
10
10
1
10
–
10
5
5
1
11.8
19.2
45.6
61.7
5
1
5
1
5
1
in the number-average molecular weight Mn of the
polymer and to noticeable narrowing of MWD. As
seen from the table, Mn of the polymer prepared in the
presence of BTC, irrespective of the temperature and
initial concentrations of the components, increases
with the conversion, which is characteristic of pseudo-
living radical processes. The polydispersity coef-
ficients Mw/Mn throughout the process remain lower
than in classical radical polymerization.
to phase separation and, on the other hand, to an
increase in the viscosity of the reaction mixture, so that
the process becomes diffusion-controlled, the rate
constant of quadratic decay of the macroradicals
decreases, and the gel effect appears [19].
In the presence of BTC, the polymerization occurs
by the pseudo-living mechanism, which results not
only in formation of the narrow-MWD polymer of
lower molecular weight, but also in suppression of the
quadratic termination of the propagating radicals [20].
This fact is apparently responsible for elimination of
the above-mentioned phenomena.
Presumably, preservation of the system homo-
geneity in AN polymerization in the presence of BTC
(AN/DMSO volume ratio 4/6), suppression of the gel
effect at AN/DMSO volume ratio of 1/1, and at-
tainment of higher conversions of the monomer in the
presence of BTC are due to different process
mechanisms.
Thus, in our experiments we determined the
optimal conditions for RAFT polymerization of AN:
[AIBN] ≤ 10–3 M, AN/DMSO volume ratio 4/6, 10–3 <
[RAFT] ≤ 10–2 M, Т = 80°С. Under these conditions,
70–80% conversion of the monomer was attained in
approximately 5 h, and the reaction mixture remained
transparent.
In the absence of BTC, the polymerization occurs
by the classical radical mechanism, and the high-
molecular-weight polymer is formed at once. At a high
content of the monomer in the reaction mixture,
accumulation of PAN rapidly leads, on the one hand,
Then we examined the performance of various
trithiocarbonates (BTC, METC, MPTC) in AN poly-
RUSSIAN JOURNAL OF APPLIED CHEMISTRY Vol. 84 No. 6 2011

CONTROLLED RADICAL POLYMERIZATION OF ACRYLONITRILE
(a) (b)
1035
P, %
P, %
Fig. 2. (a) Number-average molecular weight Mn and (b) polydispersity coefficient Mw/Mn as functions of conversion Р in AN
polymerization in DMSO solution, initiated by AIBN, in the presence of RAFT agents. 80°С, AN/DMSO volume ratio 4/6.
Concentration, M: (1) [MPTC] = 5 × 10–3, [AIBN] = 5 × 10–4; (2) [METC] = 5 × 10–3, [AIBN] = 5 × 10–4; (3) [BTC] = 5 × 10–3,
[AIBN] = 1 × 10–3; (4) [BTC] = 5 × 10–3, [AIBN] = 5 × 10–4; and (5) [MPTC] = 1 × 10–3, [AIBN] = 5 × 10–4.
merization under the chosen conditions; the results are
shown in Fig. 2.
It should be noted that the AIBN concentration
(under the conditions of relatively rapid initiator
decomposition, as is the system under consideration)
can affect the molecular-weight characteristics of PAN
only when it is comparable with the concentration of
the RAFT agent. In this case, the chains generated by
the initiator decomposition noticeably contribute to
Mn. As can be seen, at [RAFT]/[AIBN] > 5 the
generation of inactive (“dead”) chains does not play a
noticeable role and does not lead to MWD broadening
(Fig. 2b, dependences 1–4): The polydispersity
coefficients at high conversions are 1.26–1.32. With a
further decrease in the concentration of the RAFT
agent, the polydispersity coefficient of PAN formed at
high conversions increases (Fig. 2b, dependence 5) and
reaches 1.59.
As can be seen, in all the systems studied Mn of the
polymer increases linearly with conversion (Fig. 2a).
The dependence of Mn on the conversion for effective
RAFT agents is described by the equation [21]
q[M]0
Mm,
Mn = MRAFT
+
[RAFT]0 + f[I]0(1 − e−k t
)
i
where [М]0, [I]0, and [RAFT]0 are the initial
concentrations of the monomer, initiator, and RAFT
agent, respectively; МRAFT and Мm are the molecular
weights of the RAFT agent and monomer; q is the
conversion; f is the initiation efficiency; ki is the
initiation rate constant; and t is the polymerization
time.
Polymerization of acrylonitrile in the presence of
polymeric RAFT agents . As the initial RAFT agent,
low-molecular-weight trithiocarbonate, is consumed, a
polymeric RAFT agent is accumulated in the system.
Participating in the formation of an intermediate [15],
is “revives” the macroradicals and ensures occurrence
of the polymerization by the pseudo-living mechanism.
Apparently, the structure of the polymeric RAFT agent
is independent of the structure of the initial
trithiocarbonate. In other words, PAN prepared in the
presence of BTC, METC, or MPTC should have the
same structure and show the same behavior in RAFT
polymerization.
According to Eq. (1), with the AIBN, RAFT agent,
and monomer concentrations being the same, the
conversion dependences of Mn for chemically different
RAFT agents should coincide. However, as seen from
Fig. 2a, straight lines 1, 2, and 4 are parallel to each
other. This fact means that the performance of BTC,
METC, and MPTC in AN polymerization is different.
Using the Mayo [22] and Chong [23] approaches,
we estimated the constant Сt of chain transfer to these
RAFT agents, which is a measure of the performance
of RAFT agents in polymerization [12, 20]: Сt,BTC ~ 2,
Сt,METC ~ 4, and Сt,MPTC ~ 8. Thus, the performance of
RAFT agents decreases in the order MPTC > METC >
BTC.
Previous studies showed that the polymeric RAFT
agent is always by 1–2 orders of magnitude more
RUSSIAN JOURNAL OF APPLIED CHEMISTRY Vol. 84 No. 6 2011

1036
CHERNIKOVA et al.
(a)
(b)
Conversion
Conversion
PANC1
PANC2
M
M
Fig. 3. Normalized GPC curves for PAN prepared by AN polymerization in DMSO solution, initiated by AIBN, in the presence of
RAFT agents at 80°С. (a) AN/DMSO volume ratio 3/7; [PANC1] = 2 × 10–3, [AIBN] = 1 × 10–3 M; (b) AN/DMSO volume ratio
4/6; [PANC2] = 5 × 10–3, [AIBN] = 5 × 10–4 M.
(a)
(b)
P, %
P, %
Fig. 4. (a) Number-average molecular weight Mn and (b) polydispersity coefficient Mw/Mn as functions of conversion Р in AN
polymerization in DMSO solution, initiated by AIBN, in the presence of RAFT agents at 80°С. (1) AN/DMSO volume ratio 3/7,
[PANC1] = 2 × 10–3, [AIBN] = 1 × 10–3 M; (2) AN/DMSO volume ratio 4/6, [PANC2] = 5 × 10–3, [AIBN] = 5 × 10–4 M.
effective than the low-molecular-weight RAFT agent
[14, 15, 21]. This conclusion was checked in the
following experiment. We synthesized two PAN
samples: one in the presence of BTC (PANC1, Mn =
24800, Mw/Mn = 1.59) and the other, in the presence of
MPTC (PANC2, Mn = 11500, Mw/Mn = 1.30), after
which we added them to a fresh portion of a solution
of the monomer and AIBN in DMSO. The results are
shown in Figs. 3 and 4.
which for PANC1 and PANC2 is Сt > 50. With
conversion, the mode in the GPC curves is suc-
cessively shifted toward high molecular weights,
which unambiguously proves the occurrence of the
polymerization by the pseudo-living mechanism.
As seen from Fig. 4a, the quantity Mn linearly
increases with conversion in accordance with Eq. (1)
for both systems. It is important that the portion
intercepted on the ordinate coincides with Mn of the
introduced polymeric RAFT agent; this fact confirms
higher performance of both polymeric RAFT agents
compared to the low-molecular-weight agents (BTC
and MPTC). The polydispersity coefficient for PAN
even at [PANC1]/[AIBN] = 2 remains lower than in
classical radical polymerization. As the PANC
concentration in the initial reaction mixture relative to
As seen from Fig. 3, the reaction product is
characterized by unimodal MWD, i.e., both polymeric
RAFT agents introduced into AN polymerization are
fully consumed already in the initial steps of the
conversion. From the rate of consumption of the
polymeric RAFT agent, we can estimate the lower
boundary of the chain transfer constant Сt,PANC [23],
RUSSIAN JOURNAL OF APPLIED CHEMISTRY Vol. 84 No. 6 2011

CONTROLLED RADICAL POLYMERIZATION OF ACRYLONITRILE
1037
AIBN is increased (Fig. 4b, dependence 2), the frac-
tion of the formed “dead” chains decreases. This leads
to a sharp decrease in the polydispersity coefficient of
the PAN formed at low conversions. However, with an
increase in conversion the polydispersity coefficient
increases, although it remains appreciably lower than
in classical radical polymerization. Apparently, slight
broadening of MWD may be due to the occurrence of
side reactions in the system, e.g., of the chain transfer
to the polymer. The contribution of this reaction
increases with the monomer conversion, i.e., with an
increase in the PAN content in the reaction mixture
[17, 19].
We found that the thermal behavior of the PAN
sample prepared by classical radical polymerization
(Figs. 5a, 5c) appreciably differs, both in air and in an
inert atmosphere, from that of the PAN sample
prepared by the RAFT method (Figs. 5b, 5d). Pre-
sumably, the pyrolysis and oxidation of these PAN
samples occur by different mechanisms.
As seen from Fig. 5a, heating of PAN sample pre-
pared by classical radical polymerization is accom-
panied by a single narrow exothermic peak with a high
enthalpy of the process, 320 ± 1 J g–1, associated with
the cyclization. The process starts at a relatively high
temperature (~280°С) with a maximum at 292°С. This
pattern is caused by strong steric braking of the
reacting nitrile groups. Only at relatively high tem-
peratures the PAN molecules become sufficiently
mobile and the polymer starts to decompose with the
release of active radicals initiating avalanche-like
cyclization throughout the polymer volume [29].
DSC studies showed that the PAN sample (Mw =
33200, Mw/Mn = 1.54) obtained by RAFT poly-
merization with PANC2 is characterized by two glass
transition points: Тg,1 = 72 ± 1 and Тg,2 = 150 ± 1°C.
The commercial PAN sample obtained by classical
radical polymerization and having higher MW (Mw =
190100, Mw/Mn = 2.27) is characterized by higher
Тg,1 = 100 ± 1°C, with the same Тg,2, 150 ± 1°C. It
should be noted that the change in the specific heat
capacity at these relaxation transitions is close for both
samples: 0.15 ± 0.01 (Тg,1) and 0.05 ± 0.01 J g–1 K–1 (Тg,2).
Heating in an inert atmosphere of a PAN sample
synthesized in the presence of a RAFT agent is
characterized by a different pattern (Fig. 5b). In the
DSC curves there are three ill-resolved peaks cor-
responding to different physicochemical processes; the
total heat of this process is 470 ± 1 J g–1. The high-
temperature peak (~272°С) corresponds to cyclization
of the nitrile groups.
It is known that, as the temperature is increased
further to 400°С, irreversible physicochemical pro-
cesses occur in the PAN structure, leading to
transformation of the flexible linear structure of the
polymer into a rigid framework (so-called “ladder”
polymer mainly formed by cyclization of the nitrile
groups [24]). The formation of this framework is the
first step in the formation of the turbostratic structure
of the carbon material in the course of further heat
treatment. Therefore, study of the thermal behavior of
PAN is very important.
When PAN samples are heated in air, carbonization
and oxidation processes are more exothermic. Their
total enthalpy increases and reaches approximately
4720 and 860 J g–1 for PAN prepared by classical and
RAFT polymerization, respectively. Note that the DSC
curves of these samples, recorded in air, also
noticeably differ.
Figure 5c shows that, for PAN prepared by classical
radical polymerization, the peak becomes asymmetric,
appreciably broadened, and shifted toward higher
temperatures (325°С). This result is consistent with the
published data and is associated with the occurrence of
oxidative degradation along with cyclization [25].
It should be noted that the processes that occur in
the course of the heat treatment of PAN strongly
depend on the experiment conditions. When PAN is
heated in an inert atmosphere, the major process is
strongly exothermic cyclization. The process is
accompanied by the appearance in DSC curves of a
narrow exothermic peak with the maximum in the
range 260–270°С depending on the PAN preparation
procedure [25]. When PAN is heated in air, the peak in
the DSC curves is usually broadened and shifted to
higher temperatures, which is attributed to the
occurrence of new processes: oxidation, dehydrogena-
tion, degradation, and cross-linking of PAN polymer
chains [26].
The DSC curve describing the thermal behavior in
air of PAN prepared by RAFT polymerization is more
complex (Fig. 5d), with four ill-resolved peaks. The
peak with a maximum at Т = 272°С corresponds to
cyclization of nitrile groups, and high-temperature peaks
with a smeared maximum at Т = 313–340°С, to oxi-
dative degradation. The nature of the low-tem-perature
peak with a maximum at Т ~240°С is yet unclear.
RUSSIAN JOURNAL OF APPLIED CHEMISTRY Vol. 84 No. 6 2011

1038
CHERNIKOVA et al.
(a)
(b)
exo
exo
T, °C
T, °C
(c)
(d)
exo
exo
T, °C
T, °C
Fig. 5. DSC curves of PAN samples, recorded (a, b) in an argon atmosphere and (c, d) in air. 80°С, heating rate 10 deg min–1, gas
flow rate 50 ml min–1. (Q) Thermal effect and (Т) temperature. (a, c) Commercial PAN sample (Aldrich, Mw = 190100, Mw/Mn =
2.27); (b, d) PAN (Mw = 33200, Mw/Mn = 1.54) prepared by RAFT polymerization of AN in DMSO solution, initiated by AIBN, in
the presence of PANC2 at AN/DMSO volume ratio of 4/6, [AIBN] = 5 × 10–4, [PANC2] = 5 × 10–3 M.
However, the processes occurring in an inert
atmosphere and in air in PAN samples prepared by
RAFT polymerization have certain common features.
In the temperature intervals 220–240 and 250–280°С,
the processes (Figs. 5b, 5d) have close thermal charac-
teristics, suggesting that the structural transformations
of macromolecules are independent of secondary
processes induced by the atmosphere in which the
experiment is performed. It is known [27] that the
presence of other exothermic effects in addition to the
major effect corresponding to the cyclization may be
due to inhomogeneous structure of PAN, leading to
certain delay in cyclization processes in amorphous
and ordered (quasi-crystalline) areas of the polymer.
sulfoxide solution in the presence of three low-
molecular-weight trithiocarbonates acting as reversible
addition–fragmentation
chain
transfer
agents.
Introduction of trithiocarbonates does not affect the
polymerization rate.
(2) Conditions were found for preparing poly-
acrylonitrile with controllable molecular weight (from
10000 to 200000) and relatively narrow molecular-
weight distribution (Mw/Mn = 1.2–1.6).
(3) Irrespective of the performance of the initial
trithiocarbonates in reversible chain transfer, the
polymers isolated from the reaction mixture and added
to a fresh portion of the monomer and initiator cause
the process to occur by the pseudo-living mechanism.
Their performance is higher than that of low-
molecular-weight trithiocarbonates. The process yields
polymers with controllable molecular weight and
relatively narrow molecular-weight distribution.
CONCLUSIONS
(1) Polymerization of acrylonitrile, initiated by
azobis(isobutyronitrile), was performed in a dimethyl
RUSSIAN JOURNAL OF APPLIED CHEMISTRY Vol. 84 No. 6 2011

CONTROLLED RADICAL POLYMERIZATION OF ACRYLONITRILE
1039
Macromol. Chem. Phys., 2003, vol. 204, pp. 1160–1168.
(4) Introduction of the trithiocarbonate group into
the polymer chain affects the thermal behavior of
acrylonitrile in the course of heating both in an inert
atmosphere and in air.
14. Chernikova, E.V., Terpugova, P.S., Trifilov, M.Yu.,
et al., Vysokomol. Soedin., Ser. A, 2009, vol. 51, no. 6,
pp. 983–991.
15. Chernikova, E.V., Terpugova, P.S., Garina, E.S., and
Golubev, V.B., Vysokomol. Soedin., Ser. A, 2007,
vol. 49, no. 2, pp. 208–217.
REFERENCES
16. Bahman, T. and Kiasat, A.R., J. Chem. Res. (S), 1998,
1. Encyclopedia of Polymer Science and Technology,
Kroschwitz, J.I., Ed., Hoboken: Wiley, 2003, 3rd ed.,
p. 456.
pp. 454–455.
17. Polymer Handbook, Brandrup, J., Immergut, E.H., and
Crulue, E.A., Eds., New York: Wiley, 1999.
2. Shchukin, D.G., Sukhorukov, G.B., and Mohwald, H.,
Angew. Chem., 2003, vol. 42, pp. 4472–4475.
18. Chernikova, E.V., Pseudo-Living Radical Homo- and
Copolymerization by the Reversible Chain Transfer
Mechanism, Doctoral Dissertation, Moscow, 2010.
3. Che, G., Lakshmi, B.B., Fiser, E.R., and Martin, C.R.,
Nature, 1998, vol. 393, pp. 346–349.
19. Odian, G., Principles of Polymerization, Hoboken, New
4. Morris, C.A., Anderson, M.L., Stroud, R.M., et al.,
Jersey: Wiley–Interscience, 2004, 4th ed.
Science, 1999, vol. 284, pp. 622–624.
20. Handbook of RAFT Polymerization, Barner-Kowollik, C.,
5. Hou, C., Qu, R., Liu, J., and Wang, C., J. Appl. Polym.
Ed., Weinheim: Wiley–VCH, 2008.
Sci., 2006, vol. 100, pp. 4463–4492.
21. Chernikova, E.V., Tarasenko, A.V., Garina, E.S., and
Golubev, V.B., Vysokomol. Soedin., Ser. A, 2006,
vol. 48, no. 10, pp. 1787–1794.
6. Ono, H., Hisatani, K., and Kamide, K., Polym. J., 1993,
vol. 25, pp. 245–265.
7. Handbook of Radical Polymerization, Matyjaszewski, K.
22. Mayo, F.R., J. Am. Chem. Soc., 1943, vol. 65, pp. 2324–
and Davis, T.P., Eds., Hoboken: Wiley, 2002.
2329.
8. Hawker, C.J., Acc. Chem. Res., 1997, vol. 30, no. 9,
23. Chiefari, J., Chong, Y.K., Ecrole, F., et al., Macro-
pp. 373–382.
molecules, 1998, vol. 31, no. 16, pp. 5559–5562.
9. Matyjaszewski, K. and Xia, J., Chem. Rev., 2001, vol. 101,
24. Morgan, P., Carbon Fibers and Their Composites, New
pp. 2921–2990.
York: Taylor and Francis, 2005.
10. Mayadunne, R.T.A., Rizzardo, E., Chiefari, J., et al.,
Macromolecules, 1999, vol. 32, pp. 6977–6986.
25. Bahrami, S.H., Bajaj, P., and Sen, K., J. Appl. Polym.
Sci., 2003, vol. 88, pp. 685–698.
11. Dong, H., Tang, W., and Matyjaszewski, K.,
Macromolecules, 2007, vol. 40, no. 9, pp. 2974–2977.
26. Fitzer, E., Frohs, W., and Heine, M., Carbon, 1986,
vol. 24, no. 4, pp. 387–395.
12. Moad, G., Rizzardo, E., and Thang, S.H., Polymer,
2008, vol. 49, pp. 1079–1131.
27. Gupta, A. and Harisson, I.R., Carbon, 1996, vol. 34,
pp. 1427–1455.
13. Stenzel, M.H., Cummins, L., Roberts, G.E., et al.,
RUSSIAN JOURNAL OF APPLIED CHEMISTRY Vol. 84 No. 6 2011