Anionic ring-opening polymerization of alkyl 1- cyanocyclopropanecarboxylates

Article abstract of DOI:10.1021/ma0484025

Poly(alkyl 1-cyanotrimethylene-1-carboxylate)s (CH₂CH ₂C(CN)(COOR))_n are the next higher chain homologues of poly(α-cyanoacrylate)s. They were synthesized via the anionic ring-opening polymerization of the corresponding alkyl 1-cyanocyclopropanecarboxylate monomers initiated with thiophenolate salts (PhSM, with M = Li, Na, K, and NBu₄) and at temperatures above 50°C. Precipitation of the growing polymer chains at an early stage during the polymerization did not prevent further propagation, probably via polymerization in the solid or at the solid surface. The propagating cyanoacetate carbanion is very stable in a normal air atmosphere, surviving for long periods of time in the absence of strong acids. Although monodisperse polymers were obtained in most experiments (M _w/M_n < 1.10), strong experimental evidence for a living polymerization could only be obtained when K⁺ and NBu ₄⁺ counterions were used. Nonquantitative initiations characterized polymerizations initiated by PhSLi and PhSNa. Attempted polymerizations of ethyl 1-cyanocyclobutanecarboxylate failed under similar or slightly more energetic experimental conditions, with the thiophenolate initiator attacking the pendant ester rather than the cyclic methylene group.

Full text of DOI:10.1021/ma0484025

4588
Macromolecules 2005, 38, 4588-4594
Anionic Ring-opening Polymerization of Alkyl
1
-Cyanocyclopropanecarboxylates
Lawino C. Kagumba and Jacques Penelle*,†
Department of Polymer Science and Engineering, University of Massachusetts,
Amherst, Massachusetts 01003-4530
Received August 3, 2004; Revised Manuscript Received April 1, 2005
2 2 n
ABSTRACT: Poly(alkyl 1-cyanotrimethylene-1-carboxylate)s (CH CH C(CN)(COOR)) are the next higher
chain homologues of poly(R-cyanoacrylate)s. They were synthesized via the anionic ring-opening
polymerization of the corresponding alkyl 1-cyanocyclopropanecarboxylate monomers initiated with
4
thiophenolate salts (PhSM, with M ) Li, Na, K, and NBu ) and at temperatures above 50 °C. Precipitation
of the growing polymer chains at an early stage during the polymerization did not prevent further
propagation, probably via polymerization in the solid or at the solid surface. The propagating cyanoacetate
carbanion is very stable in a normal air atmosphere, surviving for long periods of time in the absence of
strong acids. Although monodisperse polymers were obtained in most experiments (Mh
w
/Mh
n
< 1.10), strong
4
+
+
experimental evidence for a living polymerization could only be obtained when K and NBu
counterions
were used. Nonquantitative initiations characterized polymerizations initiated by PhSLi and PhSNa.
Attempted polymerizations of ethyl 1-cyanocyclobutanecarboxylate failed under similar or slightly more
energetic experimental conditions, with the thiophenolate initiator attacking the pendant ester rather
than the cyclic methylene group.
Introduction
Scheme 1. Structures of the Alkyl
1
-Cyanocyclopanecarboxylates Monomers 1a-d (R )
We recently reported that cyclopropanes activated by
two ester groups placed on a geminal position of the
three-membered carbon ring can readily polymerize via
an anionic ring-opening polymerization mechanism.1,2
Under controlled conditions, neither termination nor
chain-transfer reactions were observed, and features
typical of a living polymerization were obtained up to
ethyl (a), isopropyl (b), n-butyl (c), and n-octyl (d))
Investigated in This Study and of the Corresponding
Polymers
2
temperatures as high as 200 °C. The increased reactiv-
ity compared to normal, unactivated cyclopropanes and
the high stability of the propagating carbanions both
result from the presence of the two esters acting as
strong electron-withdrawing and carbanion-stabilizing
substituents.
In this report, we extend the initial concept of using
activated 1,1-disubstituted cyclopropanes as precursors
to carbon-chain polymers substituted by polar substit-
uents on every third carbon along the backbone. The
effect of a pair of ester- and nitrile-activating substit-
uents on the reactivity of cyclopropane and cyclobutane
monomers is examined as an analogy to well-known
highly electrophilic vinyl monomers such as R-cyanoacry-
lates (super glue). The structure of the investigated
monomers and resulting polymers is depicted in Scheme
and 1,3-dibromopropane (99%) were purchased from Aldrich
and used without further purification. Isopropyl cyanoacetate
(
(
purum, g98.0%) and tetrabutylammonium thiophenolate
>97%) were obtained from Fluka, and anhydrous potassium
carbonate (p.a.) from Acros. Dimethyl sulfoxide (DMSO) used
for the polymerization was purified by distillation under
vacuum (40 °C/2 mmHg), discarding the first and last 25%,
and then drying over 4 Å molecular sieves. All other solvents
were reagent grade and used without further purification.
Synthesis of Alkyl 1-Cyanocyclopropanecarboxylates
1
. Procedures developed to synthesize the required
ultrapure monomers and methodologies to accurately
characterize these polymers with unusual structures are
presented first. The influence of several polymerization
conditions (in particular, the temperature and the
nature of initiators and counterions) on the polymeri-
zation is described and discussed.
1a-d. A heterogeneous mixture of alkyl cyanoacetate (0.1 mol
equiv), 1,2-dibromoethane (28 g, 0.15 mol), potassium carbon-
ate (40 g, 0.29 mol), and 80 mL DMSO was vigorously stirred
using a mechanical stirrer for 24 h at room temperature. Water
(
250 mL) was added to the reaction mixture, and the product
Experimental Section
was extracted with 3 × 150 mL portions of diethyl ether. The
combined ether layers were concentrated by evaporation,
Materials. Ethyl, butyl, and octyl cyanoacetates (98+%,
5%, and 99% purity, respectively), 1,2-dibromoethane (99%),
4
washed with water (30 mL), and dried overnight over MgSO .
9
The crude products were distilled under vacuum to yield the
corresponding cyclopropane monomers 1a-d. Yields, boiling
*
Author to whom correspondence should be addressed. E-
3
points, and spectroscopic data (300 MHz NMR, CDCl ) are
mail: penelle@mail.pse.umass.edu or penelle@glvt-cnrs.fr.
†
provided below for each monomer.
Current address: Laboratoire de Recherche sur les Polym e` res,
1
CNRS-Universit e´ Paris 12, 2-8 rue H. Dunant, F-94320 Thiais,
France.
1a: yield 69%, bp 45 °C/0.4 mmHg. H NMR: δ 1.34 (t, CH
3
),
1
3
1.59-1.72 (m, ring 2CH
2
), 4.27 (q, OCH
2
). C NMR: δ 13.3
1
0.1021/ma0484025 CCC: $30.25 © 2005 American Chemical Society
Published on Web 05/04/2005

Macromolecules, Vol. 38, No. 11, 2005
Polymerization of Alkyl 1-Cyanocyclopropanecarboxylates 4589
(
ring CH
2
), 14.1 (CH
3
), 19.0 (Cquart.), 63.0 (OCH
2
), 118.7 (CN).
Scheme 2. Synthesis of Alkyl
1-Cyanocyclopanecarboxylates 1a-d (R ) ethyl (a),
isopropyl (b), n-butyl (c), and n-octyl (d))
-
1
IR (liquid film, cm ): 3115 (cyclopropyl CH
2
stretch), 2986,
2
9
1
941, 2250 (CN), 1737 (CdO), 1371, 1312, 1279, 1190, 1025,
7 9 2
72, 859, 747. Anal. Calcd for C H NO (139.16): C, 60.41; H,
0.06; N, 6.53. Found: C, 60.16; H, 10.00; N, 6.60.
1
1
b: yield 54%, bp 45 °C/0.3 mmHg. H NMR: δ 1.31 (t,
3
1
3
2
CH
), 1.58-1.71 (m, ring 2CH
2
), 5.1 (m, OCH). C NMR: δ
1
3.5 (ring CH
2
), 21.6 (CH ), 19.1 (Cquart.), 71.0 (OCH), 118.8
3
1
-
(
CN). IR (liquid film, cm ): 3121 (cyclopropyl CH
2
stretch),
Measurements. H NMR and 13C NMR spectra for the
1
2
1
986, 2940, 2248 (CN), 1736 (CdO), 1361, 1309, 1281, 1190,
monomers were recorded on a Bruker DPX 300 spectrometer
8 2
106, 977, 933, 899, 749. Anal. Calcd for C H11NO (153.19):
in CDCl
3
(at room temperature) and a Bruker AMX-2 500
C, 62.72; H, 9.14; N, 7.14. Found: C, 62.50; H, 9.04; N, 7.32.
1
spectrometer in DMSO-d at 115 °C for the polymers. IR
6
1
c: yield 51%, bp 60 °C/0.3 mmHg. H NMR: δ 0.93 (t, CH
3
),
), 4.20
), 15.3 (CH
), 118.7 (CN).
2
stretch), 2962,
spectra were recorded on a BioRad FTS 175C FT-IR spectrom-
eter. Elemental analysis was performed at the Micro Analysis
Laboratory of the University of Massachusetts, Amherst. Gel
permeation chromatography (GPC) analyses were obtained
using a Polymer Laboratory PL-GPC ultrahigh-temperature
chromatographic system with a Hewlett-Packard 1100 series
isocratic pump and two PLgel 5 µm MIXED-D columns (linear
range of molecular weights: 200-400000; column efficiency
1
(
CH
.42 (m, CH
2
), 1.58-1.74 (m, ring 2CH
2
and OCH
), 13.6 (CH
O), 67.0 (OCH
2 2
CH
1
3
t, OCH
2
). C NMR: δ 13.3 (ring CH
2
3
2
-
3
), 19.0 (Cquart.), 30.4 (CH CH
2
2
2
-
1
IR (liquid film, cm ): 3115 (cyclopropyl CH
2
9
934, 2879, 2249 (CN), 1740 (CdO), 1466, 1311, 1184, 1059,
36, 747. Anal. Calcd for C H13NO (167.23): C, 64.64; H, 8.38;
9 2
N, 7.85. Found: C, 64.50; H, 8.36; N, 8.02.
1
1
d: yield 31%, bp 98 °C/0.6 mmHg. H NMR: δ 0.89 (t, CH
3
),
),
-
1
>
50000 plates m ). Dimethylformamide (DMF) was used as
1
.24-1.44 (m, 5CH
2
), 1.57-1.75 (m, ring 2CH
2
and alkyl CH
2
-1
the eluent at 100 °C at a flow rate of 1 mL min . The columns
were calibrated with polystyrene standards.
1
3
4
.19 (t, OCH
CH ), 19.0 (Cquart.), 22.6 (CH
), 31.7 (CH CH O), 66.7 (OCH
stretch), 2960, 2928, 2857,
249 (CN), 1740 (CdO), 1468, 1312,1176, 944, 747. Anal. Calcd
for C13 (223.34): C, 69.91; H, 9.50; N, 6.27. Found: C,
9.85; H, 9.64; N, 6.29.
2
). C NMR: δ 13.3 (ring CH
2
), 14.0 (CH
2
), 15.3
(
CH
CH
2
3
2
), 25.7 (CH
2 2
), 28.4 (CH ), 29.1
(
2
2
2
2
), 118.7 (CN). IR (liquid
Results and Discussion
-
1
film, cm ): 3120 (cyclopropyl CH
2
2
Synthesis of Monomers 1a-d. A procedure initially
4
H
21NO
2
developed by Zefirov et al. and previously identified in
6
our group as particularly effective for the synthesis of
Synthesis of Ethyl 1-Cyanocyclobutanecarboxylate 2.
1
ultrapure cyclopropane-1,1-dicarboxylates was used for
The same general procedure as described above for 1a-d was
the synthesis of ethyl 1-cyano-cyclopropanecarboxylate
followed using 1,3-dibromopropane instead of 1,2-dibromo-
1
1
a. The same procedure was extended to other alkyl
1
ethane. Yield: 8%, bp 96-100 °C/ 1 mmHg. H NMR: δ 1.34
-cyanocyclopropanecarboxylate monomers with isopro-
(
t, CH
butyl 2CH
153.19): C, 62.72; H, 7.25; N, 9.14. Found: C, 62.51; H, 7.23;
N, 9.11.
3
), 2.09-2.36 (m, cyclobutyl CH
2
), 2.58-2.77 (m, cyclo-
pyl (1b), n-butyl (1c), and n-octyl (1d) groups on the
ester substituent (Scheme 2). Vigorous mixing to finely
disperse the solid potassium carbonate into the hetero-
geneous mixture was achieved by mechanical stirring
and proved to be crucial to the success of the experi-
ments. The monomers obtained were of high purity as
2
), 4.27 (q, 2H, OCH
2
). Anal. Calcd for C H11NO
8 2
(
Synthesis of Thiophenolate Initiators. Potassium and
sodium thiophenolate initiators were synthesized according to
1
a reported procedure. Lithium thiophenolate was obtained
3
1
using a slight modification of a previously reported procedure:
indicated by high-field H NMR and elemental analysis,
-
1
a 1.3 mol L solution of n-BuLi (4.2 mL) in cyclohexane was
slowly added to a solution of diphenyl disulfide (1.2 g, 0.005
mol) in 30 mL dry hexane at room temperature. The white
precipitate was washed several times with hexane and dried
in vacuo at 50 °C (0.9 mmHg). The final product was stored
with no residual starting reagents (cyanoacetates) that
could act as chain-transfer agents during anionic po-
lymerization. All monomers are colorless liquids and
stable at room temperature.
Anionic Polymerization of 1a-d. Freshly synthe-
sized sodium thiophenolate initiates the ring-opening
polymerization of dialkyl cyclopropane-1,1-dicarboxy-
lates.1 In addition to their thermal stability and ease
of drying, the use of soft bases such as thiophenolates
of alkali metals increases the selectivity of the nucleo-
philic attack on the cyclopropane ring and prevents
attack on the carbonyl of the ester. Other traditional
anionic initiators such as organolithium compounds or
Grignard reagents are hard bases that preferentially
attack the carbonyl site on the ester and should be
under argon. Yield: 1.05 g (88%). Anal. Calcd for C
O)0.2 (119.65): C, 60.20; H, 4.55; S, 26.78; Li, 5.80. Found:
C, 60.56; H, 4.94; S, 24.60; Li, 5.90.
6 5
H SLi-
(
H
2
,2
Polymerization Procedure. Potassium and sodium thio-
phenolate initiators were dried just before use (200 °C/0.9
mmHg) using a B u¨ chi Kugelrohr apparatus. While purging
with nitrogen, weighed amounts of the initiator were dissolved
in a specific volume of DMSO. The cyclopropane monomer was
added to the initiator solution, and the polymerization tube
was placed in an oil bath at 30 °C or 60 °C for a specified time.
The polymer was precipitated in methanol, then filtered and
washed several times with water and acetone. The polymers
were dried in vacuo at 60 °C for 24 h.
5
avoided as initiators. The assumption was made at the
beginning of this study that thiophenolates would also
efficiently initiate the polymerization of cyclopropanes
A typical example is as follows: Sodium thiophenolate
0.018 g, 0.136 mmol) was dissolved under nitrogen in 5.0 mL
(
1
a-d, and a series of polymerizations were carried out
of dry DMSO. Ethyl 1-cyanocyclopropanecarboxylate 1a (0.500
g, 3.59 mmol) was added to this solution at room temperature
and under nitrogen. The polymerization tube was placed in
an oil bath at 60 °C for 15 min. The polymer was obtained by
precipitation into 50 mL of methanol, filtered, washed with
small aliquots of water (3× ∼50 mL) and acetone (3× ∼50 mL),
and then dried under vacuum at 60 °C for 1 day. Yield: 0.1049
1
as shown in Table 1. Analysis of the structure by H
and 13C NMR of the obtained powders established that
the polymerization proceeded via a ring-opening mech-
anism, yielding polymers with cyanoester substituents
on every third carbon (Table 2).
Structural Characterization of Poly(1a-d). The
polymers were all isolated as white powders after
precipitation in methanol. Poly(1a-c) were only soluble
in highly polar solvents such as DMSO and DMF at
temperatures above 100 °C. Poly(1d) with a longer alkyl
side chain (n-octyl) was significantly more soluble,
g (21%). GPC (DMF, 100 °C, polystyrene calibration): Mh
n
1
)
3
1
6
.9 × 10 , Mh
w
/Mh
n
) 1.06. H NMR: δ 4.3 (broad quartet, 2 H,
1
COO-CH
2
), 1.9-2.2 (broad signal, 4 H, backbone CH
2
), 1.3
). C NMR: δ 167.7 (COO),
), 49.0 (C(CN)COOEt), 32.4 (back-
-CH ).
1
13
(
2 3
broad triplet, 3 H, COOCH -CH
1
18.3 (CN), 63.0 (COO-CH
2
bone CH
2
), 14.1 (COOCH
2
3

4590 Kagumba and Penelle
Macromolecules, Vol. 38, No. 11, 2005
Table 1. Ring Opening Polymerization of Monomers
1a-d
run
no. monomer
time yield
Mh n
(GPC)
initiatora
(h)
(%)
b
Mh w/ Mh nb
1
2
3
4
5
6
7
8
9
0
1
2
3
1a
1a
1b
1c
1d
1a
1a
1b
1a
1b
1a
1b
1c
PhSK
0.25
0.25
0.25
0.25
0.25
1.0
39
21
7000
6900
1.07
1.06
1.06
1.09
1.07
1.05
1.06
1.07
1.06
1.15
1.44
1.20
1.02
PhSNa
PhSNa
PhSNa
PhSNa
PhSK
34
6800
10
4400
4600
21
84
12400
13400
16100
15000
21900
8600
PhSNa
PhSNa
1.0
64
1.0
72
PhSN(Bu)4 24
100
100
100
88
1
1
1
1
PhSN(Bu)4 24
c
NMP
NMP
NMP
24
24
24
c
c
18200
5300
81
a
3
.59 mmol monomer in 0.3 mL DMSO, 3.8 mol % initiator, 60
°
C. b Determined in DMF at 100 °C, calibrated with polystyrene
standards. N-methyl pyrrolidine.
c
dissolving in toluene, THF, and chlorinated solvents at
room temperature. However, higher molecular weight
samples of poly(1d) (>3000) required moderate heating
(30-60 °C) to dissolve in the same solvents. When
solutions of poly(1a-d) in DMSO at 100 °C were cooled
to ambient temperature, the polymers did not precipi-
tate out of solution, but formed a clear, homogeneous
gel.
1
13
Figure 1. H NMR and C NMR spectrum in DMSO-d
6
of a
typical sample of poly(1a) (poly(ethyl 1-cyanotrimethylenecar-
boxylate)) (GPC: Mh ) 7 × 10
3
n
w n
, Mh / Mh ) 1.06).
The polymers were characterized by IR spectroscopy
1
13
and high-temperature H NMR and C NMR. NMR
1
13
spectra of poly(1a) ( H NMR and C NMR in DMSO-
d6 at 100 °C) are provided in Figure 1 as an illustration
of the main features observed for the entire family of
ester-substituted poly(1a-d). Full data corresponding
to the individual polymers are summarized in Table 2.
The observed signals are fully consistent with the
polymer structure expected of the ring-opened product,
with the aromatic signal attributable to the phenylthiyl
group from the thiophenolate initiators visible. The
multiple peaks obtained for the backbone methylenes
(
from 1.9 to 2.2) produces a more complicated pattern
than the single signal previously observed for poly-
Figure 2. IR spectrum of poly(ethyl 1-cyanotrimethylenecar-
boxylate) (poly(1a)); the two curves in the inset correspond to
1
(
cyclopropane-1,1-dicarboxylate)s. This is due to the
(
a) the absorption arising from the propagating end-group
presence of quaternary C(CN)COOR stereocenters on
every third carbon along the backbone and the resulting
absence of a plane of symmetry along that backbone.
The presence of complex second-order AA′BB′ structures
for the CH2CH2 subunits at the diad level, with the
possibility of meso and dl arrangements, yields a
complex pattern for the methylene units in the back-
bone. A complete analysis of the tacticity of these
polymers on the basis of structurally related model
compounds will be discussed in a forthcoming publica-
tion.
-
C(CN)(COOEt)- in a “living” polymer isolated in the regular
way, i.e., the absence of any strong acid and (b) after reaction
with trifluoroacetic acid.
strong peak at 1190 cm- in the monomer spectra (C-C
stretching in the cyclopropane ring) has completely
disappeared, providing further evidence that nucleo-
philic attack on the ester or cyano pending groups does
not compete with the ring-opening reaction under the
experimental conditions investigated in this study.
Other spectroscopic features of interest that can be
observed by IR will be discussed in the next section.
Determination of Absolute Molecular Weights.
Poly(1a-c) are not soluble at room temperature in
1
The main IR bands of the polymers include a nitrile
-1
-1
stretching at 2249 cm , a CdO stretching at 1740 cm
and a C-O stretching at 1222 cm (Figure 2). The
,
-
1
Table 2. H NMR and 13C NMR Data (δ in ppm)a for Poly(1a-d)
1
1H NMR
13C NMR
backbone
ester
backbone CH2
R (ester)
CH2
CN
C
COO
R
1
1
1
1
a
b
c
1.9-2.2
1.9-2.2
1.9-2.2
1.9-2.2
1.3, 4.3
1.3, 5.1
32.4
32.0
32.3
32.2
118.3
118.8
118.7
118.7
49.0
49.0
49.1
49.1
167.7
167.2
167.8
167.8
14.1; 63.0
21.6; 71.0
0.9; 1.4; 1.6; 4.2
13.6; 15.3; 30.4; 67.0
d
0.9; 1.2-1.4 (broad), 1.6, 4.2
14.0; 15.3; 22.6; 25.7; 28.4; 29.1; 31.7; 66.7
a
TMS reference, DMSO-d6 at 100 °C. b Poly(1a-d) initiated with PhSNa.

Macromolecules, Vol. 38, No. 11, 2005
Polymerization of Alkyl 1-Cyanocyclopropanecarboxylates 4591
Figure 3. Typical GPC chromatogram of poly(ethyl 1-cyano-
trimethylenecarboxylate) (poly(1a)) before treatment with
trifluoroacetic acid.
organic solvents such as THF, chloroform, or even DMF.
As a result, the molecular weights were initially ob-
tained by GPC analysis in DMF at 100°C using a
calibration based on polystyrene standards. GPC chro-
matograms of the polymers characterized under these
conditions showed either single or bimodal distributions,
without clear trends between the type of distribution
observed and the polymerization conditions used to
synthesize the polymer sample. The two peaks in the
distribution were usually well separated, and both had
narrow molecular weight distributions ( Mh w/ Mh n < 1.10).
A typical example of a GPC trace for a sample of poly-
n
Figure 4. Comparison of Mh values obtained from NMR and
GPC for poly(ethyl 1-cyanotrimethylenecarboxylate) (poly(1a)).
such as those of polystyrene or poly(methyl methacry-
late) must be carefully protected against agents such
as water, oxygen, alcohols, and carbon dioxide, but
cyanoacetate carbanions are highly stabilized by two
electron-withdrawing groups and, as a result, are
extremely weak C-H bases (pK ) 16.4 in DMSO at
a
25 °C, in contrast to a pK of 32 for H O under the same
a
2
7
conditions ). The stability of the propagating cyano-
acetate carbanion during the polymerization of R-cy-
anoacrylates (super glue) has been reported.8 Weak
acids such as acetic acid only retard the polymerization
of R-cyanoacrylates, while the rapid termination of fully
initiated chains requires strong acids such as p-tolu-
enesulfonic acid.
,9
(1a) displaying a bimodal distribution is shown in
Figure 3. In this example, peak A corresponds to the
4
lower molecular weight fraction ( Mh n ) 1.8 × 10 ), while
peak B corresponds to a much higher molecular weight
5
(
Mh n ) 2.6 × 10 ). For samples displaying a bimodal
distribution, the area under peak B (high molecular
weights) was typically 3-5% of the area under peak A
The narrow distribution of molecular weights ob-
served for peak B in the GPC chromatograms suggests
that the obtained aggregates have a relatively well-
defined stoichiometry. The exact aggregation numbers
could not be determined, however, because of the lack
of a suitable calibration curve for these starlike struc-
tures and the possibility of nonspecific interactions
between the chromatography columns and the ionic core
of the aggregates (analogous to the behavior observed
for inverse micelles made of ionic macrosurfactants).
As a way to calibrate results obtained by GPC,
molecular weight analysis of the polymers was also
(
low molecular weight).
The very high molecular weights obtained for peak
B is difficult to reconcile with the chemistry and
experimental conditions used in the above experiments.
Among other hypotheses to explain this behavior, it was
suspected that the high molecular weight fraction was
due to aggregation of a few polymer chains driven by
the formation of ionic bridges between persistent
(-)
cyanoacetate carbanions -C(CN)(COOEt)
possibly
present as end groups on the polymers. To check this
hypothesis, polymer samples displaying a bimodal
distribution were redissolved in DMF and treated with
trifluoroacetic acid (rather than the methanol used in
the first precipitation step). Subsequent analysis by
GPC showed the area of peak B diminished to less than
1
performed by end-group analysis via H NMR spectros-
copy. The peak area of the signal attributable to the
aromatic protons of the PhS end group (7.2 ppm) was
compared to the methylene protons (OCH , 4.2 ppm) of
2
the ester group on the polymer (Figure 1). The absolute
1
% in relation to peak A, which provided convincing
degree of polymerization using this method was com-
pared to the apparent degree of polymerization obtained
evidence that the initial hypothesis was indeed correct.
The persistence of the propagating carbanions on
some of the polymer chains at room temperature, even
when no special storage precautions against humidity
are taken and after precipitation was carried out in a
protic solvent such as methanol, was further demon-
strated by IR spectroscopy. IR spectra of a poly(1a)
sample obtained during the polymerization prior to
termination by treatment with a strong acid is shown
in an inset in Figure 2 (curve a). The spectrum contains
by GPC. Mh values obtained by NMR for poly(1a) were
n
plotted against those obtained by GPC as shown in
Figure 4. The slope of the straight line obtained by this
methodology indicates that GPC overestimates the
actual molecular weights by a factor of 3.
NMR spectroscopy of the polymers initiated with the
thiophenolate initiators supports clean initiation and
propagation steps, with no evidence for side reactions
such as a Krapcho reaction (nucleophile-assisted decar-
-
1
10
a band at 2137 cm that fully disappears upon termi-
nation with a strong acid (curve b) and can be correlated
to a -C(CN)(COOEt)( moiety whose assignment had
been described in the literature on the basis of the IR
spectrum of the ethyl cyanoacetate tetrabutylammo-
boxylation) or backbiting reactions (Scheme 3).
Kinetic Analysis of 1a Polymerization. The po-
-)
+
lymerization of 1a initiated by potassium (K ), sodium
+
+
+
(Na ), lithium (Li ), and tetra-n-butylammonium (NBu )
4
thiophenolate was analyzed kinetically, first at 60 °C.
nium salt (absorption at 2140 cm- ). This 2137 cm
1
6
-1
Experimental results are shown in Figures 5 and 6. In
these figures, individual data points correspond to
separate experiments. The degrees of polymerization
were obtained by high-temperature SEC using a poly-
styrene calibration, and corrected to obtain absolute xj n
according to the relationship provided in Figure 4.
C≡N stretching band was also visible for a sample of
1
a polymerized under exposure to atmospheric oxygen,
carbon dioxide, and moisture.
The high stability of the propagating carbanion is not
completely unexpected. Typical propagating carbanions

4592 Kagumba and Penelle
Macromolecules, Vol. 38, No. 11, 2005
Table 3. Polymerization of 1a at Varying Temperatures
rn uo n. monomer initiatora (h)
time temp yield
(°C)
(%) Mh
(GPC)b Mh
/Mh
w n
b
n
1
2
3
1a
1a
1a
PhSNa 0.25
PhSNa 0.25 100
PhSNa 0.25 120
60
c
12
72
8300
17900
1.08
1.07
a
.59 mmol 1a in 5 mL DMSO, 3.8 mol % PhSNa. b Determined
3
c
in DMF at 100 °C, calibrated with polystyrene standards. No
polymer isolated via precipitation
hypothesis is correct; for all four initiators, the molec-
ular weight distributions of the entire polymer sample
Figure 5. Dependence of the conversion (%) with time for the
polymerization of ethyl 1-cyanocyclopropanecarboxylate 1a
(
combined soluble and insoluble fractions) is indeed very
narrow (PDI ) 1.02-1.07) over the entire conversion
range. This is also the case for the slow, PhSLi-initiated
polymerization for which a PDI of 1.06 was observed at
full conversion (24 h reaction time). Dialkyl cyclopro-
pane-1,1-dicarboxylates, the diester equivalent of the
cyanoester cyclopropane monomers investigated here,
have previously been shown to display the opposite
behavior, the early precipitation of the propagating
polymer chains to a semicrystalline solid preventing any
further propagation.1 In contrast, there are many
indications in the literature that R-cyanoacrylates can
polymerize anionically from a solid polymer mass as
seen in vapor deposition polymerizations or when used
-
+
with thiophenolate initiators: PhS X [X ) K ([), Na (2), Li
(
9), (NBu)
4
(b)]; 60 °C, 3.8 mol % initiator, 3.59 mmol 1a in
0
.3 mL DMSO.
,2
for the fast development of latent fingerprints from
monomer vapors.1
1,12
It thus appears on that basis that
our experimental observations are not completely un-
foreseen.
Figure 6. Evolution of the degree of polymerization with
conversion for the polymerization of ethyl 1-cyanocyclopro-
-
+
panecarboxylate 1a using PhS X [X ) K ([), Na(2), Li (9),
In the cases of PhSK and PhSNBu4, the molecular
weight increases linearly with conversion, reaching the
theoretical limit at full conversion (Figure 6). These
data, combined with the very low Mh w/ Mh n observed for
each sample, strongly suggests that the polymerization
is living as previously observed for the dialkyl cyclo-
propane-1,1-dicarboxylate monomers. Results summa-
rized in Figure 6 for PhSLi and PhSNa display some
significant upper deviation from the theoretical line. The
low Mh w/ Mh n observed at low and high conversions and
the almost linear relationship between degree of po-
lymerization and conversion suggest that the deviation
results from a nonquantitative initiation. As indicated
by elemental analysis, the good initial purity of the
starting initiators implies that some currently unidenti-
fied deactivation mechanism is responsible for the
observed partial initiation.
(
4 o o n
NBu) (b)]; (3.8 mol % initiator; [M ]/[I ] ) 27/1, 60 °C, Xh
obtained by GPC and corrected by NMR). The theoretical
behavior expected for a living polymerization under the above
experimental conditions is indicated by a solid line.
Scheme 3. Possible Side Reactions during the
Polymerization of 1a-d (R ) ethyl (a), isopropyl (b),
n-butyl (c), and n-octyl (d))
This possible deficiency in the initiation of PhSLi- and
PhSNa-initiated polymerizations makes it difficult to
reach a firm conclusion on the effect of the counterion
on kp. In the case of dialkyl cyclopropane-1,1-dicarboxy-
late monomers such as diisopropyl cyclopropane-1,1-
dicarboxylate, polymerizations initiated by PhSK pro2-
ceeded about twice as fast as those initiated by PhSNa.
Such an order of magnitude is compatible with the
results obtained for 1a, the experimental data showing
that, at the onset, the PhSK-initiated polymerization
proceeds twice as fast with little or no precipitation
observed.
Analysis of 1a polymerization at varying tempera-
tures showed that the rate of polymerization increased
with temperature, as expected, while maintaining the
living character of the polymerization up to at least 120
°C (Table 3).
The reaction mixture was initially homogeneous.
After about 15 min, some polymer precipitated, consis-
tent with the poor solubility of poly(1a) in DMSO below
1
00 °C even at low degrees of polymerization. Despite
the precipitation, quantitative conversions were achieved
in every case, albeit very slowly, when the counterion
was lithium (quantitative polymer recovery was ob-
tained for each four counterions after 24 h (1440 min);
data point not included in Figure 5 for clarity).
The above observations can be rationalized on the
basis of three hypotheses: (a) remaining soluble active
species (thiophenolate initiator and/or propagating oli-
gomers) maintain the polymerization after the precipi-
tation has occurred, (b) the propagation can also occur
in the solid polymer or at its surface, or (c) a fast
equilibrium between soluble (active) and insoluble
(
inactive) propagating chains can develop in the range
R-Cyanoacrylates react with neutral nucleophiles
such as amines and phosphines, generating 1,3-zwitter-
ions by Michael addition on the CdC double bond.1
These intermediate species have been identified by
of molecular weights covered by the experiments. The
evolution of the molecular weight distribution and
degree of polymerization suggests the second or third
3-19

Macromolecules, Vol. 38, No. 11, 2005
Polymerization of Alkyl 1-Cyanocyclopropanecarboxylates 4593
Table 4. Ring-opening Polymerization of 1a with Neutral
Initiators
Scheme 4. Main Reaction Observed between
Thiophenolate Anions and Cyclobutane 2
run
time yield
Mh n
(GPC)
no. monomer initiatora
(h)
(%)
b
Mh w/ Mh n b
1
2
3
4
5
6
7
8
9
1a
1a
1a
1a
1a
1a
1a
1a
1a
PhSK
1.0
1.0
1.0
1.0
1.0
3.0
84
43
59
e
12400
8800
8800
1.05
1.20
1.16
c
NMP
DBUd
Et3N
pyridine
PhSK
NMP
and the GPC data also showed no changes in Mh n.
Tetrabutylammonium salts of cyanoacetates effectively
initiate the polymerization of MMA under conditions
identical to ours,6 suggesting that the observed failure
of the propagating carbanion to react with MMA is due
to the physical inaccessibility of the MMA monomer to
the propagating chains as a result of the precipitation
of poly(1d) from solution at ambient temperature. If
correct, this conclusion would support hypothesis (b)
over (c), i.e., a direct polymerization in the solid state
or at the solid surface rather than a fast equilibrium
between soluble and insoluble propagating species.
Attempted Polymerization of Ethyl 1-Cyanocy-
clobutanecarboxylate 2. Because of the expected
lower reactivity of cyclobutane monomers over the
e
100
100
94
e
14700
8600
3700
1.05
1.44
1.15
24
,21
Et3N
pyridine
24
24
a
3
.59 mmol of 1a in 0.3 mL DMSO, 3.8 mol % initiator at 60
°
C. b Determined in DMF at 100 °C, calibrated with polystyrene
standards. N-methyl pyrrolidine. d 1,8-diazabicyclo-[5.4.0]undec-
c
7
-ene (DBU)]. e No polymer isolated by precipitation
Pepper et al. as the key species in the polymerization
of R-cyanoacrylates by neutral nucleophiles. To test
whether these nucleophiles were also able to initiate the
polymerization of activated cyclopropanes via a ring-
opening attack and the formation of hypothetical 1,4-
zwitterions, a study of the reactivity of various nucleo-
philic initiators toward the ring-opening polymerization
of 1a was conducted. The results are summarized in
Table 4. Of the selected nucleophilic initiators, PhSK
used here as a reference point is the most reactive,
achieving quantitative conversion in 3 h. Lower yields
and Mh n were obtained for N-methyl pyrrolidine (NMP),
22
analogous cyclopropanes, higher reaction tempera-
tures between 140 and 180 °C were employed for the
attempted polymerization of ethyl 1-cyanocyclobutane-
carboxylate 2. NMR analysis of the reaction mixture (4.5
mmol 2, 4.6 mol % PhSNa, 0.5 mL DSMO, 160 °C)
provided evidence that the main reaction is the attack
of PhSNa on the ethyl carbon of the ester group
1
,8-diazabicyclo-[5.4.0]undec-7-ene (DBU), and triethyl-
10
(
Krapcho reaction, Scheme 4), yielding ethyl phenyl
1
amine (Et3N) under the same conditions. A H NMR
analysis of the reaction mixture confirmed that pyridine
does not initiate the polymerization of 1a at 60 °C. The
entire set of observations is globally consistent with the
general trend of increasing the reactivity toward po-
lymerization with an increase in the nucleophilic strength
of the initiator.
thioether and the salt of 1-cyanocyanocyclobutanecar-
boxylic acid as the main products. No polymers/oligo-
mers could be identified for any of the reactions of 2
with either PhSK or PhSNa at 180 °C, even after long
reaction times.
This sharp decrease in reactivity for the cyclobutane
ring is not unexpected. Despite the almost identical
strain energies of three- and four-membered rings,
cyclobutane rings are much less reactive than cyclopro-
panes, in agreement with the reactivity ratio observed
for their heterocyclic counterparts (oxetanes vs oxiranes,
thietane vs thiiranes, azetidines vs aziridines, etc.).22
As far as we are aware, the only example reported in
the literature for the ring-opening polymerization of a
cyclobutane is on the basis of 1,1,2,2-tetracyano-
Re-initiation from the Macroinitiator. As ex-
pected from a living polymerization, reinitiation from
the macromolecular carbanion is also possible. Polym-
erization of 1a at 60 °C for 4h (3.59 mmol 1a, 0.3 mL
DMSO, 3.8 mol % PhSNa) yielded poly(1a)1, a polymer
of Xh n ) 41 (100% yield). When a second aliquot of 1a (7
mmol) was added to the solid poly(1a)1 prior to termina-
tion and allowed to run for 24 h, a second sample, poly-
(
1a)2, was obtained with a Xh n ) 102 (50% yield). Despite
3
-ethoxycyclobutane, for which polymerization was
the precipitation of the polymer from the reaction
mixture during the polymerization of 1a, the chains
remained active and could propagate in the solid state
in the presence of additional monomer, although (as
expected) the reactivity was lower.
achieved at ambient temperature using various initia-
tors including triethylamine.23 In this case, the cyclobu-
tyl ring is further activated toward the ring-opening
nucleophilic substitution of the intra-annular C-C bond
by the presence of a strong electron-donating substitu-
ent on the attacked carbon.
Tetrabutylammonium salts of cyanoacetates, ethyl
dialkyl malonates, methyl malononitrile, and nitro-
alkanes have been used as efficient initiators for the
living polymerization of acrylates and (methyl) meth-
Conclusions
6,20-21
acrylates.
As a result, attempts were made to grow
The synthesis of poly(alkyl trimethylene-1-cyano-1-
carboxylate)s (CH2CH2C(CN)(COOR))n, the next higher
homologues of poly(R-cyanoacrylate)s, can be achieved
via the anionic ring-opening polymerization of the
corresponding alkyl 1-cyanocyclopropanecarboxylate
monomers 1. Despite the early precipitation of propa-
gating polymer chains in the DMSO/monomer mixture
used for the experiments, polymerizations were ob-
served to proceed to full conversion, probably via addi-
tion of the monomer to propagating centers located in
the solid or at its surface. Molecular weights obtained
under the investigated experimental conditions (high
initiator-to-monomer ratios) were rather low, as ex-
a block of PMMA from the polymeric carbanion of 1a
under conditions similar to those used for the polym-
6,21
erization of MMA.
The polymerization of 1d was first
carried out to quantitative conversion using PhSN(Bu)4
2.24 mmol 1d, 3.8 mol %) at 60 °C in THF. Monomer
d was selected over 1a-c because of the good solubility
of poly(1d) in THF over the entire polymerization at 60
C. After the monomer 1d had been completely con-
(
1
°
sumed, the solution was cooled to ambient temperature
and MMA (2.24 mmol) was added. Analysis of the
polymer mixture was performed by NMR and GPC, but
no evidence was obtained for the presence of PMMA,

4594 Kagumba and Penelle
Macromolecules, Vol. 38, No. 11, 2005
pected. Polymerizations initiated with thiophenolate
salts, in particular potassium and tetra-n-butylammo-
nium thiophenolates, were highly efficient, with no
evidence for competing side reactions. Polymers with
narrow molecular weight distributions ( Mh w/ Mh n < 1.10)
were obtained in most cases, and the polymerizations
(6) Reetz, M. T.; Hutte, S.; Goddard, R. J. Am. Chem. Soc. 1993,
115, 9339.
(7) Bordwell, F. G. Acc. Chem. Res. 1988, 21, 456.
(8) Pepper, D. C. J. Polym. Sci., Polym. Symp. 1978, 62, 65.
(9) Pepper, D. C. Polymer J. 1980, 12, 629.
(
(
10) Krapcho, A. P. Synthesis 1982, 805.
11) Woods, J. G.; Rooney, J. M. U.S. Patent 4,675,270, 1987.
(12) Howorka, H.; Kretschmer, K. Forensic Sci. Int. 1990, 46, 31.
(13) Cronin, J. P.; Pepper, D. C. Makromol. Chem. 1988, 189, 85.
(14) Eromosele, I. C.; Pepper, D. C. Makromol. Chem., Rapid
Commun. 1986, 7, 531.
(
within the limits of this investigation) displayed char-
acteristics typical of living systems, in particular for
PhSK- and PhSNBu4-initiated polymerizations. IR spec-
troscopy supported a very high stability of the poly(1a)
carbanion, even under wet air atmospheres. Attempts
to ring-open ethyl 1-cyanocyclopropanecarboxylate 2
were unsuccessful, even at elevated temperatures,
because of the poor kinetics associated with the ring-
opening of cyclobutanes.
(
15) Johnston, D. S.; Pepper, D. C. Macromol. Chem. Phys. 1981,
82, 407.
16) Johnston, D. S.; Pepper, D. C. Macromol. Chem. Phys. 1981,
82, 421.
(17) Johnston, D. S.; Pepper, D. C. Macromol. Chem. Phys. 1981,
82, 393.
1
(
1
1
(
(
18) Pepper, D. C. Polym. J. 1980, 12, 629.
19) Donnelly, E. F.; Johnston, D. S.; Pepper, D. C.; Dunn, D. J.
References and Notes
J. Polym. Sci., Polym. Lett. Ed. 1977, 15, 399.
(
1) Penelle, J.; H e´ rion, H.; Xie, T.; Gorissen, P. Macromol. Chem.
(20) Raj, D. J. A.; Wadgaonkar, P. P.; Sivaram, S. Macromolecules
1992, 25, 2774.
Phys. 1998, 199, 1329.
(
(
2) Penelle, J.; Xie, T. Macromolecules 2000, 33, 4667.
3) Besten, R. H., S.; Brandsma, L. J. Organomet. Chem. 1990,
(21) Reetz, M. T.; Knauf, T.; Minet, U.; Bingel, C. Angew. Chem.,
Int. Ed. Eng. 1988, 27, 1373.
385, 153.
(22) Stirling, C. J. M. Tetrahedron 1985, 41, 1613.
(4) Zefirov, N. S.; Kuznetsova, T. S.; Kozhushkov, S. I.; Surmina,
(
23) Yokozawa, T.; Miyamoto, Y.; Futamura, S. Makromol. Chem.,
L. S.; Rashchupkina, Z. A. Zh. Org. Khim. (Engl. Transl.)
Rapid Commun. 1993, 14, 245.
1983, 19, 474.
(5) Danishefsky, S. Acc. Chem. Res. 1979, 12, 66.
MA0484025