Oxygen/sulfur scrambling during the copolymerization of cyclopentene oxide and carbon disulfide: Selectivity for copolymer vs cyclic [Thio]carbonates

Article abstract of DOI:10.1021/ma4015438

The catalytic coupling of cyclopentene oxide with carbon disulfide has been investigated utilizing (salen)CrCl in the presence of added onium salts. Both polymeric and cyclic materials were produced, with oxygen/sulfur atom scrambling observed in both instances. This atom redistribution process was found to require (salen)CrCl and excess epoxide, though an increase in the rate of atom scrambling was noted upon the addition of the onium salt. Cyclopentene sulfide was observed as a side product of the coupling reaction and was found to be unreactive toward both CS₂ and CO₂, instead undergoing desulfurization to cyclopentene under the conditions of the reaction. Of the 12 cyclic cyclopentene [thio]carbonates possibly produced by this coupling reaction, eight were observed, and the crystal structure of trans-cyclopentene trithiocarbonate is reported herein. Computational studies reveal that the cis- cyclic materials are more stable than their trans-counterparts by >5 kcal/mol of enthalpy, and there is a 10-25 kcal/mol preference for the formation of a C=O vs C=S double bond. When trans-cyclohexene trithiocarbonate was exposed to the catalyst system in the presence of excess cyclopentene oxide, mixed-species scrambling was observed, whereby cyclic [thio]carbonate compounds displaying both cyclopentyl and cyclohexyl backbones were produced. A proposed mechanistic pathway for atom scrambling involves nucleophilic alkoxide attack at a [thio]carbonyl center to induce oxygen/sulfur atom exchange.

Full text of DOI:10.1021/ma4015438

Article
pubs.acs.org/Macromolecules
Oxygen/Sulfur Scrambling During the Copolymerization of
Cyclopentene Oxide and Carbon Disulfide: Selectivity for Copolymer
vs Cyclic [Thio]carbonates
Donald J. Darensbourg,* Stephanie J. Wilson,† and Andrew D. Yeung
Department of Chemistry, Texas A&M University, College Station, Texas 77843, United States
S
* Supporting Information
ABSTRACT: The catalytic coupling of cyclopentene oxide
with carbon disulfide has been investigated utilizing (salen)-
CrCl in the presence of added onium salts. Both polymeric
and cyclic materials were produced, with oxygen/sulfur atom
scrambling observed in both instances. This atom redistrib-
ution process was found to require (salen)CrCl and excess
epoxide, though an increase in the rate of atom scrambling was
noted upon the addition of the onium salt. Cyclopentene
sulfide was observed as a side product of the coupling reaction
and was found to be unreactive toward both CS2 and CO2, instead undergoing desulfurization to cyclopentene under the
conditions of the reaction. Of the 12 cyclic cyclopentene [thio]carbonates possibly produced by this coupling reaction, eight were
observed, and the crystal structure of trans-cyclopentene trithiocarbonate is reported herein. Computational studies reveal that
the cis- cyclic materials are more stable than their trans-counterparts by >5 kcal/mol of enthalpy, and there is a 10−25 kcal/mol
preference for the formation of a CO vs CS double bond. When trans-cyclohexene trithiocarbonate was exposed to the
catalyst system in the presence of excess cyclopentene oxide, mixed-species scrambling was observed, whereby cyclic
[thio]carbonate compounds displaying both cyclopentyl and cyclohexyl backbones were produced. A proposed mechanistic
pathway for atom scrambling involves nucleophilic alkoxide attack at a [thio]carbonyl center to induce oxygen/sulfur atom
exchange.
polymer product which displayed a mix of different backbone
repeat units, instead of only the expected −O−(CS)−S−
dithiocarbonate repeat unit (Scheme 1). Oxygen and sulfur
INTRODUCTION
■
While the coupling of epoxides and carbon dioxide to form
polycarbonates has garnered widespread attention, the
analogous reaction of epoxides and carbon disulfide, CS2, to
form polythiocarbonates has received little consideration.1
However, there are reports of the selective coupling of epoxides
and CS2 to afford cyclic trithiocarbonate and thiocarbonate
materials that have applications in materials science and
biologically active systems.2 Notable among these studies is
that by North and co-workers utilizing bimetallic aluminum
(salen) complexes along with tetrabutylammonium bromide (n-
Bu4NBr) as effective catalysts for this process.3 Previously, this
system has been demonstrated to be very active for the
coupling of CO2 and epoxides to cyclic carbonates, under both
academic and industrial laboratory settings.4,5 In 2007, Nozaki
and co-workers investigated the coupling of propylene sulfide
with CS2 to yield poly(propylene trithiocarbonate) and cyclic
propylene trithiocarbonate.6 This process was optimized to
provide a high molecular weight copolymer with 92%
selectivity. Importantly, the poly(propylene trithiocarbonate)
was found to be completely alternating with a high refractive
index, a desirable property for plastic lenses.
Scheme 1. General Catalytic Coupling of Epoxides and CS2
Displaying O/S Scrambling
scrambling was also noted in the cyclic carbonate byproducts.
The presence of scrambled COS produced in the process was
detected by GC−MS, though no trace of propylene sulfide was
observed. It was hypothesized that propylene sulfide was
quickly consumed in reactions with available CS2, COS, or CO2
within the system. The poly[thio]carbonates produced in this
process displayed broad molecular weight distributions, with
PDI varying between 1.24 and 3.50.
Although the structure of the DMC could be surmised with
some degree of merit, the detailed nature of the specific modes
In 2008, Qi and co-workers investigated the coupling of
propylene oxide and CS2 using a heterogeneous cobalt/iron
double metal cyanide (DMC) catalyst with the aim of
synthesizing copolymers.7 This catalyst system afforded a
Received: July 22, 2013
Revised: September 20, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ma4015438 | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
of reaction of these catalytic materials is poorly defined.8
Hence, more recently our group described the analogous
copolymerization reaction between cyclohexene oxide and CS2
utilizing the well-characterized (salen)CrCl/PPNCl system in
hope of elucidating the mechanistic aspects of this process with
more clarity.9 In this instance, oxygen/sulfur atom scrambling
was observed as well in both the polymeric and cyclic carbonate
products, with the copolymer being oxygen enriched and the
cyclic carbonate being sulfur enriched. Nevertheless, the
pathway for O/S scrambling was inadequately understood.10
In this report, we have investigated the catalytic coupling of
carbon disulfide with an alternative alicyclic epoxide, cyclo-
pentene oxide, employing (salen)CrCl in the presence of
onium salts. It is noteworthy that these two alicyclic epoxide
monomers, cyclopentene and cyclohexene oxides, have been
shown to exhibit grossly different reactivity patterns with CO2
under the influence of these catalyst systems.11 This behavior is
anticipated to lead to additional and more definitive
information with regard to the mechanistic details of the
observed oxygen/sulfur atom scrambling process. Because
sulfur-containing copolymer and cyclic materials are projected
to possess attractive properties, this atom exchange process can
provide a convenient route to these products, avoiding the need
for expensive and generally commercially unavailable thiiranes.
The reactivity of cyclopentene sulfide with both CO2 and CS2
was also examined in an effort to provide a more
comprehensive analysis of these processes.
particularly when a limiting amount of CS2 is employed in the
reaction. When CO2 is added to the reaction, production of cis-
cyclopentene carbonate is achieved very quickly.
Elemental analysis was performed on the purified polymer
samples in order to better determine the chemical makeup
(Table 1). The hydrogen atoms were set at 8.00/repeat unit
Table 1. Reaction of Cyclopentene Oxide with CS2 with
(salen)CrCl and PPNN3 in a 300 mL Stainless Steel Parr
a
Reactor
PCO
MPa
,
equiv of
PPNN3
equiv of
CS2
polymer repeat
2
b
c
entry
unit
Tg, °C
1
2
3
4
2
1
2
2
1500
1500
750
0
C
5.69H8S1.56O0.76
5.58H8S1.15O0.88
5.22H8S0.80O1.22
5.81H8S1.08O1.81
14.4
9.6
0
C
C
C
0
−17.6
32.3
750
0.7
a
Reaction conditions 9.64 g of cyclopentene oxide (10.0 mL, 115
mmol, 1500 equiv), 0.0483 g of (salen)CrCl (0.076 mmol, 1 equiv),
and the appropriate amounts of PPNN3, CS2, and CO2 pressure,
b
where applicable. 80 °C, 24 h. H set at 8.00, and other elements were
calculated from elemental analysis of the purified polymer sample.
c
Midpoint Tg from the second heating cycle utilizing Differential
Scanning Calorimetry.
because this value is consistent for both poly[thio]carbonate
and poly[thio]ether linkages (Figure 1). In each case, the
RESULTS AND DISCUSSION
■
Coupling of Cyclopentene Oxide and Carbon Disul-
−
fide Utilizing PPNN3. Azide, N3 , is often used as an initiator
in the copolymerizations of epoxides and CO2 due to its
distinct νN bond stretching patterns that can be monitored by
3
Figure 1. Repeat units for [thio]carbonate and [thio]ether linkages. X
infrared spectroscopy. Clear shifts in the azide stretch can be
observed depending on the chemical environment. Because of
this valuable spectroscopic trace, bis(triphenylphosphine)-
iminium azide, PPNN3, was originally chosen as the cocatalyst
for this investigation. Following the aforementioned coupling of
CPO with CS2 utilizing (salen)CrCl and PPNN3, a wide array
of polymeric and cyclic [thio]carbonate materials were
produced. Scrambling of the oxygen and sulfur atoms is
observed in both the reactants and the products. In an effort to
quickly identify the various materials produced during the
coupling process, basic information including mass of produced
polymer versus [thio]cyclic carbonate was not collected for
these reactions. Regardless, both polymer and cyclic carbonates
were produced in each case, and differences in scrambling
products can be noted. These reactions were monitored
utilizing ATR-FTIR, and three-dimensional stack plots and
reaction profiles can be found in the Supporting Information.
Because of the overlapping nature of the 1H and FT-IR of these
species, spectral assignments were difficult.
In general, peak production of the −O−(CO)−S−
polymer peak9 at 1712 cm−1 is maximized after ∼4 h, after
which it decays away, presumably because of backbiting or
rearrangement reactions. The spectral region from 1000 to
1400 cm−1 is difficult to assign, as many different peaks are
overlapping. CS2 can be observed as the strong band at 1514
cm−1. The unchanging absorbance in the early stages of the
reaction is caused by the saturation of CS2 within the solution.
Only after sufficient consumption of CS2 can the absorbance
change be observed. CO2 production can be seen at 2338 cm−1,
= O or S.
carbon content is significantly below the expected 6.0 for
[thio]carbonate linkages, indicating a large amount of ether or
thioether linkages in the polymer backbone. Indeed, when a
limiting amount of CS2 was employed, only 5.22 carbons exist
in the idealized repeat unit, indicating that the polymer contains
mostly [thio]ether linkages (entry 3). The decrease of
[thio]carbonate content is mirrored by a decrease in Tg, from
+14.4 to −17 °C. Upon the addition of 0.7 MPa CO2, both the
[thio]carbonate content and Tg greatly increase, with 5.81
carbons/repeat unit and Tg = 32.3 °C.
Notably, H and 13C NMRs (CDCl3) of the crude reaction
1
mixtures each clearly display the production of cyclopentene
sulfide (peaks at 3.28, 30.0 ppm, respectively) and cyclopentene
(peaks at 5.76, 32.4 ppm, respectively). At the end of the 24 h
reaction period, the cyclopentene oxide has been either
completely or almost completely consumed, as evident by the
disappearance of the peaks at 3.39 and 27.1 ppm in the 1H and
13C NMRs, respectively.
Reaction of N3− with CS2. Though we were unaware prior
to the start of this project, the facile reaction of azide with CS2
to form the 5-membered heterocyclic anion azidodithiocar-
12
−
bonate, CS2N3 , is well-defined in the literature (Scheme 2).
Caution! Azidodithiocarbonate salts and derivatives are potentially
explosive, and appropriate safety precautions should be taken. Only
small quantities of the azidodithiocarbonate compounds were
B
dx.doi.org/10.1021/ma4015438 | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
after 6 h, after which it drops off quickly, presumably due to
depolymerization (Figure 3). Thus, all subsequent reactions
were monitored for this amount of time to maximize the
amount of copolymerization (Table 2).17
Scheme 2. Reaction of Azide with Carbon Disulfide To Yield
the Azodithiocarbonate Anion
a
a
Upon heating, this anionic heterocycle will breakdown to dinitrogen,
thiocyanate, and elemental sulfur.
generated in these experiments, and they were normally
generated in situ inside of a stainless steel reactor.
As a test to confirm the presence of the CS2N3 anionic
−
heterocycle, tetrabutylammonium azide, n-Bu4NN3, was dis-
solved in toluene at ambient temperature, and the characteristic
free N3− peak was observed at 2000 cm−1.13 Upon addition of 5
−
equiv of CS2, the N3 peak immediately disappeared, and no
−
distinct peaks were visible, consistent with reports that CS2N3
complexes have only weak infrared stretching patterns.14
−
CS2N3 salts and compounds are known to be thermally
unstable,15 and so the reaction was heated to the typical
reaction temperature, 80 °C, to observe any potential
decomposition. Indeed, almost immediately upon heating a
sharp peak appeared at 2055 cm−1 which indicates the
formation of n-Bu4NSCN (Figure 2).16 As such, the CS2N3
−
Figure 3. Trace of the −O−(CO)−S− monothiocarbonate
heterocycle is not long-lived in the reaction medium, instead
polymer at 1712 cm−1 displaying peak growth after 6 h. Reaction
decomposing to thiocyanate, SCN−. It is unclear at this time if
loading is (salen)CrCl:PPNCl:cyclopentene oxide:CS2
=
−
CS2N3 is able to initiate polymer growth prior to this
1:2:1500:3000.
decomposition.
Coupling of Cyclopentene Oxide and Carbon Disul-
−
−
fide Utilizing PPNCl. Because of the unclear N3 /CS2N3 /
SCN− nature of the cocatalyst in the previous studies, it was
desirable to investigate this system with the well-defined
PPNCl cocatalyst. Initial ATR-FTIR monitoring of a reaction
between cyclopentene oxide and CS2 revealed that peak −O−
(CO)−S− polymer production at 1712 cm−1 was achieved
As previously stated, the reaction time for those experiments
with PPNCl cocatalyst was reduced to 6 h in order to maximize
the growth of the known −O−(CO)−S− monothiocar-
bonate polymer product at 1712 cm−1. The largest amount of
polymer and greatest polymer selectivity occurred when equal
parts cyclopentene oxide and CS2 were employed (Table 2,
Figure 2. At left, three-dimensional stack plot of the reaction of azide (2000 cm−1) with CS2. CS2 was added after 3 min. After ∼2 h, the reaction was
heated to 80 °C, and the growth of thiocyanate was observed at 2055 cm−1. At top right, traces of the disappearance and growth peaks of free
−
N3 (red), and SCN− (blue). Bottom right, zoomed in view from 0 to 2.75 h reaction time.
C
dx.doi.org/10.1021/ma4015438 | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
Table 2. Reaction of Cyclopentene Oxide with CS2 with
(salen)CrCl and PPNCl in a 25 mL Stainless Steel Parr
a
Reactor
equiv of
mass
mass cyclic
carbonates, g
polymer repeat
b
c
d
entry
CS2
polymer, g
unit
e
1
2
3
4
5
2.5
2.0
1.5
1.0
0.5
1.0
0.20
1.11
0.22
0.91
1.26
0.83
0.49
1.50
n.d.
C5.92H8S1.70O1.43
C5.97H8S1.65O1.24
C5.96H8S1.64O1.32
n.d.
1.27
2.04
Figure 4. Carbonate region of the 13C NMR of poly[thio]carbonate
from Table 2, entry 2. CDCl3, 125 MHz.
trace
0.17
f
6
n.d.
a
Reaction conditions: 24.2 mg (salen)CrCl (1 equiv), 43.9 mg PPNCl
(2 equiv), 5.0 mL cyclopentene oxide (1500 equiv), 80 °C, 6 h in a 25
b
c
mL stainless steel reactor. Relative to cyclopentene oxide. Mass of
purified polymer sample. H set at 8.00 and other elements were
d
calculated from elemental analysis of the purified polymer sample.
e
f
Not determined. 0.7 MPa CO2 added.
entry 4). When the amount of CS2 was limiting, only trace
amounts of polymer and cyclic [thio]carbonates were produced
(entry 5). Similarly, polymer and cyclic production were
inhibited by large excesses of CS2, presumably due to dilution
of the reaction (entry 1). The addition of 0.7 MPa CO2
prevented significant polymer growth, though a relatively large
amount (1.50 g) of cyclic material was produced (entry 6).
Though other mixed O/S [thio]carbonates were produced,
NMR and GC−MS confirmed that cis-cyclopentene carbonate
formed from the coupling of cyclopentene oxide and CO2 was
the major product for this reaction, highlighting the large
difference in reactivity for cyclopentene oxide with CO2 and
CS2.
Figure 5. FT-IR of poly[thio]carbonate from Table 2, entry 2. CH2Cl2
Elemental analysis of multiple polymers give a consistent
repeat unit of C6H8S1.6O1.3, indicating that there is very little
[thio]ether content in the backbone of the polymer (Table 2,
entries 2 and 4). The polymer is therefore oxygen-enriched
from the expected (C6H8S2O)n repeat unit. We believe that the
increased time for the 24 h PPNN3 reactions led to backbiting
reactions, the loss of CS2, CO2, and COS, and corresponding
increase in ether and thioether linkages in the polymer
backbone. Similar observations have been made for the
decomposition of polycarbonates.18 Additionally, no trans-
cyclopentene trithiocarbonate was observable in any of the
produced cyclic materials by GC−MS, which is starkly different
than the 24-h PPNN3 reactions in which trans-cyclopentene
trithiocarbonate was typically the major cyclic product. This
indicates that trans-cyclopentene trithiocarbonate likely forms
from late-stage polymer backbiting reactions after CS2 has been
largely consumed.
solvent.
1712 cm−1 in the 13C NMR and FT-IR, respectively. Notably, a
small amount of the all-oxygen polycarbonate can be seen both
in the NMR (∼153 ppm) and IR spectra (1750 cm−1), and
other minor contributions from the trithiocarbonate (∼220
ppm) and dithiocarbonate (∼193 ppm) can be observed by
NMR.
Coupling of Cyclopentene Sulfide with CS2 and CO2.
Since the generation of cyclopentene sulfide has clearly been
observed, it is believable that cyclopentene sulfide is also being
consumed during the reaction. It is therefore desirable to
independently determine the relative reactivity of cyclopentene
sulfide with CS2 and CO2.
Several attempts were made to synthesize cyclopentene
sulfide according to recent literature reports through the
reaction of cyclopentene oxide with thiourea,20 potassium
thiocyanate,20 and ammonium thiocyanate21 as sulfur sources,
though none were successful. A 1951 report from van Tamelen
states that thiocyanate is not effective for the synthesis of
cyclopentene sulfide, though identical conditions are quite
effective at the synthesis of cyclohexene sulfide in a 73% yield
from cyclohexene oxide (Scheme 3).22 Indeed, the reaction
between cyclopentene oxide and SCN− does not proceed
beyond the first equilibrium step, and cyclopentene oxide is the
only recovered material. This is significant, as the SCN−
The produced poly(cyclopentene [thio]carbonate)s are
semicrystalline with Tg near 50 °C and Tm near 125 °C.19
Cyclopentene sulfide and cyclopentene are again observed in
1
the H and 13C NMRs of the crude reaction mixtures, though
significant amounts of unreacted cyclopentene oxide also exist
in each case. Notably, 13C NMR revealed that the major
[thio]carbonate component is actually at 211 ppm, caused by
the unscrambled −S−(CS)−O− polymeric product (Figure
4).9 The infrared spectrum of the corresponding polymer
shows several large peaks in the 1000−1400 cm−1 range
expected for thiones, but overlapping from the various
thiocarbonates as well as C−O and C−S single bond stretches
has prevented absolute assignment of peaks (Figure 5). The
next largest contributor to the polymeric backbone is −O−
(CO)−S−, with corresponding to peaks at 169 ppm and
−
produced by the decomposition of CS2N3 will therefore not
directly react with cyclopentene oxide to form the episulfide.
An alternative, 15 day stepwise synthesis of cyclopentene
sulfide by Harding and Owen23 and Goodman, Benitez, and
Baker24 was instead followed (Scheme 4). The foul-smelling
D
dx.doi.org/10.1021/ma4015438 | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
Scheme 3. Thiocyanate, SCN−, Is Able To Convert
Cyclohexene Oxide to Cyclohexene Sulfide with a 73% Yield,
though this Same Process Is Not Effective for Cyclopentene
Oxide to Cyclopentene Sulfide22
Table 4. Coupling Reactions of Cyclopentene Sulfide with
CO2
a
cyclopentene
b
bc
,
b
entry cocatalyst
sulfide,
%
cyclic,
1.3
0.3
%
cyclopentene, %
1
2
PPNCl
98.1
99.4
0.6
0.3
n-Bu4NN3
a
All reactions performed in 10 mL stainless steel reactors. 0.250 g
cyclopentene sulfide (2.5 mmol, 250 equiv), 1.0 mL toluene-d8, 6.3 mg
(salen)CrCl (0.01 mmol, 1 equiv), 2 equiv of the appropriate
b
cocatalyst, and 3.4 MPa CO2 at 80 °C for 24 h. Relative product yield
c
1
from crude H NMR. Unidentified coupling products.
product was distilled over CaH2 under reduced pressure to
yield pure cyclopentene sulfide as a clear, colorless liquid.
process. Cyclopentene sulfide is synthesized in situ, but it has
limited further reactivity after it is generated. This lack of
reactivity for the episulfide appears to be unique to the
cyclopentyl derivative, as both cyclohexene sulfide9 and
propylene sulfide6 will effectively couple with CS2 to yield
poly(trithiocarbonate) and cyclic trithiocarbonate materials
utilizing chromium catalysts.
Scheme 4. Three Step Synthesis of Cyclopentene Sulfide
Cyclic [Thio]carbonates. In the crude 13C NMR of cyclic
materials, three distinct compounds were consistently noted as
the major cyclic products. Peaks at 234.7, 228.3, and 197.6 ppm
(CDCl3) have been assigned to trans-cyclopentene trithiocar-
bonate, cis-cyclopentene trithiocarbonate, and cis-cyclopentene
dithiocarbonate (−S−(CO)−S− connectivity), respectively.
Following purification by column chromatography, single
crystals of trans-cyclopentene trithiocarbonate were grown
and analyzed by X-ray crystallography (Figure 6). The trans-
Several coupling studies of cyclopentene sulfide with CS2
were performed (Table 3). At 80 °C, the same temperature
used for the coupling of cyclopentene oxide and CS2, the
reactivity of cyclopentene sulfide can be observed to be very
poor, with only limited formation of cis-cyclopentene
trithiocarbonate (entries 1−3). A trace amount of the
desulfurization product, cyclopentene, is also observed. This
desulfurization is undoubtedly the source of cyclopentene
within the coupling reactions, as similar tests with cyclopentene
oxide were unsuccessful. By increasing the temperature and
reaction time to 110 °C and 24 h, respectively, ∼20%
conversion to cis-cyclopentene trithiocarbonate and ∼40%
conversion to cyclopentene is observed (entries 4−5).
Cyclopentene is the major product (36%) when only
(salen)CrCl is employed, though 7% coupling products are
observed (entry 6). In the absence of both catalyst and
cocatalyst, 4% desulfurization is observed with no noticeable
coupling product (entry 8). Similarly, cyclopentene sulfide
Figure 6. Thermal ellipsoid representation of trans-cyclopentene
trithiocarbonate with ellipsoids at 50% probability surfaces. Inset
image is looking down the S1−C4 plane to show torsion caused by the
trans-configuration. S2−C2−C2′−S2′ angle = 46.44°.
remains relatively unreactive with CO2 with both Cl− and N3
−
cocatalysts (Table 4). It is thus apparent that cyclopentene
sulfide formation is a thermodynamic sink within this reaction
a
Table 3. Coupling Reactions of Cyclopentene Sulfide with CS2
b
c
c
c
entry
reaction loading
cocat.
temp, °C
t, h
cyclopentene sulfide, %
cis-cyclopentene trithiocarbonate, %
cyclopentene, %
1
2
1:2:500:1000
1:2:500:1000
1:2:500:1000
1:2:250:500
1:2:250:1000
1:0:250:1000
0:2:250:1000
0:0:250:500
PPNCl
PPNN3
n-Bu4NN3
PPNCl
PPNCl
−
80
80
6
6
98.8
99.1
99.1
39
0.9
0.8
0.4
19
0.3
0.1
0.5
42
39
36
6
d
3
80
12
24
24
24
24
24
d
4
110
110
110
110
110
d
5
39
21
d
e
6
d
56
0.7 (6.3 )
7
PPNCl
86
8
d
8
PPNCl
96
−
4
a
b
All reactions performed in 10 mL stainless steel reactors. Reaction loading refers to the relative equivalents of (salen)CrCl:cocatalyst:cyclopentene
c
d
sulfide:CS2 loaded into the reactor. Relative product yield from crude 1H NMR. Toluene solvent employed; total volume of CS2 + toluene = 1.0
mL. Unidentified coupling products.
e
E
dx.doi.org/10.1021/ma4015438 | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
assignment is consistent with a previous literature report.25
Attempts were made to separate cis-cyclopentene trithiocar-
bonate and the −S−(CO)−S− cis-cyclopentene dithiocar-
bonate by physical and chemical means, but these experiments
were unsuccessful. When a limiting amount of CS2 or added
CO2 was employed, cis-cyclopentene carbonate formation
could also be clearly observed at 155.6 ppm.
Table 5. Calculated Relative Enthalpies of Cyclic
[Thio]carbonates
a b
,
Gas chromatography−mass spectrometry (GC−MS) proved
a useful tool for the identification of the identities of the cyclic
[thio]carbonate materials, as the GC was able to separate out
the various cyclic materials by their different elution times, and
then mass spectrometry could be employed to identify the
molecular weight of the cyclic materials. Additional information
regarding the identity and spectral data of the various cyclic
[thio]carbonate materials can be found in the Supporting
Information. GC−MS analysis showed that eight of the
possible 12 cyclic [thio]carbonates are produced from the
coupling of CS2 and cyclopentene oxide. Cis-cyclopentene
carbonate, both isomers of cyclopentene trithiocarbonate, and
cis-cyclopentene dithiocarbonate (−S−(CO)−S− connectiv-
ity) were each positively identified by matching traces from
pure samples. Only one isomer of cyclopentene monothiocar-
bonate was observed. All four structural and geometric isomers
of cyclopentene dithiocarbonate were observed, though only
the cis-(−S−(CO)−S−) connectivity was isolated. Without
calibrating for the GC−MS response factor, relative concen-
trations of the individual cyclic carbonates were unable to be
determined.
a
b
As defined in Figure 7. CBS-QB3, gas phase.
In order to better hypothesize the identity of each material,
the relative enthalpies of each of the possible cyclopentene
[thio]carbonate species were calculated. Slight (1−2 kcal/mol)
energy differences can be observed between the cis-boat and cis-
chair conformations of the cyclic systems (Figure 7). Generally,
Scrambling Mechanism. Scrambling of the oxygen and
sulfur atoms has been observed in both the polymer and cyclic
[thio]carbonate samples. In the literature, several similar
mechanisms involving the ring-opening of preformed cyclic
carbonates have been proposed. In Endo’s 2001 investigation of
cyclic trithiocarbonate formation utilizing a titanium catalyst,
scrambling was observed in the presence of excess CS2.2d In our
2009 report, the Darensbourg group proposed that the
insertion of CO2, COS, and CS2 is reversible, and it is this
reversibility that brings about the scrambling.9 A similar
mechanism has been proposed by North for the coupling of
propylene oxide and CS2 in which propylene sulfide is first
eliminated and then reenters the mechanism as a new reagent.3a
Tetrahedral spiro intermediates have been proposed and
isolated by Barbero et al. for propylene oxide/sulfide
scrambling systems utilizing HBF4·Et2O,26 but these were not
directly observable by NMR or FT-IR in this investigation.
Several test reactions were run in order to determine the
conditions necessary to induce scrambling for reactions
involving cyclopentene oxide, (salen)CrCl, and an onium salt
cocatalyst (Table 6). In three separate reactions, toluene
solutions containing a mix of cyclic cyclopentene [thio]-
carbonates were heated to 80 °C with (salen)CrCl and n-
Bu4NCl in the presence of CO2, CS2, and a control with no
other added reagents. After 48 h, no scrambling was observed
Figure 7. Calculated local minima for cyclopentene carbonate (OOO)
species: cis-boat, cis-chair, and trans-.
the cis-systems are lower in energy than their trans-counterparts
(Table 5; see also Supporting Information). For cyclopentene
carbonate, the difference is stark at ∼20 kcal/mol.11a This is
understandable, given that only cis-cyclopentene carbonate has
ever been isolated, and all attempts to make trans-cyclopentene
carbonate have failed. The enthalpy difference is much smaller
for cyclopentene trithiocarbonate, ∼5 kcal/mol, as the larger
sulfur atoms allow for less-strain in the fused 5-membered rings.
The mixed O/S species also show ∼10 kcal/mol preference for
the cis-isomers over trans-. Notably, there is also a large (10−25
kcal/mol) preference for the formation of the carbon−oxygen
double bond over the carbon−sulfur double bond. CO is
known to be a stronger bond than CS, but this preference
may also be due to the greater ability of the large sulfur atoms
to relax the molecule and therefore reduce the strain caused by
the fused 5-membered rings.
1
by H NMR or GC−MS for any of the cases (entries 1−3).
Conversely, a solution of the same cyclic [thio]carbonates,
catalyst, and cocatalyst heated to 80 °C in the presence of
cyclopentene oxide displayed distinct changes in composition,
indicating scrambling had taken place (entry 4). Further
investigations revealed that the presence of (salen)CrCl catalyst
was necessary to induce scrambling, though added n-Bu4NCl
cocatalyst increased the activity (Table 7). Thus, the catalyst is
required to activate and ring-open the epoxide, and the more
nucleophilic alkoxide then attacks the [thio]carbonyl to induce
F
dx.doi.org/10.1021/ma4015438 | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
Table 6. Conditions Needed for O/S Scrambling Utilizing a
Purified Mix of Cyclic [Thio]carbonate Materials Utilizing
(salen)CrCl and n-Bu4NCl
a
b
entry
added reagent
time, h scrambling observed?
c
1
2
3
4
none
48
48
48
16
no
no
no
yes
c
100 psi CO2
c
1 equiv of CS2
d
1 equiv of cyclopentene oxide
a
Average MW of cyclic materials arbitrarily assigned to be 160 g/mol
for calculations. Catalyst loading held at (salen)CrCl:n-Bu4NCl:cyclic
b
mix =1:2:50. 80 °C. Scrambling observed by changes in the 1H NMR
and GC−MS traces of the starting material and reaction products.
c
d
0.875 M solution in toluene. 1.4 M solution.
Table 7. Conditions Needed for O/S Scrambling Utilizing
trans-Cyclohexene Trithiocarbonate and Cyclohexene Oxide
a
Entry
catalyst
cocatalyst
scrambling observed?
1
2
3
4
−
−
−
no
Figure 9. Time lapsed 13C NMR of the reaction between 4 equiv of
cyclopentene oxide (red) and 1 equiv trans-cyclopentene trithiocar-
bonate (purple) to produce cyclopentene sulfide (green) and cis-
cyclopentene carbonate (blue) in toluene-d8. Broadening of cyclo-
pentene oxide’s carbon at 56.3 ppm is likely due to ring-opening and
ring-closing of the epoxide.
(salen)CrCl
yes, slight
no
−
n-Bu4NCl
(salen)CrCl
n-Bu4NCl
yes
a
Analyzed by in situ ATR-FTIR spectroscopy, with scrambling being
assigned by the growth of cyclohexene carbonate at 1809 cm−1.
scrambling (Figure 8). The trithiocarbonate species can be
produced from the coupling of the ring-opened thiirane
produced during the scrambling process and CS2. In addition
to the pathways outlined in Figure 8, it is possible that
cyclopentene sulfide is produced via the elimination of the RO-
(CS)-O anion from (RO-(CS)-O)C5H8−S−. It is unlikely
that cyclopentene sulfide will be separately ring-opened to
undergo coupling. The possibility of spiro intermediates has
not yet been ruled out at this time.
NMR Scrambling Studies. In addition to being studied by
ATR-FTIR, it was desirable to monitor the scrambling
reactions by NMR. Figure 9 shows the time lapsed 13C
NMRs for the reaction of four equivalents cyclopentene oxide
with one equivalent trans-cyclopentene trithiocarbonate utiliz-
ing (salen)CrCl and n-Bu4NCl at 80 °C in toluene-d8. After 15
h, the growth of cyclopentene sulfide is clearly visible with
peaks at 41.1, 29.0, and 18.3 ppm. Growth of cis-cyclopentene
carbonate can also be clearly observed by 22.5 h. After 48 h,
faint traces of cyclopentene can be observed at 130.7 ppm.
Notably, cyclopentene oxide’s signal at 56.3 ppm broadens
during the course of the reaction, likely because the epoxide is
undergoing reversible ring-opening and ring-closing processes.
Other coupling products are visible but have not been
identified. Continued experiments to identify these compounds
are underway in our laboratory.
For completeness, the reaction of cyclopentene sulfide with
cis-cyclopentene carbonate at 80 °C was also monitored by 13
C
NMR (Figure 10). The desulfurization product cyclopentene
(32.7, 23.2 ppm) is the only observable product after 48 h
reaction time, indicating that either the catalyst is incapable of
ring-opening the episulfide or that the thiolate is not a strong
enough nucleophile to attack the carbonate. A proposed
mechanism for these observed epoxide/cyclic scramblings can
be seen in Figure 11. It is expected that a mixture of the
mechanisms proposed in Figures 8 and 11 are both operational
during the course of the reaction.
Figure 8. Proposed mechanism for scrambling in the coupling of cyclopentene oxide and CS2.
G
dx.doi.org/10.1021/ma4015438 | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
observable in both the polymer and cyclic carbonate products.
When PPNN3 is employed as cocatalyst, the temperature-
−
sensitive five membered heterocyclic anion CS2N3 is formed.
It is unclear at this time whether this pseudohalogen is able to
initiate the coupling reaction before decomposing to
thiocyanate, SCN−. Under certain conditions, cyclopentene
sulfide is actually produced during the coupling of cyclopentene
oxide and CS2, and it is thus regarded as a thermodynamic sink
for this process. The coupling of cyclopentene sulfide and both
CO2 and CS2 was examined, but it was determined that
cyclopentene sulfide is highly unreactive under the conditions
studied in this report. Cyclopentene production can be
observed from the desulfurization of cyclopentene sulfide.
The Tg of the polymers is highly dependent on the
[thio]carbonate content of the polymer backbone. When a
large percentage of [thio]ether is observed, the Tg drops
significantly. [Thio]ether content is believed to increase with
increasing reaction time, as backbiting reactions lead to the loss
of CS2, COS, and CO2. The largest amount of [thio]carbonate
copolymer is produced utilizing (salen)CrCl and PPNCl at 80
°C for 6 h. The major repeating units of the polymer backbone
are dithiocarbonate −S−(CS)−O− and monothiocarbonate
−S−(CO)−O−.
Figure 10. Time-lapsed 13C NMR of the reaction between 4 equiv of
cyclopentene sulfide (green) and 1 equiv cis-cyclopentene carbonate
(blue) in toluene-d8. Cyclopentene (orange) is the first observable
product.
Mixed Species Scrambling. O/S scrambling is not limited
to epoxides and episulfides of the same backbone connectivity.
Indeed, when trans-cyclohexene trithiocarbonate was reacted
with cyclopentene oxide, (salen)CrCl, and n-Bu4NCl at 80 °C,
six cyclic [thio]carbonate materials were observed by 13C NMR
and GC−MS including cis-cyclopentene carbonate and residual
trans-cyclohexene trithiocarbonate. Mixed species cyclopentene
thiocarbonate (SO2), cyclohexene thiocarbonate (SO2), and
cyclohexene dithiocarbonate (S2O) were observed, though their
exact connectivity cannot be commented on. Cyclohexene
sulfide was also clearly produced during the course of the
reaction. The reaction mixture was concentrated in vacuo prior
to GC−MS and NMR analysis, so it is unclear whether
cyclopentene sulfide was also generated. However, production
of this thermodynamic sink is likely to have occurred.
A wide range of cyclic [thio]carbonate materials are
produced, and they have been identified utilizing NMR and
GC−MS. Single crystals capable of X-ray diffraction were
grown of trans-cyclopentene trithiocarbonate, and the S−C−
C−S carbonate dihedral angle displays 46.44° of torsion.
Computational analysis of the cyclic [thio]carbonates deter-
mined that, in general, the cis-isomers are more stable than their
trans- counterparts by >5 kcal/mol, and there is a 10−25 kcal/
mol preference for the formation of a CO double bond
rather than CS.
Investigations into the mechanism of O/S scrambling were
also performed. It was determined that scrambling requires the
presence of excess epoxide and catalyst. Added cocatalyst will
greatly increase the rate of the reaction, but it is not a
requirement. Two mechanisms have been proposed for the
observed scrambling processes, though some aspects still
remain poorly understood.
CONCLUSION
■
The catalytic coupling of cyclopentene oxide with carbon
disulfide has been examined utilizing (salen)CrCl and onium
salt cocatalyts. Scrambling of the oxygen and sulfur atoms is
Figure 11. Proposed mechanism for observed scrambling between cyclopentene oxide and trans-cyclopentene trithiocarbonate.
H
dx.doi.org/10.1021/ma4015438 | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
(8) (a) Darensbourg, D. J.; Adams, M. J.; Yarbrough, J. C. Inorg.
Chem. 2001, 40, 6543−6544. (b) Darensbourg, D. J.; Adams, M. J.;
Yarbrough, J. C.; Phelps, A. L. Inorg. Chem. 2003, 42, 7809−7818.
(c) Qin, Z.; Thomas, C. M.; Lee, S.; Coates, G. W. Angew. Chem., Int.
Ed. 2003, 42, 5484−5487. (d) Robertson, N. J.; Qin, Z.; Dallinger, G.
C.; Lobkovsky, E. B.; Lee, S.; Coates, G. W. Dalton Trans. 2006,
5390−5395.
ASSOCIATED CONTENT
■
S
* Supporting Information
Experimental and computational methods, other experimental
data, and the .cif crystallographic data file for trans-cyclo-
pentene trithiocarbonate. This material is available free of
charge via the Internet at http://pubs.acs.org.
(9) Darensbourg, D. J.; Andreatta, J. R.; Jungman, M. J.; Reibenspies,
J. H. Dalton Trans. 2009, 8891−8899.
AUTHOR INFORMATION
■
(10) This exchange process is of great significance in that it allows for
the synthesis of materials where the corresponding thioepoxides are
either commercially unavailable or expensive.
Corresponding Author
*(D.J.D.) E-mail: djdarens@chem.tamu.edu.
(11) (a) Darensbourg, D. J.; Yeung, A. D.; Wei, S.-H. Green Chem.
2013, 15, 1578−1583. (b) Darensbourg, D. J.; Wei, S.-H.; Wilson, S. J.
Macromolecules 2013, 46, 3228−3233.
Present Address
†Department of Chemistry and Biochemistry, Rose-Hulman
Institute of Technology, Terre Haute, IN 47803
(12) (a) Lieber, E.; Oftedahl, E.; Rao, C. N. R. J. Org. Chem. 1963,
28, 194−199. (b) Perman, C. A.; Gleason, W. B. Acta Crystallogr., Sect.
C: Cryst. Struct. Commun. 1991, 47, 1018−1021.
Notes
The authors declare no competing financial interest.
(13) Darensbourg, D. J.; Moncada, A. I. Inorg. Chem. 2008, 47,
10000−10008.
ACKNOWLEDGMENTS
■
(14) Crawford, M.-J.; Klapotke, T. M.; Mayer, P.; Vogt, M. Inorg.
̈
Chem. 2004, 43, 1370−1378.
We gratefully acknowledge the financial support of the National
Science Foundation (CHE-1057743) and the Robert A. Welch
Foundation (A-0923). This research is supported in part by the
Department of Energy Office of Science Graduate Fellowship
Program (DOE-SCGF), made possible in part by the American
Recovery and Reinvestment Act of 2009, administered by
ORISE-ORAU under contract no. DE-AC05-06OR23100. We
also thank Dr. Yohannes Renzom for assistance with GC−MS
data collection and Dr. Nattamai Bhuvanesh for his assistance
with X-ray crystallography.
(15) Sommer, F. Ber. Dtsch. Chem. Ges. 1915, 48, 1833−1841.
(16) Bailey, P. J.; Bell, N. L.; Nichol, G. S.; Parsons, S.; White, F.
Inorg. Chem. 2012, 51, 3677−3689.
(17) It is currently unclear as to why the 1712 cm−1 polymer peak
begins to grow again at ∼14 h reaction time, but it is possible that a
variety of equilibrium processes may contribute to this trace.
(18) Darensbourg, D. J.; Bottarelli, P.; Andreatta, J. R. Macromolecules
2007, 40, 7727−7729.
(19) This DSC data comes from the first scan at a heating rate of 5
°C/min. Attempts to remove the thermal history by using multiple
scans were unsuccessful, as no thermal transitions are observed after
heating above 125 °C for the first time.
(20) Das, B.; Reddy, V. S.; Krishnaiah, M. Tetrahedron Lett. 2006, 47,
8471−8473.
(21) Wu, L.; Yang, L.; Fang, L.; Yang, C.; Yan, F. Phosphorus, Sulfur
Silicon Relat. Elem. 2010, 185, 2159−2164.
(22) van Tamelen, E. E. J. Am. Chem. Soc. 1951, 73, 3444−3448.
(23) Harding, J. S.; Owen, L. N. J. Chem. Soc. 1954, 1528−1536.
(24) Goodman, L.; Benitez, A.; Baker, B. R. J. Am. Chem. Soc. 1958,
80, 1680−1686.
(25) Iqbal, S. M.; Owen, L. N. J. Chem. Soc. 1960, 1030−1036.
(26) Barbero, M.; Degani, I.; Dughera, S.; Fochi, R.; Piscopo, L. J.
Chem. Soc., Perkin Trans. 1 1996, 289−294.
REFERENCES
■
(1) Several comprehensive reviews cover the copolymerization of
epoxides and CO2. For example, see (a) Darensbourg, D. J.; Holtcamp,
M. W. Coord. Chem. Rev. 1996, 153, 155. (b) Coates, G. W.; Moore,
D. R. Angew. Chem., Int. Ed. 2004, 43, 6618. (c) Sugimoto, H.; Inoue,
S. J. Polym. Sci., Part A: Polym. Chem. 2004, 42, 5561. (d) Darensbourg,
D. J.; Mackiewicz, R. M.; Phelps, A. L.; Billodeaux, D. R. Acc. Chem.
Res. 2004, 37, 836. (e) Chisholm, M. H.; Zhou, Z. J. Mater. Chem.
2004, 14, 3081. (f) Darensbourg, D. J. Chem. Rev. 2007, 107, 2388.
(g) Klaus, S.; Lehenmeier, M. W.; Anderson, C. E.; Rieger, B. Coord.
Chem. Rev. 2011, 255, 1460. (h) Kember, M. R.; Buchard, A.;
Williams, C. K. Chem. Commun. 2011, 141. (i) Lu, X. B.; Darensbourg,
D. J. Chem. Soc. Rev. 2012, 41, 1462. (j) Lu, X. B.; Ren, W. M.; Wu, G.
P. Acc. Chem. Res. 2012, 45, 1721. (k) Darensbourg, D. J.; Wilson, S. J.
Green Chem. 2012, 14, 2665.
(2) (a) Durden, J. A.; Stansbury, H. A.; Catlette, W. H. J. Am. Chem.
Soc. 1960, 82, 3082. (b) Taguchi, Y.; Yanagiya, K.; Shibuya, I.; Suhara,
Y. Bull. Chem. Soc. Jpn. 1988, 61, 921−925. (c) Taguchi, Y.; Yasumoto,
M.; Shibuya, I.; Suhara, Y. Bull. Chem. Soc. Jpn. 1989, 62, 474−478.
(d) Motokucho, S.; Takeuchi, D.; Sanda, F.; Endo, T. Tetrahedron
2001, 57, 7149−7152.
(3) (a) Clegg, W.; Harrington, R. W.; North, M.; Villuendas, P. J.
Org. Chem. 2010, 75, 6201−6207. (b) North, M.; Villuendas, P. Synlett
2010, 623−627.
(4) (a) Melendez, J.; North, M.; Pasquale, R. Eur. J. Inorg. Chem.
2007, 3323−3326. (b) North, M.; Pasquale, R. Angew. Chem., Int. Ed.
́
2009, 48, 2946−2948. (c) Melendez, J.; North, M.; Villuendas, P.
Chem. Commun. 2009, 2577−2579. (d) North, M.; Pasquale, R.;
Young, C. Green Chem. 2010, 12, 1514−1539.
(5) (a) North, M.; Villuendas, P.; Young, C. Chem.Eur. J. 2009,
11454−11457. (b) Metcalfe, I. S.; North, M.; Pasquale, R.; Thursfield,
A. Energy Environ. Sci. 2010, 3, 212−215.
(6) Nakano, K.; Tatsumi, G.; Nozaki, K. J. Am. Chem. Soc. 2007, 129,
15116−15117.
(7) Zhang, X.-H.; Liu, F.; Sun, X.-K.; Chen, S.; Du, B.-Y.; Qi, G.-R.;
Wan, K. M. Macromolecules 2008, 41, 1587−1590.
I
dx.doi.org/10.1021/ma4015438 | Macromolecules XXXX, XXX, XXX−XXX