Mechanistic studies of the copolymerization reaction of oxetane and carbon dioxide to provide aliphatic polycarbonates catalyzed by (salen)CrX complexes

Article abstract of DOI:10.1021/ja800302c

Chromium salen derivatives in the presence of anionic initiators have been shown to be very effective catalytic systems for the selective coupling of oxetane and carbon dioxide to provide the corresponding polycarbonate with a minimal amount of ether linkages. Optimization of the chromium(III) system was achieved utilizing a salen ligand with tert-butyl groups in the 3,5-positions of the phenolate rings and a cyclohexylene backbone for the diimine along with an azide ion initiator. The mechanism for the coupling reaction of oxetane and carbon dioxide has been studied. Based on binding studies done by infrared spectroscopy, X-ray crystallography, kinetic data, end group analysis done by ¹H NMR, and infrared spectroscopy, a mechanism of the copolymehzation reaction is proposed. The formation of the copolymer is shown to proceed in part by way of the intermediacy of trimethylene carbonate, which was observed as a minor product of the coupling reaction, and by the direct enchainment of oxetane and CO₂. The parity of the determined free energies of activation for these two processes, namely 101.9 kJ·mol^-1 for ring-opening polymerization of trimethylene carbonate and 107.6 kJ·mol^-1 for copolymerization of oxetane and carbon dioxide supports this conclusion.

Full text of DOI:10.1021/ja800302c

Published on Web 04/29/2008
Mechanistic Studies of the Copolymerization Reaction of
Oxetane and Carbon Dioxide to Provide Aliphatic
Polycarbonates Catalyzed by (Salen)CrX Complexes
Donald J. Darensbourg,* Adriana I. Moncada, Wonsook Choi, and
Joseph H. Reibenspies
Department of Chemistry, Texas A&M UniVersity, College Station, Texas 77843
Received January 14, 2008; E-mail: djdarens@mail.chem.tamu.edu
Abstract: Chromium salen derivatives in the presence of anionic initiators have been shown to be very
effective catalytic systems for the selective coupling of oxetane and carbon dioxide to provide the
corresponding polycarbonate with a minimal amount of ether linkages. Optimization of the chromium(III)
system was achieved utilizing a salen ligand with tert-butyl groups in the 3,5-positions of the phenolate
rings and a cyclohexylene backbone for the diimine along with an azide ion initiator. The mechanism for
the coupling reaction of oxetane and carbon dioxide has been studied. Based on binding studies done by
1
infrared spectroscopy, X-ray crystallography, kinetic data, end group analysis done by H NMR, and infrared
spectroscopy, a mechanism of the copolymerization reaction is proposed. The formation of the copolymer
is shown to proceed in part by way of the intermediacy of trimethylene carbonate, which was observed as
a minor product of the coupling reaction, and by the direct enchainment of oxetane and CO
2
. The parity of
-1
the determined free energies of activation for these two processes, namely 101.9 kJ ·mol for ring-opening
-
1
polymerization of trimethylene carbonate and 107.6 kJ·mol for copolymerization of oxetane and carbon
dioxide supports this conclusion.
Introduction
improved catalytic activity, selectivity, and importantly a better
understanding of the mechanistic aspects of this process for both
The utilization of carbon dioxide in the production of value-
added chemicals has received intense scrutiny over the past
alicyclic and aliphatic epoxide monomers. Nevertheless, the
production of polycarbonates from aliphatic epoxides and CO2
has suffered from the tendency of this coupling reaction to
generally favor production of the corresponding five-membered
1
decade. This is a consequence of carbon dioxide being an
abundant, nontoxic, inexpensive, and biorenewable resource. In
addition, the utilization of carbon dioxide either as a source of
carbon or as a solvent is motivated in part by the health and
safety benefits derived from its substitution for toxic chemicals.
Conforming with these objectives, the synthesis of polycarbon-
ates Via the alternative coupling of carbon dioxide and epoxides
was first demonstrated by Inoue and co-workers in 1969.
Subsequent to Inoue’s early discovery significant advances have
been made from multinational research programs in the design
of a wide variety of well-defined metal catalysts for this
5
cyclic carbonate. This general propensity of aliphatic epoxides
and CO2 for selectively affording cyclic carbonates is somewhat
unfortunate because their corresponding copolymers are bio-
degradable materials which possess a wide variety of potential
applications, including those in biomedical areas such as drug
2
3
6
delivery devices and tissue engineering applications. Notwith-
standing, recent ligand modifications have afforded catalyst
systems with greatly enhanced selectivity for copolymer produc-
7
tion from the coupling of propylene oxide and CO2. Hopefully,
4
important transformation. These studies have led to greatly
in the future these catalysts will be useful for copolymer
synthesis from a wide variety of functionalized aliphatic
epoxides.
An optional synthetic methodology for the production of
aliphatic polycarbonates involves the ring-opening polymeri-
(
(
1) (a) Aresta, M.; Dibenedetto, A. Catal. Today 2004, 98, 455–462. (b)
Sakakura, T.; Choi, J.-C.; Yasuda, H. Chem. ReV. 2007, 107, 2365–
2
387. (c) Aresta, M.; Dibenedetto, A. Dalton Trans. 2007, 2975–2992.
2) Bottenbruch, L. Engineering Thermoplastics: Polycarbonates, Poly-
acetals, Polyesters, Cellulose Esters; Hanser Publishers: New York,
1
996; p 112.
236–240. (j) Cohen, C. T.; Chu, T.; Coates, G. W. J. Am. Chem. Soc.
2005, 127, 10869–10878.
(
(
3) Inoue, S.; Koinuma, H.; Tsuruta, T. J. Polym. Sci., Part B: Polym.
Phys. 1969, 7, 287–292.
(5) (a) Kruper, W. J.; Dellar, D. V. J. Org. Chem. 1995, 60, 725–727. (b)
Kuran, W.; Listos, T. Macromol. Chem. Phys. 1994, 195, 1011–1015.
(6) (a) Albertsson, A. C.; Eklund, M. J. Polym. Sci., Part A: Polym. Chem.
1994, 32, 265–279. (b) Pêgo, A. P.; Siebum, B.; VanLuyn, M. J. A.;
Gellego, K. J.; Van Seijen, Y.; Poot, A. A.; Grijpma, D. W.; Feijen,
J. Tissue Eng. 2003, 9, 981–994.
4) (a) Chisholm, M. H.; Zhou, Z. J. Mater. Chem. 2004, 14, 3081–3092.
(
b) Darensbourg, D. J. Chem. ReV. 2007, 107, 2388–2410. (c)
Darensbourg, D. J.; Holtcamp, M. W. Coord. Chem. ReV. 1996, 153,
1
55–174. (d) Darensbourg, D. J.; Mackiewicz, R. M.; Phelps, A. L.;
Billodeaux, D. R. Acc. Chem. Res. 2004, 37, 836–844. (e) Kuran, W.
Prog. Polym. Sci. 1998, 23, 919–992. (f) Moore, D. R.; Coates, G. W.
Angew. Chem., Int. Ed. 2004, 43, 6618–6639. (g) Paddock, R. L.;
Nguyen, S. T. Macromolecules 2005, 38, 6251–6253. (h) Sugimoto,
H.; Inoue, S. J. Polym. Sci., Part A: Polym. Chem. 2004, 42, 5561–
(7) (a) Nakano, K.; Kamada, T.; Nozaki, K. Angew. Chem., Int. Ed. 2006,
45, 7274–7277. (b) Noh, E. K.; Na, S. J.; Sujith, S.; Kim, S.-W.; Lee,
B. Y. J. Am. Chem. Soc. 2007, 129, 8082–8083. (c) Cohen, C. T.;
Coates, G. W. J. Polym. Sci., Part A.: Polym. Chem. 2006, 44, 5182–
5191.
5
573. (i) Super, M. S.; Beckman, E. J. Trends Polym. Sci. 1997, 5,
10.1021/ja800302c CCC: $40.75

2008 American Chemical Society
J. AM. CHEM. SOC. 2008, 130, 6523–6533 9 6523

A R T I C L E S
Darensbourg et al.
zation of six- and seven-membered cyclic carbonates. Unlike
five-membered cyclic carbonates which are thermodynamically
stable toward polycarbonate formation in the absence of CO2
loss, six-membered cyclic carbonates can under certain catalytic
conditions provide aliphatic polycarbonates with complete
Another alternative route to aliphatic polycarbonates involves
the copolymerization of four-membered cyclic ethers, such as
oxetane, and carbon dioxide (eq 2).
8
retention of their CO2 contents (eq 1). It is noteworthy that
trimethylene carbonate is readily prepared from 1,3-propanediol
9
and ethylchloroformate or diethylcarbonate. Although, 1,3-
propanediol is currently produced industrially from petroleum
derivatives, in the context of sustainability it can also be
synthesized by microbiological or biochemical routes.
Surprisingly, this reaction has not been widely studied, but
it is of particular interest, since in this case the cyclic carbonate
byproduct, TMC, unlike five-membered cyclic carbonates, can
be ring-opened and transformed into the alternating copolymer
by means of the reaction defined in eq 1.
1
0
The first report on this reaction appeared in 1977 by
21
Koinuma, who employed a ternary catalytic system composed
of triethylaluminium, water, and acetylacetone to copolymerize
oxetane and carbon dioxide, producing poly(TMC) with a
significant quantity of ether linkages. In this case, an anionic
coordination mechanism was proposed for the formation of
For the ring-opening polymerization (ROP) of trimethylene
carbonate (TMC or 1,3-dioxan-2-one), salts of aluminum and
22
poly(TMC). Later, Matsuda reported the use of organotin
1
1
12
tin have shown to be very effective catalysts. Our group, as
halide complexes in the presence of a base, to produce low
molecular weight polycarbonates. Tetraphenylstibonium iodide
was also employed by Matsuda to selectively synthesize
1
3
well as Cao and co-workers, has reported the use of effective
salen complexes of aluminum as catalysts for this transforma-
tion. More recently, we also have shown the use of biometal
derivatives as catalysts for the ring-opening polymerization of
2
3
trimethylene carbonate from oxetane and carbon dioxide. In
addition, Matsuda later on employed organotin iodide complexes
with phosphines or phosphine oxides, to catalyze the addition
of carbon dioxide to oxetane, yielding TMC and poly(TMC)
products. It was demonstrated in this instance that the choice
of the ligand was very important; all complexes with Bu3P
3
1
4,15
TMC and lactide.
The latter catalytic systems are of
particular importance because the use of biocompatible metals,
such as calcium, magnesium, and zinc, eliminates the difficulty
of removing trace amounts of metal residues from the produced
1
6
polycarbonates. Homoleptic lanthanide amidinate complexes,
provided poly(TMC), but in the presence of Bu PdO, TMC
1
7
24
samarium borohydride complexes, and 2,2-dibutyl-2-stanna-
,3-oxepane have also recently been investigated as catalysts
was produced exclusively in good yields. Moreover, Matsuda
1
8
1
proposed a mechanism for the formation of poly(TMC), where
TMC was formed first, and subsequent ring-opening polymer-
for the ring-opening polymerization of trimethylene carbonate
and its copolymerization with ε-caprolactone and L-lactide. In
addition, organocatalysts in the presence of benzyl alcohol were
reported by Hedrick as catalytic systems for the ROP of TMC,
where high polymerization control, low polydispersities, and
2
4
ization of the preformed TMC produced poly(TMC).
Of importance, these early reports on the coupling reaction
of oxetane and carbon dioxide suggest that a full understand-
ing of the mechanistic aspects of this process is lacking.
Recently, our group has reported a variety of catalytic systems
based on metal salen complexes for the copolymerization of
19
high end group fidelity could be obtained. Similarly, Bowden
and co-workers have utilized 2-(dimethylamino)ethanol as an
20
4b
effective catalyst for the ring-opening polymerization of TMC.
cyclohexene oxide or propylene oxide and carbon dioxide.
Optimizations of the catalytic systems and catalytic conditions
have led us to produce copolymers with greater than 99%
carbonate linkages, low polydispersities, and high molecular
weights. This success has motivated us to examine the efficiency
of metal salen complexes for the copolymerization reaction of
oxetane and carbon dioxide to afford aliphatic polycarbonates.
In a recent communication we presented preliminary studies
on this reaction catalyzed by metal salen complexes of
chromium and aluminum, where it was found that a (salen)-
Cr(III)Cl complex (N,N′-bis(3,5-di-tert-butylsalicylidene)-1,2-
ethylenediimine chromium(III) chloride) was more active than
its aluminum salen analogue to catalyze this coupling reaction
(
(
8) Clements, J. H. Ind. Eng. Chem. Res. 2003, 42, 663–674.
9) (a) Ariga, T.; Takata, T.; Endo, T. Macromolecules 1997, 30, 737–
7
3
44. (b) Carothes, W. H.; Natta, F. J. V. J. Am. Chem. Soc. 1930, 52,
14–326.
(
10) (a) Tong, I. T.; Cameron, D. C. Appl. Biochem. Biotechnol. 1992,
3
4/35, 149–159. (b) Haynie, S.; Wagner, L.; Winona, L. (i.e. DuPont
de Nemours and Co.)Patent WO/035799, 1996.
(
(
11) Rokicki, G. Prog. Polym. Sci. 2000, 25, 259–342.
12) Darensbourg, D. J.; Ganguly, P.; Billodeaux, D. R. Macromolecules
2
005, 38, 5406–5410.
(
(
(
(
(
(
13) Yang, J.; Yu, Y.; Li, Q.; Li, Y.; Cao, A. J. Polym. Sci., Part A: Polym.
Chem. 2005, 43, 373–384.
14) Darensbourg, D. J.; Choi, W.; Ganguly, P.; Richers, C. P. Macro-
molecules 2006, 39, 4374–4379.
2
5
in the presence of n-Bu4NN3 as the cocatalyst. In all the
instances, high selectivity for copolymer formation was obtained
even at high temperatures (110 °C). Based on circumstantial
15) Darensbourg, D. J.; Choi, W.; Richers, C. P. Macromolecules 2007,
4
0, 3521–3523.
16) Li, C.; Wang, Y.; Zhou, L.; Sun, H.; Shen, Q. J. Appl. Polym. Sci.
2
006, 102, 22–28.
17) Palard, I.; Schappacher, M.; Belloncle, B.; Soum, A.; Guillaume, S. M.
Chem.sEur. J. 2007, 13, 1511–1521.
(21) Koinuma, H.; Hirai, H. Makromol. Chem. 1977, 178, 241–246.
18) Pospiech, D.; Komber, H.; Jehnichen, D.; Häussler, L.; Eckstein, K.;
Scheibner, H.; Janke, A.; Kricheldorf, H. R.; Petermann, O. Biomac-
romolecules 2005, 6, 439–446.
(22) Baba, A.; Meishou, H.; Matsuda, H. Makromol. Chem. Rapid Commun.
1984, 5, 665–668.
(23) Baba, A.; Kashiwagi, H.; Matsuda, H. Tetrahedron Lett. 1985, 26,
(
19) Nederberg, F.; Lohmeijer, B. G. G.; Leibfarth, F.; Pratt, R. C.; Choi,
J.; Dove, A. P.; Waymouth, R. M.; Hedrick, J. L. Biomacromolecules
1323–1324.
(24) Baba, A.; Kashiwagi, H.; Matsuda, H. Organometallics 1987, 6, 137–
2
007, 8, 153–160.
20) Mindemark, J.; Hilborn, J.; Bowden, T. Macromolecules 2007, 40,
515–3517.
140.
(
(25) Darensbourg, D. J.; Ganguly, P.; Choi, W. Inorg. Chem. 2006, 45,
3
3831–3833.
6
524 J. AM. CHEM. SOC. 9 VOL. 130, NO. 20, 2008

Copolymerization Reaction of Oxetane and CO
2
A R T I C L E S
evidence, it was suggested that formation of copolymer did not
proceed Via the intermediacy of TMC which was observed as
a minor product of the coupling reaction.
Herein, we have extended these investigations by further
optimizing the (salen)CrCl catalytic system for this reaction.
Efforts have been made to gain a greater insight into the
mechanistic aspects of this important reaction based on kinetic
and copolymer end-group analysis studies performed utilizing
1
in situ infrared and H NMR spectroscopies.
Experimental Section
Reagents and Methods. Unless otherwise specified, all syntheses
and manipulations were carried out on a double-manifold Schlenk
vacuum line under an atmosphere of argon or in an argon filled
glovebox. Toluene and tetrahydrofuran were freshly distilled from
Figure 1. Structures of the (salen)Cr(III) chloride complexes utilized as
catalysts for the copolymerization of oxetane and CO2.
sodium/benzophenone. Ethanol was freshly distilled from Mg/I
,1,2,2-tetrachloroethane (TCE) was freshly distilled from CaH
Diethyl ether, dichloromethane, and pentane were purified by an
MBraun Manual Solvent Purification System packed with Alcoa
F200 activated alumina desiccant. Oxetane (Alfa Aesar) was freshly
2
.
.
Table 1. Selectivity for Copolymer Formation using Complex 1 in
1
2
a
the Presence of Various Cocatalysts
b
b
b
entry
cocatalyst
% TMC
% poly(TMC)
% ether linkages
c
1
2
3
4
5
6
n-Bu4NCl
c
0
0
11.7
5.9
2.3
21.1
100
100
88.2
94
97.6
79
3.0
2.9
7.2
3.6
1.4
distilled from CaH
2
and stored in the freezer of the glovebox.
n-Bu4NN3
n-Bu NBr
Trimethylene carbonate (Boehringer Ingelheim) was recrystallized
from tetrahydrofuran and diethyl ether, dried under vacuo, and
4
PPNCl
PPNN3
P(Cy)3
stored in the glovebox. Tricyclohexylphosphine (Alfa Aesar) was
+
21.5
recrystallized from distilled ethanol before use. PPNCl (PPN
)
+
(Ph P)
3 2
N ) (Aldrich) was recrystallized from dichloromethane/ether
was synthesized according to a published
a
Copolymerization conditions: 17 mg of catalyst (0.15 mol%), 1.15 g
before use, and PPNN
3
2
6
of oxetane, M/I ) 675:1, 2 equiv of cocatalyst, 10 mL of toluene, 35
procedure. Tetra-n-butylammonium bromide (Aldrich) was re-
crystallized from acetone/diethyl ether before use. Tetra-n-buty-
lammonium azide (TCI) was stored in the freezer of the glovebox
upon arrival. Bone-dry carbon dioxide supplied in a high-pressure
cylinder and equipped with a liquid dip tube was purchased from
Scott Specialty Gases. The corresponding salen ligands and
b
bar of CO2, at 110 °C for 24 h. Product distributions were determined
1
c
25
by H NMR spectroscopy. Previous published results. Note for the
first entry in Table 1 in ref 25, the % TMC is incorrectly reported due
1
to misassigned H NMR resonances.
ppm, respectively. As is readily seen in Table 1, the yield of
poly(TMC) is much greater than that of the cyclic product,
TMC, at the end of a 24 h reaction period in all instances. Of
importance here, the azide salts are slightly better than their
2
7
chromium complexes were synthesized as previously described.
1
H NMR spectra were acquired on Unity+ 300 MHz and VXR
00 MHz superconducting NMR spectrometers. IR spectra were
3
recorded on a Mattson 6021 Fourier Transform (FT) IR spectrom-
eter with an MCT detector. TGA measurements were performed
with SDT Q600 V7.0 Build 84. Analytical elemental analysis was
provided by Canadian Microanalytical Services Ltd. Molecular
+
chloride analogues, with little difference between the PPN and
n-Bu4N+ salts being observed. The bromide anion was shown
to be significantly less selective for copolymer formation than
the chloride anion, and tricyclohexylphosphine displayed the
lowest selectivity toward copolymer formation of the cocatalysts
weight determinations (M
Modular GPC apparatus equipped with ViscoGEL I-series columns
H + L) and a model 270 dual detector comprised of refractive
n w
and M ) were carried out with a Viscotek
(
28
studied.
index and light scattering detectors. High-pressure reaction kinetic
measurements were performed using an ASI ReactIR 1000 reaction
analysis system with a stainless steel Parr autoclave modified with
a permanently mounted ATR crystal (SiComp) at the bottom of
the reactor (purchased from Mettler Toledo).
Subsequent studies were carried out to interrogate the effects
of changing the nature of (i) the substituents on the phenolate
rings and (ii) the diimine backbone of the salen ligand in the
(
salen)CrCl derivative. In this instance the copolymerization
reactions were performed at a monomer/catalyst/cocatalyst ratio
of 1292:1:2, with a CO2 pressure of 35 bar at 110 °C for 24 h.
The results of this inquiry are provided in Table 2, where the
TOFs (mol of oxetane consumed/mol of initiator ·h) were
determined from the isolated copolymer obtained upon precipi-
tation from dichloromethane with a 1 M HCl solution in
methanol. It is important to note here that minimal quantities
of ether linkages were observed in the copolymer samples
resulting from consecutive oxetane ring-opening processes;
however, in all cases the CO2 content was very high (>95%).
Note that 100% carbon dioxide incorporation defines a com-
pletely alternating copolymer of oxetane and CO2. The molec-
ular weights of the copolymers were determined in THF solution
by gel permeation chromatography using RI and light scattering
detectors and polystyrene standards. In general, the observed
Results and Discussion
Initially we chose to employ the (salen)CrCl catalyst, complex
1
in Figure 1, in the presence of various cocatalysts to optimize
the selectivity for copolymer formation from the coupling of
oxetane and carbon dioxide. The copolymerization reactions
were performed under identical reaction conditions, i.e., 110
°
C and 35 bar CO2 pressure. The results are summarized in
Table 1, where the counterions of the anionic initiators were
+
+
+
either PPN [(Ph3P)2N ] or n-Bu4N . The product mixtures were
1
analyzed by H NMR spectroscopy, with the quantities of
poly(TMC), TMC, and ether linkages in the poly(TMC)
determined by integrating the resonances at 4.23, 4.43, and 3.50
(
(
26) Demadis, K. D.; Meyer, T. J.; White, P. S. Inorg. Chem. 1998, 37,
3
610–3619.
27) Darensbourg, D. J.; Mackiewicz, R. M.; Rodgers, J. L.; Fang, C. C.;
(28) (a) Darensbourg, D. J.; Mackiewicz, R. M. J. Am. Chem. Soc. 2005,
127, 14026–14038. (b) Darensbourg, D. J.; Mackiewicz, R. M.;
Rodgers, J. R. J. Am. Chem. Soc. 2005, 127, 17565.
Billodeaux, D. R.; Reibenspies, J. H. Inorg. Chem. 2004, 43, 6024–
6
034.
J. AM. CHEM. SOC. 9 VOL. 130, NO. 20, 2008 6525

A R T I C L E S
Darensbourg et al.
a
Table 2. Copolymerization of Oxetane and CO2 Catalyzed by (salen)Cr(III)Cl Complexes
b
c
d
d
entry
complex
R
1
R
2
R
3
R
4
TON
TOF
% CO
2
content
% conversion
1
2
3
4
2
3
1
4
-C4H4-
-C4H4-
H
tert-butyl
OCH3
tert-butyl
tert-butyl
tert-butyl
tert-butyl
tert-butyl
tert-butyl
588
382
775
835
24.5
15.9
32.3
34.8
97.1
97.5
95.5
95.9
54.6
33.3
67.9
71.6
H
(1R,2R)-C4H8-
a
Copolymerization conditions: Catalyst loading ) 0.077 mol %, 4 g of oxetane, 2 equiv of n-Bu4NN3, M/I ) 1292, 35 bar of CO2, at 110 °C for
b
c
d
1
2
4 h. Mol of oxetane consumed/mol of initiator. Mol of oxetane consumed/(mol of initiator ·h). Estimated by H NMR spectroscopy.
a
Table 3. Molecular Weights and Polydispersities of Poly(TMC)
b
c
entry
M/I
% conversion
M
n
(GPC)
M
n
(theoretical)
w n
M /M (GPC)
1
2
3
4
150
275
350
475
150
99.5
90.7
48.5
29.7
100.0
11 200
9500
6700
7050
14 500
15 200
25 500
17 300
14 400
15 300
1.26
1.45
1.43
1.60
1.30
d
5
a
Copolymerization conditions: Reactions carried out in toluene at
equiv of n-Bu4NN3 as
1
10 °C using (salen)Cr(III)Cl catalyst, b
2
1
cocatalyst at c 35 bar CO2 pressure.
Estimated by
H NMR
spectroscopy. Mn(theoretical) ) M/I × mol wt(oxetane + CO2) × %
d
conversion. Done in the absence of solvent using PPNN3 as cocatalyst.
Scheme 1
Figure 2. Spectra of TCE solutions of chromium salen azide complex with
equiv of n-Bu4NN3(A), after addition of 100 equiv of oxetane at room
1
temperature (B), after heating the reaction solution at 110 °C for 3 h (C).
the diimine backbone from ethylene to cyclohexylene, with the
chromium salen complex with the cyclohexylene backbone
displaying slightly higher catalytic activity.
Mn values were found to be much lower than the theoretical
values; e.g., in entry 4 of Table 2 the observed Mn 11 050 is
considerably lower than the theoretical value 85 000. This is
most likely due to a chain transfer mechanism arising from the
Substrate Binding and Ring-Opening Steps Examined by
Infrared Spectroscopy. Fundamental to a better understanding
of the mechanism of the coupling reaction of oxetane and carbon
dioxide is an investigation of the initiation step of this process.
Since it is known that oxetane has less ring strain compared to
epoxides, e.g., the heat of polymerization of ethylene oxide
(-∆Hp ) 104 kJ/mol) differing from that of oxetane by 23
7a,29
presence of trace quantities of water in the system.
The PDI
was in general about 1.5, and a copolymer with an Mn of 11 050
was found to be stable up to 260 °C by TGA. A more
comprehensive compilation of molecular weights as a function
of the monomer/initiator ratio, along with polydispersities, can
be found in Table 3. Included in Table 3 is a catalytic run done
under superanhydrous conditions, i.e., no solvent and in the
presence of the less hydroscopic PPNN3 salt. As is evident under
these conditions the molecular weight more closely tracks the
predicted value.
3
0
kJ/mol, its ease of ring opening should depart significantly
from that of epoxides. In order to address this issue we have
conducted cocatalyst, oxetane binding, and subsequent ring-
opening studies Via infrared spectroscopy using the (salen)CrN3
complex containing di-tert-butyl substituents in the 3,5-positions
of the phenolate rings and an ethylene backbone for the diimine.
We have employed the azide derivatives for these studies
Retaining the salen ligand with the phenylene backbone while
changing the substituents in the 3,5-positions of the phenolate
rings (entries 1 and 2, Table 2) reveals the Cr(III) salen
derivative containing the bulky di-tert-butyl groups to be the
more active. This is consistent with previous observations
reported for the ROP of TMC catalyzed by aluminum and
because the νN stretching vibration provides accessible probes
3
for both cocatalyst binding and anion ring-opening steps. The
results of these studies are depicted in Scheme 1 and Figure 2.
As indicated in Scheme 1, upon addition of 1 equiv of
n-Bu4NN3 to (salen)CrN3, the anionic six-coordinate bis-azide
species (salen)Cr(N3)2- readily forms at ambient temperature.
1
2,14,15
calcium salen complexes.
On the other hand, for the
copolymerization of cyclohexene oxide and CO2 catalyzed by
chromium salen complexes, higher catalytic activity was
obtained in complexes containing methoxy and tert-butyl groups
This is apparent in the νN stretching region where the infrared
3
-
1
band of (salen)CrN3 shifts from 2083 cm to a band at 2047
cm with a shoulder at 2057 cm upon addition of n-Bu4NN3.
-1
-1
2
7
in the 3- and 5-positions of the phenolate rings. We have also
studied the effects of altering the diimine backbone of the Cr(III)
salen complex while maintaining the di-tert-butyl groups in the
+
It should be noted here that the n-Bu4N salts of numerous
(
salen)CrX2- anions have been fully characterized by X-ray
crystallography and these studies will be reported elsewhere.
For example, Figure 3 illustrates the solid-state structure of the
bis-azide anion of one such derivative. Addition of a 100-fold
excess of oxetane to the bis-azide complexes displaces some
3
2
,5-positions of the phenolate moiety (entries 3 and 4, Table
). As can be seen in Table 2, the catalytic behavior of the
chromium salen complexes is not much affected by changing
(
29) (a) Darensbourg, D. J.; Fitch, S. B. Inorg. Chem. 2007, 47, 5474–
476. (b) Sugimoto, H.; Ohtsuka, H.; Inoue, S. J. Polym. Sci., Part
5
(30) Sawada, H. Macromol. Sci., ReV. Macromol. Chem. 1970, C5, 151–
A: Polym. Chem. 2005, 43, 4172–4186.
174.
6
526 J. AM. CHEM. SOC. 9 VOL. 130, NO. 20, 2008

Copolymerization Reaction of Oxetane and CO
2
A R T I C L E S
Table 4. Crystallographic Data for Complexes 5 and 6
5
6
empirical formula
C80H110ClCr2N4O6
C34.40H48.80Cl3.20
Cr0.80N1.60O2.40
678.19
fw
1364.17
110(2)
temperature (K)
crystal system
space group
a (Å)
b (Å)
c (Å)
110(2)
triclinic
triclinic
j
j
P1
P1
12.266(5)
14.702(5)
21.630(5)
102.839(5)
95.784(5)
100.245(5)
3702(2)
1.223
16.175(5)
16.394(5)
17.126(5)
89.781(4)
88.053(4)
79.249(4)
4459(2)
1.263
R (deg)
Figure 3. Ball-and-stick representation of the X-ray defined structure of
the anion of the [n-Bu4N][(salen)Cr(N3)2] complex, where the salen ligand
ꢀ
(deg)
t
γ (deg)
contains -OMe and Bu substituents in the 3,5-positions of the phenolates,
3
V (Å )
respectively, with a phenylene diimine backbone.
3
Dc (Mg/m )
Z
2
5
abs coeff (mm-
1
)
0.384
0.535
of the azide ligand as can be seen by an increase in the free
reflections collected
independent reflections
10 087
41 341
-1
azide ion concentration by its νN band at 2009 cm with a
3
10 392 [ (int) )
0.0755]
51/856
15 610 [R(int) )
0.0612]
48/979
-
concomitant decrease in the concentration of (salen)Cr(N3)2
(
2
spectrum B). Moreover, a new νN stretching band appears at
restrains/parameters
3
2
061 cm- which is assigned to (salen)Cr(N3)·oxetane. Upon
1
GOF on F
1.029
1.077
a
a
final R indices [I > 2σ(I)]
R1 ) 0.0755
R1 ) 0.0579
stirring this reaction mixture for 24 h at ambient temperature
no changes in the infrared spectrum resulted, indicative of the
ring-opening process of oxetane requiring higher temperatures.
Indeed, heating the reaction mixture for 3 h at 110 °C led to
oxetane ring opening by azide as indicated by the organic azide
band at 2100 cm- . It should be recalled that an analogous
experiment carried out with cyclohexene oxide results in epoxide
ring-opening by azide at ambient temperature, although the
b
b
Rw ) 0.1995
Rw ) 0.1433
a
a
final R indices (all data)
R1 ) 0.1239
Rw ) 0.2321
R1 ) 0.0908
Rw ) 0.1643
b
b
a
b
2
2 2
2 2 1/2
R ) Σ|Fo| - |Fc||/Σ|Fo|. Rw ) {[Σw(Fo - Fc ) ]/[Σw(Fo ) ]} ; w
2
2
2
2
2
)
1/[σ (Fo ) + (aP) + bP], where P ) [(max(Fo ), 0) + 2(Fc )]/3.
1
-
equilibrium between (salen)Cr(N3)2 and cyclohexene oxide lies
2
7
much farther to the left. That is, cyclohexene oxide has less
of a penchant for displacing the azide from the chromium(III)
center than the more basic oxetane monomer.
X-ray crystallography was utilized in conjunction with the
νN infrared spectral data (Vide supra) to verify that oxetane
3
binding to the chromium center occurs without ring-opening at
ambient temperature. Two (salen)CrCl complexes with oxetane
bound to the chromium centers were successfully characterized
by X-ray crystallography. To the best of our knowledge these
represent the only crystal structures of oxetane bound to a metal
center thus far reported in the literature. Crystallographic data
pertaining to these two crystal structures are provided in Table
4
.
Figure 4. Thermal ellipsoid plot of complex 5. Ellipsoids are at the 50%
level. H atoms are omitted for clarity. One pentane molecule was crystallized
in the unit cell and is omitted for clarity.
In our first attempt at isolating single crystals of a (salen)CrCl
complex with an axially coordinated oxetane, the presence of
trace quantities of water led to hydrolysis of the Cr-Cl bond.
Nevertheless, crystals of a hydroxo-bridged structure with
oxetane bound to one of the chromium centers were obtained
which were suitable for X-ray analysis (Figure 4 and Table 5).
Similar solid-state structures of hydroxo-bridged chromium(III)
derivatives have been reported by us resulting from the complete
a
Table 5. Selected Bond Distances and Angles for Complex 5
Cr(1A)-Cl(1)
2.342(3)
2.007(6)
1.949(6)
2.054(9)
90.5(8)
90.5(8)
91.1(7)
87.6(8)
Cr(1A)-O(1)
Cr(1B)-O(1)
Cr(1B)-O(11)
O(11)-C(11)-C(21)
O(11)-C(31)-C(21)
C(11)-O(11)-C(31)
C(31)-C(21)-C(11)
3
1
hydrolysis of (acacen)CrCl or (salen)CrOMe derivatives.
Complex 5 clearly shows that oxetane is capable of binding to
the (salen)Cr(III) center without undergoing ring-opening at
ambient temperature. The oxetane molecule is disordered in
complex 5. Upon modeling the disorder two different positions
for the oxetane ligand were found, where the dihedral angles
of the plane C-O-C and C-C-C were determined to be 4.5°
and 10.8° at 110 K. That is, the oxetane molecule is not planar,
which is in good agreement with the structure of free oxetane
a
Units of bond angles and bond distances are (deg) and (Å),
respectively.
reported by Luger and Buschmann where the dihedral angle
3
2
was found to be 10.7° at 90 K and 8.7° at 140 K.
A successful isolation of single crystals of an oxetane adduct
as depicted in Scheme 1 was achieved upon utilizing superdry
conditions and low temperature. Complex 6 was fully character-
ized by X-ray crystallography, and a thermal ellipsoid repre-
sentation of this derivative is shown in Figure 5, with selected
(
(
31) Darensbourg, D. J.; Frantz, E. B.; Andreatta, J. R. Inorg. Chim. Acta
007, 360, 523–528.
32) Luger, P.; Buschmann, J. J. Am. Chem. Soc. 1984, 106, 7118–7121.
2
J. AM. CHEM. SOC. 9 VOL. 130, NO. 20, 2008 6527

A R T I C L E S
Darensbourg et al.
Figure 5. Thermal ellipsoid plot of complex 6. Ellipsoids are at the 50% level. H atoms are omitted for clarity. Three dichloromethane and one pentane
molecules were crystallized in the unit cell and are omitted for clarity.
a
Table 6. Selected Bond Distances and Angles for Complex 6
carbon dioxide to the 110 °C reaction mixture, as well as a
slow initiation step.
Cr(1)-Cl(1)
2.3167(6)
2.3036(7)
2.0456(10)
2.0528(11)
90.75(11)
90.88(12)
85.72(12)
90.08(11)
Cr(2)-Cl(2)
The copolymerization reactions were demonstrated to be first-
order in oxetane and catalyst (complex 4) concentrations in the
presence of 2 equiv of n-Bu4NN3. This is shown in a representa-
tive case in Figure 7 where a plot of ln[(A∞ - At)/(A∞ - Ao)]
vs time for the formation of copolymer is found to be linear,
with A∞ and At being the absorbance of the νCdO band of the
copolymer at t ) ∞ and t ) time, respectively. It is important
to note here that the data plotted in Figure 7 are for the early
portion of the reaction because at higher levels of conversion
Cr(1)-O(3)
Cr(2)-O(6)
O(6)-C(76)-C(77)
O(6)-C(78)-C(77)
C(78)-C(77)-C(76)
C(78)-O(6)-C(76)
a
Units for bond distances and bond angles are (Å) and (deg),
respectively.
2
9
bond distances and bond angles listed in Table 6. Two molecules
crystallized in the asymmetric unit, where the sum of the angles
in the oxetane ligands is 359.6° and 357.4°, respectively. The
dihedral angles of the planes C-O-C and C-C-C in the
oxetane ligands were found to be 10.5° and 11.8° at 110 K,
which clearly demonstrates that the oxetane molecule when
bound to the metal center maintains its degree of nonplanarity.
Metric parameters for free oxetane and bound oxetane in
complex 6 are quite comparable as seen in Table 7.
(>65%) the reaction solution becomes more viscous. Simi-
larly, the order of the reaction with regard to complex 4 was
observed to be first-order based on the linear relationship
between initial rate vs complex 4 concentration (Figure 8). As
is readily seen in Figure 8, the linear plot has a nonzero intercept
of the x-axis of ∼0.005 M, suggesting that a small quantity of
the catalyst is degraded in the copolymerization process.
Relevant to this latter point, when the copolymerization reaction
is carried out under identical conditions except in neat oxetane
at a catalyst concentration of 0.005 M the system is active for
copolymer formation.
The dependence of the coupling of oxetane and carbon
dioxide on the concentration of the anionic initiator (cocatalyst)
was examined in the instance of n-Bu4NN3. As is evident in
Figure 9 the copolymerization rate was found to be first-order
in [cocatalyst] up to approximately 2 equiv of n-Bu4NN3. Hence,
in the range of [n-Bu4NN3] investigated between 0.5 and 2.0
equiv, the copolymerization process can be described as being
first-order in oxetane, (salen)CrX, and cocatalyst concentrations,
with zero-order dependence on [cocatalyst] greater than 8–10
equiv. This latter behavior is to be expected for a reaction where
the initiation step is not significantly faster than the propagation
step.
Kinetic Studies of the Copolymerization of Oxetane and
Carbon Dioxide. Kinetic measurements of the coupling reaction
of oxetane and carbon dioxide were performed in toluene
solution in the presence of complex 4 along with 2 equiv of
n-Bu4NN3. These reactions were monitored by in situ infrared
spectroscopy by observing the growth of the copolymer’s νCdO
band at 1750 cm- as a function of time. A typical reaction
1
-
1
profile of the absorbance of the 1750 cm infrared band with
time is illustrated in Figure 6. It should be noted as well in
Figure 6 that at the early stages of the coupling reaction a νCdO
-
1
band at ∼1770 cm assigned to TMC was observed which
subsequently disappeared. Furthermore, the presence of the
induction period seen in Figure 6 can be attributed to the drop
in temperature of ∼20 °C observed following the addition of
6
528 J. AM. CHEM. SOC. 9 VOL. 130, NO. 20, 2008

Copolymerization Reaction of Oxetane and CO
2
A R T I C L E S
a
Table 7. Selected Bond Distances and Angles for Oxetane Molecules
a
Units for bond distances and bond angles are (Å) and (deg), respectively.
Figure 7. First-order plot of the conversion of oxetane and CO2 to
poly(TMC). Reaction carried out in toluene at 110 °C with complex 4
(
0.0327 M), 2 equiv of n-Bu4NN3 at 35 bar CO2 pressure. Oxetane
2
concentration ) 4.92 M. Slope ) -0.0183, y intercept ) 0.6525 with R
)
0.9946.
Figure 6. Three-dimensional stack plot reaction profile of the IR spectra
collected every 3 min during the copolymerization reaction of oxetane and
carbon dioxide. Reaction carried out at 110 °C in toluene at 35 bar CO2
pressure.
Figure 8. Dependence of copolymerization reaction on [catalyst]. Reactions
carried out in toluene at 110 °C with complex 4, 2 equiv of n-Bu4NN3 at
The copolymerization of oxetane and CO2 was performed at
several temperatures between 80 and 110 °C, and the observed
rate constants for the process kp, where rate ) kp[oxetane][cata-
lyst][cocatalyst], were determined. These kp values as a function
of temperature are listed in Table 8, with the corresponding
Eyring plot depicted in Figure 10. The calculated activation
3
5 bar CO2 pressure. Oxetane concentration ) 4.92 M. Initial rate vs
-5 2
[
complex 4] provided a y-intercept of -2.10 with R ) 0.9931.
8.2 J/mol · deg, respectively. In this instance the enthalpy of
‡
activation (∆H ) is very similar to that observed for the
‡
‡
parameters ∆H and ∆S were 45.6 ( 3 kJ/mol and -161.9 (
copolymerization of cyclohexene oxide and CO2 (45.5 kJ/mol)
J. AM. CHEM. SOC. 9 VOL. 130, NO. 20, 2008 6529

A R T I C L E S
Darensbourg et al.
Figure 9. Initial rates for production of poly(TMC) as a function of number
of equivalents of cocatalyst. Reactions carried out in toluene at 110 °C
with complex 4 (0.014 M) and n-Bu4NN3 as cocatalyst, at 35 bar CO2
pressure, oxetane concentration ) 4.92 M.
Table 8. Variable Temperature Rate Constants for the
a
Copolymerization Reaction
2
(L /mol ·s)
2
T (K)
k
p
3
3
3
3
53
63
73
83
0.00449
0.00741
0.0103
0.0171
a
Each experiment was performed in toluene at 110 °C with 0.014 M
of complex 4 and 2 equivalents of n-Bu4NN3 at 35 bar CO2 pressure.
Oxetane concentration ) 4.92 M.
Figure 11. (A) Plot of monomer conversion vs time. (B) First-order plot
of the conversion of TMC to poly(TMC). Reaction carried out in TCE at
1
20 °C with complex 4 (0.005 06 M) and 2 equiv of n-Bu4NN3 (0.0098
M). TMC concentration ) 0.98 M, Slope ) 0.0366, y-intercept ) -0.0092
2
with R ) 0.9975.
Table 9. Rate Constant Dependence on the Concentrations of
a
Catalyst and Cocatalyst for the ROP of TMC
-
1
)
entry
[catalyst] (mol/L)
equiv of n-Bu NN
4
3
temp (°C)
kobsd (h
1
2
3
4
5
6
7
8
9
0.002 53
0.003 80
0.005 06
0.010 10
0.005 06
0.005 06
0.005 06
0.005 06
0.005 06
0.005 06
0.005 06
0.005 06
2
2
2
2
120
120
120
120
120
120
120
120
120
120
120
120
0.0169
0.0295
0.0366
0.071
0.0055
0.0168
0.0225
0.0275
0.0366
0.0414
0.0473
0.0530
0
0.5
0.75
1
2
3
Figure 10. Eyring plot for the formation of poly(TMC) in the presence of
complex 4/n-Bu4NN3 catalyst system in toluene. Slope ) -5502.5 with
2
R ) 0.992.
1
1
1
1
2
3
6
8
and lower than that for the process involving the monomers
propylene oxide and CO2 (66.2 kJ/mol) employing a similar
3
3,34
a
catalyst system.
Monomer concentration held at 0.98 M. Reactions carried out in
1
,1,2,2,-tetrachloroethane at atmospheric pressure.
Solution Kinetic Studies of the Ring-Opening Polymeriza-
tion of Trimethylene Carbonate. Herein we wish to describe in
detail kinetic measurements for the ring-opening polymerization
of trimethylene carbonate to poly(TMC), eq 1, under comparable
conditions to that reported above for the copolymerization of
1
were monitored by H NMR spectroscopy. Figure 11a displays
a typical monomer consumption vs time plot, whereas the
semilogarithmic plot of -ln([monomer]0/[monomer]t) vs time
is displayed in Figure 11b, and as might be expected the
polymerization reaction was found to be first-order with respect
to [TMC]. Table 9 summarizes the determined rate constants
for the ROP of TMC as a function of [catalyst] and [cocatalyst].
Log–log plots of the rate constants vs [catalyst] and [cocatalyst]
reveal relationships between ln kobsd vs ln[catalyst] or ln[co-
catalyst] with slopes close to unity, thus, indicative of the
polymerization reaction being first-order in [catalyst] and first-
order with respect to [cocatalyst] up to 1 equiv of cocatalyst.
Figures 12 and 13 illustrate the effect of excess quantities of
cocatalyst (n-Bu4NN3) on the rate constant of the ring-opening
3
5
oxetane and CO2 leading to poly(TMC). In this instance the
use of 1,1,2,2-tetrachloroethane was found to be advantageous
because of its high boiling point and the high solubility of both
monomer and polymer in a chlorinated solvent. These reactions
(
(
(
33) Darensbourg, D. J.; Yarbrough, J. C. J. Am. Chem. Soc. 2002, 124,
6
335–6342.
34) Darensbourg, D. J.; Yarbrough, J. C.; Ortiz, C.; Fang, C. C. J. Am.
Chem. Soc. 2003, 125, 7586–7591.
35) We have previously described the ROP of TMC catalyzed by
1
2,14,15
complexes dervied from biocompatible metals.
6
530 J. AM. CHEM. SOC. 9 VOL. 130, NO. 20, 2008

Copolymerization Reaction of Oxetane and CO
2
A R T I C L E S
Figure 12. Rate constant for production of poly(TMC) as a function of
number of equivalents of cocatalyst (n-Bu4NN3). Data taken from Table 8.
Figure 14. Eyring plot for the formation of poly(TMC) in the presence of
2
complex 4/n-Bu4NN3 catalyst system in TCE. Slope ) -8910.7 with R )
0
.9997.
Figure 13. Double reciprocal plot of the rate constant dependence of the
ROP process with [cocatalyst]. Data taken from Table 8. Slope ) 0.1044
2
and y-intercept ) 16.792 with R ) 0.9966.
Table 10. Variable Temperature Rate Constants for the
a
Polymerization Reaction
1
Figure 15. H NMR spectrum in CDCl3 of poly(TMC) obtained by way
of oxetane/CO2, in the presence of (salen)Cr(III)Cl/n-Bu4NN3 as the catalytic
system. Polymer was purified from dichloromethane and 1 M HCl solution
in methanol.
2
(L /mol ·s)
2
T (K)
k
p
3
3
3
4
78
83
93
03
0.07621
0.10469
0.19854
0.34825
the two processes are energetically quite similar. Hence, the
formation of polycarbonate from the oxetane and CO2 coupling
reaction could be occurring Via two different or concurrent
pathways, i.e., the intermediacy of TMC formation and subse-
quent polymerization and/or the direct enchainment of oxetane
and CO2. Indeed, in situ infrared spectroscopic monitoring of
the copolymerization reaction suggests both pathways are
operative (Vide supra). Further studies detailed below based on
a
Each experiment was performed in TCE with complex
0.00506 M) and n-Bu4NN3 as cocatalyst (0.0098 M), TMC
4
(
concentration ) 0.98 M.
polymerization process. As can be readily seen, the ROP
reaction ultimately becomes independent of [cocatalyst] loadings
as would be expected. A double reciprocal plot of these data
reveals a linear relationship with a limiting rate constant (kobsd)
of 0.059 h- at 120 °C.
The ring-opening polymerization of TMC was carried out
over the temperature range 105–130 °C in order to obtain the
activation parameters for this process. The rate constants of the
ring-opening reaction, kp, are listed in Table 10. The activation
parameters ∆H and ∆S calculated from the Erying plot shown
in Figure 14 were determined to be 74.1 ( 3.0 kJ/mol and -72.3
2.3 J/mol ·K, respectively. These parameters are consistent
with a reaction pathway involving the attack of a nucleophilic
center (polymer chain end) to a metal-bound cyclic carbonate
Vide infra). The ∆G value of 101.9 kJ/mol at 110 °C for the
ROP of TMC is quite similar to the comparable ∆G found for
the copolymerization of oxetane and CO2 of 107.6 kJ/mol. This
small difference in ∆G for oxetane/CO2 copolymerization and
ring-opening polymerization of TMC clearly demonstrates that
1
monitoring of the coupling of oxetane and CO2 by H NMR
1
spectroscopy, accompanied by end-group analysis of the low
molecular weight copolymers produced, are designed to more
definitively address this issue.
Further Mechanistic Insight into the Oxetane and Carbon
Dioxide Coupling Process. As mentioned above, an inquiry of
the mechanistic aspects of the oxetane and carbon dioxide
‡
‡
1
coupling reaction was undertaken utilizing H NMR spectros-
(
copy studies which should provide a better assessment of the
role of trimethylene carbonate in this process. The copolymer-
ization reactions were performed under identical conditions as
those previously described herein, i.e., complex 4/2 equiv of
n-Bu4NN3, 35 bar CO2, 110 °C, and catalyst loading of 0.28
mol %. For these investigations, analysis of the reaction mixture
was done by manually sampling a small aliquot withdrawn from
the stainless steel reactor with subsequent quenching of the
‡
(
‡
‡
J. AM. CHEM. SOC. 9 VOL. 130, NO. 20, 2008 6531

A R T I C L E S
Darensbourg et al.
Scheme 2
1
Figure 16. H NMR spectrum in CDCl3 of poly(TMC) sample obtained
by ROP of TMC in the presence of (salen)Cr(III)Cl/n-Bu4NN3 as the
catalytic system. Polymer was purified from dichloromethane and 1 M HCl
solution in methanol.
1
TCE. Finally, the polycarbonate displayed a H NMR resonance
reaction’s progress by cooling the solution to 10 °C. As was
observed during this process by in situ infrared spectroscopic
monitoring, formation of TMC was detected prior to poly(TMC)
formation. That is, the initial formation of TMC was detected
within 30 min by the appearance of a triplet at 4.43 ppm and a
quintet at 1.98 ppm, with its consequent consumption to produce
polycarbonate. After 4 h of reaction no TMC was observed in
the reaction mixture. Hence, it is apparent that in the early stages
of the reaction a portion of the polycarbonate results from the
ring-opening of trimethylene carbonate. This observation,
coupled with the presence of some ether linkages in the
copolymer, strongly supports the conclusions that both oxetane/
CO2 enchainment and TMC ring-opening are occurring simul-
taneously. End-group analyses of the polycarbonates produced
from oxetane/CO2 copolymerization and ROP of TMC were
carried out in an effort to further corroborate these claims.
at 3.79 ppm (s, 3H) attributed to the -OC(O)OCH3 end group
resulting from methanolysis of the original -OC(O)N3 end group
36
following copolymer purification from MeOH.
Based on the above observations we can conclude that two
initiation pathways are operative following the initial generation
of some trimethylene carbonate Via a back-biting process
involving the carbonate species afforded from Pathway 1
(
Scheme 2).
In a similar fashion, a low molecular weight polycarbonate
obtained from the ring-opening polymerization of trimethylene
1
carbonate was analyzed by H NMR spectroscopy. Figure 16
1
illustrates the H NMR spectrum of the resulting poly(TMC)
sample in CDCl3, which exhibits, as expected, two major
3
resonances at 4.23 ppm (t, 4H, JHH ) 6.0 Hz, -OCH2) and
3
2
.05 ppm (quint, 2H, JHH ) 6.3 Hz, -CH2). Importantly, no
1
ether linkages were obserVed in this polycarbonate. A -CH OH
end group was observed before and after polymer purification,
with H NMR resonances appearing at 4.29 ppm (t, 2H, JHH
2
Figure 15 illustrates the H NMR spectrum of a purified
polycarbonate sample obtained from the copolymerization of
oxetane and carbon dioxide. Purification of the copolymer was
achieVed by precipitation from a dichloromethane solution with
1
3
3
)
6.0 Hz, -CH2), 3.73 ppm (t, 2H, JHH ) 6.0 Hz, -CH2),
3
and 1.90 ppm (quint, 2H, JHH ) 6.3 Hz, -CH2). The presence
1
M HCl in methanol, followed by Vacuum drying. In CDCl3
of an organic azide end group was also seen in the polymer
the copolymer exhibits two major signals at 4.23 ppm (t, 4H,
3
3
3
sample at 3.43 ppm (t, 2H, J ) 6.3 Hz, -CH N ) and 2100
HH
2
3
JHH ) 6.3 Hz, -OCH2) and 2.05 ppm (quint, 2H, JHH ) 5.9
-1
1
cm
in the H NMR and infrared spectra, respectively.
Hz, -CH2). Ether linkages were observed in the copolymer at
3
2
3
Similarly, the presence of an -OC(O)OCH end group at 3.79
3
.50 ppm (t, 4H, JHH ) 5.9 Hz, -OCH2) and 1.90 ppm (quint,
H, JHH ) 5.9 Hz, -CH2). A -CH2OH end group was observed
3
ppm (s, 3H) was observed following polymer purification from
methanol.
before and after purification of the copolymer sample. These latter
3
The latter observations, i.e., the presence of -CH N and
2
3
proton resonances appeared at 4.29 ppm (t, 2H, JHH ) 6.0 Hz,
3
-OC(O)OCH end groups in the polymer produced from TMC,
3
-
CH2), 3.73 ppm (t, 2H, JHH ) 6.3 Hz, -CH2), and 1.90 ppm (t,
3
17,20,36
established that the ring-opening of TMC under these catalytic
2
H, JHH ) 5.9 Hz, -CH2).
The presence of an organic azide
1
conditions occurs Via both acyl-oxygen and alkyl-oxygen bond
cleavage modes (Scheme 3). Previously, we and others have
end group (-CH2N3) was also seen in the H NMR spectrum of
the copolymer at 3.43 ppm (t, 2H, JHH ) 6.3 Hz, -CH2), with
3
the other two resonances being obscured by the intense polymer
signals at 4.23 and 2.05 ppm. This assignment was made based
(
36) Kricheldorf, H. R.; Weegen-Schulz, B. Polymer 1995, 36, 4997–5003.
1
(37) Ling, J.; Shen, Z.; Huang, Q. Macromolecules 2001, 34, 7613–7616.
on the H NMR spectrum in CDCl3 of a model compound (3-
(
38) It is however felt that only one copolymer chain grows from a single
chromium center. In other words the second azide ligand in the anionic
chromium species is probably displaced by oxetane following the ring-
opening process, with subsequent incorporation into the copolymer
as an end group. Evidence for this is seen in Figure 2, where upon
complete ring-opening of oxetane (spectrum C) there is no metal bound
or free azide present in the spectrum.
3
azido-propan-1-ol), which showed signals at 3.74 ppm (t, 2H, JHH
3
)
1
6.0 Hz, -CH2OH), 3.47 ppm (t, 2H, JHH ) 6.6 Hz, -CH2N3),
3
.83 ppm (quint, 2H, JHH ) 5.9 Hz, -CH2CH2CH2), and 1.77
ppm (s, 1H, -OH). In addition, the infrared spectrum of this
-1
copolymer exhibited an organic azide νN mode at 2100 cm in
3
6
532 J. AM. CHEM. SOC. 9 VOL. 130, NO. 20, 2008

Copolymerization Reaction of Oxetane and CO
2
A R T I C L E S
38
Scheme 3
both are most likely involved. As pointed out in the transition
state illustrated in Scheme 4 of the initiation step, the azide ion
has some association with the chromium center during the ring-
opening process, as would be expected for the growing copolymer
chain. Following the initial ring-opening step and CO2 insertion
into the resultant chromium-oxygen bond, two pathways are open
for the intermediate. Route (1) involves consecutive additions of
oxetane and CO2 to yield the alternating copolymer, whereas route
(2) leads to TMC formation by a back-biting process with ring
Scheme 4
closure. Once TMC is formed it can enter the copolymer chain by
a coordination-insertion mechanism. The portion of route (2)
shown in red should be highly dependent on the nature of the
anionic leaving group. Indeed, we haVe noted for the bromide ion
that this pathway is competitiVe with oxetane enchainment and may
proVide a means for tuning the selectiVity of the two pathways for
poly(TMC) formation. An advantage of proceeding exclusively Via
route (2) is the absence of ether linkages in the afforded
polycarbonate.
Concluding Remarks
In summary, we have shown that chromium(III) salen
complexes along with anionic initiators are effective catalysts
for the copolymerization of oxetane and carbon dioxide.
Optimization of the chromium catalyst was achieved utilizing
a salen ligand with tert-butyl substituents in the 3,5-positions
of the phenolate rings and a cyclohexylene backbone in the
diimine, along with an azide ion as an initiator. Structural studies
showed that oxetane binding to the chromium center occurred
with little changes in its metric parameters as compared with
the free cyclic ether. In particular, it remained nonplanar with
the dihedral angle of the planes C-O-C and C-C-C being
1
0.5° at 110 K. Kinetic measurements performed employing in
situ infrared monitoring showed the oxetane/CO2 coupling
reaction to be first-order in oxetane, metal catalyst, and anionic
initiator, with the latter exhibiting zero-order dependence at high
1
concentrations. Furthermore, both infrared and H NMR spec-
troscopy demonstrated the production of trimethylene carbonate
in the early stages of the copolymerization process. Nevertheless,
the presence of ether linkages in the copolymer clearly revealed
that direct enchainment of oxetane and CO2 into the growing
polymer chain occurs. The similarity of the free energy of
activation of the copolymerization reaction of CO2 and oxetane
and the ring-opening polymerization of trimethylene carbonate
supports these findings. Additional studies are being directed
at tuning the selectivity of the oxetane and CO2 coupling process
for cyclic carbonate or copolymer by altering the anionic initiator
or by using metal salen derivatives in oxidation state +2.
Acknowledgment. We gratefully acknowledge the financial
support from the National Science Foundation (CHE-0543133) and
the Robert A. Welch Foundation (A-0923).
reported that the mechanism for ring-opening polymerization
of TMC occurs exclusively by acyl-oxygen bond cleavage.
12,16,37
Supporting Information Available: Experimental details for
the synthesis and polymerization reactions, and X-ray crystal-
lographic files in CIF format for the structure determination of
However, the presence of an organic azide end group in the
poly(TMC) sample obtained by this route utilizing the (salen)-
CrCl/n-Bu4NN3 catalytic system suggests that under these higher
temperatures the ring-opening polymerization of TMC could
also be initiated by an alkyl-oxygen bond cleavage.
Scheme 4 summarizes the proposed mechanistic aspects for the
copolymerization of oxetane and carbon dioxide based on our
current experimental findings. Although we have indicated only
one of the chromium bound azide ligands taking part in the reaction,
1
complexes 5 and 6. H NMR spectra of low molecular weight
polymers produced from oxetane/CO2 and TMC for reactions
catalyzed by complex 4 in the presence of n-Bu4NCl. This
material is available free of charge via the Internet at http://
pubs.acs.org.
JA800302C
J. AM. CHEM. SOC. 9 VOL. 130, NO. 20, 2008 6533