Carbon dioxide/epoxide coupling reactions utilizing Lewis base adducts of zinc halides as catalysts. Cyclic carbonate versus polycarbonate production

Article abstract of DOI:10.1021/ic0259641

The reactions of zinc halides with 2,6-di-methoxypyridine or 3-trifluoromethylpyridine in dichloromethane have led to the formation of quite different complexes. Specifically, reactions involving pyridine containing electron donating methoxy substituents have provided salts of the type [Zn(2,6-dimethoxypyridine)₄][Zn₂X₆], as revealed by elemental analysis and X-ray crystallography. On the other hand, simple bis-pyridine adducts of zinc halides were isolated from the reactions involving the pyridine ligand with electron withdrawing substituents and characterized by X-ray crystallography, for example, Zn(3-trifluoromethylpyridine)₂Br₂. These zinc complexes were shown to be catalytically active for the coupling of carbon dioxide and epoxides to provide high molecular weight polycarbonates and cyclic carbonates, with the order of reactivity being Cl ≥ Br > 1, and 2,6-di-methoxypyridine > 3-trifluoromethylpyridine. Polycarbonate production from carbon dioxide and cyclohexene oxide was shown to be first-order in both metal precursor complex and cyclohexene oxide, as monitored by in situ infrared spectroscopy at 80 °C and 55 bar pressure. For reactions carried out in CO₂ swollen epoxide solutions in the absence of added quantities of pyridine, the copolymer produced contained significant polyether linkages. Alternatively, reactions performed in the presence of excess pyridine or in hydrocarbon solvent, although slower in rate, afforded completely alternating copolymers. For comparative purposes, zinc chloride was a very effective homopolymerization catalyst for polyethers. Additionally, zinc chloride afforded copolymers with 60% carbonate linkages in the presence of high carbon dioxide pressures. In the case of cyclohexene oxide, the copolymer back-biting reaction led exclusively to the production of the trans cyclic carbonate as shown by infrared spectroscopy in ν(C=O) region and X-ray crystallography. The unique feature of these catalyst systems is their simplicity.

Full text of DOI:10.1021/ic0259641

Inorg. Chem. 2003, 42, 581−589
Carbon Dioxide/Epoxide Coupling Reactions Utilizing Lewis Base
Adducts of Zinc Halides as Catalysts. Cyclic Carbonate versus
Polycarbonate Production
Donald J. Darensbourg,* Samuel J. Lewis, Jody L. Rodgers, and Jason C. Yarbrough
Department of Chemistry, Texas A&M UniVersity, College Station, Texas 77843
Received August 21, 2002
The reactions of zinc halides with 2,6-di-methoxypyridine or 3-trifluoromethylpyridine in dichloromethane have led
to the formation of quite different complexes. Specifically, reactions involving pyridine containing electron donating
methoxy substituents have provided salts of the type [Zn(2,6-dimethoxypyridine)4][Zn2X ], as revealed by elemental
6
analysis and X-ray crystallography. On the other hand, simple bis-pyridine adducts of zinc halides were isolated
from the reactions involving the pyridine ligand with electron withdrawing substituents and characterized by X-ray
crystallography, for example, Zn(3-trifluoromethylpyridine)2Br2. These zinc complexes were shown to be catalytically
active for the coupling of carbon dioxide and epoxides to provide high molecular weight polycarbonates and cyclic
carbonates, with the order of reactivity being Cl g Br > I, and 2,6-di-methoxypyridine > 3-trifluoromethylpyridine.
Polycarbonate production from carbon dioxide and cyclohexene oxide was shown to be first-order in both metal
precursor complex and cyclohexene oxide, as monitored by in situ infrared spectroscopy at 80 °C and 55 bar
pressure. For reactions carried out in CO swollen epoxide solutions in the absence of added quantities of pyridine,
2
the copolymer produced contained significant polyether linkages. Alternatively, reactions performed in the presence
of excess pyridine or in hydrocarbon solvent, although slower in rate, afforded completely alternating copolymers.
For comparative purposes, zinc chloride was a very effective homopolymerization catalyst for polyethers. Additionally,
zinc chloride afforded copolymers with 60% carbonate linkages in the presence of high carbon dioxide pressures.
In the case of cyclohexene oxide, the copolymer back-biting reaction led exclusively to the production of the trans
cyclic carbonate as shown by infrared spectroscopy in ν(CdO) region and X-ray crystallography. The unique
feature of these catalyst systems is their simplicity.
Introduction
oxide.2 Specifically, these more recent catalytic processes
have been generally performed in subcritical or supercritical
carbon dioxide where the catalysts are soluble in the CO2
expanded epoxide, thus requiring no added organic solvent.
In addition, other metal complexes, for example, aluminum-
The utilization of carbon dioxide in the coupling reaction
with epoxides to afford polycarbonates and/or cyclic carbon-
ates has received much attention recently. This alternative
method of synthesizing polycarbonates to the currently
employed industrial process, which involves interfacial
polycondensation of phosgene and diols, was first demon-
strated by Inoue and co-workers using heterogeneous zinc
catalysts in organic solvents.1 Since this initial discovery,
several soluble zinc catalysts have been developed which
possess high activity toward the copolymerization of carbon
dioxide and alicyclic epoxides, for example, cyclohexene
(2) (a) Darensbourg, D. J.; Holtcamp, M. W. Macromolecules 1995, 28,
7577. (b) Super, M.; Berluche, E.; Costello, C.; Beckman, E.
Macromolecules 1997, 30, 368. (c) Super, M.; Beckman, E. J.
Macromol. Symp. 1998, 127, 89. (d) Cheng, M.; Lobkovsky, E. B.;
Coates, G. W. J. Am. Chem. Soc. 1998, 120, 11018. (e) Beckman, E.
Science 1999, 283, 946. (f) Darensbourg, D. J.; Holtcamp, M. W.;
Struck, G. E.; Zimmer, M. S.; Niezgoda, S. A.; Rainey, P.; Robertson,
J. B.; Draper, J. D.; Reibenspies, J. H. J. Am. Chem. Soc. 1999, 121,
107. (g) Darensbourg, D. J.; Wildeson, J. R.; Yarbrough, J. C.;
Reibenspies, J. H. J. Am. Chem. Soc. 2000, 122, 12487. (h) Cheng,
M.; Moore, D. R.; Reczek, J. J.; Chamberlain, B. M.; Lobkovsky, B.
E.; Coates, G. W. J. Am. Chem. Soc. 2001, 123, 8738. (i) Darensbourg,
D. J.; Yarbrough, J. C. J. Am. Chem. Soc. 2002, 124, 6335. (j) Moore,
D. R.; Cheng, M.; Lobkovsky, E. B.; Coates, G. W. Angew. Chem.,
Int. Ed. 2002, 41, 2599.
* To whom correspondence should be addressed. E-mail: DJDarens@
mail.chem.tamu.edu.
(1) Inoue, S.; Koinuma, H.; Tsuruta, T. J. Polym. Sci., Part B: Polym.
Phys. 1969, 7, 287.
10.1021/ic0259641 CCC: $25.00 © 2003 American Chemical Society
Published on Web 12/20/2002
Inorganic Chemistry, Vol. 42, No. 2, 2003 581

Darensbourg et al.
(2,6-lutidine), 2,6-di-tert-butylpyridine, and 2,4,6-tri-tert-butylpy-
ridine which were purchased from Aldrich Chemical Co. All
isotopically labeled solvents for NMR studies were purchased from
Cambridge Isotope Laboratories. Infrared spectra and kinetic
measurements were monitored on ASI’s ReactIR 1000 system
equipped with a MCT detector and a 30 bounce SiCOMP in situ
(III) or chromium(III) porphyrinates, have been investigated
as catalysts for the coupling of carbon dioxide and epoxides
to afford polycarbonates.3 On the other hand, these and other
catalyst systems have been very effective at coupling carbon
dioxide and aliphatic epoxides, especially propylene oxide,
to produce mainly cyclic carbonates.4
1
probe. H and 13C NMR spectra were recorded on Unity + 300
Nevertheless, synthesis of cyclic carbonates by carbon
dioxide/epoxide coupling reactions is of interest because of
the use of these compounds in a variety of applications,
including high-boiling solvents,5 additives for hydraulic
fluids,6 and the curing of phenol-formaldehyde resins,7 plus
many more.8 In this regard, there have been several reports
of effective homogeneously catalyzed coupling of epoxides
and carbon dioxide to produce cyclic carbonate,3d,9 including
the conversion of chiral oxiranes to chiral cyclic carbonates.9f
Prominent among these is the recent work of Kim and co-
workers.10 These researchers have described a catalyst system
derived from the addition of various pyridine bases to zinc
halides salts which effectively catalyzes the coupling of
propylene oxide and carbon dioxide to produce cyclic
carbonates. Indeed, Kim’s study, along with their isolation
of a stable pyridinium zinc derivative of the ring opened
epoxide, was the inspiration for our investigation of these
zinc halides-pyridine adducts for coupling reactions of
cyclohexene oxide and carbon dioxide.
MHz or VXR 300 MHz superconducting high resolution spec-
trometers.
Reaction of Zinc Bromide with 2,6-Dimethoxypyridine, 1a.
To a 15 mL CH2Cl2 suspension of ZnBr2 (0.448 g, 1.99 mmol)
was added 2 equiv of 2,6-dimethoxypyridine (0.552 g, 3.97 mmol).
The slurry was stirred at room temperature for approximately 12 h
after which a fine white powder had precipitated in place of the
granular ZnBr2. The solvent was removed under vacuum and any
unreacted pyridine removed by heating to 40 °C to yield 0.945 g
of product (95%). Alternatively, ZnBr2 was dissolved in diethyl
ether, and the product was precipitated upon the addition of the
pyridine, followed by washing with ether. This method, however,
leads to lower yields presumably due to a sparing solubility of the
complex in ether.
Elemental analysis of the white powder was consistent with its
formulation as Zn3Br6(2,6-dimethoxypyridine)4. Anal. Calcd for
C28H36O8N4Br6Zn3: C, 27.29; H, 2.94; N, 4.55. Found: C, 26.29;
H, 2.92; N, 4.37. Colorless, block crystals of 1a suitable for X-ray
analysis were obtained by layering a solution of zinc bromide in
diethyl ether with a solution of 2,6-dimethoxypyridine in CH2Cl2
contained in a 1/8 in. sealed glass tube. Alternatively, colorless, block
crystals were obtained by slow evaporation of solvent from a
concentrated THF solution of the white powder into toluene at -20
°C. X-ray analysis revealed this material to be a THF adduct of
Zn3Br6(2,6-dimethoxy-pyridine)4, 1a‚2THF.
Experimental Section
Methods and Materials. Unless otherwise specified, all syn-
theses and manipulations were carried out on a double manifold
Schlenk vacuum line under an argon atmosphere or in an argon
filled glovebox. Glassware was flame dried before use. All solvents
were freshly distilled prior to being used. Cyclohexene oxide was
purchased from Lancaster Synthesis and was distilled from calcium
hydride. Anhydrous ZnCl2 and ZnI2 were purchased from Aldrich
Chemical Co., and anhydrous ZnBr2 was purchased from Lancaster
Synthesis. 3-Trifluoromethylpyridine, 2,6-dimethoxypyridine, 2-phe-
nylpyridine, and 2,6-diphenylpyridine were purchased from Lan-
caster Synthesis and used as received, as were 2,6-dimethylpyridine
The chloride (1b) and iodide (1c) analogues of complex 1a were
prepared and isolated as white powders in an identical manner to
that of 1a and were shown by elemental analysis to be of similar
composition. For example, Anal. Calcd for Zn3Cl6(2,6-dimethoxy-
pyridine)4 (1b): C, 34.83; H, 3.76; N, 5.80. Found: C, 32.83; H,
3.66; N, 5.47. Furthermore, reactions carried out in a similar manner
t
employing 2,6-substituted pyridines with phenyl, methyl, and Bu
groups, as well as 2-phenylpyridine, provided correspondent zinc
derivatives.
Employing the published synthetic method for Zn(py)2X2 and
Zn(bipy)X2, which involves the slow addition of the pyridine ligand
to a hot absolute alcohol solution of the zinc halide, provided
complexes of the same formulation as 1a.11 For example, from the
reaction of ZnBr2 and 2,6-dimethoxypyridine, a white powder was
obtained which analyzed to be Zn3Br6(dimethoxypyridine)4. Anal.
Calcd for C28H36O8N4Br6Zn3: C, 27.29; H, 2.94; N, 4.55. Found:
C, 27.07; H, 3.05; N, 4.40.
(3) (a) Inoue, S. J. Polym. Sci., Part A: Polym. Chem. 2000, 38, 2861.
(b) Mang, S.; Cooper, A. I.; Colclough, M. E.; Chauhan, N.; Holmes,
A. B. Macromolecules 2000, 33, 303. (c) Stamp, L. M.; Mang, S. A.;
Holmes, A. B.; Knights, K. A.; de Miguel, Y. R.; McConvey, I. F.
Chem. Commun. 2001, 2502. (d) Kruper, W. J.; Dellar, D. V. J. Org.
Chem. 1995, 60, 725.
(4) Darensbourg, D. J.; Holtcamp, M. W. Coord. Chem. ReV. 1996, 153,
155.
(5) Behr, A. Carbon Dioxide ActiVation by Metal Complexes; VCH:
Weinheim, 1988; p 7.
(6) Nankee, R. J.; Avery, J. R. Schrems, J. E. Dow Chem. Patent FR.
15,72,282, 1967; Chem Abstr. 1970, 72, 69016p.
Synthesis of (3-Trifluoromethylpyridine)2ZnX2, X ) Cl (2a),
Br (2b), and I (2c). The preparation of these derivatives was
initiated in a manner similar to that utilized for the synthesis of the
other pyridine derivatives. For example, to a 15 mL CH2Cl2
suspension of ZnBr2 (0.217 g, 0.964 mmol) was added 2 equiv of
3-trifluoromethylpyridine (0.283 g, 1.92 mmol). After stirring the
reaction mixture for 6 h, the solution became clear and slightly
yellow. Upon removal of the solvent under vacuum, 0.386 g (77.2%
yield) of (3-trifluoro-methylpyridine)2ZnBr2 was isolated. Anal.
Calcd for C12H8N2F6ZnBr2: C, 27.75; H, 1.55; N, 5.39. Found:
C, 27.60; H, 1.62; N, 5.35. Yellow, X-ray quality crystals of the
(7) Pizzi, H.; Stephanou, A. J. Appl. Polym. Sci. 1993, 49, 2157.
(8) Shaikh, A.-A. G.; Sivaram, S. Chem. ReV. 1996, 96, 951-976.
(9) (a) Matsuda, H.; Ninagawa, A.; Nomura, R. Chem. Lett. 1979, 1261.
(b) Matsuda, H.; Ninagawa, A.; Nomura, R.; Tsuchida, T. Chem. Lett.
1979, 573. (c) Ratzenhofer, M.; Kisch, H. Angew. Chem., Int. Ed.
Engl. 1980, 19, 317. (d) Nomura, R.; Ninagawa, A.; Matsuda, H. J.
Org. Chem. 1980, 45, 3735. (e) Nomura, R.; Kimura, M.; Teshima,
S.; Ninagawa, A.; Matsuda, H. Bull. Chem. Soc. Jpn. 1982, 55, 3200.
(f) Kisch, H.; Millini, R.; Wang, I. J. Chem. Ber. 1986, 119, 1090.
(g) Du¨mler, W.; Kisch, H. Chem. Ber. 1990, 123, 277. (h) Paddock,
R. L.; Nguyen, S. T. J. Am. Chem. Soc. 2001, 123, 11498.
(10) (a) Kim, H. S.; Kim, J. J.; Lee, B. G.; Jung, O. S.; Jang, H. G.; Kang,
S. O. Angew. Chem., Int. Ed. 2000, 39, 4096. (b) Kim, H. S.; Kim, J.
J.; Kwon, H. N.; Chung, M. J.; Lee, B. G.; Jang, H. G. J. Catal. 2002,
205, 226.
(11) Postmus, C.; Ferraro, J. R.; Wozniak, W. Inorg. Chem. 1967, 6, 2030.
582 Inorganic Chemistry, Vol. 42, No. 2, 2003

Carbon Dioxide/Epoxide Coupling Reactions
Table 1. Crystallographic Data for Complexes 1a, 1a‚2THF, 2b, and 8
1a
1a‚2THF
2b
8
empirical formula
fw (g/mol)
T (K)
C29H38Br6Cl2N4O8Zn3
1317.10
110(2)
C36H52Br6N4O10Zn3
1376.39
293(2)
C12H10Br2F6N2Zn
521.41
293(2)
C12H10O3
202.20
293(2)
wavelength (Å)
cryst syst
space group
a (Å)
b (Å)
c (Å)
0.71073
triclinic
P1
0.71073
monoclinic
Cc
26.607(2)
14.4148(12)
15.9052(13)
0.71073
orthorhombic
Fdd2
25.348(4)
26.884(4)
4.7868(8)
0.71073
orthorhombic
P212121
6.4041(10)
9.2575(15)
11.7214(19)
14.201(3)
15.419(3)
20.423(4)
86.34(3)
87.36(3)
89.29(3)
4457.9(15)
4
1.962
7.145
23,761
15,492
R (deg)
â (deg)
125.4240(10)
γ (deg)
cell V (Å3)
Z
4971.0(7)
4
1.839
6.311
15,402
8314
3262.1(9)
8
694.91(19)
4
D
calcd (mg/m3)
2.123
6.463
3399
1131
8.31
1.933
0.139
4455
1670
3.52
abs coeff (mm-1
obsd reflns
)
unique reflns (I > 2σ)
R,a % [I > 2σ]
9.10
19.82
6.94
17.97
Rw,a % [I > 2σ]
20.65
10.69
a R ) ∑||Fo| - |Fc||/∑|Fc||/∑|Fo|; Rw ) {[∑w(Fo - Fc )2]/[∑w(Fo )2]}1/2
.
2
2
2
product were obtained by slow evaporation of a concentrated CH2-
Cl2 solution of the complex into toluene at -20 °C.
cooling in a dry ice/acetone bath, a 15-mL solution of triphosgene
(1.08 g, 3.64 mmol) was added dropwise to the reaction mixture
resulting in an orange solution followed by slow warming to room
temperature. Workup of the product was analogous to the prepara-
tion of 1 but was recrystallized in ether/pentane to yield 0.45 g
Copolymerization of Cyclohexene Oxide/Propylene Oxide and
Carbon Dioxide. A 0.100 g sample of the catalyst complex was
dissolved in 20 mL of the appropriate epoxide. The solution was
rapidly loaded via an injection port into a 300 mL stainless steel
Parr autoclave that had been previously dried overnight at 80 °C
under vacuum. The reactor was pressurized to 700 psi with CO2,
heated to 80 °C, and stirred for 24 h. After that time, the autoclave
was cooled and the CO2 vented into a fume hood. The reactor was
opened, and the polymer was isolated from the viscous/solid mixture
by dissolution in small amounts of methylene chloride followed
by precipitation from methanol.
Isolation of trans-Cyclic Cyclohexylcarbonate, 8. This com-
pound was isolated from the reaction mixture of a polymerization
process utilizing complex 1a as a catalyst precursor. Following the
removal of high molecular weight polymer by precipitation in
methanol, the remaining solution of MeOH and CH2Cl2 was allowed
to slowly evaporate in the air to provide crystals suitable for X-ray
analysis of the cyclic carbonate.
Synthesis of trans-Cyclic Cyclohexylcarbonate, 8.12 An excess
of pyridine (19.5 g, 246 mmol) was added to a 10-mL CH2Cl2
solution of trans-1,2-cyclohexanediol (2.49 g, 21.4 mmol) followed
by cooling in a dry ice/acetone bath. A 15-mL CH2Cl2 solution of
triphosgene (4.17 g, 14.1 mmol) was subsequently added dropwise
to give an orange solution, which was allowed to warm slowly to
room temperature. The reaction mixture was quenched using
saturated aqueous NH4Cl followed by extraction of the aqueous
phase with CH2Cl2. The organic extracts were washed with 1 M
HCl, saturated NaHCO3, saturated NaCl, H2O, dried with MgSO4,
and filtered. Solvent and excess pyridine were removed in vacuo
followed by product isolation using a silica gel column and
recrystallized in benzene/hexanes to yield 1.52 g (76%). 1H NMR
(CDCl3): δ 1.35-2.25 (m, 8H, CH2), 3.96 (m, 2H, CH-carbonate).
13C NMR (CDCl3): δ 18.58, 26.12, 75.26, 154.70. IR ν(CdO),
CH2Cl2: 1881 cm-1 (w), 1817 cm-1 (m), 1803 cm-1 (s).
Synthesis of cis-Cyclic Cyclohexylcarbonate, 9.12 This com-
pound was synthesized by the same procedure as described for 1
by adding excess pyridine (3.75 g, 47.4 mmol) to a 10-mL CH2Cl2
solution of cis-1,2-cyclohexanediol (0.60 g, 5.2 mmol). Upon
1
(87%). H NMR (CDCl3): δ 1.35-1.70 (m, 4H, CH2), 1.90 (q,
4H, CH2), 4.71 (m, 2H, CH-carbonate). 13C NMR (CDCl3): δ
23.01, 28.02, 83.32, 154.80. IR ν(CdO), CH2Cl2: 1869 cm-1 (w),
1823 cm-1 (sh), 1802 cm-1 (s), 1793 cm-1 (sh).
Polymer Characterization. Polymer samples were first char-
1
acterized by H NMR and IR spectroscopy. The amount of ether
1
linkages was determined via H NMR by integrating the peaks
corresponding to the methine protons of the polyether at ∼3.45
ppm and the polycarbonate at ∼4.6 ppm. The presence or absence
of the corresponding cyclic carbonate was investigated by monitor-
ing the presence or absence of the ν(CdO) of the cyclic carbonate
at ∼1825 cm-1. Finally, Mw and Mn measurements were carried
out at Exxon-Mobil using GPC.
High Pressure in Situ Kinetic Measurements. High pressure
reaction kinetic measurements were carried out using a stainless
steel Parr autoclave modified with a silicon probe to connect to
the ASI ReactIR 1000 system. The rate of polymer and cyclic
carbonate formation was monitored by repeating a typical polym-
erization run, except reducing the CO2 pressure to 400 psi. The
formation rates of the polycarbonate and the cyclic carbonate were
monitored by following the ν(CdO) of the polycarbonate at ∼1750
cm-1 and of the cyclic carbonate at ∼1825 cm-1
.
X-ray Crystallography. A Bausch and Lomb 10× microscope
was used to identify suitable colorless crystals of 1a, 1a‚2THF,
2b, and 8 from a representative sample of crystals of the same habit.
The representative crystal was coated in a cryogenic protectant (i.e.,
mineral oil, paratone, or apezeon grease) and was then fixed to a
glass fiber, which in turn was fashioned to a copper mounting pin.
The mounted crystals were then placed in a cold nitrogen stream
(Oxford) maintained at 110 K on a Bruker SMART 1000 three
circle goniometer.
Crystal data and details of data collection for the complexes are
provided in Table 1. The X-ray data were collected on a Bruker
CCD diffractometer and covered more than a hemisphere of
reciprocal space by a combination of three sets of exposures; each
exposure set had a different æ angle for the crystal orientation,
(12) Burk, R. M.; Roof, M. B. Tetrahedron Lett. 1993, 34, 395.
Inorganic Chemistry, Vol. 42, No. 2, 2003 583

Darensbourg et al.
Figure 2. Thermal ellipsoid representation of complex 1a‚2THF. Average
bond distances and bond angles in the cation are Zn-N (2.052[10]Å) and
N-Zn-N (109.6[4]°), whereas those in the anion are Zn-Br (2.371[18]
Å), Zn-O (2.110[8] Å), Br-Zn-Br (116.04[7]°), and O-Zn-Br (101.3-
[2]°).
Figure 1. Thermal ellipsoid representation of complex 1a. A. Zn(2,6-
dimethoxypyridine)42+. B. Zn2Br62-. Average bond distances and bond
angles in one of the cations are Zn-N (2.063[16] Å) and N-Zn-N (109.7-
[7]°), whereas those in one of the anions are av terminal Zn-Br (2.350[3]
Å), av bridging Zn-Br (2.501[3] Å), Br(1)-Zn(3)-Br(1A) (98.18(11)°),
Br(2)-Zn(3)-Br(3) (120.97(13)°), Br(3)-Zn(3)-Br(1A) (107.76(11)°), Br-
(3)-Zn(3)-Br(1) (109.07(11)°), Br(2)-Zn(3)-Br(1) (107.67(12)°), and Br-
(2)-Zn(3)-Br(1A) (110.78(11)°).
and each exposure covered 0.3° in ω. The crystal-to-detector
distance was 4.9 cm. Crystal decay was monitored by repeating
the data collection for 50 initial frames at the end of the data set
and analyzing the duplicate reflections; crystal decay was negligible.
The space group was determined on the basis of systematic absences
and intensity statistics.
Figure 3. Thermal ellipsoid representation of complex 2b. Zn-Br and
Zn-N bond distances are 2.3746(17) Å and 2.084(12) Å, respectively,
whereas the bond angles are N-Zn-N (117.3(6)°), Br-Zn-Br (127.44-
(13)°), and av N-Zn-Br (103.3[3]°).
The structures were solved by direct methods. Full-matrix least-
squares anisotropic refinement for all non-hydrogen atoms yielded
R(F) and Rw(F2) values as indicated in Table 1 at convergence.
Hydrogen atoms were placed in idealized positions with isotropic
thermal parameters fixed 1.2 or 1.5 times the value of the attached
atom. Neutral atom scattering factors and anomalous scattering
factors were taken from the International Tables for X-ray Crystal-
lography Vol. C.
and 1a‚2THF with partial atom labeling schemes are
depicted in Figures 1 and 2, where selected bond distances
are also provided. On the other hand, the corresponding
reactions of the zinc halides with 3-trifluoromethylpyridine
yielded the simple ZnX2(3-trifluoromethylpyridine)2 com-
plexes. It was possible to isolate X-ray suitable crystals of
the derivative where X ) Br (2b) by slow evaporation of
methylene chloride from a concentrated CH2Cl2 solution of
the complex into toluene at -20 °C. The thermal ellipsoid
drawing of 2b may be found in Figure 3, along with selected
bond distances and angles. As is evident from Figure 3, the
zinc center in 2b is of distorted tetrahedral geometry as
expected for this type of complex.
Results and Discussion
The reactions of zinc halides (X ) Cl, Br, or I) and various
2,6-disubstituted pyridines have been carried out in the
absence of highly coordinating solvents to afford salts of
the form [Zn(pyridine)4][Zn2X6]. For example, X-ray quality
crystals of [Zn(2,6-dimethoxypyridine)4][Zn2Br6] (1a) were
obtained from diethyl ether, whereas, in THF, crystals of
[Zn(2,6-dimethoxypyridine)4][ZnBr3(THF)]2 (1a‚2THF) were
isolated. Thermal ellipsoid representations of complexes 1a
Because of the effectiveness of pyridine derivatives of zinc
halides for catalytically coupling carbon dioxide and pro-
pylene oxide to provide propylene carbonate, it was of
interest to examine these complexes for the copolymerization
584 Inorganic Chemistry, Vol. 42, No. 2, 2003

Carbon Dioxide/Epoxide Coupling Reactions
Scheme 1
Table 2. Catalytic Activity for the Copolymerization of Carbon
Dioxide and Cyclohexene Oxidea
catalyst
% carbonate TOF (g poly/g Zn/h)b
Zn3Cl6(2,6-dimethoxypyridine)4
Zn3Br6(2,6-dimethoxypyridine)4
Zn3I6(2,6-dimethoxypyridine)4
Zn3Br6(2-phenylpyridine)4
ZnBr2(3-trifluoromethylpyridine)2
ZnCl2
80
85
80
80
91
60
27.1 (13.5)
27.5 (13.5)
15.7 (7.9)
14.0 (7.1)
1.87 (0.9)
22 (11.6)
a Catalyst loading (0.100 g), 20.0 mL of cyclohexene oxide. Reaction
conditions: 80 °C at a total pressure of 50 bar for a 24 h period. b TOF
values in parentheses are in units of mol of epoxide/mol of Zn/h.
of CO2 and cyclohexene oxide. Indeed, regardless of the
solid-state form of these zinc derivatives, that is, [Zn-
(pyridine)4][Zn2X6] or Zn(pyridine)2X2, all complexes ex-
amined exhibited significant reactivity for affording poly-
(cyclohexenylene carbonate) from CO2/cyclohexene oxide
(Scheme 1). As noted in Scheme 1, these catalysts afford
noticeable quantities of polyether linkages in the copolymers,
as well as production of the minor product, trans-cyclic
cyclohexylcarbonate. The copolymerization reactions were
monitored in the ν(CdO) region of the infrared spectrum of
the polycarbonate utilizing an in situ ASI 1000 ReactIR probe
fitted to a modified 300 mL stainless steel Parr reactor. Figure
4 depicts a typical reaction profile as monitored by infrared
spectroscopy for a reaction carried out at 80 °C in cyclo-
hexene oxide at a CO2 pressure of 30 bar. The major infrared
absorption in the ν(CdO) region at ∼1750 cm-1 is for the
polycarbonate, whereas the minor peak at ∼1825 cm-1 is
due to the cyclic carbonate.
Table 2 summarizes the activity of several of these zinc
halide catalysts, where complex 1a displays the highest
activity with a TOF of 27.5 g of polymer/g of zinc/h.
Consistent with previously reported results for propylene
carbonate production, the general trend observed is that the
catalytic activity varies with halide, Cl ≈ Br > I, and the
activity increases with increasing electron donating ability
of the substituents on the pyridine ligands.10 The highly active
representative catalyst, complex 1a, was employed to
investigate the copolymerization process in more detail. As
noted in Table 2, in neat cyclohexene oxide/CO2, the
copolymer produced contained ∼85% carbonate and ∼15%
ether linkages (see Figure 5). However, if the copolymeri-
zation process is carried out in a 2.5 M solution of
cyclohexene oxide in toluene, the resulting copolymer
contains essentially 100% carbonate linkages as indicated
Figure 4. In situ infrared monitoring of the poly(cyclohexylene carbonate)
and cyclic cyclohexylcarbonate formation as a function of time. Reaction
run using complex 1a as catalyst in CO2 swollen cyclohexene oxide solution
at 30 bar pressure and 80 °C.
out in a 10-fold excess of added pyridine ligand, the TOF is
drastically reduced to ∼5 g of polymer/g of zinc/h, but the
carbonate content of the copolymer is ∼100%. In these
instances where copolymers were produced with 100%
carbonate linkages, there was a concomitant increase in the
molecular weights of the polymers. That is, the average Mn
and Mw values were found to be 44 000 and 313 000,
respectively, as compared to the corresponding molecular
weights of 13 000 and 99 000 for the copolymers with 85%
carbonate linkages.
1
by the absence of the ether resonance in the H NMR
spectrum at 3.45 ppm (Figure 5). As expected under these
conditions, the TOF is reduced to ∼20 g of polymer/g of
zinc/h. Similarly, if the copolymerization reaction is carried
As illustrated in Scheme 1, cyclic cyclohexylcarbonate (8)
is observed as a minor product from the carbon dioxide/
Inorganic Chemistry, Vol. 42, No. 2, 2003 585

Darensbourg et al.
Figure 5. 1H NMR of poly(cyclohexylene carbonate). A. Reaction carried
out in cyclohexene oxide. B. Reaction carried out in cyclohexene oxide
with 10-fold excess of 2,6-dimethoxypyridine added or in a 2.5 M solution
of cyclohexene oxide in toluene.
Figure 7. Infrared spectra in ν(CdO) region of cyclic cyclohexylcarbonate.
A. Cis isomer in CH2Cl2. B. Trans isomer in CH2Cl2. C. Trans isomer in
toluene.
copolymer versus cyclic carbonate observed when utilizing
cyclohexene oxide as monomer. On the other hand, the
corresponding five-membered ring of the cyclic carbonate
derived from propylene oxide/CO2, propylene carbonate, is
perfectly planar.2f Consistent with this observation, the CO2/
propylene oxide coupling process carried out under similar
reaction conditions utilizing 1a as catalyst afforded mostly
cyclic carbonate, with a minor product of copolymer. Indeed,
this is consistent with what has been previously observed
for other soluble catalyst systems.2
To definitively identify the stereochemistry of the cyclo-
hexylcarbonate in solution by infrared and 13C NMR
spectroscopy, we have independently synthesized both cis
and trans isomers of cyclic cyclohexylcarbonate from the
corresponding 1,2-cyclohexanediol and triphosgene by the
established methods.12 These data have allowed us to clearly
establish that the trans isomer, compound 8, is the sole
carbonate produced from the reaction defined in Scheme 1.
Figure 7 compares the infrared spectra in the ν(CdO) region
for the cis and trans cyclohexylcarbonates in methylene
chloride solvents. It should be noted, however, that the
infrared spectra of cyclic cyclohexylcarbonates are highly
solvent dependent.13 For example, see the spectrum of trans-
cyclohexylcarbonate obtained in toluene in Figure 7C, where
the principal ν(CdO) frequency is ∼1825 cm-1, much like
that observed in the CO2 swollen cyclohexene oxide solution
(Figure 4).
Figure 6. Thermal ellipsoid representation of trans-cyclic cyclohexyl-
carbonate. Selected bond distances and angles are the following: C(1)-
O(2), 1.193(2) Å; C(1)-O(1), 1.346(2) Å; C(1)-O(3), 1.347(2) Å; C(2)-
C(7), 1.501(6) Å, all other C-C bonds av 1.516[5] Å, and O(1)-C(1)-
O(2), 124.14(18)°; O(2)-C(1)-O(3), 124.4(19)°; O(1)-C(1)-O(3),
111.44(16)°; O(3)-C(7)-C(2), 103.1(4)°; O(1)-C(2)-C(7), 101.2(4)°.
cyclohexene oxide coupling reaction. We were able to obtain
crystals of 8 suitable for X-ray diffraction analysis by the
slow evaporation of the methanol/CH2Cl2 solution which
remained upon removal of the high molecular weight
copolymer. The thermal ellipsoid drawing of compound 8
is illustrated in Figure 6, along with the atomic numbering
scheme and selected bond distances and angles. As expected,
the cyclohexane ring carbon atoms are disordered (see
Supporting Information). As noted in Figure 6, 8 is the trans
isomer where there is significant strain in the five-membered
ring of the carbonate. That is, the average deviation from
the least squares plane defined by atoms O(1), O(3), C(1),
C(2), C(7) is 0.116 Å, with the C(2) and C(7) atoms of the
cyclohexane ring exhibiting the greater variations at -0.1642
and 0.1753 Å, respectively. This ring strain is most likely
the major factor which leads to the favorable production of
(13) The infrared spectrum in the ν(CdO) region in methylene chloride
of the cyclohexylcarbonate produced during the copolymer synthesis
is identical to that for the pure trans isomer depicted in Figure 7B. In
cyclohexene oxide and in toluene solutions, the spectra of trans-
cyclohexylcarbonate are quite different from that observed in MeCl2
as is seen in Figures 7C and 8.
586 Inorganic Chemistry, Vol. 42, No. 2, 2003

Carbon Dioxide/Epoxide Coupling Reactions
Figure 9. Plots of the ln[initial rate] of copolymer (poly(cyclohexylcar-
bonate) formation versus (A) ln [1a] and (B) ln[cyclohexene oxide]. Slope
in part A ) 1.03, and that in part B ) 0.96.
Figure 10. Plot of the ln[initial rate] of cyclic cyclohexylcarbonate
formation versus ln [1a]. Slope ) 1.50.
reactions is provided in Figure 8, where the ν(CdO)
stretching vibrations are seen at ∼1750 and 1825 cm-1 for
the polycarbonate and cyclic carbonate, respectively. Figure
8 also illustrates the peak profiles as a function of time for
production of both the major (poly(cyclohexenylene carbon-
ate)) and minor (cyclic cyclohexylcarbonate) products. As
is readily obvious from a comparison with the reactions
monitored in the absence of organic solvent (Figure 4), in
this instance there is a somewhat enhanced production of
cyclic carbonate versus polycarbonate. The initial rates for
the formation of poly(cyclohexylene carbonate) were deter-
mined as a function of [catalyst] and [cyclohexene oxide].
Plots of ln[initial rate] versus ln[catalyst] or ln[cyclohexene
oxide] yielded slopes of unity, indicative of copolymer
formation being first-order in both [catalyst] and [epoxide]
(Figure 9). On the other hand, the initial rate of cyclic
cyclohexylcarbonate formation is mixed-order in catalyst
Figure 8. In situ infrared monitoring of the poly(cyclohexylene carbonate)
and cyclic cyclohexylcarbonate formation as a function of time. Reaction
performed using complex 1a as catalyst in 2.5 M cyclohexene oxide in
toluene at 55 bar pressure and 80 °C.
In an effort to better assess the mechanistic aspects of the
copolymerization reaction involving cyclohexene oxide and
carbon dioxide in the presence of one of these soluble zinc
halide derivatives, specifically complex 1a, we have per-
formed kinetic measurements. These investigations were
carried out at a total pressure of 30 bar and a temperature of
80 °C with the catalyst (or catalyst precursor) dissolved in
a solution of cyclohexene oxide in toluene of known
molarity. Reducing the CO2 loading pressure to 10 bar did
not significantly affect the rate of the reactions. An organic
solvent was employed in these studies to ensure a single
phase reaction system, where the concentrations of reagents
could be more accurately controlled.14 Additionally, under
these conditions, copolymers containing 100% carbonate
linkages are obtained. A representative series of infrared
spectra for the cyclohexene oxide/carbon dioxide coupling
(14) From studies of the phase behavior of the cyclohexene oxide-CO2
binary,2c at 55 bar CO2 pressure, there exist cyclohexene oxide-rich
(XCHO at 90 °C ≈ 0.70) and cyclohexene oxide-poor phases, the former
being where the catalyst resides and the phase monitored by our in
situ infrared probe.
Inorganic Chemistry, Vol. 42, No. 2, 2003 587

Darensbourg et al.
Scheme 2
2-
(Figure 10), that is, 1.50, suggesting involvement of both
dimeric and monomeric zinc species as catalysts. This result
is consistent with the enhanced rate of cyclic carbonate
production in hydrocarbon solvent as compared to that in
neat epoxide; that is, dimer formation would be favored in
toluene (see Figures 4 and 8).
where Zn2X6 reacts with cyclic ether (THF) to afford
ZnX3(THF)1-, or Zn(pyridine)42+ undergoes ligand substitu-
tion with an oxygen donor ligand (H2O) yielding Zn-
(pyridine)3(H2O)2+, have been fully characterized. In addition
to the structures previously discussed, a related complex, [Zn-
(2,6-dimethoxypyridine)3H2O][Zn2Br6], 9, was obtained from
the reaction of ZnBr2 and 2,6-dimethoxy-pyridine where
adventitious water was present. The crystal structure of this
species was found to be similar to complex 1a with the
exception that one of the pyridine ligands on the zinc cation
was replaced by a water molecule. A ball-and-stick drawing
of the cation of this salt is provided in the Supporting
Information. Nevertheless, some of these potential intermedi-
ates are more credible than others on the basis of known
reactivity patterns. Schemes 2 and 3 contain possible
candidates resulting from the intermediacy of Zn(pyridine)3-
(epoxide)2+, ZnX3(epoxide)1-, or ZnX2(pyridine)(epoxide).
For precursor complexes of the type [Zn(2,6-R2C5H3N)4]-
[Zn2X6], potential intermediates in the copolymerization
reactions are complexes A, B, and C as indicated in Scheme
2. These zinc species are derived from Zn(2,6-R2C5H3N)3-
(epoxide)2+ and ZnX3(epoxide)- which result from ligand
substitution reactions at the zinc centers of the parent salts.
Recall that analogues of these zinc deriVatiVes haVe been
identified for the oxygen donors H2O and THF, respectiVely.
Whereas the process affording A would not be expected to
vary significantly with the nature of the halide counterion
as is observed in Table 2, the process leading to B and C
would be consistent with pyridine enhancing reaction rates
and the nature of the copolymer thus produced as well as
the fact that ZnCl2 alone is a good homopolymerization
catalyst. Both processes would be first-order in complex 1a.
Furthermore, the cationic nature of A is not conducive to
promoting the CO2 enchainment step. Furthermore, [Zn-
(pyridine)4][BF4]2 was shown to be inactiVe as a catalyst for
the copolymerization process. Alternatively, starting with
Finally, because zinc halides are molecular solids which
are soluble in cyclic ethers, it was of interest to examine
their ability in the absence of pyridine ligands to catalyze
the homopolymerization of cyclohexene oxide to polyether
or the copolymerization of cyclohexene oxide/CO2 to poly-
carbonates. Relevant to these processes, it is well-established
that ZnX2/(n-Bu4N)X salts catalyze the coupling of mono-
substituted oxiranes to cyclic carbonates.9g In our study, the
zinc halides were observed to readily homopolymerize
cyclohexene oxide to polyether at ambient temperature. Upon
reacting cyclohexene oxide and carbon dioxide in the
presence of the zinc halides at 90 °C and 55 bar pressure,
high molecular weight polycarbonates with substantial poly-
ether linkages were produced with TOFs as high as 20 g of
polymer/g of Zn/h. An analogous reaction involving ZnCl2
in the presence of the sterically demanding 2,6-di-tert-
butylpyridine base gave identical results as the ZnCl2 only
catalyzed process. Because of the high propensity of these
salts for forming homopolymers, a sample of ZnCl2 was
purified and sealed in a glass ampule which was broken after
the addition of CO2 to the reaction system in order to
minimize the production of polyethers. Even under these
conditions, the resulting polymer contained ∼30% ether
linkages and ∼70% carbonate linkages. These observations
are consistent with the effectiveness of the pyridine ligands
to increase the CO2 incorporation and TOFs of this catalyst
system.
An array of possible intermediates in this carbon dioxide/
epoxide coupling process can be proposed. In this regard, it
should be noted that derivatives of [Zn(pyridine)4][Zn2X6],
588 Inorganic Chemistry, Vol. 42, No. 2, 2003

Carbon Dioxide/Epoxide Coupling Reactions
Scheme 3
ZnX2N2 precursors leads to species D which should readily
lead to enchainment of both CO2 and epoxides. Indeed, the
dimeric form of D, where the epoxide is propylene oxide,
has been isolated and X-ray crystallographically characterized
for the CO2/propylene oxide coupling processes leading to
cyclic carbonate.10 The effect of added pyridine on the CO2
content of the polycarbonate is attributed to maintaining the
active zinc center coordinatively saturated, thereby avoiding
a consecutive epoxide ring-opening process. The rate de-
crease experienced in excess pyridine is readily accom-
modated in Schemes 2 and 3.
Finally, it is noteworthy that the zinc salts, [Zn(pyridine)4]-
[Zn2X6], do not undergo major ligand redistribution reactions
at ambient temperature in cyclic ethers. For example, [Zn-
(2,6-di-MeOC5H3N)4][ZnBr3(THF)]2 was isolated from a
THF solution of complex 1a. Nevertheless, under the more
rigorous reaction conditions of catalysis, it is possible that
the zinc species in solution are quite different from the
precursor complex.
(pyridine)4][Zn2X6] depending on the nature of the pyridine
ligand. Because of a lack of good spectral probes (1H and
13C NMR) of these various zinc species in solution, all
derivatives were characterized in the solid state by X-ray
crystallography. These zinc complexes were shown to be
catalytically active for the coupling of carbon dioxide and
epoxides to provide high molecular weight polycarbonates
and cyclic carbonates, with activity being greater for deriva-
tives containing electron donating pyridine ligands. Although
this investigation has not given rise to significantly more
active or selective catalysts for the copolymerization of
epoxides and carbon dioxide than those previously published,
the unique feature of these zinc catalyst precursors is their
synthetic simplicity.
Acknowledgment. Financial support from the National
Science Foundation (Grants CHE-99-10342 and CHE 98-
07975 for the purchase of X-ray equipment), and the Robert
A. Welch Foundation, is acknowledged..
Supporting Information Available: Complete details (CIF
format) of the X-ray diffraction studies on compounds 1a, 1a‚
2THF, 2b, 8, and 9. Ball-and-stick representation of the cation of
compound 9. This material is available free of charge via the
Internet at http://pubs.acs.org.
Summary
We have demonstrated herein that zinc halides in the
presence of substituted pyridines lead to a variety of
complexes involving zinc centers which contain quite dif-
ferent ligand sets, for example, ZnX2(pyridine)2 or [Zn-
IC0259641
Inorganic Chemistry, Vol. 42, No. 2, 2003 589