Bis 2,6-difluorophenoxide dimeric complexes of zinc and cadmium and their phosphine adducts: Lessons learned relative to carbon dioxide/cyclohexene oxide alternating copolymerization processes catalyzed by zinc phenoxides

Article abstract of DOI:10.1021/ja002855h

Dimeric phenoxide derivatives of zinc and cadmium have been synthesized from the reaction of the corresponding metal bistrimethylsilylamide and two equivalents of 2,6-F₂C₆H₃OH in tetrahydrofuran. The zinc analogue, [Zn(O-2,6-F₂C₆H₃) ₂·THF]₂ (1), has been characterized in the solid state via X-ray crystallography, where the zinc centers are shown to possess distorted tetrahedral geometry containing two bridging phenoxides and a terminal phenoxide and THF ligand. The distance between the metal centers (Zn···Zn) was found to be 3.05 A, and the THF ligands lie on opposite sides of the plane formed by the two zinc atoms and two bridging phenoxide ligands' oxygen atoms. There are several Zn···F nonbonding distances involving the bridging phenoxide ligands that are less than the van der Waals internuclear distance. In addition, both the zinc and cadmium dimeric derivatives have been prepared such that the labile THF ligands are replaced by the sterically encumbering basic phosphine, PCy₃. The solid-state structures of [Zn(O-2,6-F₂C₆H₃)₂ ·PCy₃]₂ (2) and [Cd(O-2,6-F₂C₆H₃)₂ ·PCy₃]₂ (5) are similar to that of complex 1, where the tricyclohexylphosphine ligands, like the THF ligands, are accommodated in a trans configuration. The ³¹P NMR spectrum of complex 2 in C₆D₆ upon addition of free PCy₃ exhibits sharp resonances assigned to both the complex (9.58 ppm) and free PCy₃ (10.6 ppm), which is indicative of slow exchange of the phosphine ligands. On the other hand, the phosphine ligands on the cadmium derivative (5) are involved in an exchange process with free PCy₃ via a rapid equilibrium between 5 and two equivalents of Cd(O-2,6-F₂C₆H₃)₂ (PCy₃)₂. The equilibrium reaction strongly favors the monomer cadmium bisphosphine complex at low temperature (-80°C). As expected, the ¹¹³Cd and ³¹P NMR spectra of complex 5 in solution in the absence of excess PCy₃ is quite similar to that determined in the solid state by CP/MAS. Complex 1 and its chloro- and bromophenoxide analogues were shown to be effective catalysts for the copolymerization of cyclohexene oxide and CO₂, the terpolymerization of cyclohexene oxide/ propylene oxide/CO₂, and the homopolymerization of cyclohexene oxide. In the case of the copolymerization process (80 °C and 55 bar), the polycarbonate copolymer that was produced is completely alternating, with no polyether linkages. At the same time, the homopolymerization of cyclohexene oxide to afford polyether in the presence of 1 as catalyst is much more facile than the copolymerization process. Importantly, for both copolymerization and homopolymerization processes catalyzed by complex 1, the initiator of the polymer chain growth is a difluorophenoxide unit, as revealed by ¹⁹F NMR, with both CO₂ insertion and epoxide ring-opening being involved in the initiation step. At 80°C and 55 bar, the coupling of propylene oxide and CO₂ led almost exclusively to propylene carbonate. On the other hand, at lower temperatures (i.e., 40°C), copolymer formation was favored over cyclic carbonate production. Because of the relative rates of copolymerization of cyclohexene oxide and carbon dioxide as a function of the halogen atom in the [Zn(O-2,6-X₂C₆H₃)₂ THF]₂ catalysts, that is, F > Cl > Br, activation of epoxide by the zinc center is proposed to be rate-limiting relative to the CO₂ insertion process.

Full text of DOI:10.1021/ja002855h

J. Am. Chem. Soc. 2000, 122, 12487-12496
12487
Bis 2,6-difluorophenoxide Dimeric Complexes of Zinc and Cadmium
and Their Phosphine Adducts: Lessons Learned Relative to Carbon
Dioxide/Cyclohexene Oxide Alternating Copolymerization Processes
Catalyzed by Zinc Phenoxides
Donald J. Darensbourg,* Jacob R. Wildeson, Jason C. Yarbrough, and
Joseph H. Reibenspies
Contribution from the Department of Chemistry, Texas A&M UniVersity, P.O. Box 30012,
College Station, Texas 77842
ReceiVed August 2, 2000
Abstract: Dimeric phenoxide derivatives of zinc and cadmium have been synthesized from the reaction of the
corresponding metal bistrimethylsilylamide and two equivalents of 2,6-F2C6H3OH in tetrahydrofuran. The
zinc analogue, [Zn(O-2,6-F2C6H3)2‚THF]2 (1), has been characterized in the solid state via X-ray crystallography,
where the zinc centers are shown to possess distorted tetrahedral geometry containing two bridging phenoxides
and a terminal phenoxide and THF ligand. The distance between the metal centers (Zn‚‚‚Zn) was found to be
3.059 Å, and the THF ligands lie on opposite sides of the plane formed by the two zinc atoms and two bridging
phenoxide ligands’ oxygen atoms. There are several Zn‚‚‚F nonbonding distances involving the bridging
phenoxide ligands that are less than the van der Waals internuclear distance. In addition, both the zinc and
cadmium dimeric derivatives have been prepared such that the labile THF ligands are replaced by the sterically
encumbering basic phosphine, PCy3. The solid-state structures of [Zn(O-2,6-F2C6H3)2‚PCy3]2 (2) and [Cd(O-
2,6-F2C6H3)2‚PCy3]2 (5) are similar to that of complex 1, where the tricyclohexylphosphine ligands, like the
THF ligands, are accommodated in a trans configuration. The 31P NMR spectrum of complex 2 in C6D6 upon
addition of free PCy3 exhibits sharp resonances assigned to both the complex (9.58 ppm) and free PCy3 (10.6
ppm), which is indicative of slow exchange of the phosphine ligands. On the other hand, the phosphine ligands
on the cadmium derivative (5) are involved in an exchange process with free PCy3 via a rapid equilibrium
between 5 and two equivalents of Cd(O-2,6-F2C6H3)2(PCy3)2. The equilibrium reaction strongly favors the
monomer cadmium bisphosphine complex at low temperature (-80 °C). As expected, the 113Cd and 31P NMR
spectra of complex 5 in solution in the absence of excess PCy3 is quite similar to that determined in the solid
state by CP/MAS. Complex 1 and its chloro- and bromophenoxide analogues were shown to be effective
catalysts for the copolymerization of cyclohexene oxide and CO2, the terpolymerization of cyclohexene oxide/
propylene oxide/CO2, and the homopolymerization of cyclohexene oxide. In the case of the copolymerization
process (80 °C and 55 bar), the polycarbonate copolymer that was produced is completely alternating, with no
polyether linkages. At the same time, the homopolymerization of cyclohexene oxide to afford polyether in the
presence of 1 as catalyst is much more facile than the copolymerization process. Importantly, for both
copolymerization and homopolymerization processes catalyzed by complex 1, the initiator of the polymer
chain growth is a difluorophenoxide unit, as revealed by 19F NMR, with both CO2 insertion and epoxide
ring-opening being involved in the initiation step. At 80 °C and 55 bar, the coupling of propylene oxide and
CO2 led almost exclusively to propylene carbonate. On the other hand, at lower temperatures (i.e., 40 °C),
copolymer formation was favored over cyclic carbonate production. Because of the relative rates of
copolymerization of cyclohexene oxide and carbon dioxide as a function of the halogen atom in the [Zn(O-
2,6-X2C6H3)2‚THF]2 catalysts, that is, F > Cl > Br, activation of epoxide by the zinc center is proposed to be
rate-limiting relative to the CO2 insertion process.
Introduction
report, the known zinc bis(alkoxides) or bis(aryloxides) were
insoluble in common organic solvents and, hence, were pre-
sumed to be polymeric in structure.2 Recently, we carried out
rather comprehensive studies of the synthesis and structural
characterization of soluble bisaryloxides of zinc3,4 and their
Utilizing the sterically encumbered aryloxide ligand contain-
ing tert-butyl substituents in the 2,6-positions of the aryl ring,
Caulton and co-workers prepared the first well-characterized,
soluble zinc aryloxide complexes.1 That is, (2,6-di-tert-
butylphenoxide)2Zn(THF)2 was obtained upon reacting Zn-
[N(SiMe3)2]2 with 2,6-di-tert-butylphenol in tetrahydrofuran, and
its structure was defined crystallographically. Prior to their
(2) Mehrotra, R. C. AdV. Inorg. Chem. Radiochem. 1983, 26, 269.
(3) Darensbourg, D. J.; Holtcamp, M. W. Macromolecules 1995, 28,
7577.
(4) 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.
(1) Geerts, R. L.; Huffman, J. C.; Caulton, K. G. Inorg. Chem. 1986,
25, 1803.
10.1021/ja002855h CCC: $19.00 © 2000 American Chemical Society
Published on Web 12/02/2000

12488 J. Am. Chem. Soc., Vol. 122, No. 50, 2000
Darensbourg et al.
phosphine adducts.5 These investigations were stimulated by
the potential use of these derivatives as catalysts for the coupling
reaction of carbon dioxide and epoxides to afford polycarbonates
and/or cyclic carbonates. Indeed, these complexes represent the
first reported catalysts for the coupling of CO2 and cyclohexene
oxide to provide high-molecular-weight polycarbonates which
exhibited greatly enhanced reactivity over earlier published
work.6 Since these studies, several other active zinc catalysts
for these processes have been discovered.7-10
Chemical Co. and purified by distillation over calcium hydride. Bone
dry carbon dioxide was purchased from Scott Specialty Gases, Inc.
2,6-difluorophenol, 2,6-dichlorophenol, 2,6-dibromo-4-methylphenol,
and tricyclohexylphosphine were purchased from Aldrich Chemical Co.
and were sublimed and stored in a glovebox prior to use. Zn-
[N(SiMe3)2]2 and Cd[N(SiMe3)2]2 were prepared according to published
literature,13 stored in the glovebox, and used immediately after removal
from the box. K(2,6-difluorophenoxide) was synthesized and isolated
from the reaction of KH with the corresponding phenol in THF. Infrared
spectra were recorded on a Mattson 6081 spectrometer with DTGS
and mercury cadmium telluride (MCT) detectors. All isotopically
labeled solvents for NMR experiments were purchased from Cambridge
Isotope Laboratories. 1H and 13C NMR spectra were recorded on Varian
XL-200E, Unity +300 MHz, and VXR 300 MHz superconducting high-
resolution spectrometers. 19F and 31P data were acquired on a Unity
+300 MHz superconducting NMR spectrometer operating at 282 and
121 MHz, respectively. All 19F NMR data are referenced to 10% CFCl3
and 1% CClH2CClF2 in acetone-d6, whereas all 31P NMR data are
referenced to H3PO4 (85% in D2O). Solution-state 113Cd spectra were
recorded on a Varian XL-400E superconducting high-resolution
spectrometer operating at 88 MHz and using an external 0.1 M Cd-
(ClO4)2/D2O reference. Solid-state 113Cd spectra were recorded on a
Bruker MSL 300 superconducting spectrometer using an external Cd-
(ClO4)2 reference and operating at 66.546 MHz. Elemental analyses
were carried out by Galbraith Laboratories, Inc.
By way of contrast, when the bis-(2,6-di-tert-butylphenolate)-
zinc complex is isolated from a hydrocarbon solvent, it
crystallizes as a dimeric species containing two trigonal zinc
metal centers, each with two bridging and one terminal
aryloxides.1,11 This structure is similar to those we reported for
the monomeric, anionic derivatives of zinc and cadmium, for
example, [K(THF)6][Zn(O-2,6-t-Bu2C6H3)3]. In this latter in-
stance, there are no bound THF molecules, even though several
are contained in the crystal lattice. Evidently, the arrangement
of three bulky phenoxides will not allow for the additional
binding of bases. Consistent with this conclusion is the
observation that this latter zinc derivative is ineffective at
binding(activating) epoxides for CO2 coupling reactions.
In continuation of our previous work on bis-phenoxides of
zinc and their role as catalysts for coupling reactions of carbon
dioxide and epoxides, we report here on several zinc complexes
that were derived from phenols that were substituted in the 2,6-
positions with sterically less-crowded, electron-withdrawing
halogen groups. In these instances, we have generated dimeric
species from THF solution which have been shown by X-ray
crystallography to contain tetrahedrally coordinated zinc centers.
That is, in addition to the two bridging and one terminal
phenolate ligands, there is a THF molecule bound to each zinc.
These zinc derivatives represent the most effective phenoxide
catalysts for producing polycarbonates from CO2/epoxides that
we have thusfar uncovered. Furthermore, unlike the other
reported complexes, these remain stable and maintain their
catalytic activity when stored in moist air. Although the increase
in catalytic effectiveness of these phenolates of zinc is not of
major significance, there are several important lessons to be
learned with regard to the mechanistic aspects of this process
from their use. These include questions of initiator(polymer end
group) and polyether formation. We have described here, as
well, complimentary investigations of the cadmium analogues.
Incorporated into these latter studies are solution and solid-state
113Cd NMR measurements which compare the structures of these
complexes in both states to each other as well as to the
corresponding zinc derivatives.
Note! Cadmium compounds and their wastes are extremely toxic
and must be handled carefully. Cadmium waste products should be
stored in a separate clearly-marked container.
Synthesis of [(2,6-Difluorophenoxide)2Zn]2(THF)2, (1). A 5-mL
THF solution of 2,6-difluorophenol (0.135 g, 1.04 mmol) was added
to a 5-mL THF solution of Zn[N(SiMe3)2]2 (0.20 g, 0.52 mmol),
resulting in a clear colorless solution which was stirred at room
temperature for 2 h. The solution was then concentrated to 5 mL and
placed in a freezer at -20 °C. Colorless block crystals formed after
several days. The supernatant liquid was removed by cannula, and the
crystals were dried under vacuum to yield 0.140 g of product (69%).
Anal. Calcd for C32H28O6F8Zn2: C, 48.56; H, 3.57. Anal. Calcd. for
C24H12O4F8Zn2: C, 44.54; H, 1.87. Found: C, 42.53; H, 2.32. The
inability to obtain a close match in elemental analysis is a result of the
1
lability of the THF coordinated to the zinc center. H NMR (C5D5N):
δ 1.78 [m, 8 H{THF}], 3.74 [m, 8 H, {THF}], 6.45 [m, 4 H], 6.90 [t,
8 H{3,5-H}]. 13C {H} NMR (C5D5N): δ 26.31 {THF}, 68.32 {THF},
111.54-112.34 [m, {3,5-C6H3} {4-C6H3}], 145.38 [t, JC-F ) 15.6 Hz
{ipso-C6H3}], 156.98 [dd, JC-Fl ) 235.95 Hz, JC-F2 ) 18.61 Hz {2,6-
19
C6H3}]. F {H} NMR(C5D5N): δ -135.57.
Synthesis of [(2,6-Difluorophenoxide)2Zn]2(PCy3)2, (2). A THF
solution of 2,6-difluorophenol was added to a THF solution of Zn-
[N(SiMe3)2]2 as described in the synthesis for complex 1. The resulting
clear colorless solution was stirred at room temperature for 30 min. A
5-mL THF solution of PCy3 (0.145 g, 0.52 mmol) was subsequently
added to the reaction mixture, and the resulting solution was allowed
to stir for an additional 90 min. Colorless block crystals appeared upon
cooling a concentrated THF solution (∼3 mL) of the complex at -20
°C for several days. The supernatant liquid was removed by cannula,
and the crystals were dried under vacuum to yield 0.181 g of product
(58%). Anal. Calcd for C60H78O4F8P2Zn2: C, 59.65; H, 6.52. Found:
Experimental Section
Methods and Materials. 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. Glassware was flamed out thoroughly prior to use. Solvents
were freshly distilled from sodium benzophenone before use. Cyclo-
hexene oxide and propylene oxide were purchased from Aldrich
(12) Darensbourg, D. J.; Niezgoda, S. A.; Draper, J. D.; Reibenspies, J.
H. Inorg. Chem. 1999, 38, 1356.
(13) Burger, H.; Sawodny, W.; Wannaget, V. J. Organomet. Chem. 1965,
3, 113.
(5) Darensbourg, D. J.; Zimmer, M. S.; Rainey, P.; Larkins, D. L. Inorg.
Chem. 2000, 39, 1578.
(6) Rokicki, A.; Kuran, W. J. Macromol. Sci., ReV. Macromol. Chem.
1981, C21, 135.
(7) Super, M.; Berluche, E.; Costello, C.; Beckman, E. Macromolecules
1997, 30, 368.
(14) SMART 1000 CCD; Bruker Analytical X-ray Systems: Madison,
WI; 1999.
(15) SAINT-Plus, version 6.02; Bruker: Madison, WI; 1999.
(16) Sheldrick, G. SHELXS-86 Program for Crystal Structure Solution;
Institut fur Anorganische Chemie der Universitat: Tammanstrasse 4, D-3400
Gottingen, Germany; 1986.
(8) Cheng, M.; Lobkovsky, E. B.; Coates, G. W. J. Am. Chem. Soc. 1998,
120, 11018.
(9) Beckman, E. Science 1999, 283, 946.
(17) Sheldrick, G. SHELXL-97 Program for Crystal Structure Refine-
ment; Institut fur Anorganische Chemie der Universitat: Tammanstrasse
4, D-3400 Gottingen, Germany; 1997.
(10) Darensbourg, D. J.; Zimmer, M. S. Macromolecules 1999, 32, 2137.
(11) Kunert, M.; Bra¨uer, M.; Klobes, O.; Go¨rls, H.; Dinjus, E.; Anders,
E. Eur. J. Inorg. Chem. 2000, 1803.
(18) SHELXTL, version 5.0; Bruker: Madison, WI; 1999.
(19) Darensbourg, D. J.; Zimmer, M. S.; Rainey, P.; Larkins, D. L. Inorg.
Chem. 1998, 37, 2852.

Dimeric Complexes of Zinc and Cadmium
J. Am. Chem. Soc., Vol. 122, No. 50, 2000 12489
Table 1. Crystallographic Data for Complexes 1-3 and 5
1
2
3
5
empirical formula
FW
crystal system
space group
V, Å3
C32H28F8O6Zn2
791.36
triclinic
P1(bar)
1554.4(4)
2
C72H102F8O7P2Zn2
1424.14
orthorhombic
Pna2(1)
7000.8(8)
4
C32H28F8K2O6Zn
803.2
monoclinic
C 2/c
3772.6(6)
4
C72H102F8O7P2Cd2
1520.39
triclinic
P1(bar)
3498.1(8)
2
Z
a, Å
b, Å
c, Å
9.4696(15)
10.2093(16)
16.492(3)
78.116(3)
87.977(4)
85.147(4)
110(2)
1.691
3.379
7199
2593
16.9185(11)
15.2890(10)
27.0648(18)
90.0
90.0
90.0
110(2)
1.328
0.803
9409
16.8412(16)
18.3146(17)
13.5747(13)
90.0
115.707(2)
90.0
110(2)
1.229
0.939
4559
14.946(2)
15.056(2)
15.609(2)
86.456(2)
86.197(2)
89.733(3)
110(2)
1.441
2.050
16009
11 128
R, deg
â, deg
γ, deg
T, K
d
calcd, g/cm3
absorp coeff, mm-1
unique reflections
observed reflections
[I > 2σ]
4369
3356
R,a % [I > 2σ(I)]
Rw,a %
5.82
9.11
8.08
18.71
5.77
15.51
4.19
10.11
a R ) Σ|Fo| - |Fc|/ΣFo. Rw ) {[Σw(Fo2 - Fc2)2/[Σw(Fo2)2]}1/2
.
C, 59.64; H, 6.73. 1H NMR (C5D6): δ 0.86-2.20 [m, 66 H, P(C6H11)3],
6.27 [m, 4 H], 6.80 [t, 8 H{3,5-H}]. 13C {H} NMR (C5D5N): δ 27.31-
32.66 P(C6H11)3, 111.69-112.12 [m, {3,5-C6H3} {4-C6H3}], 145.32
[t, JC-F ) 14.6 Hz {ipso-C6H3}], 156.94 [dd, JC-Fl ) 235.95 Hz, JC-F2
) 8.55 Hz {2,6-C6H3}]. 19F{H} NMR(C6D6): δ -133.48 [bridging
-OC6H3F2], -127.38 [terminal -OC6H3F2]. 31P {H} NMR (C6D6): δ
9.576.
111.94 [m, {3,5-C6H3}], 148.16 [t, JC-F ) 15.6 Hz {ipso-C6H3}], 157.83
[dd, JC-F1 ) 235.45 Hz, JC-F2 ) 11.07 Hz {2,6-C6H3}]. 19F {H} NMR-
31
113
(C6D6): δ -134.74. P {H} NMR (C6D6): δ 30.63 [dd, J
)
Cd-P
111
2508 Hz, J
) 2390 Hz].
Cd-P
Synthesis of [(2,6-Dichlorophenoxide)2Zn]2(THF)2, (6). A 5-mL
THF solution of 2,6-dichlorophenol (0.170 g, 1.04 mmol) was added
to a 5-mL THF solution of Zn[N(SiMe3)2]2 (0.20 g, 0.52 mmol), leading
to a clear colorless solution which was stirred at room temperature for
2 h. The product was isolated by precipitation with hexanes, and the
supernatant liquid was removed by cannula. The solid was then
vacuumed to dryness, yielding 0.199 g (69%). Anal. Calcd for C24H12O4-
Cl8Zn2‚(THF)2: C, 41.64; H, 3.06. Anal. Calcd for C24H12O4Cl8Zn2:
C, 37.01; H, 1.56. Found: C, 37.29; H, 2.18. As with 1, THF lability
causes an inaccurate match in the elemental analysis. 1H NMR
(C5D5N): δ 1.86 [m, 8 H {THF}], 3.61 [m, 8 H, {THF}], 6.47 [t, 4 H
{4-H}], 7.35 [d, 8 H{3,5-H}]. 13C {H) NMR (C5D5N): δ 26.31 {THF},
68.32 {THF}, 115.42 [s, {4-C6H3}], 125.74 [s, {3,5-C6H3}], 128.90
[s, {2,6-C6H3}], 158.52 [s, {ipso-C6H3}].
Synthesis of [(2,6-Dibromo-4-methylphenoxide)2Zn]2(THF)2, (7).
A 5-mL THF solution of 2,6-dibromo-4-methylphenol (0.275 g, 1.04
mmol) was added to a 5-mL THF solution of Zn[N(SiMe3)2]2 (0.20 g,
0.52 mmol), leading to a clear colorless solution which was stirred at
room temperature for 2 h. The product was isolated by precipitation
with hexanes, and the supernatant liquid was removed by cannula. The
solid was then vacuumed to dryness, yielding 0.275 g (75%). Anal.
Calcd for C36H40O6Br8Zn2: C, 32.39; H, 2.72. Found: C, 30.90; H,
2.53. 1H NMR (C5D5N): δ 1.61 [m, 4 H{THF}], 2.02 [s, 12 H{-CH3}],
3.64 [m, 4 H, {THF}], 7.30 [s, 8 H{3,5-H}]. 13C {H} NMR (C5D5N):
δ 20.05 [s, {-CH3}], 26.31 {THF}, 68.32 {THF}, 115.72 [s, {4-C6H3}],
123.01 [s, {3,5-C6H3}], 126.21 [s,{2,6-C6H3}], 157.29 [s, {ipso-C6H3}].
X-ray Crystallography. A Bausch and Lomb 10× microscope was
used to identify suitable colorless crystals of 1-3 and 5 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.
Synthesis of K2[(2,6-Difluorophenoxide)4Zn](THF)2, (3). After the
addition of 2,6-difluorophenol to Zn[N(SiMe3)2]2, as in the preparation
of 1, and following an hour of stirring, a 5-mL THF solution of K(O-
2,6-F2C6H3) (0.174 g, 1.04 mmol) was added to the reaction mixture.
The resulting solution was allowed to stir for an additional 90 min and
then was concentrated to approximately 10 mL and placed in a freezer
at -20 °C. Colorless block crystals formed after several days. The
supernatant liquid was removed by cannula, and the crystals were dried
under vacuum to yield 0.214 g of product (63%). Anal. Calcd for
1
ZnC24H12F8O4K2: C, 43.94; H, 1.85. Found: C, 43.64; H, 2.02. H
NMR (CD3CN): δ 6.19 [m, 4 H], 6.62 [m, 8 H{3,5-H}]. 13C {H} NMR
(C5D5N): δ 111.27-111.81 [m, {3,5-C6H3} {4-C6H3}], 144.87 [t, JC-F
) 27.155 Hz {ipso-C6H3}], 156.69 [dd, JC-F1 ) 233.84 Hz, JC-F2
)
9.05 Hz {2,6-C6H3}]. 19F {H} NMR(CD3CN): δ -136.782.
Synthesis of [(2,6-Difluorophenoxide)2Cd]2(THF)2, (4). This com-
plex was synthesized by a method similar to that described for 1 by
adding a 5-mL THF solution of 2,6-difluorophenol (0.120 g, 0.92 mmol)
to a 5-mL THF solution of Cd[N(SiMe3)2]2 (0.20 g, 0.46 mmol). A
white solid precipitated out immediately after the addition. This
precipitate is assumed to be a dimeric species of bis-2,6-difluorophen-
oxide cadmium containing bridging phenoxides. This precipitate was
washed several times with hexanes and dried under vacuum to yield
0.173 g (85%). Anal. Calcd for C24H12O4F8Cd2: C, 38.89; H, 1.64.
Found: C, 38.97; H, 1.69. 1H NMR (C5D5N): δ 1.60[m, 8 H {THF}],
3.63 [m, 8 H, {THF}], 6.30 [m, 4 H], 6.94 [t, 8 H {3,5-H}. 13C {H}
NMR (C5D5N): δ 26.08 {THF}, 68.40 {THF}, 108.69 [t, {4-C6H3}],
112.2 [{3,5-C6H3}], 148.43 [t, JC-F ) 15.6 Hz {ipso-C6H3}], 157.91
[dd, JC-Fl ) 235.44 Hz, JC-F2 ) 11.57 Hz {2,6-C6H3}]. 19F{H} NMR-
(C5H5N): δ -134.97.
Synthesis of [(2,6-Difluorophenoxide)2Cd]2(PCy3)2, (5). A 10-mL
THF solution of 2,6-difluorophenol (0.120 g, 0.92 mmol) and PCy3
(0.130 g, 0.46 mmol) were added concurrently to a 5-mL THF solution
of Cd[N(SiMe3)2]2 (0.20 g, 0.46 mmol), leading to a clear colorless
solution which was stirred at room temperature for 2 h. The solution
was then concentrated to approximately 8 mL and placed at -20 °C.
Colorless block crystals formed after several days. The supernatant
liquid was removed by cannula, and the crystals were dried under
vacuum to yield 0.183 g of product (61%). Anal. Calcd for C60H78O4F8P2-
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, 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.14
1
Cd2: C, 55.34; H, 6.05. Found: C, 55.14; H, 5.49. H NMR (C5D6):
δ 0.97-2.25 [m, 66 H, PC6H11], 6.18 [m, 4 H], 6.73 [t, 8 H{3,5-H}].
13C {H} NMR (C5D5N): δ 27.26-32.65 (PC6H11), 108.86 [t, {4-C6H3}],

12490 J. Am. Chem. Soc., Vol. 122, No. 50, 2000
Darensbourg et al.
The structures were solved by direct methods. Full-matrix least-
yields of >60% by the general synthetic route described by
Caulton and co-workers,1 and by previous work in our labora-
tories,3,4 for the synthesis of a large variety of monomeric
(ArO)2Zn(base)2 complexes. The reactions, depicted specifically
in eq 1 for the 2,6-difluorophenol derived product, were carried
out in THF solution and employing purified starting materials.
Although vacuum-dried samples of complex 1 exhibited varying
levels of THF content as indicted by C,H elemental analysis,
1H NMR of the isolated complex revealed the presence of one
THF molecule per zinc. Crystals of complex 1 suitable for X-ray
analysis were obtained from a cooled (-20 °C) concentrated
solution of the complex in THF.
squares anisotropic refinement for all non-hydrogen atoms yielded RF
2
and wRF values as indicated in Table 1 at convergence. Hydrogen
atoms were place 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 Crystallography, Vol. C.
For the title compound: data reduction, SAINTPLUS (Bruker15);
program(s) used to solve the structure, SHELXS-86 (Sheldrick16);
program(s) used to refine the structure, SHELXL-97 (Sheldrick17);
program(s) used for molecular graphics, SHELXTL version 5.0
(Bruker18); software used to prepare material for publication, SHELXTL
version 5.0 (Bruker18).
Solid-State 113Cd NMR. The solid-state NMR spectra were acquired
utilizing a Bruker MSL 300 superconducting spectrometer equipped
with a magnet operating at 7.05 T (Larmor frequency of 66.546 MHz
for 113Cd). The samples were ground and packed into zirconium oxide
rotors with Kel-F end caps for use in a 7-mm supersonic probe from
Bruker. Spinning speeds were regulated by a Bruker spin rate controller.
All chemical shifts and tensor elements were referenced to an external
sample of 0.1 M Cd(C1O4)2 in D2O solution at 25 °C, with positive
shifts denoting movement of resonances to lower shielding. Cd(NO3)2‚
4H2O was used as a secondary standard relative to the cadmium
perchlorate solution in D2O to account for the cross polarization (CP)
with proton decoupling. The recycle delay used was 15 s, with a 1 H
7π/2 pulse width, 5.25 µs, and a contact time of 15 ms for all samples.
Principal elements of the shielding tensor were extracted utilizing
WINFIT software package from Bruker Instruments running on a
Adosea pentium personal computer.
High-Pressure Copolymerization of CO2 with Cyclohexene
Oxide. A sample of the active catalyst (0.100 g) was dissolved in 20.0
mL of cyclohexene oxide. The solution was loaded via an injection
port into a 300-mL stainless steel Parr autoclave which had previously
been dried overnight under vacuum at 80 °C. The reactor was
pressurized to 600 psi with C02 and heated to 80 °C, which increased
the pressure to 750-800 psi, and left to react between 24 and 48 h.
After 24-48 h of reaction time, the reactor was cooled and opened,
and the viscous/solid mixture was isolated by dissolution in CH2Cl2
Subsequent reaction of complex 1 with two equivalents of
the sterically encumbering basic tertiary phosphine, PCy3, at
ambient temperature resulted in the formation of complex 2, in
which the THF ligands are simply displaced by the better donor
phosphine ligands. On the other hand, the addition of four
equivalents of K(O-2,6-F2C6H3) in THF to complex 1 led to
disruption of the dimeric structure with concomitant production
of the monomeric dianion, 3. In this instance, the proclivity of
the nonbulky phenoxide to bridge two zinc centers is overcome
by the addition of strong nucleophiles. These processes are
represented in Scheme 1. Colorless crystals suitable for X-ray
crystallographic studies of both derivatives, 2 and 3, were readily
obtained from concentrated, cooled (-20 °C) solutions of the
respective complex in THF.
The cadmium analogue of complex 1 was synthesized in a
manner similar to that described for the zinc derivative.
However, upon adding 2,6-difluorophenol to Cd[N(SiMe3)2]2,
the product [Cd(O-2,6-F2C6H3)2‚THF]2, 4, immediately pre-
cipitated from the THF solution. Nevertheless, complex 4 is
concluded to be dimeric, as evident by 1H NMR in C5D5N and
elemental analysis, which show 1 equiv of THF per cadmium
center. Upon addition of PCy3 simultaneously with the phenol
to the cadmium amide in THF, the complex formed remains in
solution to afford the analogue to complex 2, [Cd(O-2,6-
F2C6H3)2‚PCy3]2 (5). Well-formed crystals of this latter deriva-
tive were obtained from a cooled, concentrated THF solution
of 5.
1
and precipitated out in MeOH. The polymer was analyzed by H and
13C NMR spectroscopy and by GPC measurements. We have deter-
mined physical and mechanical properties of the poly(cyclohexenylene
carbonate) polymer produced herein, as well as a comparison of similar
characteristic of it with the widely applied bisphenol A-polycarbonate
polymer. These latter experiments were carried out in collaboration
with the research group of Dr. Cor Koning at DSM, The Netherlands.
This work is being published elsewhere.
High-Pressure Copolymerization of CO2 with Propylene Oxide.
A 0.100-g amount of active catalyst was dissolved in 20.0 mL of
propylene oxide. The resulting solution was added through the injection
port to a predried 300-mL autoclave, and the reactor was pressurized
to 600 psi with CO2. The reactor was heated at 55 °C, raising the
pressure to 650-700 psi, for 48 h. After this period of time, the reaction
mixture was diluted with CH2Cl2 (1:10) and analyzed by infrared
spectroscopy in the ν(CO) region.
High-Pressure Terpolymerization of CO2 with Propylene Oxide
and Cyclohexene Oxide. A 0.100-g amount of active catalyst was
dissolved in a solution composed of 10.0 mL (∼50 mol %) of
cyclohexene oxide and 7.0 mL (∼50 mol %) of propylene oxide. The
resulting solution was added through the injection port of a 300-mL
predried autoclave, and the reactor was pressurized to 600 psi with
CO2. The reactor was then heated to 55 °C, raising the pressure to
650-700 psi, for between 24 and 48 h. After this period of time, the
reactor was opened, and the viscous/solid mixture was dissolved in
CH2Cl2 and precipitated out in MeOH. The polymer was analyzed by
1H and 13C NMR spectroscopy and by DSC measurements.
In addition, it was possible to prepare the corresponding
chloro- and bromophenolate derivatives of complex 1 in a
similar manner. However, after several attempts at crystalliza-
tion, it has not thus far been possible to obtain crystals suitable
1
for definitive X-ray structure determinations. Again, H NMR
Results and Discussion
and elemental analysis imply that these species are also dimeric,
in that one equivalent of THF per zinc metal was found. Recall
that monomeric zinc phenoxides exhibit 2 equivs of THF per
zinc center.
Synthesis, Spectral, and Structural Characterization of
Dimeric Zinc and Cadmium Phenoxides. The dimeric zinc
phenoxide derivatives were synthesized and isolated in purified

Dimeric Complexes of Zinc and Cadmium
J. Am. Chem. Soc., Vol. 122, No. 50, 2000 12491
Scheme 1
Complexes 1-3 and 5 have been characterized in the solid
state by X-ray crystallography, and a list of selected bond lengths
and angles is provided in Table 2. In the crystallographic
independent unit of complexes 1 and 5, two half molecules of
1 and 5 were found; therefore, for each complex, quite similar
bond distances and angles are separately compiled in Table 2
for the two independently generated molecules. Complex 1
crystallized from THF as a dimer in which the zinc centers are
coordinated in distorted tetrahedral geometry, each containing
two bridging phenoxides and a terminal phenoxide and THF
ligand. Figure 1 displays a thermal ellipsoid rendering of 1,
along with a partial atom-labeling scheme for one of the
molecules generated from the two half molecules in the unit
cell (See Supporting Information for the other labeling scheme).
Because the molecular parameters are so similar, only those
for one of the independently generated molecules in the unit
cell will be discussed. The structure of 1 consists of a planar
arrangement of the zinc atoms and the oxygen atoms of the
bridging phenoxide ligands which form a parallelogram with
an O-Zn-O bond angle of 80.07(19)° and a Zn-O-Zn bond
angle of 99.93(19)°. The Zn-O bond lengths for the bridging
phenoxide ligands average 1.992(4) Å, in which the terminal
phenoxide Zn-O bond length is over 0.1 Å shorter at 1.869(4)
Å. The distance between the metal centers (Zn‚‚‚Zn) in the dimer
is 3.059 Å. By way of comparison, the parallelogram formed
by the two bridging 2,6-di-tert-butylC6H3O and two zinc atoms
of the previously reported dimer, [Zn(O-2,6-t-Bu2C6H3)2]2,
where the zinc centers are coordinated in a distorted trigonal
planar geometry, is more rectangular, with an average O-Zn-O
angle of 82.4(1)°. This, coupled with the shorter average Zn-O
bond length for the more basic bridging phenoxide of 1.961(1)
Å, places the metal centers closer, with a Zn‚‚‚Zn distance of
2.9484(6) Å.11 The two THF ligands in complex 1, with Zn-O
bond lengths of 2.003(4) Å, are arranged in a trans configuration.
The average Zn-O-C(aryl) bond angles are 129.7 and 124.0°
for bridging and terminal phenoxides, respectively.
and 2.7019 Å, respectively. This is illustrated in Figure 2 for
the closest Zn‚‚‚F interaction comprising F(1A), which leads
to an incipient planar five-membered chelate ring. These
separations are slightly shorter than the nonbonded van der
Waals internuclear distance of 2.86 Å. The distances for the
nearest fluorine atom of the terminal phenoxide to the zinc center
(Zn(1)‚‚‚F(3) and Zn(2)‚‚‚F(7)) are 2.961 Å and 2.728 Å.
As indicated in Figure 3, which represents the thermal
ellipsoid drawing of complex 2, and in Table 2, there is little
perturbation to the dimeric unit of complex 1 upon replacing
the THF donor ligands with the sterically encumbering, basic
phosphine, PCy3. As was the case for the THF ligands, the
tricyclohexylphosphine ligands are accommodated in a trans
configuration. The average Zn-P bond distance of 2.410(4) Å
is slightly shorter than that of 2.433(2) Å, which is found in
the monomeric Zn(O-2,6-t-Bu2C6H3)2‚PCy3 derivative.5,19 In
complex 2, because of the bulky phosphine ligands, the fluorine
atoms of the terminal phenoxide ligands are significantly closer
to the metal centers, with Zn(1)‚‚‚F(7) and Zn(2)‚‚‚F(1) separa-
tions of 2.840 and 2.856 Å, respectively. The bridging phenox-
ides provide Zn(1)‚‚‚F(4) and Zn(2)‚‚‚F(6) distances of 3.191
and 3.158 Å. Three molecules of THF are found in the crystal
lattice per independent unit.
Complex 3, which results from the addition of the strong
-
nucleophile, O-2,6-F2C6H3 to 1, leads to a disruption of the
dimeric structure by formation of a monomeric tetrakisphenoxide
zinc dianion of distorted tetrahedral geometry. Figure 4 contains
a thermal ellipsoid representation of complex 3. Two of the
phenoxide ligands are bound to zinc with Zn-O bond lengths
of 1.949(3) Å, whereas the corresponding other two Zn-O bond
lengths are slightly longer at 1.974(3) Å. The O-Zn-O bond
angles range from 97.1 to 117.6°, with an average value of
109.7°. Previously, we reported trisphenoxide derivatives of zinc
and cadmium which were produced by the addition of the
sterically demanding phenoxide, KO-2,6-t-Bu2C6H3, to the
monomeric M(O-2,6-t-Bu2C6H3)2‚2THF derivatives.20 The struc-
tures of these anions are trigonal planar, and the potassium
cations are surrounded by six THF molecules. On the other hand,
in complex 3, the potassium cations exhibit interaction with both
fluorine atoms (K‚‚‚F ) 2.723(3) Å) and THF molecules.
Interestingly, there are several short Zn‚‚‚F nonbonding
distances in 1, with those involving the fluorine atoms of the
two bridging phenoxides being the shorter. That is, the Zn(1)‚
‚‚F(2) and Zn‚‚‚F(1A) separations were found to be 2.837 and
2.768 Å, respectively. Comparable distances in the other
molecule, that is, Zn(2)‚‚‚F(6A), were determined to be 2.746
(20) Darensbourg, D. J.; Niezgoda, S. A.; Draper, J. D.; Reibenspies, J.
H. Inorg. Chem. 1999, 38, 1356.

12492 J. Am. Chem. Soc., Vol. 122, No. 50, 2000
Darensbourg et al.
Table 2. Selected Bond Distances (Å) and Bond Angle (deg) for
Complexes 1-3, and 5a,b
Complex 1
Zn(1)-O(1)
Zn(1)-O(2)
1.986(4)
1.869(4)
Zn(1)-O(lA)
Zn(1)-O(3)
1.997(4)
2.003(4)
Zn(1)-O(l)-Zn(1A) 99.93(19) O(1)-Zn(1)-O(lA) 80.07(19)
O(2)-Zn(1)-O(1)
131.3(2)
Zn(1)-O(1)-C(aryl) 129.7(4)
O(2)-Zn(1)-O(1A) 122.31(17)
Zn(1)-O(2)-C(aryl) 124.0(4)
Zn(2)-O(4)
Zn(2)-O(5)
1.980(4)
1.877(4)
Zn(2)-O(4A)
Zn(2)-O(6)
1.986(4)
2.008(4)
Zn(2)-O(4)-Zn(2A) 101.34(19) O(4)-Zn(2)-O(4A) 78.66(19)
O(5)-Zn(2)-O(4) 124.81(19) O(5)-Zn(2)-O(4A) 136.89(17)
Zn(2)-O(4)-C(aryl) 128.2(4) Zn(2)-O(5)-C(aryl) 125.1(4)
Complex 2
Zn(1)-O(1)
Zn(1)-O(2)
Zn(2)-O(1)
Zn(2)-O(4)
2.001(8)
Zn(1)-O(3)
Zn(1)-P(2)
Zn(2)-O(2)
Zn(2)-P(1)
1.916(9)
2.419(4)
2.022(8)
2.401(4)
2.013(9)
2.006(9)
1.878(8)
Zn(1)-O(2)-Zn(2) 99.3(4)
Zn(1)-O(1)-Zn(2) 100.3(4)
O(1)-Zn(1)-O(2)
O(1)-Zn(2)-O(4)
O(2)-Zn(1)-P(2)
80.4(3)
O(1)-Zn(2)-O(2)
O(1)-Zn(1)-P(2)
80.1(3)
Figure 1. Thermal ellipsoid representation of complex 1.
124.6(4)
117.1(3)
Zn(1)-O(2)-C(aryl) 128.0(8)
114.7(3)
Zn(1)-O(1)-C(aryl) 131.8(8)
Zn(2)-O(1)-C(aryl) 127.6(8)
Zn(2)-O(2)-C(aryl) 131.8(8)
O(2)-Zn(1)-O(3)
O(2)-Zn(2)-P(1)
123.6(3)
114.8(3)
O(1)-Zn(2)-P(1)
Zn(1)-O(1)
117.4(2)
1.949(3)
Complex 3
Zn(1)-O(2)
1.974(3)
O(1)-Zn(1)-O(lB) 114.49(17) O(1)-Zn(1)-O(2B) 97.10(11)
117.58(11)
O(2)-Zn(1)-O(2B) 114.34(15)
O(2)-Zn(1)-O(1B) 97.10(11) O(1)-Zn(1)-O(2)
Zn(1)-O(2)-C(aryl) 121.6(2)
Zn(1)-O(1)-C(aryl) 132.8(3)
Complex 5
Cd(1)-O(1)
Cd(1)-O(2)
2.222(2)
2.124(2)
Cd(1)-O(lA)
Cd(1)-P(1)
2.255(2)
2.5426(10)
O(1)-Cd(1)-O(lA) 77.17(10) Cd(l)-O(1)-Cd(lA) 102.83(9)
O(2)-Cd(1)-O(1)
116.00(10) O(2)-Cd(1)-O(lA) 118.25(10)
Cd(1)-O(1)-C(aryl) 126.3(2)
Cd(1)-O(2)-C(aryl) 126.4(2)
O(1)-Cd(1)-P(1)
O(1A)-Cd(1)-P(1) 112.32(7)
124.15(7) O(2)-Cd(1)-P(1)
106.88(8)
Cd(2)-O(3)
Cd(2)-O(4)
2.244(2)
2.127(2)
Cd(2)-O(3A)
Cd(2)-P(2)
2.229(3)
2.5470(10)
O(3)-Cd(2)-O(3A) 77.66(10) Cd(2)-O(3)-Cd(2A) 105.24(10)
Figure 2. Closet Zn‚‚‚F interaction in complex 1. Insert illustrates
planarity of the incipient five-membered chelate ring.
O(4)-Cd(2)-O(3)
129.29(10) O(4)-Cd(2)-O(3A) 123.21(10)
Cd(2)-O(3)-C(aryl) 128.8(2)
Cd(2)-O(4)-C(aryl) 128.1(2)
O(3)-Cd(2)-P(2)
114.79(7) O(4)-Cd(2)-P(2)
102.98(8)
O(3A)-Cd(2)-P(2) 119.08(7)
a Estimated standard deviations are given in parentheses. b Symmetry-
generated atoms designated by (number A).
The cadmium analogue of complex 2 is also dimeric in the
solid state,that is, [Cd(O-2,6-F2C6H3)2‚PCy3]2 (5). A thermal
ellipsoid representation of complex 5 is shown in Figure 5 for
one of the independently generated molecules in the unit cell.
Because molecular parameters in the two independent molecules
are so similar (see Table 2), we will discuss only one here. Three
molecules of THF are found in the crystal lattice per independent
unit. The parallelogram formed by the two cadmium centers
and the bridging oxygen atoms of the phenoxide ligands has
an O-Cd-O angle smaller by ∼3° (77.17(10)°) and a Cd-
O-Cd angle larger by ∼3° (102.83(9)°) than those of complex
2. The nonbonding Cd‚‚‚Cd separation of 3.499 Å, as compared
to 3.059 Å for the zinc analogue, is consistent with the difference
in the radii of cadmium vs zinc. The Cd-O bridging phenoxide
bond lengths average 2.239(2) Å, whereas the terminal Cd-O
bond distances are 2.124(2) Å. By way of contrast to complex
2, the Cd-P bond length at 2.5426(9) Å is slightly longer than
that in the monomeric species Cd(O-2,6-t-Bu2C6H3)2‚PCy3 of
2.5274(12) Å, and the shortest Cd‚‚‚F nonbonding distances are
Figure 3. Thermal ellipsoid representation of complex 2.
to the fluorine atoms of the bridging phenoxides at 2.858 Å.
Again, this latter separation is shorter than the nonbonded
Cd‚‚‚F van der Waals internuclear distance of 3.05 Å.
In addition, these zinc and cadmium bisphenoxide derivatives
readily lend themselves to characterization in solution via
multinuclear NMR spectroscopy. For example, the 19F NMR

Dimeric Complexes of Zinc and Cadmium
J. Am. Chem. Soc., Vol. 122, No. 50, 2000 12493
in the 19F NMR spectrum at 134.97 ppm, which is indicative
of only terminal phenoxide ligands. On the other hand, the
isotropic 113Cd chemical shift of 4 obtained in the solid state at
9.26 ppm (chemical shift anisotropy and asymmetry parameter
of -245 and 0.20 ppm, respectively) is consistent with a dimeric
structure similar to that of the zinc derivative, that is, a cadmium
center containing 4 oxygen donor ligands.21
The CP/MAS solid-state 113Cd NMR spectrum of complex
5, [Cd(O-2,6-F2C6H3)2‚PCy3]2, displays an isotropic shift (σiso)
113Cd-P
of 252.6 ppm with J
) 2297 Hz. The chemical shift
anisotropy and asymmetry parameter were determined to be
-253 and 0.70 ppm, respectively. The solid-state 31P NMR
spectrum of 5 was composed of a resonance at 26.5 ppm with
coupling to cadmium. However, due to limitations on spinning
speeds, the resonances due to 111Cd-31P and 113Cd-31P coupling
could not be distinguished, thereby not allowing for an accurate
determination of coupling constants. The 113Cd NMR spectrum
of complex 5 in THF solution is similar to that obtained in the
solid state, displaying a chemical shift (δ) of 232.2 ppm with
113
J
Cd-P ) 2480 Hz at -80 °C. A complex cadmium spin-spin
Figure 4. Thermal ellipsoid representation of the anion of complex 3.
coupling pattern to the fluorine nuclei of two types of phenoxide
ligands was also observed, with J
≈ 66 Hz. The 31P
113Cd-F
resonance of complex 5 in THF solution appears at 30.7 ppm,
with a 113Cd-31P coupling constant of 2508 Hz. This resonance
is significantly downfield from that of free PCy3 and is slightly
upfield from the value for the monomer Cd(O-2,6-t-Bu2C6H3)2‚
PCy3 (δ ) 31.3 ppm with JCd-P ) 2274 Hz).19 Two 19F
resonances were noted in THF solution at -80 °C, -135.8 and
-136.7 ppm for bridging and terminal phenoxide ligands.
Curiously, the 113Cd, 31P, and 19F NMR spectra of complex
5 in solution were very dependent on the presence of small
quantities of free PCy3 impurities in the sample. This, in turn,
accounts for the temperature dependence of these NMR spectra
under these conditions. For example, in these instances, the 31
P
NMR spectrum at ambient temperature is slightly broadened,
and it is not possible to differentiate between the 113Cd-P and
111Cd-P coupling constants. However, when the temperature
is lowered, the resonance due to complex 5 sharpens and clearly
displays coupling to both cadmium isotopes (see Figure 6).
Simultaneously, a second minor resonance appears upfield,
around 24 ppm, which exhibits Cd-P coupling. Over this same
temperature range (ambient to -80 °C) the 19F NMR spectra
proceed from a single resonance at -135.4 ppm to two principal
signals of equal intensity at -135.8 and -136.7 ppm, with the
former displaying some unresolved coupling to cadmium, which
we assign to the bridging phenoxide ligand. Similarly, under
these conditions (small excess of PCy3 and ambient to -80 °C),
the 113Cd NMR spectrum shifts from a doublet centered at 218.8
ppm to one centered at 232.2 ppm.
Figure 5. Thermal ellipsoid representation of complex 5.
spectrum of complex 2 in C6D6 is consistent with its solid-
state structure, exhibiting two resonances of equal intensity that
correspond to bridging and terminal phenoxide ligands at
-133.5 and -127.4 ppm which are shifted upfield from the
free phenol. Further, the 31P NMR spectrum of complex 2 in
C6D6 displayed a small upfield shift from that of free PCy3,
going from 10.6 ppm for unbound PCy3 to 9.58 ppm in the
bound complex. Both bound and free 31P resonances are sharp
at ambient temperature for complex 2 in the presence of 0.1
equiv of free PCy3 (vide infra). Conversely, the dimeric nature
of the solid-state structures of the dichloro- and dibromophe-
nolate analogues of complex 1 (complexes 6 and 7) is confirmed
31P
The identity of the minor cadmium phosphine species (δ
111
113
) 22.4 ppm with J
and J
) 1590 and 1665 Hz at
Cd-P
Cd-P
-80 °C, respectively) present in samples of complex 5 contain-
ing small quantities of free PCy3 becomes apparent when excess
PCy3 is added to 5. That is, the addition of 6 equivs of PCy3 to
a sample of pure [Cd(O-2,6-F2C6H3)2‚PCy3]2 (5) resulted in a
31P NMR spectrum at ambient temperature which consisted of
broad, weak signals due mainly to 5 and free PCy3. When the
sample was cooled incrementally, the 31P resonance due to
complex 5 initially sharpened and decreased in intensity with
concomitant increases in intensity of the signals at ∼24 ppm
(which displays unresolved 111Cd and 113Cd coupling) and that
of free PCy3. Finally, at -80 °C the latter two resonances are
1
by the H NMR spectra of these derivatives in C5D5N, which
showed that 1 equiv of THF was bound per zinc center prior to
dissolution in C5D5N, whereas monomeric zinc bisphenoxides
show 2 equivs of bound THF per zinc center. On the other hand,
complex 1 appears to be monomeric in the strongly coordinating
solvent C5D5N, with only one 19F resonance being observed at
-136.8 ppm. Presumably because of their close similarity,
complexes 6 and 7 are also monomeric in C5D5N solution.
The cadmium analogue of complex 1, [Cd(O-2,6-F2C6H3)2‚
THF]2 (4), is soluble only in strong bases such as pyridine, in
which it is most likely dissociated into monomeric species. That
is, complex 4 dissolved in C5D5N exhibits a single resonance
(21) Darensbourg, D. J.; Niezgoda, S. A.; Draper, J. D.; Reibenspies, J.
H. J. Am. Chem. Soc. 1998, 120, 4690.

12494 J. Am. Chem. Soc., Vol. 122, No. 50, 2000
Darensbourg et al.
Table 3. Catalytic Activity for the Copolymerization of Carbon
Dioxide and Cyclohexene Oxidea
turnover no
(g polym/g Zn)
turnover freq
(g poly/g Zn/hr)
catalyst
[Zn(O-2,6-F2C6H3)2‚THF]2b 1
[Zn(O-2,6-Cl2C6H3)2‚THF]2 6
[Zn(O-2,6-Br2, 4-MeC6H2)2‚THF]2 7
790
539
329
16.5
11.2
7.2
a Catalyst loading (0.100 g), 20.0 mL of cyclohexene oxide, CO2 at
ambient temperature 41.3 bar. Reaction conditions: 80 °C at a total
pressure of 55 bar. b Molecular weight distribution of copolymer:
absolute Mn and Mw ) 42 000 and 252 000 g/mole, or a PDI ) 6.0.
explained by a rapid equilibrium process, as described in eq 2,
which is shifted far to the right as the temperature is lowered.
This is consistent, as well, with the conclusion based on the
spectrum of complex 4 when it is dissolved in the strongly
coordinating solvent, C5D5N (i.e., eq 3).
[Cd(O-2,6-F2C6H3)2‚PCy3]2 + excess PCy3 h
2Cd(O-2,6-F2C6H3)2(PCy3)2 (2)
[Cd(O-2,6-F2C6H3)2‚THF]2 + excess C5H5N h
2Cd(O-2,6-F2C6H3)2(C5H5N)2 (3)
By way of contrast, in the presence of sterically encumbering
phenoxides such as 2,6-Ph2C6H3O- and 2,6-t-Bu2C6H3O-, the
monomeric cadmium derivatives have been fully characterized
as monophosphine complexes of PCy3 which exhibit no
exchange with free PCy3 in solution. This is indicative of the
bisphosphine complex’s being inaccessible in these instances.
That is, rapid phosphine exchange occurs via an associative
pathway. Conversely, these bulky phenoxide derivatives of
cadmium in the presence of smaller phosphines, such as PMe3,
readily form bisphosphine complexes. Indeed, Cd(O-2,6-t-
Bu2C6H3)2(PMe3)2 has been characterized in the solid state by
X-ray crystallography.22 Furthermore, in these cases, the bisphe-
noxide complexes were shown to be in rapid equilibrium with
free phosphine and the monophosphine derivatives, with the
equilibrium position being shifted in favor of the bisphosphine
derivative at low temperature.
Copolymerization Reactions of Epoxides and Carbon
Dioxide. Complex 1, [Zn(O-2,6-F2C6H3)2‚THF]2, has been
shown to be an effective catalyst for the copolymerization of
cyclohexene oxide and carbon dioxide to afford high-molecular-
weight alternating copolymer, poly(cyclohexenylene carbonate),
eq 4. At 80 °C and 55 bar, this catalyst exhibits an average
Figure 6. Variable-temperature 31P NMR spectra of complex 5 in the
presence of 6 equivs of PCy3. Insert A shows 31P NMR spectrum of
complex 5 with small quantities of free PCy3 at -60 °C. Insert B
represents the 113Cd NMR spectrum of 5 in the presence of 6 equivs of
PCy3 at -40 °C.
turnover number of 790 g polym/g Zn for a 48-h reaction period,
which translates into a turnover frequency (TOF) of 16.5 g
polym/g Zn/hr (see Table 3). Unlike previous zinc phenoxide
catalysts and other recently developed catalysts,4,8 complex 1
is stable in moist air and does not lose activity when it is allowed
to stand in a moist oxygen atmosphere for prolonged periods
of time. Complex 1 is even more active at catalyzing the
homopolymerzation of cyclohexene oxide to provide the poly-
ether, poly(cyclohexenylene oxide), eq 5. This is particularly
interesting, because in the copolymerization process, the
copolymers that were produced have essentially 100% carbonate
sharp and are the only 31P signals present in the spectrum. The
113Cd NMR spectrum of this sample exhibits two broad
resonances centered at ∼220 and ∼360 ppm at ambient
temperature which ultimately are converted to a triplet centered
113Cd-31
at 395.1 ppm, with J
P value of 1678 Hz at -40 °C. These
observations are summarized in Figure 6. At the same time,
the sample displays only one 19F resonance in the 19F NMR
spectrum at -137.5 ppm. This spectral behavior is readily
(22) Darensbourg, D. J.; Rainey, P.; Larkins, D. L.; Reibenspies, J. H.
Inorg. Chem. 2000, 39, 473.

Dimeric Complexes of Zinc and Cadmium
J. Am. Chem. Soc., Vol. 122, No. 50, 2000 12495
linkages. The percentage of carbonate linkages is assessed from
the integration of the methane protons next to the carbonate
linkages (δ ) 4.60 ppm) and ether linkages (δ ) 3.45 ppm).
The copolymers that were produced herein possessed no
discernible resonance at 3.45 ppm.
We previously proposed that the lack of successive epoxide
ring-opening steps, that is, concurrent polyether formation and
polycarbonate production, is the result of the catalyst’s pos-
sessing only one epoxide binding site. For example, the
monomeric bisphenoxide zinc catalysts which contain two labile
ether ligands, for example, Zn(O-2,6-t-Bu2C6H3)2(THF)2, and
hence, two epoxide binding sites, generally afford 5-10% ether
linkages, even at high CO2 pressures.4 However, upon replacing
the THF ligands in these derivatives with a single, nonlabile
PCy3 ligand, only one epoxide binding site is present, and as a
result, the carbonate linkages in the copolymer are greatly
increased without any loss in catalytic activity.5 Furthermore,
this absence of ether linkages in the copolymers when complex
1 is used as catalyst strongly supports the fact that its dimeric
structure remains intact during catalysis. This conclusion argues,
as well, for the dimeric nature of the PCy3 zinc derivative,
complex 2, under the conditions of catalytic copolymerization
because this derivative was found to be ineffective as a catalyst
for the copolymerization of CO2 and cyclohexene oxide. Recall
that Zn(O-2,6-t-Bu2C6H3)2‚PCy3 is catalytically active for this
Figure 7. 19F NMR spectra in CDCl3 of polymers produced from
reactions catalyzed by [Zn(O-2,6-F2C6H3)2‚THF]2 (1): (a) Low-molec-
ular-weight polycarbonate from cyclohexene oxide and CO2. (b)
Polyether from cyclohexene oxide.
mixture shows two distinct glass transition temperatures at 62.3
°C and 115.2 °C for the polyethers and polycarbonates,
respectively. For the terpolymerization of propylene oxide (50
mol %), cyclohexene oxide (50 mol %), and CO2, the catalyst
exhibited an average turnover number of 201 g polym/g Zn and
a turnover frequency of 4.19 g polym/g Zn/hr. The formation
of cyclic propylene carbonate, which shows an intense asym-
metric ν(CO) stretching vibration at 1800 cm-1 and is a better
ligand for zinc than epoxides, hinders catalytic activity and
contributes to the decrease in turnover number and frequency.
The polymer that was produced has 84.76 mol % cyclohexene
oxide carbonate linkages, 11.90 mol % propylene carbonate
linkages, and 3.34 mol % propylene ether linkages. Propylene
carbonate and ether linkages show up in the 1H NMR spectrum
at 5.01 and 3.59 ppm, respectively. There are no cyclohexene
oxide linkages in the terpolymer. As a result of the incorporation
of propylene oxide into the polymer, the Tg is reduced from
115 °C for the cyclohexene oxide copolymer to 101.9 °C for
the terpolymer. For the copolymerization of propylene oxide
and CO2, 1 produced only a thin film of polymer. A comparison
of cyclic carbonates to polycarbonates produced was determined
process.5 Similarly, the fully coordinated Zn(O-2,6-F2C6H3)4
-2
dianion displays no catalytic activity for the coupling of CO2
and epoxides.
In addition to complex 1, complexes 6 and 7, which contain
phenoxides with chlorine and bromine substituents, are catalyti-
cally active for the production of high-molecular-weight,
completely alternating copolymers from CO2 and cyclohexene
oxide. Table 3 lists the turnover numbers and turnover frequen-
cies for the three derivatives for reactions carried out under
identical conditions. Similar to complex 1, complexes 6 and 7
remain active upon exposure to moist air. As seen in Table 3,
the catalytic activity for the halogenated phenoxide complexes
of zinc decrease as follows: F > Cl > Br. Because at least one
of the phenoxides remains in the coordination sphere of the zinc
center during chain growth, this trend in reactivity presumably
results from an increase in electron density at the zinc center
as the electronegativity of the halogen decreases, thereby
inhibiting the metal center’s binding ability to the epoxide
substrate. This, in turn, suggests that epoxide activation is more
important than CO2 insertion in this instance, because the CO2
insertion process is favored by a more nucleophilic metal-
alkoxide moiety.23,24
As mentioned earlier, complex 1 is a very efficient catalyst
for the polymerization of cyclohexene oxide to polyethers in
the absence of CO2. For example, upon adding the catalyst and
cyclohexene oxide to the reactor, polymerization to polyethers
takes place prior to the introduction of CO2; however, once CO2
is introduced, the copolymerization process dominates. The
consequent polymer that was produced is a mixture of poly-
(cyclohexene oxide carbonate) and poly(cyclohexene oxide
ether). Differential scanning calorimetry analysis of the polymer
by infrared spectroscopy, with propylene polycarbonate showing
-1
a ν(CO) absorption at 1750 cm
and cyclic propylene
carbonate exhibiting a ν(CO) mode at 1800 cm-1. Under similar
reaction conditions to those used in cyclohexene oxide/CO2
copolymerization, propylene cyclic carbonate is the major
product that is produced. However, in agreement with previous
work,4 when the reaction temperature is lowered to 40 °C,
propylene polycarbonate is the major product formed. However,
this reduction in the amount of cyclic propylene carbonate is
not sufficient to stop the poisoning effect of this byproduct to
the catalyst.
The presence of the fluorine substituents on complex 1 has
allowed us to verify the proposed initiation step in the
copolymerization process. Definitive support for initiation
involving nucleophile addition of the zinc-bound phenoxide to
epoxide or carbon dioxide monomers was obtained by 19F NMR
spectroscopy (eq 6). The 19F NMR spectra in CDCl3 of the
purified low-molecular-weight copolymer, as well as the poly-
ether, provided by cyclohexene oxide employing complex 1 as
catalyst are shown in Figure 7. The spectrum of the copolymer
displayed in Figure 7a contains at least four 19F resonances
between -126.8 and -127.3 ppm, whereas the corresponding
(23) Darensbourg, D. J.; Sanchez, K. M.; Reibenspies, J. H.; Rheingold,
A. L J. Am. Chem. Soc. 1989, 111, 7094.
(24) Darensbourg, D. J.; Mueller, B. L.; Bischoff, C. J.; Chojnacki, S.;
Reibenspies, J. H. Inorg. Chem. 1991, 30, 2418.

12496 J. Am. Chem. Soc., Vol. 122, No. 50, 2000
Darensbourg et al.
oxide. Although the homopolymerization process is the more
facile of the two, the copolymerization takes place in a
completely alternating manner, that is, with essentially no
polyether linkages. This is presumably due to the availability
of only one epoxide binding site at the zinc center of complex
1 and supports a dimeric structure for the active catalyst. This
latter conclusion is further sustained by the inactivity for
catalyzing these polymerization processes of the phosphine
adduct of 1, complex 5, where the epoxide binding site is
occupied by PCy3. In both polymerization processes, to afford
polycarbonates and polyethers utilizing complex 1 as catalyst,
the difluorophenoxide end group in the polymers was identified
by 19F NMR spectroscopy. The decrease in catalytic activity of
the [Zn(O-2,6-X2C6H3)2‚THF]2 complexes for the copolymer-
ization reaction as the electronegativity of the halogen substit-
uents decreases strongly supports epoxide activation as rate-
limiting for chain growth, as opposed to CO2 insertion.
Furthermore, the observation that homopolymerization of cy-
clohexene oxide to polyethers is much more facile than
copolymerization of cyclohexene oxide and CO2 leads to the
conclusion that epoxide ring-opening by a metal-bound alkoxide
functionality from the growing polymer chain is much faster
than by the corresponding carbonate functionality.
19F NMR spectrum for the polyether exhibits a broad group of
resonances centered at -126.8 ppm. The separation of the two
groups of resonances in Figure 7a suggests both CO2 insertion
(-127.3 ppm) and epoxide insertion (-126.8 ppm) as the initial
steps in the copolymerization process. This is to be contrasted
to the case of sterically encumbering substituents in the 2,6-
positions of the bisphenoxide zinc complexes, where initiation
by CO2 insertion at the metal center is prohibited.
Conclusions and Comments
The use of phenolate ligands substituted in the 2,6-positions
with sterically unhindered substituents, such as fluorines, has
provided dimeric bisphenoxides of zinc and cadmium of the
form [M(O-2,6-F2C6H3)2‚L]2. The metal centers, which are
coordinated by two bridging phenoxides and one terminal
phenoxide, possess distorted tetrahedral geometries with the
fourth ligand(L), THF or PCy3, accommodated in a trans
configuration. 31P NMR spectra of the zinc tricyclohexylphos-
phine derivative reveal no facile exchange of free and bound
phosphine in solution, whereas the cadmium analogue was
shown to undergo a rapid equilibration with free PCy3 in solution
via the intermediary of the Cd(O-2,6-F2C6H3)2(PCy3)2 complex,
the latter of which is stabilized relative to the dimer and free
PCy3 at low temperature.
Acknowledgment. Financial support from the National
Science Foundation (CHE 99-10342 and CHE 98-07975 for the
purchase of X-ray equipment) and the Robert A. Welch
Foundation is greatly appreciated. We thank Dr. Robert Taylor
for his expertise in obtaining the solid-state and solution 113Cd
NMR spectra. We are also grateful to DSM, The Netherlands,
for the DSC analysis.
Supporting Information Available: Complete details of the
X-ray diffraction study of 1-3 and 5. This material is available
free of charge via the Internet at http://pubs.acs.org.
Complex 1, [Zn(O-2,6-F2C6H3)2‚THF]2, was shown to be an
effective catalyst for the copolymerization of cyclohexene oxide
and carbon dioxide and the homopolymerization of cyclohexene
JA002855H