Mechanistic aspects of the copolymerization of CO2 and epoxides by soluble zinc bis(phenoxide) catalysts as revealed by their cadmium analogues

Article abstract of DOI:10.1021/ja9801487

A family of monomeric cadmium(II) phenoxides of the general formula Cd(O-2,6-R₂C₆H₃)₂(base)_2-3 [R = Ph, (t)Bu, Me; base = THF, THT, pyridine, propylene carbonate] has been synthesized. Determination o

Full text of DOI:10.1021/ja9801487

4690
J. Am. Chem. Soc. 1998, 120, 4690-4698
Mechanistic Aspects of the Copolymerization of CO2 and Epoxides
by Soluble Zinc Bis(phenoxide) Catalysts as Revealed by Their
Cadmium Analogues
D. J. Darensbourg,* S. A. Niezgoda, J. D. Draper, and J. H. Reibenspies
Contribution from the Department of Chemistry, Texas A&M UniVersity, P.O. Box 300012,
College Station, Texas 77842
ReceiVed January 13, 1998
Abstract: A family of monomeric cadmium(II) phenoxides of the general formula Cd(O-2,6-R2C6H3)2(base)2-3
t
[R ) Ph, Bu, Me; base ) THF, THT, pyridine, propylene carbonate] has been synthesized. Determination
of the solid-state structures of all complexes have uncovered a variety of geometries ranging from the expected
distorted tetrahedral to square planar to trigonal bipyramidal, depending upon the identity of the phenoxide
substituents and of the base ligands. By utilizing multinuclear NMR techniques available to cadmium,
particularly 113Cd NMR, it was possible to discover the identities of these complexes in solution. Reactivities
of these complexes with the small molecules CO2, COS, and CS2 were also carried out. While CO2 showed
no reactivity with the sterically encumbered Cd(II) bis(phenoxides), insertion was observed spectroscopically
to occur for COS and CS2. The product of CS2 insertion was isolated and characterized and found to be a
dimeric xanthate complex, [Cd(S2CO-2,6-Ph2C6H3)]2[µ-O-2,6-Ph2C6H3]2. These Cd(II) complexes are analogues
to Zn(II) bis(phenoxides), which are highly active catalysts for the copolymerization of CO2 with epoxides.
Although the Cd(II) bis(phenoxides) show only greatly reduced activity for this copolymerization reaction,
the results of solution studies and reactivities have proven instructive toward better understanding the mechanism
through which their Zn(II) counterparts work.
We have recently discovered that monomeric bis(phenoxides)
A general mechanism for the copolymerization process has
been proposed for zinc catalysts which have at least one anionic
oxygen-coordinating ligand, such as a carboxylate, an alkoxide,
or a phenoxide.4 As drawn in the skeletal mechanism in Scheme
1, initiation is believed to occur by coordination of the epoxide
to the electrophilic Zn center. The coordinated epoxide is ring-
opened through nucleophilic attack on the R-carbon by the
oxygen ligand, leading to an alkoxide species which can then
rapidly insert CO2. The propagation of this polymer chain
follows as the carbonate species can subsequently ring-open the
next coordinated epoxide, followed again by CO2 insertion.
Although we have illustrated in Scheme 1 that ring opening
occurs exclusively at the unsubstituted carbon center of the
epoxide, it has been demonstrated by NMR experiments that
ring opening takes place at either carbon atom.5
of zinc, Zn(O-2,6-R2C6H3)2(ether)2 [R ) Ph, tBu, iPr, Me; ether
) THF, diethyl ether], are excellent homogeneous catalysts for
the copolymerization and terpolymerization of CO2 with ep-
oxides (eq 1).1 These bis(phenoxide) zinc complexes are
(1)
monomeric and hence soluble in many organic solvents and
epoxides due to the use of the bulky substituents in the ortho
positions of the phenoxide ligands which prevent aggregation.2
High-molecular-weight copolymer can be obtained in moderate
to extremely good yields, depending upon the catalyst derivative
which is used. It has been possible to obtain full structural
characterization of all of these catalyst derivatives to determine
that all have roughly the same distorted-tertrahedral geometry
in the solid state.3 However, because of the absence of
spectroscopic tools employable for zinc complexes, it has been
difficult to elucidate the solution behavior of these complexes
pertinent to understanding the mechanism through which they
act and thus grasp the keys to optimization of the copolymer-
ization process.
To uncover experimental evidence consistent or inconsistent
with this mechanism for the Zn bis(phenoxide) catalyst system,
we have again turned to the use of cadmium analogues. The
113
use of cadmium(II) provides an accessible metal nucleus,
-
Cd, to more intimately examine the interaction and reactivity
of the ligands at the metal center through multinuclear NMR
techniques. Previously, we have been able to utilize Cd(II)
carboxylate model complexes which contained a coordinated
epoxide ligand to study the interaction of the epoxide with the
metal center and the ring opening of the coordinated epoxide
by the carboxylate ligand.6
(1) Darensbourg, D. J.; Holtcamp, M. W. Macromolecules 1995, 28,
7577.
(2) Geerts, R. L.; Huffman, J. C.; Caulton, K. G. Inorg. Chem. 1986,
25, 1803.
(3) Darensbourg, D. J.; Holtcamp, M. W.; Struck, G. E.; Zimmer, M.;
Niezgoda, S. A.; Rainey, R. P.; Robertson, J. B.; Draper, J. D.; Reibenspies,
J. H. Manuscript in preparation.
(4) (a) Soga, K.; Hyakkoku, K.; Ikeda, S. Makromol. Chem. 1978, 179,
2837. (b) Inoue, S. CHEMTECH 1976, 6, 588. (c) Kobayashi, M.; Tang,
Y.-L.; Tsuruta, T.; Inoue, S. Makromol. Chem. 1973, 169, 69. (d) Inoue,
S.; Koinuma, H.; Yokoo, Y.; Tsuruta, T. Makromol. Chem. 1971, 143, 97.
(5) Lednor, P. W.; Rol, N. C. J. Chem. Soc., Chem. Commun. 1985,
598.
S0002-7863(98)00148-6 CCC: $15.00 © 1998 American Chemical Society
Published on Web 04/29/1998

Copolymerization by Zinc Phenoxide Catalysts
J. Am. Chem. Soc., Vol. 120, No. 19, 1998 4691
for NMR experiments were purchased from Cambridge Isotope
Laboratories. Benzene-d6 was distilled over calcium hydride and stored
in the glovebox over molecular sieves. 1H NMR and 13C NMR spectra
were recorded on a Varian XL-200E superconducting high-resolution
spectrometer. 113Cd NMR spectra were recorded on a Varian XL-400E
superconducting high-resolution spectrometer operating at 88 MHz
using an external 0.1 M Cd(ClO4)2/D2O reference. Elemental analysis
were carried out by Canadian Microanalytical Services, Ltd.
Scheme 1
Synthesis of Cd(O-2,6-tBu2C6H3)2(THF)2 (1). Synthesis was
carried out in a manner similiar to the published procedure.6 Cd-
[N(SiMe3)2]2 (0.48 g, 1.11 mmol) in 2 mL of THF was stirred with a
solution of 2,6-di-tert-butylphenol (0.458 g, 2.22 mmol) in 2 mL of
THF. After the yellow solution was stirred for 2 h, the flask was placed
at -20 °C. Colorless crystals formed overnight. The supernate was
cannulated to a clean flask, and a second crop of crystals were isolated.
With both yields, after removal of the supernate, the crystals were
washed quickly with very cold THF and dried briefly under vacuum.
The combined yields totaled 0.327 g (44%) of pale yellow crystalline
solid. Anal. Calcd for C36H58O4Cd: C, 64.8; H, 8.76. Found: C,
63.19; H, 8.51. 1H NMR (C6D6): δ 1.25 [m, 4H, {THF}], 1.60 [s,
18H, {-CMe3}], 3.48 [m, 4H, {THF}], 6.78 [t, 1H, {4-H}], 7.28 [d,
2H, {3,5-H}].
Synthesis of Cd(O-2,6-tBu2C6H3)2(THT)2 (2). Cd[N(SiMe3)2]2
(0.20 g, 0.46 mmol) in 1 mL of THT was stirred while 2,6-di-tert-
butylphenol (0.19 g, 0.92 mmol) in 1 mL of THT was added. The
resulting yellow solution was stirred for 2 h, then placed at -20 °C
for 2 days while crystals of the product formed. The supernate was
removed from the crystals by cannula, and the crystals were dried in
vacuo to give 0.246 g (76%) of a yellow crystalline solid. Anal. Calcd
for C36H58O4Cd: C, 61.83; H, 8.36. Found: C, 60.47; H, 8.24. 1H
NMR (C6D6): δ 1.38 [4H, m, {THT}], 1.66 [18H, s, -tBu2}], 2.47
[4H, m, {THT}], 6.8 [1H, t, {4-C6H3}], 7.3 [2H, d, {3,5-C6H3}].
Synthesis of Cd(O-2,6-tBu2C6H3)2(pyridine)3 (3). A 2-mL pyridine
solution of 2,6-di-tert-butylphenol (0.19 g, 0.92 mmol) was added to a
solution of Cd[N(SiMe3)2]2 (0.20 g, 0.46 mmol) in 2 mL of pyridine.
After stirring for 1 h, the yellow solution was placed at -20 °C for 2
days. Pale yellow crystals formed and were isolated by removing the
supernate, washing quickly with cold fresh pyridine and drying under
vacuum to give 0.250 g (71%) of a yellow crystalline solid. Crystals
suitable for X-ray diffraction were grown by dissolving the yellow
crystals in toluene after removing the supernate and storing at -20 °C
for 2 days. Anal. Calcd for C43H57O2N3Cd: C, 67.0; H, 7.69; N, 5.53.
Found: C, 66.41; H, 7.48; N, 5.76. 1H NMR (C6D6): δ 1.66 [18H, s,
{-C-tBu}], 6.58 [3H, t, {C5H5N}], 6.88 [2H, t, {p-C6H3}], 6.9 [6H, t,
{C5H5N}], 7.42 [4H, d, {m-C6H3}], 8.53 [6H, d, {C5H5N}]. 13C{H}
NMR (C6D6): δ 31.7 {-CMe3}, 35.9 {-CMe3}, 114.5 {4-C6H3}, 124.3
{C5H5N}, 125.3 {3,5-C6H3}, 134.0 {C5H5N}, 139.0 {2,6-C6H3}, 150.4
{C5H5N}, 168.5 {ipso-C6H3}.
Herein, we have synthesized a family of Cd(II) bis(phenox-
ides) analogous to our zinc catalysts. The solid-state structures
of these complexes of the general formula Cd(O-2,6-R2C6H3)2-
(solvent)2-3 adopt a variety of solid-state conformations,
depending on the identity of the phenoxy substituents and the
coordinating base. One derivative, Cd(O-2,6-tBu2C6H3)2(THF)2
(1), previously reported by Buhro,7 had an unexpected square-
planar ligand arrangement which is considerably different from
its tetrahedral zinc analogue. Recently, we have reported the
structures of three more complexes of this family which each
took on a different solid-state geometry: Cd(O-2,6-tBu2C6H3)2-
(THT)2 (2) is another square-planar complex, Cd(O-2,6-t-
Bu2C6H3)2(py)3 (3) is trigonal bypyramidal, and Cd(O-2,6-
Ph2C6H3)2(THF)2 (4) has distorted tetrahedral geometry.8 The
syntheses and X-ray crystal structures of these and three more
derivatives are reported here, including the results of structural
redetermination of Buhro’s complex carried out at low temper-
ature. Nevertheless, it is important to note that all structural
parameters in complex 1 determined at 193 K are quite similar
to those previously reported at ambient temperature. Addition-
ally, the results of solution 1H, 13C, and 113Cd NMR experiments
to determine the identity of these species in solution and to study
the catalytic activity of these Cd(II) complexes are provided.
In particular, reactions with the small molecules CO2, COS, and
CS2 were carried out and a novel xanthate product formed from
CS2 insertion has been isolated and characterized.
Synthesis of Cd(O-2,6-Ph2C6H3)2(THF)2 (4). 2,6-Diphenylphenol
(0.455 g, 1.86 mmol) was dissolved in 4 mL of THF and added to a
solution of Cd[N(SiMe3)2]2 (0.40 g, 0.92 mmol) in 2 mL of THF. The
pale yellow solution stirred for 3 h at room temperature, then was
concentrated to less than 1 mL and placed at -20 °C. After 2-3 days,
colorless crystals formed. They were isolated by removing the
supernate and drying briefly under vacuum to give a yield of 0.360 g
(52%) of pale yellow powder. The crystals may be washed with very
cold THF to remove reaction byproducts, although the yield is variably
reduced due to the high solubility of 2 in THF. Anal. Calcd for
C44H42O4: C, 70.73; H, 5.67. Found: C, 68.79; H, 5.50. 1H NMR
(C6D6): δ 6.8-7.3 (br, 13H, 2,6-Ph2C6H3-), 3.59 (m, 4H, THF), 1.43
(m, 4H, THF). 13C{H} NMR (C6D6): δ 168.1 {ipso-C6O}, 125-132
{2,6-Ph2C6}, 67.9 {THF}, 25.9 {THF}.
Experimental Section
Methods and Materials. All complexes are extremely air- and
moisture-sensitive and were handled using Schlenk and glovebox
techniques under an argon atmosphere. Solvents were freshly distilled
before use except for tetrahydrothiophene (THT) and pyridine, which
were distilled over calcium hydride and stored over molecular sieves.
Phenols and propylene carbonate were purchased from Aldrich and
used as received. Cd[N(SiMe3)2]2 was synthesized and distilled
according to the literature procedure.9 All isotopically labeled chemicals
(6) (a) Darensbourg, D. J.; Niezgoda, S. A.; Holtcamp, M. W.; Draper,
J. D.; Reibenspies, J. H. Inorg. Chem. 1997, 36, 2424. (b) Darensbourg, D.
J.; Holtcamp, M. W.; Khandelwal, B.; Klausmeyer, K. K.; Reibenspies, J.
H. J. Am. Chem. Soc. 1995, 117, 538.
(7) Goel, S. C.; Chiang, M. Y.; Buhro, W. E. J. Am. Chem. Soc. 1990,
112, 6724.
(8) Darensbourg, D. J.; Niezgoda, S. A.; Reibenspies, J. H.; Draper, J.
D. Inorg. Chem. 1997, 36, 5686.
(9) Bu¨rger, H.; Sawodny, W.; Wannagat, U. J. Organomet. Chem. 1965,
3, 113.
Synthesis of Cd(O-2,6-Ph2C6H3)2(THT)2 (5). A 3-mL THT
solution of 2,6-diphenylphenol (0.228 g, 0.93 mmol) was added to a
neat sample of Cd[N(SiMe3)2]2 (0.20 g, 0.46 mmol), and the resulting
pale yellow solution was stirred for 1 h at room temperature. The
solution was concentrated to approximately half-volume and placed at
-20 °C for 2 days. Opaque crystals formed and were isolated by
removing the supernatant solution and drying under vacuum for 20

4692 J. Am. Chem. Soc., Vol. 120, No. 19, 1998
Table 1. Crystallographic Data for 1-6 and 8
Darensbourg et al.
1
2
3
4
5
6
8
formula
C36H58CdO4
667.22
monoclinic
P21/n
9.4957(10)
11.7841(13)
15.858(2)
C36H58CdO2S2
699.34
triclinic
P1h
9.251(2)
9.493(2)
11.516(2)
74.74(3)
88.92(3)
64.95(3)
883.4(3)
1
1.315
6.263
1.54178
2825
2639
C48H62CdN4O2
839.42
monoclinic
P21/c
19.249(5)
15.152(2)
18.085(3)
C44H42CdO4
747.18
triclinic
P1h
8.868(2)
11.066(2)
19.554(4)
78.38(3)
83.29(3)
73.51(3)
1798.4(6)
2
1.380
5.193
1.54178
5642
5322
C44H42CdO2S2
779.30
tetragonal
P4cc
16.299(2)
16.299(2)
15.497(3)
C36H54CdO8
727.19
monoclinic
P21/c
9.343(4)
22.236(8)
9.694(4)
C74H52O4S4Cd2
1356.82
orthorhombic
Pnna
17.471(4)
27.146(5)
15.246(3)
formula weight
cryst system
space group
a, Å
b, Å
c, Å
R, deg
â, deg
93.659(12)
117.46(2)
90
117.86(3)
90
γ, deg
V, Å3
1770.9(3)
2
4680(2)
4
1.191
0.505
0.71073
8512
8223
293(2)
7.44
4117.0(10)
4
1.257
0.665
0.71073
1422
1422
193(2)
8.61
1780.6(12)
2
1.356
0.661
0.71073
2027
1918
193(2)
6.67
7231(2)
4
1.246
6.165
1.54178
3559
Z
d (calcd), g/cm3
absorp coeff, mm-1
λ, Å
1.251
0.649
0.71073
3315
3115
193
no. of total reflns
no. of unique reflns
T, K
3559
163(2)
6.95
17.80
193(2)
5.86
13.14
163(2)
12.24
27.04
R,a %
6.41
17.30
Rw,a %
13.87
20.22
17.27
a R ) ∑||Fo| - |Fc||/∑Fo. Rw ) {[∑w(Fo - Fc2)2]/[∑w(Fo )2]}1/2
.
2
2
min; 0.352 g (98%) of the product was collected. Anal. Calcd for
C44H42O2S2Cd: C, 67.81; H, 5.43. Found: C, 65.63; H, 5.85. 1H NMR
(C6D6): δ 1.47 [4H, m, {THT}], 2.55 [4H, m, {THT}], 6.8-7.4 [13H,
m-br, {2,6-Ph2C6H3-}]. 13C{H} NMR (C6D6): δ 31.25 {THT}, 31.91
{THT}, 126-131 {2,6-Ph2C6H3}, 168.1 {ipso-C6H3}.
(C6H6): ν(CS2) 1152, 1058 cm-1; ν(13CS2) 1129, 1023 cm-1
.
13C NMR
(C6D6): δ 231.4 (O13CS2).
113Cd NMR Experiments. Samples of 1-6 were prepared in CD2-
Cl2, each approximately 0.05 M in Cd complex. Spectra were obtained
at 20 °C intervals between 20 and -87 °C. After the sample had
warmed to room temperature, an 8 mol excess of the respective
coordinating donor solvent was added, and another set of spectra were
collected. A 0.05 M sample of 7 was prepared in pyridine, and spectra
were collected only at 20 and -30 °C. Small molecule insertion
reactions were performed on similar samples in C6D6 or CD2Cl2 at room
temperature.
Synthesis of Cd(O-2,6-tBu2C6H3)2(propylene carbonate)2 (6). A
2-mL toluene solution of 0.19 g (0.92 mmol) of 2,6-di-tert-butylphenol
was added to a 2 mL of toluene solution of Cd[N(SiMe3)2]2 (0.20 g,
0.46 mmol). Slightly more than 2 equiv of (()-propylene carbonate
(0.1 mL, 1.16 mmol) was added to the solution, and the mixture was
stirred for 2 h. After concentrating the solution to half-volume, it was
placed at -20 °C for 2 days while colorless crystals of 7 formed. The
crystals were isolated upon removing the supernatant solution and
quickly washing the crystals 3 × 5 mL with hexanes while keeping
the flask cold. After drying for 10 min under vacuum, 0.177 g (53%)
of pale yellow crystals were obtained. Anal. Calcd for C36H54O8Cd:
C, 59.46; H, 7.48. Found: C, 57.75; H, 7.37. 1H NMR (C6D6): δ
0.48 [3H, d, {PC}], 1.36-1.56 [18H, d-br, {2,6-tBu-C6H3-}], 2.80
[1H, t, {PC}], 3.23 [1H, t, {PC}], 3.57 [1H, m, {PC}], 6.8 [1H, br,
{4-O-C6H3}], 7.2 [2H, br, {3,5-O-C6H3}]. 13C{H} NMR (C6D6): δ
18.57 {PC}, 31.45 {-CMe3}, 33.42 {-CMe3}, 69.9 {PC}, 72.9 {PC},
117 {4-C6H3}, 125 {3,5-C6H3}, 139 {2,6-C6H3}, 166 {ipso-C6H3}.
X-ray Crystallographic Study of 1-6 and 8. Crystal data and
details of the data collections are given in Table 1. Colorless block
crystals of 1-6 and 8 were mounted of a glass fiber with epoxy cement
at room temperature and cooled to 193 K (1, 3, 5, and 6) or 163 K (2,
4, and 8) in a N2 cold stream. Preliminary examination and data
collection were performed on a Siemens P4 X-ray diffractometer for
1, 3, 5, and 6 (Mo KR λ ) 0.710 73 Å radiation) and a Rigaku AFC5
X-ray diffractometer for 2, 4, and 8 (Cu KR λ ) 1.541 78 Å radiation.)
Cell parameters were calculated from the least-squares fitting of the
setting angles for 25 reflections. ω scans for several intense reflections
indicated acceptable crystal quality. Data were collected for 4.0° e
2θ e 50.0° for 1, 3, 5, and 6 and for 8.0° e 2θ e 120.0° for 2, 4, and
8. Three control reflections collected for every 97 reflections showed
no significant trends. Background measurements by stationary-crystal
and stationary-counter techniques were taken at the beginning and end
of each scan for half of the total scan time. Lorentz and polarization
corrections were applied to the total reflections for each complex, listed
in Table 1. A semiempirical absorption correction was applied to all
complexes. Totals of unique reflections for each complex (Table 1)
with |I| g 2.0sI were used in further calculations. Structures were
solved by direct methods [SHELXS, SHELXTL-PLUS program
package, Sheldrick (1993)]. Full-matrix least-squares anisotropic
refinement for all non-hydrogen atoms yielded R, Rw(F2), and S values
at convergence for all complexes, listed in Table 1. Hydrogen atoms
were placed idealized positions with isotropic thermal parameters fixed
at 0.08. Neutral-atom scattering factors and anomalous scattering
correction terms were taken from the International Tables for X-ray
Crystallography.10
Synthesis of [Cd(O-2,4,6-Me3C6H2)2]n (7). Cd[N(SiMe3)2]2 (0.52
g, 1.2 mmol) was stirred with 10 mL of THF. A solution of 2,4,6-
trimethylphenol (0.317 g, 2.6 mmol) in 5 mL of THF was added to the
cadmium amide solution, and white precipitate began forming im-
mediately. The mixture was stirred for 1 h, after which time the white
solid was collected on a glass frit and dried overnight under vacuum.
The yield, 0.445 g, of white powder is quantitative. Anal. Calcd for
C18H26O2Cd: C, 55.89; H, 6.77. Found: C, 55.22; H, 6.02. 1H NMR
(C5ND5): δ 2.58 [3H, s, {4-C6-CH3}], 2.79 [6H, s, {2,6-C6-(CH3)2}],
7.87 [2H, s, {3,5-C6H2}]. 13C NMR (C5ND5): δ 18.8 {-CH3}, 20.9
{-CH3}, 121.6 {2,6-C6H2}, 125.6 {3,5-C6H2}, 129.2 {4-C6H2}, 162.9
{ipso-C6H2}.
Synthesis of [Cd(S2CO-2,6-Ph2C6H3)]2(m-O-2,6-Ph2C6H3)2 (8) and
13CS2-Labeled Complex. A 20-mL benzene solution of complex 4
was prepared as above using 0.25 g (0.58 mmol) of Cd[N(SiMe3)2]2
and 0.228 g (1.16 mmol) of 2,6-diphenylphenol. Two equivalents of
CS2 (0.07 mL, 0.088 g) or 1 equiv of 13CS2 was added by syringe. The
solution was stirred overnight, then was filtered to remove any
precipitate. The supernatant solution was placed under vacuum to
remove the solvent and reaction byproducts, and a pale yellow
crystalline solid was obtained in 58% yield (0.180 g.) Alternatively,
crystals large enough to use for single-crystal X-ray diffraction formed
within a few days from the concentrated benzene solution. Anal. Calcd
for C74H52O4S4Cd2: C, 65.44; H, 3.86. Found: C, 65.16; H, 4.11. IR
Results and Discussion
The syntheses of complexes 1-5 are all carried out in a
similiar manner. Specifically, 2 equiv of 2,6-disubstituted
phenol is reacted with pure Cd[N(SiMe3)2]2 in the appropriate
(10) Ibers, J. A., Hamilton, W. C., Eds.; International Tables for X-ray
Crystallography; Kynoch Press: Birmingham, England, 1974; Vol. IV.

Copolymerization by Zinc Phenoxide Catalysts
J. Am. Chem. Soc., Vol. 120, No. 19, 1998 4693
Table 2. Selected Bond Distances (Å) and Bond Angles (deg) for
1-6a,b
donor solvent, as shown in eq 2 . Alternatively, complex 6 is
donor
Complex 1: Cd(O-2,6-tBu2C6H3)2(THF)2
Cd[N(SiMe3)2]2 + 2 2,6-R2C6H3OH solvent8
Cd(1)-O(1)
O(1)-C(1)
2.068(3)
1.341(5)
Cd(1)-O(2)
2.465(3)
R ) tBu (1, 2, 3, 6)
Ph (4, 5)
O(1)-Cd(1)-O(1A) 180.0
O(1)-Cd(1)-O(2)
O(1)-Cd(1)-O(2A) 89.47(11)
C(1)-O(1)-Cd(1)
126.0(2)
81.53(11) Cd(1)-O(2)-(C4H8 plane) 174.1
Cd(O-2,6-R2C6H3)2(donor)n + 2 HN(SiMe3)2 (2)
donor ) THF (1, 4) THT (2, 5), py (3), PC (6)
Complex 2: Cd(O-2,6-tBu2C6H3)2(THT)2
n ) 2 or 3
Cd(1)-O(1)
O(1)-C(6)
2.102(6)
Cd(1)-S(1)
2.768(2)
1.343(10)
formed by addition of slightly greater than 2 equiv of propylene
carbonate (PC) to a reaction solution of the two reagents (eq 2)
in toluene. The use of the bis(bis(trimethylsilylamide)) complex
of Cd as the starting material is the only route we have found
to cleanly form the bis(phenoxides), since metathesis reactions
utilizing the Cd(II) halides results in mixed halide-phenoxide
complexes. A color change from colorless to yellow or pale
yellow occurs immediately upon mixing the reagents, as the
exothermic reaction quickly takes place. The products are best
isolated through crystallization from a concentrated reaction
solution because precipitation or removal of all solvent under
vacuum will result in loss of the donor ligands. Unfortunately,
since the complexes are so soluble even at low temperature,
the yields of products obtained in this fashion are variable and
not always high, although it appears that this acid-base reaction
does proceed to completion. The donor ligands are somewhat
labile even in the solid state. Prolonged (longer than 20 min)
periods in vacuo will remove or partially remove them in most
cases, and decomposition will occur. Evidence of the loss of
the coordinated donor groups can be found in 1H NMR
integrations and leads to inexact results in elemental analyses
(Experimental Section).
The same methodology (eq 2) was employed for the prepara-
tion of a complex which had a less sterically bulky phenol, 2,4,6-
trimethylphenol, using THF as solvent. The product, [Cd(O-
Me3C6H2)2]n (7), is observed to precipitate immediately upon
mixing the reagents, regardless of the amount of solvent used.
Presumably an aggregate species with bridging phenoxides,11
7, is soluble only in pyridine and can be dissolved by addition
of approximately 4 mol of pyridine to the THF reaction mixture.
Crystals of 7 suitable for X-ray diffraction have not been
obtained thus far. It is expected that the complex crystallized
from a pyridine solution would adopt a tetrahedral geometry
with two coordinated pyridine molecules, similar to that which
has been defined for its Zn analogue, Zn(O-2,4,6-Me3C6H4)2-
(py)2.3 The smaller methyl groups on the phenoxide rings
should not impose linearity of these ligands; however, more
than two pyridines may be able to coordinate to the cadmium
center, as is found in complex 3.
O(1)-Cd(1)-O(1A) 180.0
O(1)-Cd(1)-S(1)
77.7(2)
125.3(5)
O(1)-Cd(1)-S(1A) 102.3(2) C(6)-O(1)-Cd(1)
Cd(1)-S(1)-(C4H8 plane) 125.6
Complex 3: Cd(O-2,6-tBu2C6H3)2(py)3
Cd(1)-O(1)
Cd(1)-N(1)
Cd(1)-N(3)
2.193(5)
2.342(7)
2.372(6)
Cd(1)-O(2)
Cd(1)-N(2)
O(1)-C(1)
2.190(5)
2.347(6)
1.314(9)
O(1)-Cd(1)-O(2)
O(1)-Cd(1)-N(2)
Cd(1)-O(1)-C(1)
156.1(2)
92.5(2)
165.0(5)
O(1)-Cd(1)-N(1)
O(1)-Cd(1)-N(3)
N(3)-Cd(1)-N(2)
102.4(2)
88.4(2)
169.7(2)
Complex 4: Cd(O-2,6-Ph2C6H3)2(THF)2
Cd(1)-O(1)
Cd(1)-O(3)
O(1)-C(1)
2.073(5)
2.320(6)
1.307(9)
Cd(1)-O(2)
Cd(1)-O(4)
2.074(5)
2.293(6)
O(1)-Cd(1)-O(2)
150.1(2) O(3)-Cd(1)-O(4)
92.4(2) O(1)-Cd(1)-O(4)
130.1(5) C(19)-O(2)-Cd(1)
83.1(3)
107.5(2)
120.0(4)
O(1)-Cd(1)-O(3)
C(1)-O(1)-Cd(1)
Cd(1)-O(3)-(C4H8 plane) 172.7
Cd(1)-O(4)-(C4H8 plane) 169.4
Complex 5: Cd(O-2,6-Ph2C6H3)2(THT)2
2.12(2)
Cd(1)-O(1)
C(1)-O(1)
Cd(1)-S(1)
2.634(9)
1.34(3)
O(1)-Cd(1)-O(1A)
O(1)-Cd(1)-S(1)
C(1)-O(1)-Cd(1)
93.8(11) S(1)-Cd(1)-S(1A)
157.5(5)
133(2)
84.3(4)
95.1(6)
O(1)-Cd(1)-S(1A)
Cd(1)-S(1)-(C4H8 plane) 114
Complex 6: Cd(O-2,6-tBu2C6H3)2(PC)2
Cd(1)-O(1)
C(1)-O(1)
2.038(8)
1.359(12)
Cd(1)-O(2)
O(2)-C(15)
2.393(8)
1.177(19)
O(1)-Cd(1)-O(1A)
O(1)-Cd(1)-O(2)
Cd(1)-O(1)-C(1)
180.0
83.7(3)
128.3(6)
O(2)-Cd(1)-O(2A)
O(1)-Cd(1)-O(2A)
Cd(1)-O(2)-C(15)
180.0
96.3(3)
128.4(11)
a Estimated standard deviations given in parentheses. b Symmetry-
generated atoms designated by (XA).
All complexes 1-6 have been characterized by X-ray
crystallography, and a list of selected bond lengths and angles
is provided in Table 2. Complexes 1 and 2 which contain the
same tert-butyl-substituted phenol but the different donor ligands
THF and THT, respectively, both adopt a crystallographically
imposed square-planar geometry about the metal center which
is very rare for Cd(II) complexes. A thermal ellipsoid repre-
sentation of 1 is shown in Figure 1. The Cd-O(phenoxide)
bond length in 1 of 2.058 Å is slightly shorter than that found
in 2, 2.102(6) Å. On the other hand, the Cd-O(THF) bond
length for the ether ligands at 2.465(3) Å is comparable to that
observed for the Cd-S(THT) bond if the difference in covalent
radii of oxygen and sulfur (0.3 Å) is taken into account.
Figure 1. Thermal ellipsoid representations of Cd(O-2,6-tBu2C6H3)2-
(THF)2 (1) (50% probability).
Apparently, neither THF nor THT is a strong enough donor
ligand to cause any distortion in the linear O(phenol)-Cd-
O(phenol) fragment. However, when pyridine is used as the
(11) (a) Mehrotra, R. C. AdV. Inorg. Chem. Radiochem. 1983, 26, 269.
(b) Melhotra, K. C.; Martin, R. L. J. Organomet. Chem. 1986, 239, 159.

4694 J. Am. Chem. Soc., Vol. 120, No. 19, 1998
Darensbourg et al.
Figure 2. Thermal ellipsoid representation of Cd(O-2,6-Ph2C6H3)2-
(THT)2 (5) (50% probability).
donor with the same phenoxy ligands, distortion from square
planar geometry is attained by the coordination of a third
pyridine molecule. Complex 3, Cd(O-2,6-tBu2C6H3)2(py)3,
displays a solid-state trigonal bipyramidal structure with two
axial pyridines and phenoxides in equatorial positions.8 The
average Cd-O(phenoxide) bond distance is 2.192(5) Å, and
the angle, O(1)-Cd-O(2), between the phenoxide ligands is
156.1(2)°. Solution studies addressed in the next section suggest
that complex 3 forms in a stepwise process, initially affording
either a square-planar or distorted tetrahedral bis(pyridine)
derivative, which can subsequently bind another pyridine base
at low temperatures in solution or in the solid state.
Figure 3. Thermal ellipsoid representation of Cd(O-2,6-tBu2C6H3)2-
(PC)2 (6) (50% probability) with inset of ball-and-stick drawing.
Finally, as we were exploring the effect of donor ligands on
coordination number and geometry, we recalled that cyclic
carbonates are also excellent donor solvents. Complexes of our
catalysts with this ligand and the study of its coordination ability
would be particularly interesting since cyclic carbonates are the
byproduct, or in some cases the only product, of CO2 and
epoxide coupling with the Zn and Cd catalysts. Strong
coordination could mean that these donors have the potential
to act as inhibitors in the polymerization system. Cd(O-2,6-t-
Bu2C6H3)2(propylene carbonate)2 was synthesized in toluene
solution with addition of a slight excess of 2 equiv of the high-
boiling donor base. The structure of this complex, as shown
in Figure 3, again has a crystallographically imposed square-
planar geometry about the metal center. The Cd-O(phenoxide)
bond distance of 2.038(8) Å is slightly shorter than that same
bond in 1; however, the Cd-O(PC) distance of 2.393(8) Å is
over 0.1 Å shorter than the Cd-O(THF) distance in 1. All
bond angles in 6 are comparable with those in the other square-
planar complexes. The Zn analogue to this complex has recently
been prepared and characterized in our laboratories, and it adopts
distorted-tetrahedral geometry and has a Zn-O(PC) bond
distance of 2.071(11) Å.3 The carbonyl bond length of the
coordinated propylene carbonate in 6, 1.177(19) Å, may be
compared with that in the Zn analogue, 1.20(2) Å, and with
that reported for free ethylene carbonate, 1.15 Å,12 to assess
the degree of weakening of the double bond upon coordination.
As anticipated, distortion from the unique square-planar
geometry is also obtained when the ortho groups on the phenoxy
ligands are sterically smaller and less basic than the bulky
electron-donating tert-butyl. The 2,6-diphenylphenoxide deriva-
tive of complex 1 has been found to have the distorted
tetrahedral geometry expected for a four-coordinate d10 metal
complex, similar to its zinc analogue.3 Cd(O-2,6-Ph2C6H3)2-
(THF)2 (4) and its THT derivative (5) shown in Figure 2 have
been found to have similar Cd-phenoxy and Cd-donor bond
lengths but largely different bond angles around the metal center.
The Cd-O(phenoxide) bond length of 2.12(2) Å in 5 is
approximately 0.05 Å longer than those in 4, but the angle
between these two phenoxide ligands is drastically smaller in 5
at 93.8(11)° than that in 4 of 150.1(2)°. The Cd-S(THT) bond
length, 2.634(9) Å, is comparable to the Cd-O(THF) bond
distance, again once the difference in ionic radii is taken into
account, and the angle between the two donor ligands, 84.3-
(4)°, is also very similar. The difference between the orienta-
tions of the phenoxide ligands with respect to the cadmium
center in 4 and 5 probably results from steric congestion about
the cadmium center in the latter derivative. That is, the sulfur
donor ligands in complex 5 are pyramidal (angle defined by
the Cd-S vector and the midpoint of the SC2 plane is 114.1°),
thereby causing steric interactions between the phenoxide
ligands and the heterocyclic ring systems. This steric repulsion
leads to a closing of the O(1)-Cd-O(1A) angle. On the other
hand, the THF donor ligands are more accommodating of the
bulky phenoxide groups in this d10 metal complex and adopt a
more linear arrangement: Cd-O(1)-[OC2 plane] and Cd-
O(2)-[OC2 plane] are 164.4 and 172.7°, respectively. Similar
ether and thioether binding parameters are seen in the square-
planar derivatives, 1 and 2, where the corresponding angles are
174.1 and 125.6°.
Solution Studies. In Caulton’s previous report on the soluble
zinc(II) phenoxide complexes, it was noted that the THF ligands,
while coordinated in the solid-state, are labile in solution. On
the basis of 1H NMR evidence of Zn(O-2,4,6-tBu3C6H2)2(THF)2
in benzene solutions, it was believed that only one of the THF
ligands was dissociating, leaving a three-coordinate, soluble Zn
bis(alkoxide) species that could act as a catalyst.2,13 The
experiments discussed here were done to examine the coordina-
(12) Brown, C. J. Acta Crystallogr. 1954, 7, 92.

Copolymerization by Zinc Phenoxide Catalysts
J. Am. Chem. Soc., Vol. 120, No. 19, 1998 4695
Table 3. 1H Chemical Shifts of Donor Solvents in C6D6 Solutions
at 20 °Ca,b
Table 4. 113Cd NMR Chemical Shifts Obtained in CD2Cl2
Solutions (referenced to 0.1 M Cd(ClO4)2 in D2O)
chemical shifts
(â, R, ppm)
difference from
free (ppm)
core atoms
in structure
complex (donor solvent)
complex
δ (ppm) at 20 °C
d (ppm) -80 °C
1 (THF)
2 (THT)
3 (py)
4 (THF)
5 (THT)
6 (PC)
1.25, 3.48
1.38, 2.47
6.58, 6.9, 8.53
1.43, 3.58
1.47, 2.55
0.18, 0.07
1
2
3
4
5
6
7
O4
-2
144
109
73
21
221
50
53
254
-17
0.14, 0.10
0.15, 0.15, 0.01
0, 0.03
O2S2
O2N3
O4
O2S2
O4
0.05, 0.02
76
0.48, 2.80, 3.23, 3.57 0.23, 0.33, 0.38, 0.36
5
Zn(2,6-tBu2C6H3)2(THF)2 1.13, 3.40
Zn(2,6-Ph2C6H3)2(THF)2 1.19, 3.11
0.30, 0.15
0.24, 0.44
O2Nx
113a
a Complex 7 was run in a C6D5N solution.
a The chemical shift difference from free solvent in benzene is also
given. b Free solvents in benzene solution: THF, δ 1.43, 3.55; THT, δ
1.52, 2.57; py, δ 6.73, 7.05, 8.54; PC δ 0.71, 3.13, 3.61, 3.93.
which distorts an otherwise symmetric square-planar complex.
The tert-butyl signal is similarly split in the room-temperature
spectrum of 6, which must be a consequence of coordination
(or equilibrium shifted largely toward coordination) of the
unsymmetric PC ligands, whereas warming the sample leads
to a single peak, presumably as PC dissociates. This splitting
is more pronounced in CD2Cl2 solution as the solvent induces
larger chemical shift differences between the two peaks.
Coordination of THF in 1, THT in 2, or two pyridine molecules
in 3, which is suggested by the large shifts from free solvent,
would not be expected to cause this splitting due to the
symmetry of the ligands and the final complex.
The data from the 113Cd NMR have given a better picture of
the cadmium coordination sphere. Table 4 lists the chemical
shifts of Cd complexes in CD2Cl2 solutions. These are the first
reported for Cd phenoxides or alkoxides in solution.8 The only
other example of a four-coordinate Cd complex with only O
donors for which a solution 113Cd NMR shift is reported is the
tetrahedral phosphine oxide complex, [Cd(OP(C6H11)3)4][AsF6]2,
which displayed a single peak at 94 ppm in SO2 solution.14 Data
in the literature for Cd-O complexes have been mainly from
studies of either oxyanion salts of Cd2+ in aqueous solutions or
the metal complexed by large chelating agents. In most of these
studies, the Cd complexes are presumed to be of octahedral or
higher coordination.15
Scheme 2
tion of the cadmium analogues to the donor bases in solution
with the aid of 113Cd NMR.
All complexes 1-7 were first studied by 1H NMR in benzene
or methylene chloride solutions. Even at room temperature,
the spectra exhibit chemical shifts for the donor base ligands
which are located upfield from that expected for the free solvent
in solution. Table 3 lists the chemical shifts for the base ligands
of these complexes in benzene-d6 solutions and the difference
(parts per million) from the free solvent in benzene.
There are a few trends from Table 3 to which attention should
be given. First, the chemical shift difference of the protons â
to the heteroatom is in almost all cases greater than that for the
R-protons. It would be expected that, as coordination to a metal
occurs, the atoms closest to the donor atom would experience
the greater change in electronic environment. Second, the shift
differences for Zn complexes are larger than those for the Cd
analogues, which would be anticipated as Zn is a better Lewis
acid. Third, there is practically no chemical shift difference
for complexes 4 and 5, both of which contain the 2,6-
diphenylphenoxide ligands, while those complexes which have
tert-butyl- substituted phenoxides (1, 2, 3, and 6) seem to be
interacting with the donor solvents.
It is not possible to ascertain from the ambient temperature
spectra if these peak positions of the bases are representative
of clearly coordinate or uncoordinated donor molecules or rapid
exchange between free and coordinated solvent. An equilibrium
between free and bound solvent could be imagined to exist in
one of two ways, as illustrated in Scheme 2.
To investigate the possibilities, 1H NMR spectra of some of
these complexes at variable temperatures were obtained in
benzene (higher temperature) or methylene chloride (lower
temperature) solutions. When a CD2Cl2 solution of 2 was
cooled to -87 °C, all peaks, phenoxide and base, show an
additional upfield shift of only 0.1 ppm and the peaks broaden
slightly. Alternatively, the low-temperature spectra of complex
3 is more interesting because, once below -80 °C, the peak
representative of the tert-butyl groups and the pyridine signals
each split into two broad peaks. This could be due to the
coordination of a third pyridine as the equilibrium is slowed,
The room-temperature peak positions suggest that complexes
4 and 5 have the same structure in solution, and thus, that there
is no interaction with their donor ligands. The complexes 3
and 7 are also believed to have a similar coordination sphere at
ambient temperature, based on the peak positions, most probably
that of a bis(phenoxide) with two coordinated pyridine ligands.
That the peak positions for complexes 1, 2, 3, and 6 are quite
different despite having the same phenoxide ligands is evidence
1
consistent with the H NMR data for the interaction of THF,
THT, pyridine, and PC, respectively, with the metal center. It
is interesting to note the difference in the low-temperature peak
positions for complexes which contain the same core atoms but
vary in the substituents on the phenoxides. Whether this stems
from the basicity of the phenoxide or the geometry of the
complex, or a combination of both, is not discernible from these
data alone.
It is most likely that the coordination of the donor solvents
occurs as an equilibrium process. The spectra obtained at
intervals between 20 and -80 °C for 2 and 5, illustrated in
Figure 4, are representative of the two types of substituted
phenoxide complexes. In the series of spectra for 5, a second
(14) Dean, P. A. W. Can. J. Chem. 1981, 59, 3221.
(15) (a) Armitage, I. M.; Pajer, R. T.; Schoot Uiterkamp, A. J. M.;
Chlebowski, J. F.; Coleman, J. E. J. Am. Chem. Soc. 1976, 98, 5710. (b)
Summers, M. F. Coord. Chem. ReV. 1988, 86, 43 and references therein.
(c) Ellis, P. D. In The Multinuclear Approach to NMR Spectroscopy; NATO
ASI Series C103; Lambert, J. B., Riddell, F. G., Eds.; D. Reidel Pub.:
Boston, MA, 1983; p 457 and references therein.
1
(13) Temperature-dependent H NMR studies from our laboratories on
the Zn(O-2,6-Ph2C6H3)2(THF)2 complex in CD2Cl2 have clearly demon-
strated that the THF ligands are dissociated at ambient temperature, with
exchange between free and bound THF being slowed below -85 °C.

4696 J. Am. Chem. Soc., Vol. 120, No. 19, 1998
Darensbourg et al.
Figure 4. 113Cd NMR spectra for complexes 2 and 5 at variable temperatures. Spectra for (a) complex 2 and (b) complex 5.
peak is seen to emerge as the temperature is lowered while the
intensity of the original peak decreases. This does imply that
the conversion from the 2-coordinate complex to the 4-coor-
dinate complex proceeds with no intermediate. Because two
signals can be seen in the spectra for 4 and 5, it follows that
the rate of exchange between coordinated and noncoordinated
solvent must be slower than 1.6 × 104 s-1 for 5 and 1.8 × 103
s-1 for 4. The observation of only one signal for 2 and its
analogues, along with the position of the peak, implies that the
equilibrium is shifted well toward the 4-coordinate complex even
at room temperature and that the exchange rate is very fast,
leading to only a single averaged signal.
Insertion Reactions. Insertion of CO2, COS, and CS2 into
M-O bonds is well-known to occur rapidly,16 so insertions of
the these small molecules into Cd-phenoxide bonds have been
examined by 13C and 113Cd NMR. This is of interest also to
determine if CO2 insertion occurs prior to epoxide ring opening
in the mechanistic pathway for the copolymerization. When
13C-labeled CO2 was bubbled into an NMR samples of dichlo-
romethane-d2 solutions of 4, no reaction was observed visibly
or through its 13C NMR spectrum. The spectrum showed only
the intense signal at 125 ppm for free 13CO2, even after 15 h,
by which time, insoluble white precipitate had begun to form
in the NMR tube. It was presumed that CO2 insertion did not
occur immediately because of the steric hindrance of the phenyl
groups ortho to the Cd-phenoxide bond16 but that, over time,
insertion may have taken place through some type of decon-
structive pathway, leading to insoluble polymeric carbonate
compounds and phenol. The insertion of COS into complex 4
was then tried, as the formation of a strong Cd-S bond derived
from a weaker CdS bond might facilitate the reaction to give
a thiocarbonate complex. The reaction was followed by 113Cd
NMR. Within minutes of bubbling COS through the sample, a
113Cd resonance was observed at 131 ppm, downfield from that
of 4 at 76 ppm, corresponding to the formation of a Cd-S bond.
For comparison, an identical reaction was tried with 2 equiv of
CS2 added to a sample of 4. The new 113Cd NMR signal was
at 146 ppm for the CS2-insertion product. Possibly, interaction
of the Cd center with the distal S atom could produce a more
deshielded chemical shift relative to that of the COS product;
however, the two products should have comparable structures.
The product of CS2 insertion into complex 4 was isolated
and characterized crystallographically. Although the data were
of poor quality and the structure was highly disordered, we can
determine that insertion occurs only at one M-O bond, as
shown in the drawing in Figure 5. The crystals of [Cd(S2CO-
2,6-Ph2C6H3)]2[µ-(O-2,6-Ph2C6H3)2] (8) were grown from ben-
zene solution; thus, with no donor ligands present, dimer
formation is observed. Infrared and 13C NMR spectroscopies
data were also collected to verify that the insertion reaction
yields this complex in solution. The CS2 stretches in the IR
spectrum were difficult to discern, so the complex was
synthesized again using 13CS2. The IR peaks at 1152 and 1058
cm-1 (∆ν ) 94 cm-1) in the spectrum of the nonlabeled
complex were observed to shift to 1129 and 1023 cm-1 in the
spectrum of the 13C-labeled complex (see Figure 6.) The small
difference between the asymmetric and symmetric ν(CS2)
stretching frequencies is indicative of a chelating coordination17
of the dithiocarbonate unit which is found in the solid-state
structure.
The 13C-labeled complex was also used to obtain the 13C
(16) Darensbourg, D. J.; Mueller, B. L.; Bischoff, C. J.; Chojnacki, S.
S.; Reibenspies, J. H. Inorg. Chem. 1991, 30, 2418.
NMR peak position for 8 at 231 ppm for the dithiocarbonate’s

Copolymerization by Zinc Phenoxide Catalysts
J. Am. Chem. Soc., Vol. 120, No. 19, 1998 4697
may be beyond a simple insertion product, perhaps including
bridging thiocarbonate ligands. Additional studies are needed
to better characterize the nature of this species.
Catalytic Activity. The cadmium complex 7 was examined
for its activity as a catalyst for the copolymerization of
cyclohexene oxide with CO2. This complex was chosen because
its Zn analogue has been found to be the most active of our Zn
bis(phenoxide) catalysts.3 The golden-colored solution obtained
after the high-pressure reaction was analyzed by IR spectros-
copy. The spectrum contained a peak in the carbonate stretching
region at 1750 cm-1 indicative of polycarbonate linkages.
Removal of the epoxide solvent yielded a sticky, transparent
polymer which is soluble in methanol, an indication that the
product is of low molecular weight. The starting material Cd-
[N(SiMe3)2]2 was also tried as a catalyst because it had been
found to rapidly homopolymerize cyclohexene oxide. The
product solution obtained after a run identical to the previous
one again exhibits an IR peak at 1750 cm-1 and the isolated
product is a sticky, methanol-soluble polymer film. Neverthe-
less, the Cd complex clearly catalyzed formation of polymer
product, however, with a greatly reduced activity relative to its
zinc analogue. However, the concentration of Cd used in these
catalytic runs was three times the usual amount of Zn used.
When a run was executed using 7 at a Cd concentration (0.034
g with 10 mL of epoxide) which was the same as the usual Zn
concentration, very little polymer was formed. It is of impor-
tance to note here that, although the cadmium complexes
containing the bulky phenoxide ligands (e.g., complex 4) do
not undergo insertion reactions with CO2, these derivatives
slowly catalyze the homopolymerization of epoxides. This
observation is a consequence of the fact that CO2 insertion
requires interaction with the sterically blocked lone pairs on
the phenoxide ligand, whereas the epoxide ring-opening process
involves interaction of the epoxide at the metal center, much
like the COS and CS2 insertion processes.
Figure 5. Ball-and-stick drawing of complex 8.
Figure 6. Infrared spectra for (a) CS2 insertion and (b) 13CS2 insertion
into complex 4 in benzene solution.
carbon atom. Very little 13C NMR data are available for
comparison for complexes of CS2 which might allow inferences
to structure based on peak position. Products of simple CS2
insertion into zero-valent metal-aryl oxide bonds have been
reported to display peaks in their 13C NMR spectra around 200
ppm;18 however, a more recent report of dithiocarbonate
complexes of tellurium(IV) has listed peak positions further
downfield for several complexes with anisobidentate dithiocar-
bonate ligands. Furthermore, the -S2COR peak for the
potassium salts of the ligands was seen at 233 pm and considered
Conclusions
Monomeric Cd aryloxides, Cd(O-2,6-R2C6H3)2(donor)2, are
easily synthesized by utilizing phenoxides which contain
sterically bulky groups in the ortho ring positions, a tactic which
has been found to work for many metals with aryl-chalco-
genolate ligands. The family of complexes discussed within
start with the same basic reaction but lead to a variety of Cd
complexes with coordination modes reflecting the much richer
coordination chemistry of cadmium as compared with zinc. With
the characterization of several square-planar complexes, we have
shown that while Buhro’s complex is unusual it is not unique.
As Buhro had suggested, the structure of these complexes
appears to be controlled by the coordination of the sterically
bulky phenoxide ligands.7 When the smaller, less-basic 2,6-
diphenylphenoxide is employed, distorted coordination is ob-
tained with donors of varying strength, THF and THT, but the
very bulky tert-butyl-substituted phenoxide complexes are not
altered in structure for the different base ligands. Coordination
of two donor molecules should be able to induce at least a small
amount of distortion to the linear segment to give distorted-
tetrahedral complexes,20 as is observed for zinc with the same
phenoxide ligands. However, it appears that the only way
distortion will occur with the 2,6-di-tert-butyl phenoxy ligands
is if the donor is strong enough, like pyridine, to coordinate
additional molecules and, thus, override the tendency of the
bulky phenoxides to be linear.
to be representative of complexes which have bidentate ligands.19
An NMR-scale reaction of complex 2 in benzene with
13
-
CO2 and with 13CS2 was also completed using 13C NMR to
monitor the reactions. Again, no insertion was observed for
13CO2. The reaction did occur for 13CS2, with the product peak
occurring at 234 ppm, comparable to that observed for insertion
into 4. The reaction of COS with 2 was followed by 113Cd
NMR since the 13C-labeled reagent is not readily available. A
sample of complex 9 in C6D6 displays a room-temperature signal
at 148 ppm. The spectrum obtained within minutes of bubbling
COS through this sample displayed a peak at 431 ppm which
would be consistent with at least one M-O insertion with
ligation through the S, causing the large downfield shift. This
chemical shift is much farther downfield than that observed for
the reaction of COS with 2, and the structure of this complex
(17) Nakamoto, K. Infrared and Raman Spectra of Inorganic and
Coordination Compounds, 4th ed.; John Wiley & Sons: New York, 1986;
pp 232, 343.
(18) Darensbourg, D. J.; Sanchez, K. M.; Reibenspies, J. H.; Rheingold,
A. L. J. Am. Chem. Soc. 1989, 111, 7094.
(19) Bailey, J. H. E.; Drake, J. E.; Khasrou, L. N.; Yang, J. Inorg. Chem.
1995, 34, 124.
(20) Haaland, A. Angew. Chem., Int. Ed. Engl. 1989, 28, 992.

4698 J. Am. Chem. Soc., Vol. 120, No. 19, 1998
Darensbourg et al.
The solution studies of these complexes indicate that there
is a difference in the ability of these complexes to coordinate
bases depending on the basicity of the phenoxy ligands, with
the less basic, phenyl-substituted phenoxides surprisingly show-
ing little or no interaction at room temperature. The 113Cd NMR
results for these complexes are the first reported for monomeric
phenoxides or alkoxides. The only trend which can be noted
is the deshielding observed for complexes with the phenyl-
substituted phenoxide as compared to their tert-butyl-substituted
counterparts. This could be due to either the change in geometry
of the complexes from tetrahedral to square planar or could be
caused by ring-current effects.
The insertion reactions of CS2 and COS and the lack of
reactivity of CO2 with these complexes have been very
enlightening. While the steric bulk of the subtituents of the
phenoxide ligands aids in formation of soluble monomers, it
apparently will not allow insertion of CO2.21 This means that
ring opening and insertion of at least one epoxide into the
metal-phenoxide bond would have to occur first to remove
the steric bulk from around the metal center, after which CO2
insertion should easily occur as depicted in Scheme 3. These
findings are totally consistent with earlier studies of tungsten
model complexes involving CO2, COS, and CS2 insertion
reactions into W-O bonds containing sterically bulky phenox-
ides.16
Scheme 3
cyclohexene oxide to give low-molecular-weight polymers.
Cadmium is expected to have decreased activity as compared
with zinc; however, it is much easier to propose comparisons
between the Cd models and the Zn catalysts when they show
some similarities in their catalytic function.
Acknowledgment. Financial support of this research by the
National Science Foundation (Grant 96-15866) is greatly
appreciated. S.A.N. thanks the Department of Education for a
GAANN fellowship. We are again grateful for helpful discus-
sions with Professor W. E. Buhro.
Supporting Information Available: Ball-and-stick drawings
and tables of anisotropic thermal parameters, bond lengths, and
bond angles for complexes 1, 5, 6, and 8 (26 pages, print/PDF).
See any current masthead page for ordering information and
Web access instructions.
Finally, it is notable that these cadmium phenoxides are
catalytically active toward the copolymerization of CO2 with
(21) Nevertheless, complex 7, which contains sterically less encumbered
phenoxide ligands, most likely readily inserts CO2 as has been demonstrated
for its zinc analogue dissolved in pyridine.3.
JA9801487