Catalytic activity of a series of Zn(II) phenoxides for the copolymerization of epoxides and carbon dioxide

Article abstract of DOI:10.1021/ja9826284

A series of zinc phenoxides of the general formula (2,6- R₂C₆H₃O)₂Zn(base)₂ [R = Ph, (t)Bu, (i)Pr, base = Et₂O, THF, or propylene carbonate] and (2,4,6-Me₃C₆H₂O)₂Zn(pyridine)₂ have been synthesized and characterized in the solid state by X-ray crystallography. All complexes crystallized as four-coordinate monomers with highly distorted tetrahedral geometry about the zinc center. The angles between the two sterically encumbering phenoxide ligands were found to be significantly more obtuse than the corresponding angles between the two smaller neutral base ligands, having average values of 140°and 95°, respectively. In a noninteracting solvent such as benzene or methylene chloride at ambient temperature, the ancillary base ligands are extensively dissociated from the zinc center, with the degree of dissociation being dependent on the base as well as the substituents on the phenolate ligands. That is, stronger ligand binding was found in zinc centers containing electron-donating tert-butyl substituents as opposed to electron-withdrawing phenyl substituents. In all instances, the order of ligand binding was pyridine > THF > epoxides. These bis(phenoxide) derivatives of zinc were shown to be very effective catalysts for the copolymerization of cyclohexene oxide and CO₂ in the absence of strongly coordinating solvents, to afford high-molecular-weight polycarbonate (M(w) ranging from 45 x 10³ to 173 x 10³ Da) with low levels of polyether linkages. However, under similar conditions, these zinc complexes only coupled propylene oxide and CO₂ to produce cyclic propylene carbonate. Nevertheless, these bis (phenoxide) derivatives of zinc were competent at terpolymerization of cyclohexene oxide/propylene oxide/CO₂ with little cyclic propylene carbonate formation at low propylene oxide loadings. While CO₂ showed no reactivity with the sterically encumbered zinc bis(phenoxides), e.g., (2,6-di-tert-butylphenoxide)₂Zn(pyridine)₂, it rapidly inserted into one of the Zn-O bonds of the less crowded (2,4,6- trimethylphenoxide)₂Zn(pyridine)₂ to provide the corresponding aryl carbonate zinc derivative. At the game time, both sterically hindered and sterically nonhindered phenoxide derivatives of zinc served to ring-open epoxide, i.e., were effective catalysts for the homopolymerization of epoxide to polyethers. The relevance of these reactivity patterns to the initiation step of the copolymerization process involving these monomeric zinc complexes is discussed.

Full text of DOI:10.1021/ja9826284

J. Am. Chem. Soc. 1999, 121, 107-116
107
Catalytic Activity of a Series of Zn(II) Phenoxides for the
Copolymerization of Epoxides and Carbon Dioxide
Donald J. Darensbourg,* Matthew W. Holtcamp, Ginette E. Struck, Marc S. Zimmer,
Sharon A. Niezgoda, Patrick Rainey, Jeffrey B. Robertson, Jennifer D. Draper, and
Joseph H. Reibenspies
Contribution from the Department of Chemistry, Texas A&M UniVersity, P.O. Box 30012,
College Station, Texas 77842
ReceiVed July 24, 1998
t
i
Abstract: A series of zinc phenoxides of the general formula (2,6-R2C6H3O)2Zn(base)2 [R ) Ph, Bu, Pr,
base ) Et2O, THF, or propylene carbonate] and (2,4,6-Me3C6H2O)2Zn(pyridine)2 have been synthesized and
characterized in the solid state by X-ray crystallography. All complexes crystallized as four-coordinate monomers
with highly distorted tetrahedral geometry about the zinc center. The angles between the two sterically
encumbering phenoxide ligands were found to be significantly more obtuse than the corresponding angles
between the two smaller neutral base ligands, having average values of 140° and 95°, respectively. In a
noninteracting solvent such as benzene or methylene chloride at ambient temperature, the ancillary base ligands
are extensively dissociated from the zinc center, with the degree of dissociation being dependent on the base
as well as the substituents on the phenolate ligands. That is, stronger ligand binding was found in zinc centers
containing electron-donating tert-butyl substituents as opposed to electron-withdrawing phenyl substituents.
In all instances, the order of ligand binding was pyridine > THF > epoxides. These bis(phenoxide) derivatives
of zinc were shown to be very effective catalysts for the copolymerization of cyclohexene oxide and CO2 in
the absence of strongly coordinating solvents, to afford high-molecular-weight polycarbonate (Mw ranging
from 45 × 103 to 173 × 103 Da) with low levels of polyether linkages. However, under similar conditions,
these zinc complexes only coupled propylene oxide and CO2 to produce cyclic propylene carbonate. Nevertheless,
these bis (phenoxide) derivatives of zinc were competent at terpolymerization of cyclohexene oxide/propylene
oxide/CO2 with little cyclic propylene carbonate formation at low propylene oxide loadings. While CO2 showed
no reactivity with the sterically encumbered zinc bis(phenoxides), e.g., (2,6-di-tert-butylphenoxide)2Zn(pyridine)2,
it rapidly inserted into one of the Zn-O bonds of the less crowded (2,4,6-trimethylphenoxide)2Zn(pyridine)2
to provide the corresponding aryl carbonate zinc derivative. At the same time, both sterically hindered and
sterically nonhindered phenoxide derivatives of zinc served to ring-open epoxide, i.e., were effective catalysts
for the homopolymerization of epoxide to polyethers. The relevance of these reactivity patterns to the initiation
step of the copolymerization process involving these monomeric zinc complexes is discussed.
Introduction
however, this method is plagued by low catalytic activity, thus
requiring removal of the metal from the synthesized polymer
by an acid wash.5 Other catalyst systems composed of double
The synthesis of select polycarbonates and/or cyclic carbon-
ates via metal-catalyzed coupling reactions of carbon dioxide
and epoxides (eq 1) has been extensively investigated as a
potential pathway for the effective utilization of carbon dioxide.
(2) A brief review covering this and related work has appeared: Inoue,
S. Carbon Dioxide as a Source of Carbon; Aresta, M., Forti, G., Eds.; Reidel
Publishing Co.: Dordrecht, 1987; p 331. See also an earlier, more
comprehensive review: Rokicki, A.; Kuran, W. J. Macromol. Sci., ReV.
Macromol. Chem. 1981, C21, 135. An excellent current assessment of the
status of catalysts for copolymerization processes utilizing CO2 as a
monomer has been published: Super, M. S.; Beckman, E. J. Trends Polym.
Sci. 1997, 5, 236.
(3) (a) Inoue, S.; Koinuma, H.; Tsuruta, T. Makromol. Chem. 1969, 130,
210. (b) Inoue, S.; Koinuma, H.; Yokoo, Y.; Tsuruta, T. Makromol. Chem.
1971, 143, 97. (c) Kobayashi, M.; Inoue, S.; Tsuruta, T. Macromolecules
1971, 4, 658. (d) Kobayashi, M.; Tang, Y. L.; Tsuruta, T.; Inoue, S.
Makromol. Chem. 1973, 169, 69. (e) Kobayashi, M.; Inoue, S.; Tsuruta, T.
J. Polym. Sci., Polym. Chem. Ed. 1973, 11, 2383. (f) Kuran, W.;
Pasynkiewicz, S.; Skupinska, J.; Rokicki, A. Makromol. Chem. 1976, 177,
11. (g) Kuran, W.; Pasynkiewicz, S.; Skupinska, J. Makromol. Chem. 1976,
177, 1283. (h) Soga, K.; Uenishi, K.; Hosada, H.; Ikeda, S. Makromol.
Chem. 1977, 178, 893. (i) Soga, K.; Hyakkoku, K.; Ikeda, S. Makromol.
Chem. 1978, 179, 2837.
Inoue and co-workers first reported on this process in 1969,
employing a catalyst derived from Zn(CH2CH3)2 and H2O.1,2
During the intervening years, a variety of related catalysts and
catalyst precursors containing zinc have been shown to promote
this copolymerization process.3 The most active of these
catalysts are those prepared by the reaction of zinc hydroxide
or zinc oxide with dicarboxylic acids.4 Indeed, there is a
commercial process based on these insoluble zinc catalysts;
(4) Soga, K.; Imai, E.; Hattori, I. Polym. J. 1981, 13, 407.
* To whom correspondence should be addressed. E-mail: DJDARENS@
mail.chem.tamu.edu.
(1) Inoue, S.; Koinuma, H.; Tsuruta, T. J. Polym. Sci., Part B 1969, 7,
287.
(5) (a) Rokicki, A. (Air Products and Chemicals, Inc., and Arco Chemical
Co.). U.S. Patent 4,943,677, 1990. (b) Motika, S. (Air Products and
Chemicals, Inc., Arco Chemical Co., and Mitsui Petrochemical Industries
Ltd.). U.S. Patent 5,026,676, 1991.
10.1021/ja9826284 CCC: $18.00 © 1999 American Chemical Society
Published on Web 12/19/1998

108 J. Am. Chem. Soc., Vol. 121, No. 1, 1999
Darensbourg et al.
differences.11 A 1.0-g (7.3 mmol) amount of ZnCl2 was combined with
a 2.67-g (14.4 mmol) amount of sodium bis-trimethylsilyl amide in a
50-mL Schlenk flask equipped with a reflux condensor. After the
addition of 30 mL of diethyl ether via syringe, the slurry was refluxed
for 1 h; subsequent removal of the formed sodium chloride by filtration
resulted in a colorless to pale yellow solution. A 3.61-g (14.7 mmol)
quantity of 2,6-diphenylphenol dissolved in 15 mL of diethyl ether was
added to the ZnII(bis-trimethylsilyl amide)2 solution via cannula. The
solution was then concentrated under vacuum to less than one-half its
original volume and placed in a freezer overnight. From this, 4.20 g of
product (80.8% yield) was isolated as colorless crystals. 1H NMR (CD2-
Cl2): δ 1.13 (t, 12H); 3.28 (q, 8H); 6.65-7.65 (m, 2,6-Ph2PhO-, 26H).
Free diethyl ether and bound diethyl ether are in equilibrium in solution,
with diethyl ether being predominantly free in deuterated dichloro-
methane. Anal. Calcd for C44H46O4Zn: C, 75.04; H, 6.58. Found: C,
80.79; H, 5.60. In part, the inability to obtain a close match in elemental
analyses is due to the ease with which the coordinated ether ligands
are lost.
metal cyanide complexes employed in this copolymerization
process have obvious alternative environmental shortcomings.6
Recently, soluble zinc complexes have been discovered which
possess significant catalytic activity to be attractive as catalysts
for the production of high-molecular-weight polycarbonates
from CO2 and epoxides. One of these systems represents a
modification of a process described in the patent literature which
involves catalysts produced from the reaction of zinc oxide with
acid anhydrides in alcoholic solvents or with the monoester of
a dicarboxylic acid.7 This methodology has been notably
improved upon by carrying out the esterification reaction in a
long-chain perfluorinated alcohol (tridecafluorooctanol) to
impart catalyst solubility in liquid or supercritical carbon
dioxide.8 Although the explicit structure of this latter catalyst
is unknown due to the unavailability of crystals suitable for
crystallographic analysis, it was shown to be very active for
the copolymerization of 1,2-epoxycyclohexane (cyclohexene
oxide) and carbon dioxide.
In another major development in this area, we have com-
municated our results on the copolymerization of cyclohexene
oxide/CO2 and the terpolymerization of cyclohexene oxide/
propylene oxide/CO2 catalyzed by soluble, well-defined mon-
omeric zinc complexes.9 These complexes consist of distorted
tetrahedral zinc phenoxide derivatives which possess bulky
substituents in the 2 and 6 positions of the phenolate ligands,
with the coordination sphere of zinc being completed with two
labile donor ligands. For example, (2,6-diphenylphenoxide)2-
ZnII(THF)2 is a typical representative of this group of catalysts.
Herein, we wish to describe the synthesis and characterization
of a series of these zinc(II) complexes along with the catalytic
activity of the aforementioned zinc derivatives for the copo-
lymerization and terpolymerization of carbon dioxide and
epoxides.
(2,6-Diphenylphenoxide)2Zn(THF)2, 2. This complex was prepared
in a manner analogous to that described above for complex 1 by simply
replacing diethyl ether with tetrahydrofuran as solvent for 2,6-
diphenylphenol. A 4.00-g amount (77.8% yield) of product was isolated
1
as colorless crystals. H NMR (CD2Cl2): δ 1.72 (m, THF, 8H); 3.46
(m, THF, 8H); 6.65-7.65 (m, 2,6-Ph2PhO-, 26H). Free THF and
bound THF are in equilibrium in solution, and THF appears to be
predominantly free in deuterated dichloromethane. Anal. Calcd for
C44H42O4Zn: C, 75.48; H, 6.05. Found: C, 73.91; H, 4.89.
(2,6-Diisopropylphenoxide)2Zn(THF)2, 3. This complex was pre-
pared in a manner analogous to that described above using 0.5 g (3.7
mmol) of ZnCl2, 1.4 g (7.6 mmol) of NaN(Si(CH3)3)2, and 1.3 g of
2,6-diisopropylphenol. A 1.4-g (7.8 mmol) amount (67.6% yield) of
1
product was isolated as colorless crystals. H NMR (CD2Cl2): δ 1.15
(broad, 12H); 1.78 (m, 8H); 3.15 (broad, 4H); 3.65 (m, 8H); 6.65-7.2
(broad, PhO-, 6H). Anal. Calcd for C32H50O4Zn: C, 68.13; H, 8.93.
Found: C, 66.88; H, 8.14.
(2,6-Di-tert-butylphenoxide)2Zn(THF)2, 4. A 1-mL THF solution
of 2,6-di-tert-butylphenol (0.214 g, 1.04 mmol) was added to a 1-mL
THF solution of Zn[N(SiMe3)2]2 (0.20 g, 0.52 mmol), leading to a
colorless solution which was stirred at room temperature for 1 h.
Approximately 2 mL of hexanes was added, and the clear solution was
then placed at -20 °C. Colorless block crystals formed after several
days. The supernate was removed, and the crystals were dried under
vacuum and collected to yield 0.156 g (49%). Anal. Calcd for C36H58O4-
Zn: C, 69.71; H, 9.43. Found: C, 67.38; H, 8.65. 1H NMR (C6D6): δ
1.13 [m, 4H, {THF}], 1.58 [s, 18H, {-CMe3}], 3.40 [m, 4H, {THF}],
6.82 [t, 1H, {4-H}], 7.30 [d, 2H, {3, 5-H}]. 13C{H} NMR(C6D6): δ
25.26 {THF}, 31.6 {-CMe3}, 35.6 {-CMe3}, 70.1 {THF}, 117.4 {4-
C6H3}, 125.4 {3,5-C6H3}, 139.2 {2,6-C6H3}, 163.8 {ipso-C6H3}.
[(2,4,6-Trimethylphenoxide)2Zn]n, 5. A 0.5-g (3.7 mmol) amount
of anhydrous ZnCl2 was combined with a 1.4-g (7.6 mmol) amount of
sodium bis-trimethylsilyl amide, which was refluxed in diethyl ether
as before. After removal of the sodium chloride, the addition of 1.3 g
(9.6 mmol) of 2,4,6-trimethylphenol resulted in the immediate precipi-
tation of 1.0 g (56% yield) of polymeric product. Anal. Calcd for
C18H22O2Zn: C, 64.39; H, 6.60. Found: C, 63.80; H, 6.90.
Preparation and Isolation of Zinc(bis-trimethylsilyl amide), Zn-
[N(SiMe3)2]2. Alternatively, the synthesis of these bis phenoxides of
zinc can be accomplished in a stepwise manner. That is, the Zn-
[N(SiMe3)2]2 intermediate can be isolated and purified in sizable
quantities12 and subsequently employed in the preparation of a variety
of bis(phenoxide)ZnL2 (L ) base) derivatives.
In a drybox, anhydrous zinc chloride (5.0 g, 36.7 mmol) was placed
in a 100-mL Schlenk flask along with approximately 2 equiv of NaN-
(SiMe3)2 (12.1 g, 66.1 mmol). The flask was removed from the drybox
and placed under a positive pressure of argon. Diethyl ether (60-70
mL) was added to the flask, and the reaction mixture was stirred for 1
h under argon. (NOTE: This reaction is exothermic, generating enough
heat to boil the ether. Hence, the reaction flask should be connected
Experimental Section
Methods and Materials. All manipulations were carried out under
an inert atmosphere unless otherwise stated, using glassware which
was flame-dried prior to use. The solvents were freshly distilled before
use. Cyclohexene oxide was purchased from Aldrich Chemical Co.,
and propylene oxide was purchased from Fluka Chemical Co., both
were distilled over calcium hydride before use. 2,6-Diphenylphenol,
2,4,6-tri-tert-butylphenol, 2,6-di-tert-butylphenol, 2,4,6-tri-methylphe-
nol, 2,6-di-isopropylphenol, and sodium bis-trimethylsilyl amide were
purchased from Aldrich Chemical Co. ZnCl2 dihydrate was purchased
from Fischer and was dehydrated by the published procedure.10 ZnII-
(2,4,6-tri-tert-butylphenoxide)2(THF)2 was prepared according to the
published procedure.11 Infrared spectra were recorded on a Mattson
1
6021 spectrometer with DTGS and MCT detectors. H and 13C NMR
spectra were recorded on a Varian XL-200 or Unity+ 300 supercon-
ducting high-resolution spectrometer. Elemental analyses were carried
out by Galbraith Laboratories Inc. and Canadian Microanalytical
Services, Ltd.
Synthesis of (2,6-Diphenylphenoxide)2Zn(diethyl ether)2, 1. Com-
plex 1 was prepared in a manner similar to that employed in the
synthesis of (2,4,6-tri-tert-butylphenoxide)2Zn(THF)2, with only minor
(6) (a) Kruper, W. J.; Smart, D. J. (The Dow Chemical Co.). U.S. Patent
4,500,704, 1983. (b) Chen, L.-B. Makromol. Chem., Macromol. Symp. 1992,
59, 75. (c) Yang, S.-Y.; Fang, X.-G.; Chen, L.-B. Polym. AdV. Technol.
1996, 7, 605.
(7) Sun, H.-N. (Arco Chemical Co.). U.S. Patent 4,783,445, 1988.
(8) (a) Super, M.; Berluche, E.; Costello, C.; Beckman, E. Macromol-
ecules 1997, 30, 368. (b) Super, M.; Beckman, E. J. Macromolecular
Symposia 1998, 127, 89.
(9) Darensbourg, D. J.; Holtcamp, M. W. Macromolecules 1995, 28,
7577.
(10) Pray, A. R. Inorg. Synth. 1957, 5, 153.
(11) Geerts, R. L.; Huffman, J. C.; Caulton, K. G. Inorg. Chem. 1986,
25, 1803.
(12) Burger, H.; Sawodny, W.; Wannaget, U. J. Organomet. Chem. 1965,
3, 113.

Catalytic ActiVity of a Series of Zn(II) Phenoxides
J. Am. Chem. Soc., Vol. 121, No. 1, 1999 109
Table 1. Crystallographic Data for Complexes 1-4, 6, and 7.
1
2
3
4
6
7
formula
formula weight
cryst system
space group
a, Å
b, Å
c, Å
C44H46O4Zn
704.18
orthorhombic
Aba2
19.851(13)
15.122(10)
12.424(6)
C44H42O4Zn
700.15
triclinic
P1h
8.778(3)
11.348(4)
19.078(6)
77.87(3)
81.74(3)
72.94(2)
1769.4(10)
2
1.314
0.737
6621
6279
6.81
14.71
C32H50O4Zn
564.09
monoclinic
Cc
24.079(7)
9.973(2)
14.828(4)
C36H58O4Zn
620.19
monoclinic
P21/c
17.359(2)
10.1645(12)
20.981(6)
C36H54O8Zn
680.16
orthorhombic
Aba2
21.688(9)
9.622(3)
16.897(6)
C28H32N2O2Zn
493.93
monoclinic
C2/c
16.517(3)
25.558(5)
14.638(3)
R, deg
â, deg
90
120.57(2)
113.02(2)
90
121.16(3)
γ, deg
V, Å3
3730(4)
4
1.254
0.700
1653
1653
4.90
3065.9(14)
4
1.222
0.834
1771
1767
2.96
3407.2(11)
4
1.209
0.756
3927
3720
6.10
3526(2)
4
1.281
0.745
1616
1616
8.33
5288(2)
8
1.241
0.954
4094
3942
6.91
Z
d(calcd), g/cm3
absorp coeff, mm-1
no. of total reflections
no. of unique reflections
R,a %
Rw,a %
5.15
6.07
12.98
17.08
17.43
a R ) ∑||Fo| - |Fc||/∑Fo. RW ) {[∑w(Fo - Fc2)2]/[∑w(Fo )2]}1/2
.
2
2
to a water-cooled condenser which is Vented.) Upon cooling of the
mixture to ambient temperature, NaCl precipitated out and was removed
by filtration through a glass frit under an inert atmosphere. The solid
collected was washed with two portions of ether (5 mL), and the
combined ether filtrate was evacuated at reduced pressure to provide a
clear, colorless oil of Zn[N(SiMe3)2]2. The product was purified by
vacuum distillation (2-3 mmHg) in a short-path apparatus at 103 °C;
9.4 g of product was collected (74% yield) and shown to be pure by
1H NMR (singlet at 0.5 ppm).
via an injection port into a 300-mL autoclave which had previously
been dried overnight under vacuum at 90 °C. The autoclave was then
placed under 700-800 psi of carbon dioxide and heated to the
appropriate temperature. After the allotted time (48-69 h), the autoclave
was allowed to cool to room temperature, upon which the polymer
was extracted. Unreacted monomer was removed by repeated precipita-
tion of the polymer from a dichloromethane solution with methanol.
The insoluble catalyst, Zn(2,4,6-trimethylphenoxide)2, was added to
the reactor in the drybox. The reactor was then removed from the
drybox, and the cyclohexene oxide was added through the injection
port. The very reactive catalyst for homopolymerization of epoxides,
Zn(2,4,6-tri-tert-butylphenoxide)2(THF)2, was sealed into a glass ampule
and placed into the reactor. The reactor was evacuated overnight, and
then the cyclohexene oxide was added via the injection port. The ampule
was broken after the introduction of carbon dioxide and was heard to
shatter upon stirring.
(2,6-Di-tert-butylphenoxide)2Zn(propylene carbonate)2, 6. Two
milliliters of a toluene solution of 2,6-di-tert-butylphenol (0.21 g, 1.04
mmol) was added via cannula to a solution of Zn[N(SiMe3)2]2 (0.20 g
or 0.52 mmol in 2 mL of toluene) under an atmosphere of argon. The
reaction solution was allowed to stir for 1 h at ambient temperature,
followed by the addition of 1.04 mmol (89 µL) of propylene carbonate
(PC) which was degassed and added via a microliter syringe. Upon
removal of the toluene solvent under vacuum, an oily residue remained
which was washed with hexane to provide a fine white powder of 6 in
69% yield. 1H NMR (C6D6): δ 0.38 (d, 3H, PC-Me), shifted 0.27 ppm
upfield from free PC; 1.70 (s, 12H, t-Bu); 2.68 (dd, 1H, PC methylene),
shifted 0.47 ppm upfield from free PC; 3.11 (dd, 1H, PC methylene),
shifted 0.39 ppm upfield from free PC; 3.49 (sx, 1H, PC methine),
shifted 0.36 ppm upfield from free PC; 6.80 (t, 1H, benzylic H); 7.32
(d, 2H, benzylic Hs). Infrared of ν(CO): THF (1813 cm-1), toluene
(1818 cm-1). This peak is unshifted from that in free propylene
carbonate. Anal. Calcd for C36H54O8Zn: C, 63.57; H, 8.00. Found: C,
61.58; H, 7.74. Clear, colorless crystals of 6 were obtained from a
concentrated toluene solution of the complex maintained at -20 °C.
(2,4,6-Trimethylphenoxide)2Zn(pyridine)2, 7. In a procedure analo-
gous to that described above for the preparation of 6, 0.20 g (0.52
mmol) of Zn[N(SiMe3)2]2 and 0.14 g (1.04 mmol) of 2,4,6-trimeth-
ylphenol were mixed in diethyl ether. A white solid, assumed to be an
aggregate of bis-2,4,6-trimethylphenoxide zinc containing bridging
phenoxide ligands, precipitated immediately. The reaction mixture was
stirred for about 45 min at ambient temperature. Upon addition of 84
µL (1.04 mmol) of distilled pyridine to the reaction mixture, the zinc
complex completely dissolved. Vacuum removal of all volatiles led to
a waxy, white solid which redissolved upon addition of ether solvent.
1H NMR (C6D6): δ 2.21 (s, CH3 of phenolate); 2.38 (s, CH3 of
phenolate); 6.48 (t, H on C5H5N), shifted 0.22 ppm downfield from
free C5H5N; 6.6-6.8 (benzylic protons); 8.49 (d, H on C5H5N), shifted
0.04 ppm upfield from free C5H5N. Clear, colorless crystals of 7 were
obtained from a concentrated THF solution of the complex layered
with hexane and maintained at -20 °C.
Variable-Temperature NMR Experiments. All variable-temper-
1
ature experiments were performed using H NMR on a 200E Varian
instrument equipped with a variable-temperature unit. In a typical
experiment, 20 mg of the desired Zn complex and 0.5 mL of CD2Cl2
were placed into an NMR tube, and the tube was sealed under vacuum.
CD2Cl2 was replaced with benzene-d6 for high-temperature experiments.
Further information concerning individual experiments is discussed in
the Results section.
Polymer Characterization. Polymer samples were first analyzed
by means of IR and NMR spectroscopy for the absence of cyclic
carbonates (ν(CO) ) 1800 cm-1) and percentages of ether and carbonate
linkages (1H: 3.45 and 4.60 ppm respectively). 1H NMR was also used
to check for the presence of cyclohexene oxide, which has a distinct
singlet at 3.13 ppm. 13C NMR was utilized for determining the
stereoregularity or tacticity of the polymer. Mw and Mn determinations
were carried out by Exxon Research using GPC. Finally, trans
stereochemistry was determined to be present in the polymer through
a cleavage reaction. This was accomplished by refluxing a sample of
the polymer in a methanolic solution of NaOH (2 g in 8 mL), according
to the procedure described in the literature.3b After the mixture was
refluxed overnight, the solution was neutralized with hydrochloric acid
and analyzed by GC/MS.
X-ray Crystallographic Studies of Complexes 1-4, 6, and 7.
Crystal data and details of the data collection are provided in Table 1.
Colorless block crystals of the six complexes were mounted on a glass
fiber with epoxy cement at ambient temperature, and complexes 2 and
4 were cooled to 163 and 193 K, respectively, in a N2 cold stream.
Preliminary examination and data collection were performed on a
Siemens P4 X-ray diffractometer (Mo KR, λ ) 0.710 73 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°. Three control reflections collected for every 97 reflections
Copolymerizations of Epoxides and Carbon Dioxide. A typical
copolymerization run was as follows: 0.125 g of the respective zinc
bisphenoxide complex was dissolved in cyclohexene oxide and/or
propylene oxide. We typically used 5-20 mL of epoxide in various
amounts depending on the desired copolymer. The solution was loaded

110 J. Am. Chem. Soc., Vol. 121, No. 1, 1999
Darensbourg et al.
Scheme 1
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 refine-
ment 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 in 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.13
1
Figure 1. Variable-temperature H NMR spectra of (2,6-diphenyl-
phenoxide)Zn(THF)2 (2) dissolved in CD2Cl2 in the THF region of the
spectrum.
weak binding ability of epoxides as ligands. Indeed, in the base
adducts of η3-HB(3-Phpz)3Cd(acetate), we have quantified the
order of ether binding as THF > dioxane > cyclohexene oxide
g propylene oxide.16 On the other hand, in the case involving
the zinc phenoxide derivative containing 2,6-tert-butyl substit-
uents, homopolymerization of the epoxide occurred rapidly at
ambient temperature, thereby prohibiting the isolation of an
epoxide adduct.
Results and Discussion
Although the solid-state structures of these monomeric zinc
phenoxides all reveal coordination of the various bases to the
zinc center (vide infra), in solution the extent of base binding
is greatly dependent on the temperature, solvent, base, and nature
of phenoxide ligands coordinated to zinc. For example, upon
dissolving (2,6-diphenylphenoxide)2Zn(THF)2 (2) in C6D6 or
CD2Cl2, the THF ligands are essentially completely dissociated
from the metal center at ambient temperature, as demonstrated
Synthesis, Spectral, and Structural Characterization of
Monomeric Zinc Phenoxides. The monomeric zinc phenoxide
derivatives were synthesized in good, purified yields (50-80%)
by the general synthetic route described by Caulton and co-
workers11 for the synthesis of the (ArO)2Zn(THF)2 complexes,
where OAr ) 2,4,6-tri-tert-butylphenoxide and 2,6-di-tert-
butylphenoxide. We have performed the process outlined in
Scheme 1 both in a one-pot synthesis and by isolating/purifying
the Zn[N(SiMe3)2]2 intermediate prior to its reaction with the
various phenols. It is important to remove the NaCl precipitant
formed in the initial step prior to carrying out the subsequent
acid-base reaction in order to avoid the formation of complexes
which incorporate sodium chloride, e.g., Na[(2,6-diisopropyl
phenoxide)4Zn2Cl]‚3THF.14 The X-ray structure of this dimeric
species shows the sodium ion to be six-coordinate, involved in
an η6-interaction with the phenoxide π-system, and bound to
two THF ligands and a zinc-bridged chloride ligand. In the
instance of R ) R′ ) Me, i.e., the 2,4,6-trimethylphenol-derived
product, the zinc complex immediately precipitates out of
solution. This behavior is indicative of aggregrate formation in
the absence of steric inhibition by this less bulky phenoxide
ligand. However, the aggregation, which presumably involves
bridging phenoxide ligands, is readily disrupted by strong bases
such as pyridine to afford monomeric, soluble adducts.
1
by H NMR spectroscopy. That is, the proton resonances for
the R-CH2 and â-CH2 groups of THF in the case of complex 2
dissolved in CD2Cl2 occur at 3.46 and 1.72 ppm, respectively,
whereas the corresponding resonances are observed at 3.69 and
1.83 ppm for free THF in CD2Cl2. Similarly, at ambient
temperature, these proton resonances are observed at 3.53 and
1.39 ppm for complex 2 dissolved in C6D6 compared to 3.55
and 1.43 ppm for free THF in C6D6. Upon lowering the
1
temperature of a CD2Cl2 solution of 2, the two H resonances
assigned to THF shifted upfield and between -40 and -60 °C
are seen to broaden. Eventually, four resonances at 3.57 and
1.73 ppm (free THF)17 and 2.49 and 1.30 ppm (bound THF)
appeared at -85 °C, indicative of slow THF exchange (Figure
1). Unlike what is seen in the 2,6-di-tert-butylphenoxide
analogue (vide infra), complex 4, the R-CH2 protons are shifted
upfield to a greater extent than the â-CH2 units, as anticipated
for coordination of the oxygen atom of THF to the zinc center,
where the CH2 unit closest to the donor atom should experience
the larger change in electronic environment.
As illustrated in Scheme 1, the ether ligands upon isolation
of these zinc complexes can be replaced by other bases; e.g.,
the relevant bases for the studies reported herein include
epoxides, propylene carbonate, pyridine, and phosphines.15 The
propylene oxide derivative was prepared by repeated dissolution
of the diethyl ether complex in the epoxide solvent. Unfortu-
nately, we have not yet been able to isolate and fully characterize
an epoxide complex of zinc as we have been able to accomplish
for the other adducts (vide infra). This, in part, is due to the
As noted in Figure 1, not all of the THF ligands are bound
to the zinc center at -85 °C in complex 2.18 Nevertheless, from
the integrated peak intensities, the ratio of bound to free THF
is greater than 1. Indeed, in the instance where the total THF to
(16) Darensbourg, D. J.; Niezgoda, S. A.; Holtcamp, M. W.; Draper, J.
D.; Reibenspies, J. H. Inorg. Chem. 1997, 36, 2426.
(17) There is a small upfield shift of the free THF proton resonances in
CD2Cl2 as the temperature is lowered from ambient to -85 °C of 0.12 and
0.10 ppm.
(18) Upon evacuating samples of the bisphenoxide complexes of zinc
at reduced pressures at ambient temperature, some THF is lost from the
isolated solid. That is, the ratio of THF to zinc is less than 2:1, generally
being about 1.8:1 if evacuation is not carried out for a prolonged period.
(13) Ibers, J. A., Hamilton, W. C., Eds. International Table for X-ray
Crystallography; Kynoch Press: Birmingham, England, 1974; Vol. IV.
(14) Darensbourg, D. J.; Yoder, J. C.; Struck, G. E.; Holtcamp, M. W.;
Draper, J. D., Reibenspies, J. H. Inorg. Chim. Acta 1998, 274, 115.
(15) Darensbourg, D. J.; Zimmer, M. S.; Rainey, P.; Larkins, D. L. Inorg.
Chem. 1998, 37, 2852.

Catalytic ActiVity of a Series of Zn(II) Phenoxides
J. Am. Chem. Soc., Vol. 121, No. 1, 1999 111
Table 2. Selected Bond Distances (Å) and Bond Angles (deg) for
Complexes 1-4, 6, and 7a,b
Complex 1: Zn(O-2,6-Ph2C6H3)2(Et2O)2
Zn(1)-O(1)
1.886(6) Zn(1)-O(2)
2.107(7)
O(1)-C(1)
1.333(11)
O(1)-Zn(1)-O(1A)
O(1)-Zn(1)-O(2)
O(1)-Zn(1)-O(2A)
140.0(4) C(1)-O(1)-Zn(1)
103.5(3) O(2)-Zn(1)-O(2A)
103.3(3) Zn(1)-O(2)-[OC2 plane] 168.0(4)
125.1(6)
94.6(4)
Complex 2: Zn(O-2,6-Ph2C6H3)2(THF)2
1.864(4) Zn(1)-O(2)
Zn(1)-O(1)
Zn(1)-O(3)
O(1)-C(1)
1.864(4)
2.036(4)
1.332(6)
106.5(2)
108.5(2)
122.3(3)
133.0(4)
2.080(4) Zn(1)-O(4)
1.335(6) O(2)-C(19)
O(1)-Zn(1)-O(2)
O(1)-Zn(1)-O(4)
O(4)-Zn(1)-O(3)
136.2(2) O(1)-Zn(1)-O(3)
106.6(2) O(2)-Zn(1)-O(4)
90.5(2) C(1)-O(1)-Zn(1)
Zn(1)-O(3)-[OC2 plane] 168.8(4) C(19)-O(2)-Zn(1)
Zn(1)-O(4)-[OC2 plane] 174.4(3)
Figure 2. Logarithmic plot of Kf as a function of temperature for
complex 2 in CD2Cl2 solution.
Complex 3: Zn(O-2,6-iPr2C6H3)2(THF)2
Zn(1)-O(1)
1.846(11) Zn(1)-O(2)
2.028(13) Zn(1)-O(4)
1.856(13)
2.074(11)
1.38(2)
106.0(5)
103.9(5)
127.0(11)
124.3(10)
Zn(1)-O(3)
O(1)-C(1)
1.29(2) O(2)-C(13)
zinc molar ratio was determined to be 1.8, the bound THF to
zinc ratio at -85 °C was shown to be 1.4. Hence, the facile
stepwise equilibria (eqs 2 and 3) which exist between free and
bound THF bases lie intermediate between the (2,6-diphenyl-
phenoxide)2Zn(THF) and (2,6-diphenylphenoxide)2Zn(THF)2
species.19 From the fast-exchange region of the 1H NMR spectra
O(1)-Zn(1)-O(2)
O(1)-Zn(1)-O(4)
O(4)-Zn(1)-O(3)
139.8(2) O(1)-Zn(1)-O(3)
104.2(5) O(2)-Zn(1)-O(4)
91.0(2) C(1)-O(1)-Zn(1)
Zn(1)-O(3)-[OC2 plane] 174.0(3) C(13)-O(2)-Zn(1)
Zn(1)-O(4)-[OC2 plane] 172.3(3)
Complex 4: Zn(O-2,6-tBu2C6H3)2(THF)2
Zn(1)-O(1)
1.873(5) Zn(1)-O(2)
1.878(5)
2.086(7)
1.360(8)
104.3(2)
99.6(2)
K1
Zn(1)-O(3)
2.103(5) Zn(1)-O(4)
[Zn] + THF y
\
z [Zn]-THF
(2)
(3)
O(1)-C(1)
1.348(8) O(2)-C(15)
O(1)-Zn(1)-O(2)
O(1)-Zn(1)-O(4)
O(4)-Zn(1)-O(3)
142.9(2) O(1)-Zn(1)-O(3)
101.1(2) O(2)-Zn(1)-O(4)
100.3(2) C(1)-O(1)-Zn(1)
K2
[Zn]-THF + THF y\z [Zn]-(THF)2
134.2(5)
131.0(4)
Zn(1)-O(3)-[OC2 plane] 174.0(2) C(15)-O(2)-Zn(1)
{[Zn]-(THF)2}
Zn(1)-O(4)-[OC2 plane] 175.5(4)
Complex 6: Zn(O-2,6-tBu2C6H3)2(PC)2
Kf )
,
{[Zn]}{THF}2
where [Zn] ) (2,6-diphenylphenoxide)2Zn
Zn(1)-O(1)
O(1)-C(1)
O(1)-Zn(1)-O(1A)
O(1)-Zn(1)-O(2)
O(1)-Zn(1)-O(2A)
1.850(11) Zn(1)-O(2)
1.30(2) O(2)-C(15)
136.4(7) C(1)-O(1)-Zn(1)
100.4(5) O(2)-Zn(1)-O(2A)
108.4(5) Zn(1)-O(2)-C(15)
2.071(11)
1.20(2)
155.7(10)
96.3(6)
(between 25 and -60 °C), the overall formation constant (Kf
) K1K2) can be determined as a function of temperature
according to the method of Popov.20 From the linear relationship
of ln Kf vs T-1 (Figure 2) for the composite equilibrium defined
in eqs 2 and 3, the values of ∆H°, ∆S°, and ∆G° were calculated
to be -28.9 kJ/mol, -87.4 J/mol, and -2.80 kJ/mol, respec-
tively. The complex formation constant (Kf) was found to be
2.86 at 25 °C. In addition, from the estimated rates of THF
exchange (40 s-1 at -75 °C) as a function of temperature, an
Ea value of 44 kJ/mol was obtained.
127.5(13)
Complex 7: Zn(O-2,4,6-Me3C6H2)2(Py)2
1.885(4) Zn(1)-N(1)
1.330(8)
Zn(1)-O(1)
O(1)-C(1)
O(1)-Zn(1)-O(1A)
O(1)-Zn(1)-N(1A)
N(1)-Zn(1)-N(1A)
2.065(5)
134.5(3) O(1)-Zn(1)-N(1)
104.8(2) C(1)-O(1)-Zn(1)
102.2(3)
103.3(2)
136.5(4)
a Estimated standard deviations are given in parentheses. b Symmetry-
generated atoms designated by (#A).
The effect of changing the electron-withdrawing 2,6-diphenyl
substituents on the phenoxide ligands to the electron-releasing
2,6-di-tert-butyl groups leads to faster bound/free THF exchange
rates. That is, starting with a sample of (2,6-di-tert-butyl-
phenoxide)2Zn(THF)x (where x ) 1.2-1.4) in toluene, it was
not possible to observe the slow exchange limit at -80 °C. The
free/bound THF exchange rate is approximated to be greater
than 200 s-1 at -80 °C. It is important to reiterate here that, in
this derivative, the â-CH2 proton signal of the THF ligands shifts
upfield to a much greater extent than the R-CH2 signal in toluene
solution. A similar observation has previously been reported
for this complex in benzene by Caulton and co-workers.11 For
the temperature range from 20 to -80 °C, the proton resonances
for the R-CH2 and â-CH2 groups in THF shift from 3.339 and
1.064 ppm to 3.108 and 0.674 ppm, respectively, for the (2,6-
di-tert-butylphenoxide)2Zn(THF)1.2 in toluene.21 From the mag-
nitude of these upfield proton shifts of the THF ligand in this
derivative, 0.211 and 0.366 ppm, there is significant THF
binding even at ambient temperature. As expected in excess
THF, where the THF:Zn ratio is equal to 4.2, these proton
signals appear much closer to those of free THF at 3.521 and
1.219 ppm at 20 °C. Hence, the Kf values for the equilibria
defined in eqs 2 and 3 are greater in this instance than those
for the 2,6-diphenylphenoxide derivative. The order of ligand
binding as a function of the nature of the phenoxide ligands is
seen in other instances for both zinc and cadmium deriVatiVes
encompassing a wide Variety of ligands.15,22 This trend in
metal-ligand binding most likely reflects the greater steric
requirements of the 2,6-diphenyl ligands. Based on electronic
arguments, the opposite effect would be anticipated, i.e., metal-
ligand binding would be enhanced by electron-withdrawing
substituents on the phenolate ligands.
Complexes 1-4, 6, and 7 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. All complexes crystallize as
(19) It should be noted here that it is not possible to isolate the THF
resonances for these individual zinc species at -85 °C in CD2Cl2.
(20) (a) Roach, E. T.; Hardy, P. R.; Popov, A. I. Inorg. Nucl. Chem.
Lett. 1973, 9, 359. (b) Popov, A. I. In Modern NMR Techniques and Their
Application in Chemistry; Popov, A. I., Hallenga, K., Eds.; Marcel
Dekker: New York, 1987; p 485.
(21) There is a small upfield shift of the â-CH2 proton resonance of free
THF in toluene of 0.13 ppm at -80 °C.

112 J. Am. Chem. Soc., Vol. 121, No. 1, 1999
Darensbourg et al.
Figure 6. Molecular structure of complex 7; thermal ellipsoids at 50%
probability.
Figure 3. Molecular structure of complex 2; thermal ellipsoids at 50%
probability.
Scheme 2
142.9°, whereas the corresponding values encompassed by φ
are 90.5-100.3°. Interestingly, the structures of the closely
related derivatives, 4 and 8,11 which differ only in the presence
of a tert-butyl substituent in the 4-position of the phenoxy ring
in the latter case, have strikingly different θ parameters (142.9-
(2)° vs 121.7(3)°) with more comparable φ values (100.3(2)°
vs 94.93(27)°). For the sterically less bulky 2,4,6-trimethylphe-
noxide derivative, which is insoluble in ether solvents but
solubilized upon addition of 2 equiv of the better donor ligand
pyridine and crystallized as complex 7, the θ angle of 134.5° is
quite similar to those found in the other derivatives. The thermal
ellipsoid drawing of 7 is displayed in Figure 6, where the
N-Zn-N angle, analogous to φ in the ether derivatives, was
observed to be 102.2°.
Figure 4. Molecular structure of complex 3; thermal ellipsoids at 50%
probability.
The average Zn-O(aryl) bond distances in all of the
structurally characterized zinc derivatives (complexes 1-4, 6,
and 8) which contain the O4 core atom ligand set are quite
similar, covering a range of only 0.037 Å, i.e., 1.850-1.887 Å.
Similarly, Zn-O(aryl) bond lengths of 1.885(5) Å are seen in
the pyridine adduct, complex 7, where the O2N2 ligand set is
present. Likewise, the Zn-O (THF, Et2O, PC) bond length
varies only slightly, from 2.051 to 2.107 Å, although these
neutral ligands differ somewhat in donor properties. The Zn-N
bond distance in 7 is 2.065(4) Å, consistent with a slightly
stronger interaction and the 0.02 Å longer effective atomic radius
of nitrogen as compared to oxygen. The average Zn-O-C(aryl)
angles (â in Scheme 2) in all of these derivatives traverse the
broad range of 125.1-155.7°. Interestingly, the oxygen donor
ligand propylene carbonate in complex 6 is pyramidal, with a
Zn-O-C angle (R) of 127.5(13)°, whereas the corresponding
angles in the ether donor ligand derivatives (defined by the
Zn-O vector and the midpoint of the OC2 plane) are much
more obtuse (168.0-175.5°). This latter observation presumably
results from the ether donor’s ability to accommodate the bulky
phenoxide ligands. Indeed, this has previously been seen in the
cadmium analogues.22
Figure 5. Molecular structure of complex 6; thermal ellipsoids at 50%
probability.
four-coordinate monomers with highly distorted tetrahedral
geometry about the zinc center, much like that exhibited by the
originally reported analogue, (2,4,6-tri-tert-butylphenoxide)2Zn-
(THF)2 (8).11 Thermal ellipsoid representations of selected
derivatives may be found in Figures 3-5, along with a common
abbreviated atom labeling scheme (other figures may be found
in the Supporting Information). The skeletal rendition, indicating
the salient solid-state structural features of these derivatives, is
displayed in Scheme 2. In all instances, the angle between the
two sterically encumbering phenoxide ligands (θ) is significantly
more obtuse than the corresponding angles observed between
the two smaller neutral base ligands (φ). For example, the values
of θ in complexes 1-4 and 6 span the narrow range of 136.2-
Copolymerization Reactions of Epoxides and Carbon
Dioxide. Because of the solution properties exhibited by the
monomeric (phenoxide)2Zn(ether)2 derivatives described in this
report, it is expected that these complexes should serve as good
(22) Darensbourg, D. J.; Niezgoda, S. A.; Draper, J. D.; Reibenspies, J.
H. J. Am. Chem. Soc. 1998, 120, 4690.

Catalytic ActiVity of a Series of Zn(II) Phenoxides
J. Am. Chem. Soc., Vol. 121, No. 1, 1999 113
Figure 7. Yield of copolymer (cyclohexene oxide/CO2) in g/g of Zn
for a 69 h run at 80 °C and 750 psi loading pressure of CO2 using
various bis(phenoxide)2Zn complexes, and effects of reaction conditions
on copolymer yield employing the (2,6-diphenylphenoxide)2Zn(Et2O)2
(1) catalyst.
Figure 8. 13C NMR spectrum of poly(cyclohexene carbonate) obtained
with complex 5 as catalyst. Recorded in CDCl3 at ambient temperature.
catalysts or catalyst precursors for ring-opening of epoxides and
CO2 coupling processes. This should be particularly true in
noninteracting solVents, where there would be little competition
for metal binding of the weakly interacting epoxide substrates.
Indeed, we have communicated preliminary results demonstrat-
ing the efficiency of these complexes for catalyzing the
copolymerization process defined in eq 4, where m . n.9 Herein,
we wish to describe these studies in greater detail.
it was necessary to initially pressurize the cyclohexene oxide
loaded reactor with CO2 prior to introducing the catalyst. This
was accomplished by placing the catalyst in a sealed glass
ampule which was broken when mechanical stirring was
started.24
1
The polymers were characterized by FTIR and H and 13C
NMR spectroscopies utilizing the previously established meth-
odology for similar copolymers produced from other zinc
catalyst systems.25 The intense asymmetric ν(CO2) stretching
vibration of the polycarbonate linkages was observed at 1750
cm-1. The percentage of carbonate linkage in the purified
1
polymer was calculated from the relative intensities of the H
NMR signals of the methine protons next to the carbonate
linkages (δ ) 4.60 ppm) and ether linkages (δ ) 3.45 ppm). In
general, the carbonate linkages were greater than 90% for
reactions carried out at CO2 pressures greater than 750 psi. A
notable exception, where carbonate linkages were as low as
53%, was observed for the very active epoxide homopolymer-
ization catalyst, (2,6-di-tert-butylphenoxide)2Zn(THF)2. 13C
NMR spectroscopy in the carbonate region was utilized to assess
the stereoregularity or tacticity of the polymer (Figure 8). That
is, consistent with previous reports,25 there are two 13C
resonances at δ ) 153.7 and 153.1 ppm which are assigned to
syndiotactic and isotactic copolymer chain tacticity on the basis
of corresponding 13C NMR data from the model compound 2,2′-
oxydicyclohexanol.26 In addition, there is at least one more
carbonate 13C resonance at 153.3 ppm which can be assigned
to heterotactic forms of the copolymer. As is readily apparent
from Figure 8, the copolymer chain tacticity is predominantly
syndiotactic. The stereochemistry of the cyclohexene oxide ring-
opening reaction in the cyclohexene oxide/carbon dioxide
Figure 7 (top-left corner) summarizes the effect of R
substituents on the phenoxide ligands of the zinc(ether)2
complexes on the yield of high-molecular-weight polymer
(characterized by its insolubility in methanol) for the copolym-
erization reaction defined in eq 4. It is important to note here
that these processes were carried out in the absence of added
solVents; i.e., the reactants, zinc phenolates/cyclohexene oxide/
CO2 constitute the reaction system. Furthermore, prior to the
production of large quantities of copolymers under the reaction
conditions specified, the solutions appeared to be homogeneous,
as evidenced by a video recording of processes performed in a
high-pressure reactor equipped with sapphire windows. As
indicated in Figure 7, the yield of high-molecular-weight
polymer increases over a reaction period of 69 h from 477 to
1441 g of polymer/g of zinc as the substituents on the phenolate
ligands of the catalyst vary as Pr < Ph < Bu < Me. Also
described in Figure 7 are the effects of time, temperature, and
pressure on the polymer yield for the (2,6-diphenylphenoxide)2Zn-
(Et2O)2 catalyst system. The polymer yield was not enhanced
for a reaction carried out at much higher pressures of carbon
dioxide, i.e., supercritical conditions.23 The 2,6-di-tert-butyl-
phenoxide derivative, complex 4, was very active as an epoxide
homopolymerization catalyst to afford high-molecular-weight
polyethers. To retard this process relative to copolymerization,
i
t
(24) Alternatively, we have recently shown that the rate of the homo-
polymerization of cyclohexene oxide catalyzed by complex 4 is greatly
retarded in the presence of 1 equiv of PCy3, where the complex (2,6-di-
tert-butylphenoxide)2ZnPCy3 is formed at ambient temperature.15 At the
same time, the copolymerization process readily occurs; e.g., under the
reaction conditions defined in Figure 7, where complex 4 provides 677 g
of polymer/g of Zn with only 53% carbonate linkages, this catalyst in the
presence of 1 equiv of PCy3 affords 375 g of polymer/g of Zn with over
92% carbonate linkages. A more definitive study of phosphine derivatives
of zinc and cadmium phenoxides and their role in copolymerization catalysis
of epoxides and CO2 will be the subject of a later publication.
(23) For a polymer run carried out with complex 2 as catalyst at 80 °C
and 6500 psi of CO2 (thanks to Dr. Christine Costello at Exxon Research,
Annandale), the yield (357 g/g of Zn) and % carbonate linkages (90) of
copolymer were essentially identical to those of an otherwise analogous
run performed at 80 °C and 800 psi pressure (366 g/g of Zn and 91%
carbonate linkages).
(25) (a) Kuran, W.; Listos, T. Macromol. Chem. Phys. 1994, 195, 977.
(b) Koinuma, H.; Hirai, H. Makromol. Chem. 1977, 178, 1283.
(26) Hasebe, Y.; Tsuruta, T. Makromol. Chem. 1987, 188, 1403.

114 J. Am. Chem. Soc., Vol. 121, No. 1, 1999
Darensbourg et al.
Scheme 3
manner analogous to that depicted in Scheme 3 for propylene
oxide/CO2 copolymerization.
Relevant to the zinc-bound propylene carbonate intermediate
indicated in Scheme 3, it is of importance to note that propylene
carbonate is a much better ligand toward zinc than cyclic ethers.
That is, in this investigation, propylene carbonate was found to
be bound to zinc when complex 6 was dissolved in C6D6 at
ambient temperature by 1H NMR spectroscopy. All proton
resonances for the propylene carbonate ligand in complex 6 are
shifted significantly upfield; e.g., the multiplet for the hydrogen
attached to the carbon bearing the CH3 group is shifted from
3.93 ppm in free propylene carbonate to 3.46 ppm. Furthermore,
the addition of excess THF to a benzene solution of 6 did not
alter the propylene carbonate binding. Hence, the stronger
affinity of zinc for propylene carbonate, a byproduct of the
copolymerization process, over cyclic ethers (particularly
epoxides) should greatly inhibit theses catalytic reactions.
Carbon Dioxide Insertion Reactions Involving Zinc Phe-
noxides. It is of interest to examine the insertion reaction of
carbon dioxide into the Zn-O bond of these bis(phenoxide)Zn
derivatives since, in the initiation process for the copolymer
synthesis, CO2 insertion (eq 5) and/or epoxide ring-opening (eq
6) could be occurring. The latter process requires prior substrate
copolymerization processes was determined to be trans by
alkaline hydrolysis of the polymer, which provided pure trans-
1,2-cyclohexanediol.27 The molecular weights (Mn and Mw) of
the methanol-insoluble copolymers were measured by gel
permeation chromatography (GPC) by comparison with poly-
styrene standards, with Mw values ranging from 45 × 103 to
173 × 103 and Mw/Mn ratios varying from 2.5 to 4.5. However,
the polydispersity indexes have been observed to be as large as
10 in a few catalytic runs. This broad range in polydispersity
may be attributed to more than one zinc species initiating the
polymerization process or to protonation being the termination
step. It is estimated, based on Mn and the yield of polymer, that
greater than 50% of the zinc atoms are active in forming high-
molecular-weight copolymers.
Under the reaction conditions where these zinc phenolates
actively catalyze the copolymerization of cyclohexene oxide and
carbon dioxide, the reaction between propylene oxide and CO2
leads predominantly to propylene carbonate. This process
presumably occurs via the back-biting reaction depicted in
Scheme 3.25a This observation is in agreement with results from
other laboratories, where in the absence of steric constraints
about the metal center the back-biting process is favored over
polymer growth when propylene oxide is the substrate. Nev-
ertheless, at much lower temperatures (e.g., 40 °C using complex
2 as catalyst), copolymer formation is slightly favored over
cyclic carbonate production. Evidently, the rate of cyclic
carbonate formation has a greater temperature dependence than
the rate of copolymer production. Furthermore, the ratio of cyclic
carbonate to copolymer afforded at these low temperatures is
influenced to a small extent by the substitutents on the phenoxide
complex employed as catalyst.
On the contrary, the terpolymerization reaction between
cyclohexene oxide/propylene oxide/CO2 readily takes place in
the presence of these zinc phenolate catalysts, with little cyclic
carbonate production at low propylene oxide loadings (<50%
mole fraction). As anticipated on the basis of similar affinity
of the group 12 metals for the two epoxides,16 the resulting
terpolymer is a random mixture of polycarbonate linkages with
small quantities of polyether linkages. The 13C NMR spectra
of these terpolymers (reported in ref 9) display, in addition to
isotactic and syndiotactic poly(cyclohexenylene carbonate)
linkages, head-to-tail and tail-to-tail poly(propylene carbonate)
linkages.28 Other less intense carbonate resonances assignable
to some combination of various cyclohexenylene carbonate-
propylene carbonate linkages were also present in the spectra.
The lack of cyclic propylene carbonate formation during this
process is attributed to steric encumberance provided by the
cyclohexene carbonate linkages which inhibit the process in a
coordination, whereas the CO2 insertion process has been
demonstrated in previous studies to be a concerted reaction.29
Furthermore, we have demonstrated that the reaction depicted
in eq 5 is greatly inhibited for bulky metal phenolates.30 For
example, CO2 rapidly and reversibly inserts into the W-OPh
bond of the W(CO)5OPh- anion to afford W(CO)5O2COPh-,
but insertion of CO2 into the W-O bond of (2,6-diphenyl-
phenoxide)W(CO)5- does not occur, even at high pressures of
CO2. That is, CO2 insertion is governed by the accessibility of
the oxygen lone pairs on the phenoxide for interaction with the
weakly electrophilic carbon center of CO2. Herein we have
examined the propensity for CO2 insertion into a sterically
encumbered vs sterically accessible metal phenoxide employing
13C NMR spectroscopy.
Upon bubbling 13CO2 into a pyridine-d5 solution of complex
7, (2,4,6-trimethylphenoxide)2Zn(C5H5N)2, in addition to the
broad peak for free 13CO2, a 13C resonance in the carbonate
region at 170.1 ppm was evident (peak c in Figure 9B).
Moreover, the 13C signal corresponding to the ipso carbon of
the phenoxide ring was split into two resonances, one at its
original position (162.5 ppm, peak a) and one shifted downfield
to 165.9 ppm (peak a* in Figure 9B). This latter observation is
indicative of CO2 insertion into only one of the metal phenolate
groups (eq 7). In addition, from the relative peak intensities of
(27) Inoue, S.; Koinuma, H.; Yokoo, Y.; Tsuruta, T. Makromol. Chem.
1971, 143, 97.
(29) Darensbourg, D. J.; Sanchez, K. M.; Reibenspies, J. H.; Rheingold,
A. L. J. Am. Chem. Soc. 1989, 111, 7094.
(28) Lednor, P. W.; Rol, N. C. J. Chem. Soc., Chem. Commun. 1985,
598.
(30) Darensbourg, D. J.; Mueller, B. L.; Bischoff, C. J.; Chojnacki, S.
S.; Reibenspies, J. H. Inorg. Chem. 1991, 30, 2418.

Catalytic ActiVity of a Series of Zn(II) Phenoxides
J. Am. Chem. Soc., Vol. 121, No. 1, 1999 115
13C resonance for free 13CO2 can be understood as follows. The
line shape of the 13C signal of CO2 in the complex is independent
of the [CO2]; i.e., the lifetime (τ) of CO2 in complex 9 equals
k-1-1, whereas the line shape of the 13C resonance of free CO2
depends on [CO2], where τ ) [CO2]{k-1[9]}-1
.
Comments and Conclusions
In this paper, we have described the preparation and structural
characterization of a series of monomeric zinc bis phenolate
derivatives containing sterically encumbering substitutents at
the 2,6-positions of the ring to inhibit aggregate formation. These
bis phenoxide complexes were isolated in the solid state as four-
coordinate, highly distorted tetrahedra, where zinc’s coordination
sphere is completed by two labile donor ligands such as THF.
In noninteracting solvents such as benzene or methylene
chloride, these ancillary ligands are extensively dissociated from
the zinc center, with the degree of ligand dissociation being a
function of the substituents on the phenoxide groups. Unexpect-
edly, stronger ligand binding is exhibited by zinc centers
containing electron-donating tert-butyl substituents on the
phenolate ligands as opposed to electron-withdrawing phenyl
substitutents.33 This is a general observation for a wide range
of ancillary ligands (from cyclic ethers to tertiary phosphines)
for both zinc and cadmium monomeric bis phenoxide deriva-
tives.15,22 In all instances, the order of ligand binding is py >
THF > epoxides, although the difference in ∆G° for binding
equilibria between ethers and epoxides is less than 5 kJ/
mol.16,34,35
Figure 9. (A) 13C NMR spectrum of complex 7 in C5D5N. (B) 13C
NMR spectrum of complex 7 in 13CO2 saturated C5D5N. (C) 13C NMR
spectrum of complex 2 in 13CO2-saturated C5D5N.
Hence, the solution properties of these monomeric zinc
species exhibited in noninteracting media, in particular liquid
epoxides and CO2, make them good catalysts for ring-opening
epoxides in either homopolymerization or copolymerization with
carbon dioxide processes to afford polyethers or polycarbonates,
respectively. Indeed, as we have demonstrated herein, these zinc
derivatives are among the very best catalysts for the copolym-
erization of cyclohexene oxide and CO2 for the production of
high-molecular-weight polycarbonates with low levels of poly-
ether linkages.36 The order of catalytic activity as a function of
the phenolate ligands on zinc was found to be 2,4,6-trimeth-
ylphenoxide . 2,6-di-tert-butylphenoxide > 2,6-diphenyl-
phenoxide > 2,6-di-isopropylphenoxide. This trend in catalytic
activity is a balance between an electron-rich metal center, which
is more reactive toward both epoxide ring-opening and CO2
insertion processes, and the steric hindrance about the metal
center, which retards both processes. Although we2,8,9 have come
a long way in developing more active, better defined catalysts
for the copolymerization of selected epoxides and CO2, a process
which affords high-molecular-weight polymers (Mw > 50 000)
with low polydispersity (Mw/Mn < 1.2) for a variety of epoxides
has yet to be uncovered (see Note Added in Proof).
the ipso carbon resonances (a and a*), which would be
anticipated to have similar T1 values, it is possible to conclude
that the insertion reaction (eq 7) lies far to the right. Because
complex 7 is four-coordinate in pyridine solution, as shown by
1H NMR, the resulting carbonate derivative is most likely
monomeric.31 This is to be contrasted with the dimer which
has been isolated and structurally characterized from the reaction
of (2,6-diphenylphenoxide)2Cd(THF)2 and CS2 in benzene.22
Furthermore, the fact that CO2 insertion readily occurs in the
coordinating solvent pyridine is totally consistent with previous
mechanistic aspects in general, where no prior coordination of
CO2 to the metal is necessary for C-O bond formation. At the
same time, strongly interacting solvents render these complexes
inert to ring-opening reaction (eq 6), where coordinating of
substrates is required, i.e., homo- or copolymerization of
epoxides does not take place in coordinating solvents. On the
other hand, upon saturating a pyridene-d5 solution of the
sterically encumbered (2,6-diphenylphenoxide)2 Zn derivative
with 13CO2, no reaction was observed (Figure 9C). This is
similar to our observation previously reported for the cadmium
analogue22 and in group 6 carbonyl derivatives.30
The broadness and slight shift downfield (126.2 vs 125.7
ppm) of the free 13CO2 peak for the reaction defined in eq 7
and depicted in Figure 9B and C as b are consistent with an
equilibrium process with rapid insertion/deinsertion of CO2
occurring.32 This observation also indirectly supports a mono-
meric structure for the insertion product (9), since a dimer
structure would be expected to be more stable toward decar-
boxylation. The sharp 13C signal for the carbonate and broad
Because of the facility of the back-biting reaction of the
intermediate resulting from CO2 insertion into the zinc-promoted
ring-opening of propylene oxide as compared to polymer
growth, these zinc derivatives were only effective at producing
(33) Molecular modeling studies indicate that the two-coordinate zinc
complexes resulting from ancillary ligand dissociation should be too
sterically hindered to afford dimer structures. Furthermore, we have generally
found these aggregate structures, when formed from smaller phenoxide
ligands, not to be soluble in nonpolar solvents such as benzene.
(34) Denisov, V. M.; Kuznetsov, Yu. P. IzV. Akad. Nauk. SSSR, Ser.
Khim. 1975, 25, 2595.
(35) Duman, P.; Guerin, P. Can. J. Chem. 1978, 56, 925.
(36) The only catalyst system which is comparable to the one presented
here is the system described in ref 8.
(31) Unlike what we have reported for the analogous cadmium bis-
(phenoxide) pyridine derivative,16 no phenoxide complex of zinc possessing
a coordination number greater than 4 has been observed.
(32) On the basis of the chemical shift differences in hertz between the
inserted and free CO2, the rate constant for insertion/deinsertion of CO2
into complex 7 in pyridine solution is considerably slower than 3300 s-1 at
ambient temperature.

116 J. Am. Chem. Soc., Vol. 121, No. 1, 1999
Darensbourg et al.
propylene carbonate.37 Nevertheless, these bis phenoxide deriva-
tives of zinc were competent at terpolymerization of cyclohexene
oxide/propylene oxide/CO2, with little cyclic propylene carbon-
ate formation as long as the propylene oxide was maintained at
less than 50% mole fraction. Of importance to a system capable
of catalyzing the production of poly(propylene carbonate), which
in general provides sizable quantities of cyclic propylene
carbonate as a byproduct, we have shown this byproduct to be
a good ligand toward zinc, and much better than an epoxide.
As a consequence, propylene carbonate should be a very
effective competitive inhibitor for these copolymerization
processes. Similarly, the preferential binding of the zinc center
in these bis phenoxide derivatives to donor solvents as compared
to epoxides is amply demonstrated by the observation that
copolymerization reactions attempted in THF or dioxane were
completely unsuccessful.
accessible lone pairs on the phenoxide’s oxygen atom. There-
fore, small substituents (e.g., Me) on the 2,6-positions of the
phenyl ring allow for CO2 insertion to provide a metal aryl
carbonate intermediate, whereas large substituents (e.g., tert-
butyl, phenyl) do not. Hence, in the latter instances, initiation
has to involve epoxide ring-opening prior to CO2 insertion. In
subsequent steps, where the newly formed metal alkoxide is
not sterically congested, CO2 insertion into the metal-O bond
readily occurs. For sure, in the absence of steric inhibition, CO2
insertion (eq 5) is more facile than epoxide ring-opening (eq
6).
Acknowledgment. Financial support of this research by the
National Science Foundation (Grant 96-15866) and the Robert
A. Welch Foundation is greatly appreciated.
Note Added in Proof
Pertinent to the initiation step of the copolymerization process,
we have confirmed that, in the absence of coordinating solvents,
both sterically hindered and nonhindered phenoxide derivatives
of zinc catalyze the homopolymerization of epoxides to poly-
ethers. That is, homopolymerization of epoxides requires a site
on zinc for epoxide binding. Indeed, the tert-butyl-substituted
phenoxide ligand was found to be the most active of the group
investigated herein for homopolymerization of epoxides. On the
other hand, CO2 insertion into the Zn-O(phenoxide) bond does
not require a site on zinc for coordination but necessitates
Subsequent to the submission of this manuscript, a well-
defined, very active zinc catalyst was reported for the copo-
lymerization of cyclohexene oxide/CO2 which provides a
copolymer of low polydispersity (Cheng, M.; Lobkovsky, E.
B.; Coates, G. W. J. Am. Chem. Soc. 1998, 120, 11018).
Supporting Information Available: Ball-and-stick drawings
and tables of anisotropic thermal parameters, bond lengths, and
bond angles for complexes 1-4, 6, and 7 (36 pages, print/PDF).
See any current masthead page for ordering information and
Web access instructions.
(37) In general, for a large variety of catalyst systems, the yields of
copolymer of propylene oxide/CO2 are much less than 50 g of polymer/g
of zinc for conditions similar to those employed here; e.g., see ref 3-7.
JA9826284