Synthesis and Structural Characterization of Double Metal Cyanides of Iron and Zinc: Catalyst Precursors for the Copolymerization of Carbon Dioxide and Epoxides

Article abstract of DOI:10.1021/ic0347900

Several synthetic approaches for the preparation of double metal cyanide (DMC) derivatives of iron(II) and zinc(II) are described. These include (1) metathesis reactions of ZnCl₂ or ZnI₂ with KCpFe(CN) ₂CO in aqueous solut

Full text of DOI:10.1021/ic0347900

Inorg. Chem. 2003, 42, 7809−7818
Synthesis and Structural Characterization of Double Metal Cyanides of
Iron and Zinc: Catalyst Precursors for the Copolymerization of Carbon
Dioxide and Epoxides
Donald J. Darensbourg,* M. Jason Adams, Jason C. Yarbrough, and Andrea L. Phelps
Department of Chemistry, Texas A&M UniVersity, College Station, Texas 77843
Received July 8, 2003
Several synthetic approaches for the preparation of double metal cyanide (DMC) derivatives of iron(II) and zinc(II)
are described. These include (1) metathesis reactions of ZnCl2 or ZnI2 with KCpFe(CN)2CO in aqueous solution,
(2) reactions of KCpFe(CN)2CO and its phosphine-substituted analogues with Zn(CH CN)4(BF ) and subsequent
3
4 2
displacement of acetonitrile at the zinc centers by the addition of a neutral (phosphine) or anionic (phenoxide)
ligand, and (3) reactions of the protonated HCpFe(CN)2(phosphine) complexes with Zn(N(SiMe3)2)2, followed by
the addition of phenols. All structures are based on a diamond-shaped planar arrangement of the Fe2(CN)4Zn2
core with various appended ligands at the metal sites. Although attempts to replace the iodide ligands in
[CpFe(µ-CN)2PPh3ZnI(THF)]2 with acetate using silver acetate failed, two novel cationic mixed-metal cyanide salts
based on the [CpFe(PPh3)(µ-CN)2Zn(NC H )]22+ framework were isolated from pyridine solution and their structures
5
5
were defined by X-ray crystallography. The anionic ligand bound to zinc in these derivatives, which serve as an
anionic polymerization initiator, was shown to be central to the catalytic copolymerization reaction of CO /epoxide
2
to provide polycarbonates and cyclic carbonates. The structurally stabilized phosphine-strapped complexes
[CpFe(µ-CN)2Zn(X)THF]2(µ-dppp), where X ) I or phenolate, were shown to be thermally stable under the conditions
(80 °C) of the copolymerization reaction by in situ infrared spectroscopy. Both of these derivatives were proposed
to serve as mimics for the heterogeneous DMC catalysts in the patent literature, with the derivative where the
initiator is a phenolate being more active for the production of polycarbonates.
Introduction
copolymerizing carbon dioxide and propylene oxide, instead
producing cyclic propylene carbonate. However, in some
The metal-catalyzed copolymerization of carbon dioxide
and epoxides to produce biodegradable polycarbonates offers
an economical and environmentally benign alternative to the
current industrially practiced method for their preparation.
This latter methodology involves the biphasic polyconden-
sation of phosgene (or one of its derivatives) and a diol.1
Much recent research activity has been directed at the design
of well-characterized transition metal complexes as homo-
geneous catalysts or catalyst precursors for the copolymer-
ization of carbon dioxide and cyclohexene oxide to produce
poly(cyclohexylene carbonate).2 Unfortunately, these well-
defined catalysts are in general much less effective at
(2) (a) Darensbourg, D. J.; Holtcamp, M. W. Macromolecules 1995, 28,
7577-7579. (b) Kruper, W. J.; Dellar, D. V. J. Org. Chem. 1995, 60,
725-727. (c) Super, M.; Berluche, E.; Costello, C.; Beckman, E.
Macromolecules 1997, 30, 368-372. (d) Super, M.; Beckman, E. J.
Macromol. Symp. 1998, 127, 89-108. (e) Cheng, M.; Lobkovsky, E.
B.; Coates, G. W. J. Am. Chem. Soc. 1998, 120, 11018-11019. (f)
Beckman, E. Science 1999, 283, 946-947. (g) Darensbourg, D. J.;
Holtcamp, M. W.; Struck, G. E.; Zimmer, M. S.; Niezgoda, S. A.;
Rainey, P.; Robertson, J. B.; Draper, J. D.; Reibenspies, J. H. J. Am.
Chem. Soc. 1999, 121, 107-116. (h) Nozaki, K.; Nakano, K.; Hiyama,
T. J. Am. Chem. Soc. 1999, 121, 11008-11009. (i) Darensbourg, D.
J.; Wildeson, J. R.; Yarbrough, J. C.; Reibenspies, J. H. J. Am. Chem.
Soc. 2000, 122, 12487-12496. (j) Cheng, M.; Darling, N. A.;
Lobkovsky, E. B.; Coates, G. W. Chem. Commun. 2000, 2007-2008.
(k) Inoue, S. J. Polym. Sci., Part A 2000, 38, 2861-2871. (l) Mang,
S.; Cooper, A. I.; Colclough, M. E.; Chauhan, N.; Holmes, A. B.
Macromolecules 2000, 33, 303-308. (m) Cheng, M.; Moore, D. R.;
Reczek, J. J.; Chamberlain, B. M.; Lobkovsky, B. E.; Coates, G. W.
J. Am. Chem. Soc. 2001, 123, 8738-8749. (n) Darensbourg, D. J.;
Yarbrough, J. C. J. Am. Chem. Soc. 2002, 124, 6335-6342. (o)
Eberhardt, R.; Allmendinger, M.; Luinstra, G. A.; Rieger, B. Orga-
nometallics 2003, 22, 211-214. (p) Eberhardt, R.; Allmendinger, M.;
Rieger, B. Macromol. Rapid Commun. 2003, 24, 194-196.
* To whom correspondence should be addressed. E-mail:
djdarens@mail.chem.tamu.edu. Fax: (979) 845-0158.
(1) Engineering Thermoplastics: Polycarbonates, Polyacetals, Polyesters,
Cellulose Esters; Bottenbruch, L., Ed.; Hanser Pub.: New York, 1996;
p 112.
10.1021/ic0347900 CCC: $25.00 © 2003 American Chemical Society
Published on Web 11/01/2003
Inorganic Chemistry, Vol. 42, No. 24, 2003 7809

Darensbourg et al.
instances it is possible to enhance the rate of polypropylene
carbonate production over cyclic propylene carbonate by
carrying out the process at lower temperatures.3,4
The patent literature contains numerous references dem-
onstrating that double metal cyanides (DMC) of the general
formula [Ma]m[Mb(CN)6]n, where Ma is a oxophilic metal and
Mb is a transition metal, show great promise for the coupling
of propylene oxide or ethylene oxide and carbon dioxide.5
Among catalysts of this composition, zinc hexacyanofer-
rate(III) has been particularly effective. It is noteworthy to
state here that the generally Very high actiVity achieVed by
these catalysts for both poly(ether polyol)s and polycarbonate
production is profoundly dependent on the details of the
actiVation steps of these heterogeneous catalysts.6 Neverthe-
less, there is one report in the open literature which describes
an active cobalt or nickel cyanide/zinc catalyst for the
synthesis of poly(ether polyol)s which appears to be less
sensitive to the method of catalyst activation.7 Despite their
impressive activity for homopolymerization of epoxides or
copolymerization of epoxide/CO2, these DMC catalysts are
poorly understood with respect to structure and reaction
pathway.
Herein, we wish to communicate our early findings aimed
at providing a better understanding of the role of double
metal cyanides of iron and zinc in the copolymerization
process of carbon dioxide and epoxides. These studies
involve the synthesis and characterization of several well-
defined, mixed-metal cyanide derivatives of iron(II) and
zinc(II), along with an examination of their reactivity toward
the CO2/epoxides coupling reaction.8
Figure 1. Proposed structure of [CpFe(CO)(µ-CN)2]2[Zn], where bridging
cyanides are drawn as lines for clarity.
cyanide and two carbonyl stretching vibrations at 2139 and
2119 cm-1 and 1986 and 1973 cm-1, respectively. The
observation that the two ν(CN) values are shifted to higher
frequency when compared with those of 1 (2085 and 2095
cm-1) is strongly indicative of both cyanide ligands being
coordinated in bridging modes, i.e., of the form Fe-CtN-
Zn. Nevertheless, this is not necessarily the case, and other
techniques would be essential to definitively describe the
nature of the cyanides bonding modes.
2[K][CpFe(CN)2CO] +
water
[Zn][CpFe(CN) CO]
insoluble precipitate
ZnCl2
8
2 + 2KCl (1)
2
After exhaustive attempts of carrying out reaction 1 under
varying reaction conditions in an effort to provide single
crystals of 2 suitable for X-ray crystallography failed, we
have undertaken solid-state 13C and 15N NMR studies to
better define the structure of complex 2. The 13C resonance
due to the cyanide ligands in 2 was observed at 158.3 ppm,
or about 15.9 ppm upfield from the parent complex 1.
Furthermore, the 15N NMR spectrum of 2 exhibited a single
chemical shift for the two cyanide ligands at 251.7 ppm,
which is shifted downfield from 1 by 64.9 ppm. Hence, the
solid-state NMR results definitively show the two cyanide
ligands to be in the same magnetic environment or to be
chemically equivalent. On the basis of the ν(CN) and 13C/
15N NMR data, coupled with the elemental analysis (i.e., Fe:
Zn of 2:1 and <0.3% Cl), 2 is assigned the structure shown
in Figure 1. The proposed structure of 2 bears a close
resemblance to that found for the zeolitic zinc ferrocyanide,
Na2Zn3[Fe(CN)6]2‚nH2O.9 The basic structure of this network
solid can be described as octahedral FeC6 units surrounded
by tetrahedral ZnN4 units, with sodium ions and water
molecules filling the interstitial cavities.
Although complex 2 lacked the desired solubility to serve
as a homogeneous catalyst, the copolymerization of propy-
lene oxide and carbon dioxide was undertaken with 2
employed as a heterogeneous catalyst for this process.
Complex 2 was activated in a manner similar to that
described in the patent literature.6 That is, 2 was homog-
enized in degassed tert-butyl alcohol. This was followed by
a water wash and a second homogenization step using tert-
butyl alcohol and poly(propylene glycol) monobutyl ether.
The resulting slurry was filtered through a 4-5.5 µm filter
Results and Discussion
Our initial investigations into the synthesis of double metal
cyanides of iron and zinc began with the reaction of [K]-
[CpFe(CN)2CO] (1) with ZnCl2 in water.8 This reaction was
performed by the dropwise addition of an aqueous solution
of 1 to 1 equiv of ZnCl2 in water with vigorous stirring.
The yellow powder isolated from this reaction was insoluble
in water and all common organic solvents. Elemental analysis
of this material proved it to be the simple product of a
metathesis reaction (eq 1), where zinc had replaced potassium
as the counterion to the CpFe(CN)2CO anion. The identical
product was obtained from the reactions of 1 and ZnI2 in
water or 1 and ZnCl2 in acetonitrile. The solid-state (KBr)
infrared spectrum of [Zn][CpFe(CN)2CO]2 (2) exhibited two
(3) Darensbourg, D. J.; Yarbrough, J. C.; Ortiz, C.; Fang, C. C. J. Am.
Chem. Soc. 2003, 125, 7586-7591.
(4) Allen, S. D.; Moore, D. R.; Lobkovsky, E. B.; Coates, G. W. J. Am.
Chem. Soc. 2002, 124, 14284-14285.
(5) (a) Kruper, W. J., Jr.; Swart, D. J. U.S. Patent 4,500,704, 1985. (b)
There is also a similar report in the open literature: Chen, L.-B.
Makromol. Chem. Macromol. Symp. 1992. 59, 75-82.
(6) (a) Milgrom, J. U.S. Patent 3,427,256, 1969. (b) Le-Khac, B. U.S.
Patent 5,482,908, 1996. (c) Ooms, P.; Hofmann, J.; Gupta, P.;
Groenendaal, L U.S. Patent 6,204,357, 2001. (d) Le-Khac, B. U.S.
Patent 6,211,330, 2001. (e) Hofmann, J.; Ooms, P.; Gupta, P.; Schafer,
W. U.S. Patent 6,291,388, 2001.
(7) Garcia, J. L.; Jang, E. J.; Alper, H. J. Appl. Polym. Sci. 2002, 86,
1553-1557.
(8) Darensbourg, D. J.; Adams, M. J.; Yarbrough, J. C. Inorg. Chem. 2001,
40, 6543-6544.
(9) (a) Gravereau, P.; Garnier, E.; Hardy, A. Acta Crystallogr. 1979, B35,
2843-2848. (b) Garnier, E.; Gravereau, P.; Hardy, A. Acta Crystallogr.
1982, B38, 1401-1405.
7810 Inorganic Chemistry, Vol. 42, No. 24, 2003

Double Metal Cyanides of Iron and Zinc
Table 1. IR and 31P NMR Data for Compounds 2-7
compd
condition
ν(CN), cm-1
ν(CO), cm-1
δ, ppm
[CpFe(CO)(µ-CN)2]2[Zn] (2)
KBr
2139, 2119
2134, 2124
2131, 2118
2135, 2120
2133, 2118
2139, 2126
1986, 1973
1982
1992
1986
1987
[CpFe(CO)(µ-CN)2ZnI(CH3CN)]2 (3)
CH3CN
CH3CN
CH3CN
CH3CN
THF
[CpFe(CO)(µ-CN)2Zn(PPh2Me)(CH3CN)]2[BF4]2 (4)
[CpFe(CO)(µ-CN)2Zn(PCy3)(CH3CN)]2[BF4]2 (5)
[CpFe(CO)(µ-CN)2Zn(P(tBu)3)(CH3CN)]2[BF4]2 (6)
[CpFe(CO)(µ-CN)2Zn(2,4,6-OC6H2(CH3)3)(CH3CN)]2 (7)
-21.8
29.6, 11.8
37.3, 31.0
a
1985
a
1H NMR chemical shifts for 7 are as follows: 2.19 (s), 2.21 (s), and 6.78 (s) ppm.
frit and dried overnight in a vacuum oven at 80 °C to provide
the activated catalyst derived from complex 2. The copo-
lymerization attempt utilizing this catalyst produced little
polypropylene carbonate, instead yielding most cyclic pro-
pylene carbonate as evidenced by its ν(CO2) band at 1799
cm-1, which was approximately one-third as intense as the
ν(CO2) absorption of the copolymer at 1749 cm-1. Although
these observations are less than extraordinary, they do
indicate that complex 2 exhibits reactivity properties in
common with the heterogeneous double metal cyanides in
the patent literature.
Since the goal of this research is to prepare well-
characterized DMC catalysts, thus requiring no preactivation
steps, a soluble double metal cyanide (DMC) of iron and
zinc was prepared upon reacting ZnI2 with KCpFe(CN)2CO
Figure 2. Ball-and-stick drawing of complex 8, [CpFe(PPh3)(µ-CN)2ZnI-
(THF)]2.
in acetonitrile (instead of water) to afford complex 3,
[CpFe(CO)(µ-CN)2ZnI(CH3CN)]2. Alternatively, it was
possible to inhibit the formation of metal aggregate and
afford soluble DMC derivatives by adding phosphines to
Zn(CH3CN)4(BF4) prior to reacting it with KCpFe(CN)2CO
in acetonitrile. In this manner the cationic PPh2Me (4), PCy3
(5), and PtBu3 (6) derivatives were synthesized. Table 1 lists
the ν(CN) and ν(CO) infrared data, as well as the 31P NMR
data for all of the complexes prepared. In general the 31P
NMR signals are broad in complexes 4-6, indicative of
exchange between bound and free phosphine groups.
Anionic ligands, such as phenoxide, bound to zinc may
protect it from affording aggregate double metal cyanides
with KCpFe(CN)2CO. Indeed, this was achieved by the
addition of 1 equiv of KCpFe(CN)2CO to a solution of the
mono(phenoxide) derivative of zinc which was prepared from
the reaction of anhydrous zinc chloride and sodium 2,4,6-
trimethylphenoxide in acetonitrile. The absence of the very
stable bis(phenoxide) derivative of zinc was visually ob-
served since it is insoluble under these conditions.2g As noted
in the case of the cationic phosphine derivatives, the ν(CN)
and ν(CO) frequencies were shifted to higher wave-
numbers compared to the parent potassium salt (1). The 1H
NMR spectrum of this phenoxide derivative, complex 7,
[CpFe(CO)(µ-CN)2Zn(2,4,6-OC6H2(CH3)3)(CH3CN)]2, shows
methyl singlets at 2.19 and 2.21 ppm and an aromatic singlet
at 6.78 ppm.
phosphine and L2 ) diphosphine.10 The reaction of ZnI2 with
KCpFe(CN)2PPh3 in acetonitrile provided the complex
[CpFe(PPh3)(µ-CN)2ZnI(THF)]2 (8) in good yield upon
isolation from a THF solution. As expected, the ν(CN)
vibrations of the parent KCpFe(CN)2PPh3 species shift to
higher wavenumbers upon complexation with zinc. Further-
more, these vibrational modes are significantly lower in
frequency at 2092 and 2082 cm-1 than those of the CO
analogue [CpFe(CO)(µ-CN)2ZnI(THF)]2, which appear at
2134 and 2124 cm-1. Presumably, this is a reflection of
enhanced interaction of the nitrogen centers with zinc in
complex 8, thereby accounting for its greater thermal
stability. Single crystals of complex 8 were obtained from a
THF/hexane solution. However, these crystals were of poor
quality and only led to a structural solution with an R value
of 20.8%. Nevertheless, the atom connectivity was ad-
equately defined indicating the iron and zinc metal centers
are linked by cyanide ligands. This structure is also consistent
with infrared and NMR spectral data in solution, as well as
elemental analysis. A ball-and-stick representation of the
structure is depicted in Figure 2.
A mixed-metal cyanide complex similar to 8 containing
slightly lower ν(CN) vibrational modes at 2088 and
2080 cm-1 was synthesized from KCpFe(CN)2PPh3 and
Zn(CH3CN)4(BF4)2 in acetonitrile. This derivative,
[CpFe(PPh3)(µ-CN)2Zn(CH3CN)2BF4]2 (9), is potentially a
Because of the lack of thermal stability (e.g., dissociation
of CO ligand) under the conditions of the copolymerization
process of these latter soluble carbonyl derivatives of DMCs
(complexes 2-7), we have directed our attention at synthe-
sizing more stable phosphine complexes. This was ac-
complished by employing phosphine-substituted derivatives
of KCpFe(CN)2L or [KCpFe(CN)2]2L2, where L ) mono-
-
convenient route to substitute other anions for BF4 , such
as alkoxides or phenoxides. These latter nucleophiles are
good initiators for the copolymerization reaction involving
(10) Darensbourg, D. J.; Adams, M. J.; Yarbrough, J. C.; Phelps, A. L.
Eur. J. Inorg. Chem. 2003, 3639-3648.
Inorganic Chemistry, Vol. 42, No. 24, 2003 7811

Darensbourg et al.
Table 2. Infrared (ν(CN)) and 31P NMR Spectral Data for Complexes 8-16
compd
condition
ν(CN), cm-1
δ, ppm
[CpFe(PPh3)(µ-CN)2ZnI(THF)]2 (8)
CH3CN
THF
CH3CN
KBr
KBr
pyridine
THF
2092, 2082
2091, 2082
2088, 2080
2088, 2084, 2072, 2063
2081, 2062
81.4
80.6
[CpFe(PPh3)(µ-CN)2Zn(BF4)(CH3CN)]2 (9)
[{CpFe(PPh3)(µ-CN)2Zn(NC5H5)}2{CpFe(PPh3)(µ-CN)2}2ZnI(NC5H5)][AgI2] (10a)
[CpFe(µ-CN)2(PPh3)Zn(NC5H5)2]2[AgI2]2 (10b)
2091, 2081
81.6
71.8
[CpFe(µ-CN)2ZnI(THF)]2(µ-dppp) (11)
[CpFe(µ-CN)2Zn(BF4)(THF)]2(µ-dppp) (12)
H[CpFe(PPh3)(CN)2] (13)
2088, 2082 (sh)
2086, 2082 (sh)
2102, 2068
CH3CN
KBr
70.52
81.5a
[HCpFe(CN)2]2(µ-dppp) (14)
[CpFe(PPh3)(µ-CN)2Zn(2,6-OC6H3(tert-butyl)2)(THF)]2 (15)
[CpFe(µ-CN)2Zn(2,6-OC6H3(tert-butyl)2)(THF)]2(µ-dppp) (16)
KBr
THF
THF
2101, 2083
2101, 2080
2101, 2078
a In methanol solvent.
carbon dioxide and epoxides. The BF4- anion is most likely
interacting at the zinc center as inferred from the similarity
of the ν(CN) vibrational modes in compounds 8 and 9. Table
2 contains ν(CN) and 31P NMR data for the mixed-metal
iron phosphine complexes prepared in this study.
In an alternate approach to preparing double metal
cyanides with other anionic nucleophiles bound to the zinc
centers, we have reacted complex 8 with silver acetate. A
precipitate of AgI was readily evident upon reacting complex
8 with 2 equiv of silver acetate. Following solvent removal
via vacuum, the resulting yellow solid was dissolved in
pyridine and the solution allowed to stand at ambient
temperature with slow evaporation of pyridine leading to
crystals of complex 10a, [{CpFe(PPh3)(µ-CN)2Zn(NC5H5)}2-
{CpFe(PPh3)(µ-CN)2}2ZnI(NC5H5)][AgI2]. The solid-state
infrared spectrum of the single crystals of 10a displayed a
broad ν(CN) region with bands at 2088, 2084, 2072, and
2063 cm-1. Upon further standing, the pyridine solution
afforded another set of single crystals, complex 10b,
[CpFe(µ-CN)2(PPh3)Zn(NC5H5)2]2[AgI2]2, which exhibited
a much simpler ν(CN) solid-state spectrum with absorptions
at 2081 and 2062 cm-1.
Figure 3. Ball-and-stick representation of the structure of complex 10a.
Because of the complicated nature of the infrared spectrum
of complex 10a, it was of particular importance to structurally
characterize this derivative via X-ray crystallography. Un-
fortunately, due to several factors, including the quality of
the isolated crystals, the structure of 10a was only refined
to an R value of 18.4%. Nevertheless, the crystal structure
is of sufficient quality with a goodness-of-fit parameter of
1.161 to well-define the important features of the structure.
As depicted in Figure 3, complex 10a was found to consist
of polymeric double metal cyanide chains comprised of zinc/
iron cyanide diamonds connected by bridged iron-zinc
linkages. The coordination sphere of zinc in the diamond
was completed by a pyridine molecule, where in the bridge
by both pyridine and iodide ligands. This polymeric chain
propagates in an undulating fashion due to the bends at the
bridging units. Silver diiodide serves as the counterion for
the complex, which along with solvent molecules has been
omitted from Figure 3 for clarity.
Figure 4. Thermal ellipsoid plot of [CpFe(µ-CN)2(PPh3)Zn(NC5H5)2]2-
[AgI2]2 (10b), with hydrogen atoms and anions removed for clarity.
ligand substitution process at the zinc centers of the diamond
structure. That is, replacement of the ‚‚‚NC-FeCp(PPh3)-
CN ligand in complex 10a by pyridine leads to the dicationic
complex 10b. As seen in all the complexes reported upon
herein and elsewhere, the Fe-C-N angles deviate only
slightly from 180° (Table 3), whereas, the C-N-Zn angles
are far from linearity at 163.4 and 157.7°. It is interesting
that these cyanide bridges are not symmetric, which allows
one of them to form a more linear orientation at the expense
The second batch of crystals obtained from the reaction
of complex 8 with silver acetate in pyridine provided the
structure illustrated in Figure 4. For descriptive purposes this
complex, 10b, can be thought to be derived from 10a by a
7812 Inorganic Chemistry, Vol. 42, No. 24, 2003

Double Metal Cyanides of Iron and Zinc
Table 3. Selected Bond Distances (Å) and Angles (deg) for
[CpFe(µ-CN)2(PPh3)Zn(NC5H5)2]2[AgI2]2, 10b, and
[CpFe(µ-CN)2ZnI(THF)2](µ-dppp), 11a
Complex 10b
Fe(1)-C(1)
Fe(1)-C(2)
Zn(1)-N(1)
Zn(2)-N(2)
Zn(1)-N(3)
Zn(1)-N(4)
1.849(4)
1.847(4)
1.927(3)
1.931(3)
2.023(3)
2.025(3)
Fe(1)-C(1)-N(1)
Fe(1)-C(2)-N(2)
Zn(1)-N(1)-C(1)
Zn(1)-N(2)-C(2)
N(1)-Zn(1)-N(2)
N(3)-Zn(1)-N(4)
Fe(1)-C(1)-N(1)
178.6(3)
176.8(3)
163.4(3)
157.7(3)
118.43(13)
112.76(12)
178.6(3)
Complex 11
Fe(1)-C(1)
Fe(1)-C(2)
Zn(1)-N(1)
Zn(2)-N(2)
Zn(1)-O(1)
Zn(2)-O(2)
1.838(10)
1.851(11)
1.943(8)
1.946(8)
2.058(10)
2.046(5)
Fe(1)-C(1)-N(1)
Fe(1)-C(2)-N(2)
Zn(1)-N(1)-C(1)
Zn(2)-N(2)-C(2)
N(1)-Zn(1)-N(1A)
N(2)-Zn(2)-N(2A)
I(1)-Zn(1)-O(1)
I(2)-Zn(2)-O(2)
178.6(9)
178.5(9)
156.0(8)
167.9(7)
110.8(5)
106.5(4)
106.7(3)
119.7(2)
Figure 6. Comparative illustration of catalytically active zinc site in 11
(top) and heterogeneous (bottom) systems.
a Estimated standard deviations are given in parentheses.
etry, which is largely due to the restrictive length of the
phosphine bridge. The resulting puckered ring provides two
unique zinc environments, as a result of one of the zinc atoms
greater deviation from linearity in the cyanide bridge.
Furthermore, the nature of the bridging phosphine relegates
both phosphine groups to the same face of the metallocycle.
This is a significant departure from the nonbridged double
metal cyanides, which all form the more symmetric and less
sterically hindered complexes with substituents on opposite
faces of the ring. Consistent with these structural observations
we were unable to prepare the analogous complex using the
[KCpFe(CN)2)]2(µ-dppe) salt, whereas, extending the phos-
phine bridge to dppb ((diphenylphosphino)butane) readily
allowed for the synthesis of an analogue of complex 11.
The structural parameters observed in complex 11 are quite
similar to the corresponding bond distances and angles in
Na2Zn3[Fe(CN)6]2‚9H2O.9 For example, selected bond lengths
(Å) and angles (deg) in complex 11 of Fe-C 1.838(10),
Zn-N 1.945(8)Å, Fe-C-N 178.6(9), Zn-N-C 156.0(8)
and 167.9(7), and N-Zn-N 110.8(5) compare well with the
analogous values in Na2Zn3[Fe(CN)6]2‚9H2O: Fe-C 1.881(3),
Zn-N 1.9685(7), Fe-C-N 178.8(7), Zn-N-C 159.7(6),
N-Zn-N 110(4). These results indicate that the model
double metal cyanide derivative, complex 11, is structurally
comparable to the activated form of the heterogeneous DMC
catalysts found in the patent literature.6 Furthermore, the
coordinated THF ligands have been shown to have binding
affinities to zinc only slightly greater than those of epoxides.11
Therefore, the dissolution of 11 in liquid epoxides should
lead directly to the formation of the activated epoxide
complex. This is illustrated in Figure 6, where the most
significant difference between the active zinc centers in the
homogeneous and heterogeneous catalysts systems is the
poorer epoxide ring-opening agent, iodide, in complex 11.
The development of methods for easily replacing the
iodide ligand in complex 11 are therefore necessary for
providing a better model homogeneous catalyst. In a manner
similar to that described for the synthesis of the Fe-PPh3
Figure 5. Thermal ellipsoid plot of [CpFe(µ-CN)2ZnI(THF)]2(µ-dppp) (11)
with the a side view inset provided for clarity.
of the other. The bond angle of N(1)-Zn(1)-N(2) is quite
large at 118°, with the angle formed by the zinc and pyridine
ligands being closer to tetrahedral at 112°. Albeit that
complexes 10a,b possess interesting structural features
(polymeric chain structure of 10a and cationic diamond
structure of 10b) not previously observed, it is apparent that
the synthetic methodology of replacing a zinc-bound halide
with the anion from a soluble silver salt is fraught with
complications. Therefore, other preparative approaches will
need to be explored.
Since the copolymerization reaction of epoxides and
carbon dioxide is generally carried out at temperatures greater
than 60 °C, it is desirable to have thermally robust iron/zinc
double metal cyanide derivatives as catalysts. To achieve
this goal we have employed the phosphine bridged diiron
species, [CpFe(CN)2PPh2(CH2)3PPh2Fe(CN)2Cp]2-, which
can serve as a single side strapping to hold the structure intact
during the catalytic reaction. Complex 11, [CpFe(µ-CN)2ZnI-
(THF)]2(µ-dppp), was prepared from the reaction of the
bridging diiron dianion and ZnI2 in acetonitrile following
isolation and recrystallization from THF.8 The X-ray derived
crystal structure of complex 11 may be seen in Figure 5.
The effects of bridging the iron centers in 11 can be seen
in the metric data for the metallocycle unit. This basket-
shaped molecule displays a nonplanar [Fe(CN)2Zn]2 geom-
(11) Darensbourg, D. J.; Niezgoda, S. A.; Holtcamp, M. W.; Draper, J.
D.; Reibenspies, J. H. Inorg. Chem. 1997, 36, 2426-2432.
Inorganic Chemistry, Vol. 42, No. 24, 2003 7813

Darensbourg et al.
derivative, 9, the tetrafluoroborate analogue of complex 11
was synthesized. Technically, complex 12, [CpFe(µ-CN)2Zn-
(BF4)(THF)]2(µ-dppp), should serve as a good source for
providing more active DMC catalysts via subsequent reaction
with various sodium salts, e.g., sodium phenoxides. However,
the difficulty of removing all of the sodium tetrafluoroborate
byproduct is a disadvantage of this synthetic methodology.
This simple metal salt contamination must be avoided due
to its potential detrimental affect on the copolymerization
process. An alternative method to prepare derivative of
complex 11 with other nucleophiles attached to zinc would
entail the use of protonated iron cyanide complexes in
conjunction with zinc amides.12
The protonated derivative, HCpFe(CO)(CN)2, was first
synthesized by Coffey13 and has since been crystallographi-
cally characterized by us where protonation is shown to occur
at the cyanide nitrogen.14 Despite the synthesis of these
derivatives in aqueous solution, no water was found in the
crystal structure of the protonated dicyanide. This is ben-
eficial since residual water is detrimental to the copolym-
erization process and would have to be removed. The
reactions of KCpFe(CN)2PPh3 and [KCpFe(CN)2]2(µ-dppp)
with excess concentrated HCl in methanol resulted in
formation of the protonated analogues, complexes 13 and
14, H[CpFe(PPh3)(CN)2] and [HCpFe(CN)2]2(µ-dppp), re-
spectively. These protonated metallodicyanides are insoluble
under neutral pH conditions in water and most organic
solvents and, therefore, preclude obtaining solution NMR
data. Solid-state infrared spectra showed that upon protona-
tion the ν(CN) vibrational modes in 13 and 14 shift to higher
frequencies as compared to their potassium salts. A similar
observation was made with the HCpFe(CO)(CN)2 derivative.
The reaction of complex 13 with 1 equiv of Zn[N(SiMe3)2]2
can be visually monitored and shown to yield the silyamide
double metal cyanide complex. This derivative is extremely
moisture sensitive and decomposes rapidly in THF solution.
It was therefore synthesized in situ and directly reacted with
1 equiv of phenol to afford the phenolate complex 15,
[CpFe(PPh3)(µ-CN)2Zn(2,6-OC6H3(tert-butyl)2)(THF)]2. The
analogous diphosphine-bridged complex 16, [CpFe(µ-CN)2Zn-
(2,6-OC6H3(tert-butyl)2)(THF)]2(µ-dppp), was prepared in a
similar fashion from the protonated iron dicyanide 14. Both
phenolate complexes, 15 and 16, have similar ν(CN)
frequencies at 2101, 2080 and 2101, 2078 cm-1, respectively.
Copolymerization Reactions. Complex 11 was explored
as a catalyst or catalyst precursor for the copolymerization
of cyclohexene oxide and carbon dioxide. Although iodide
is not a good initiator, this complex was chosen because it
was crystallographically well-defined and expected to remain
intact under the reaction conditions. This latter expectation
was examined by monitoring complex 11 dissolved in
R-pinene oxide. R-Pinene oxide was selected because of its
similarity to cyclohexene oxide in its binding properties, yet
Figure 7. In situ infrared spectra of the ν(CN) of 11 under simulated
polymerization conditions with reactant dissolution increasing over the first
four scans.
it is an epoxide which is difficult to ring-open and lead to
polyether formation.15 The ν(CN) infrared spectrum of
complex 11 at 80 °C was followed in a glass vessel fitted to
an in situ ReactIR probe over a 24 h time period. As shown
in Figure 7, once the complex completely dissolves in the
epoxide, there is no change in the ν(CN) band positions or
intensities with time, nor do additional ν(CN) bands appear
in the spectrum. Hence, this is strong evidence that complex
11 maintains its core structural integrity under the reaction
conditions of the copolymerization process.
The copolymerization of cyclohexene oxide and carbon
dioxide was carried out as described in the Experimental
Section by employing complex 11 (0.15 g, 0.12 mmol) as
catalyst. An infrared spectrum in the ν(CO2) region of the
crude reaction products obtained in methylene chloride
revealed a peak at 1802 cm-1 for the cyclic carbonate which
was much larger than that of the copolymer at 1750 cm-1
(see Figure 8). Interestingly, the cyclic carbonate produced
was exclusively the cis isomer as revealed by comparison
with an authentic sample synthesized from 1,2-cis-cyclo-
hexanediol and triphosgene.16 Degradation of the growing
polymer chain should lead to production of the trans isomer
of 7,9-dioxabicyclo[4.3.0]nonan-8-one.17 A possible mech-
(12) Darensbourg, D. J.; Phelps, A. L.; Adams, M. J.; Yarbrough, J. C. J.
Organomet. Chem. 2003, 666, 49-53.
(13) Coffey, C. E. Inorg. Nucl. Chem. 1963, 25, 179-185.
(14) Lai, C.-H.; Lee, W.-Z.; Miller, M. L.; Reibenspies, J. H.; Darensbourg,
D. J.; Darensbourg, M. Y. J. Am. Chem. Soc. 1998, 120, 10103-
10114.
(15) Darensbourg, D. J.; Wildeson, J. R.; Lewis, S. J.; Yarbrough, J. C. J.
Am. Chem. Soc. 2002, 124, 7075-7083.
(16) Burk, R. M.; Roof, M. B. Tetrahedron Lett. 1993, 34, 395-398.
(17) Kuran, W.; Listos, T. Macromol. Chem. Phys. 1994, 195, 977-984.
7814 Inorganic Chemistry, Vol. 42, No. 24, 2003

Double Metal Cyanides of Iron and Zinc
Figure 8. Crude copolymerization product afforded using 11 as the
catalyst, with spectrum of an authentic cis-cyclohexene carbonate super-
imposed.
Figure 10. Infrared spectrum of crude reaction product from the
copolymerization of cyclohexene oxide and CO2: (A) using complex 15
as catalyst; (B) using complex 16 as catalyst.
was also shown to be a good model for the heterogeneous
DMC catalyst. Figure 10B depicts the infrared spectrum of
the crude product derived from the copolymerization reaction
catalyzed by complex 16. This spectrum clearly shows trans
cyclic carbonate production, along with a sizable amount of
copolymer (85% carbonate linkages). Hence, these latter
results are similar to the behavior of the patented DMC
catalysts; however, it should be noted that they are consider-
ably less active. That is, these homogeneous systems afford
yields of copolymer < 10% with TON’s < 20 g of polymer/g
of Zn. This is to be compared with the Kruper patent (zinc
ferricyanide catalyst system) for the copolymerization of
cyclohexene oxide/CO2 carried out under similar reaction
conditions, where a 30% conversion (100% copolymer) was
reported.5a
Figure 9. Proposed mechanism for cis cyclic carbonate production.
anism for cis cyclic carbonate formation based on the work
of Kisch18 involves a double inversion at the epoxide carbon
center as illustrated in Figure 9. This process is believed to
be facilitated by the easily displaced iodide nucleophile.
Consequently, cyclic carbonate production occurs at the
expense of the enchainment step, thereby accounting for the
lack of copolymer formation.
In a similar manner the copolymerization reaction of
cyclohexene oxide and carbon dioxide was performed by
employing complex 15 as catalyst which contains the better
initiator, phenoxide, attached to zinc. As illustrated in Figure
10A the infrared spectrum of the crude reaction product in
methylene chloride shows a significant improvement in the
quantity of copolymer (88% carbonate linkages via 1H NMR)
produced relative to cyclic carbonate. Furthermore, the cyclic
carbonate is of the trans configuration as anticipated from
degradation of the growing polymer chain. For comparative
infrared spectra of cis and trans cyclic cyclohexyl carbonate,
see ref 19. The structurally stabilized strapped complex 16
Concluding Comments
Herein, we have demonstrated that it is possible to
synthesize double metal cyanide (DMC) derivatives of
iron(II)/zinc(II) which exhibit modest catalytic activity for
the coupling of carbon dioxide and epoxides to afford
polycarbonates and cyclic carbonates. With utilization of
these complexes, in particular the very thermally stable
phosphine-bridged derivative (16), a working model for the
active site in the activated DMC heterogeneous catalysts is
(19) Darensbourg, D. J.; Lewis, S. J.; Rodgers, J. L.; Yarbrough, J. C. Inorg.
Chem. 2003, 42, 581-589.
(18) Du¨mler, W.; Kisch, H. Chem. Ber. 1990, 123, 277-283.
Inorganic Chemistry, Vol. 42, No. 24, 2003 7815

Darensbourg et al.
pulse widths optimized between 8 and 6 µs. Contract times of 8 µs
were used in all cases. Delay times of both 15N and 13C were set to
6 s.
described. Indeed, this single side strapping by bridging
phosphine ligands represents a new approach to stabilizing
DMC derivatives. This model invokes a tetrahedrally coor-
dinated zinc center containing two bridging cyanide ligands
from the stable anionic iron cyanide moieties, along with an
initiator and a site for epoxide binding/activation. A major
issue in our catalyst design strategy is the incorporation of
the appropriate initiator group. Thus far, we have only been
able to fully characterize a DMC derivative containing an
iodide as initiator. Hence, there is a need for developing
synthetic methods for preparing and structurally defining
DMC complexes possessing better initiators, such as azide
or methoxide groups. Using heterogeneous zinc glutarate
catalysts, Chisholm and co-workers20 have reported hydroxyl
end groups in polypropylene carbonates produced from
propylene oxide and CO2. Similarly, the DMC catalysts in
the patent literature are likely to employ OH- or OR-
initiators on some zinc sites on the basis of their method of
activation. It is hoped with proper catalyst design incorporat-
ing good initiators all zinc sites could be catalytically active
in the homogeneous DMC catalysts. It may also be necessary
to fine-tune the electronics of the bridging cyanide ligands
by way of the iron center, i.e., using modified cyclopenta-
dienyl ligands and various phosphine/phosphite ligands, to
more closely mimic the situation in the heterogeneous
catalytic system. These investigations are ongoing in our
laboratory.
Synthesis of [CpFe(CO)(µ-CN)2]2[Zn]‚H2O, 2. Method I. A
0.2 g (0.83 mmol) amount of KCpFe(CN)2(CO) in 25 mL of H2O
was added dropwise into a vigorously stirring solution of 0.113 g
(0.83 mmol) ZnCl2 over a period of 2 h at room temperature. The
turbid yellow reaction mixture was filtered and washed twice with
water to remove any remaining starting material salts. Residual
water was removed by heating the product under vacuum overnight
to yield the yellow powder 2. Attempts to dissolve 2 in all common
solvents failed, and the same results were noted when using CH3CN
instead of water as the reaction solvent. Anal. Calcd for [CpFe(CO)-
(µ-CN)2]2[Zn]‚H2O: C, 39.59; H, 2.49; N, 11.54; O, 9.89. Found:
C, 39.30; H, 2.15; N, 12.16; O, 10.30; Cl, <0.3.
Method II. Complex 2 can also be synthesized from the ZnI2
salt under aqueous conditions by following the same procedure as
described for method I.
Synthesis of [CpFe(CO)(µ-CN)2ZnI(CH3CN)]2, 3. A slurry of
0.133 g (0.42 mmol) of ZnI2 in 15 mL of CH3CN was added to a
Schlenk flask containing 0.1 g (0.42 mmol) of KCpFe(CN)2(CO)
at room temperature via cannula. The reaction mixture was stirred
for 60 min to afford a yellow solution along with a white precipitate.
The solution was filtered through a frit containing Celite to remove
the KI byproduct. CH3CN was removed under vacuum to yield a
fine yellow powder of 3. Anal. Calcd for C20H16O2N6Fe2Zn2I2 (Mr
) 868.6): C, 27.65; H, 1.86; N, 9.68. Found: C, 28.02; H, 1.92;
N, 10.32.
Synthesis of [CpFe(CO)(µ-CN)2Zn(PR3)(CH3CN)]2[BF4]2,
Where PR3 ) PPh2Me (4), PCy3 (5), and PtBu3 (6). These
derivatives were all prepared in a similar manner from the
appropriate phosphine. In a typical synthesis a solution of 0.4 g
(0.99 mmol) of Zn(CH3CN)4(BF4)2 in 15 mL of CH3CN was added
to 0.2 g (0.99 mmol) of PPh2Me in a Schlenk flask via cannula.
The reaction mixture was stirred at ambient temperature for 1 h
prior to being transferred to a flask containing 0.24 g (0.99 mmol)
of KCpFe(CN)2CO in 10 mL of acetonitrile. An immediate reaction
occurred as evidenced by the formation of a fine white precipitation
of KBF4. The reaction mixture was stirred for an additional 1 h
and filtered through a frit containing Celite to remove the salt
byproduct. Vacuum removal of the acetonitrile solvent afforded a
fine yellow powder of complex 4. Anal. Calcd for C46H42Fe2O2N6-
Zn2P2B2F8 (Mr ) 1188.9): C, 46.47; H, 3.56; N, 7.07. Found: C,
47.02; H, 3.71; N, 7.22.
Synthesis of [CpFe(CO)(µ-CN)2Zn(2,4,6-OC6H2(CH3)3)-
(CH3CN)]2, 7. A 0.079 g (0.5 mmol) amount of sodium 2,4,6-
trimethylphenoxide in acetonitrile was added to a Schlenk flask
containing 0.068 g (0.5 mmol) of ZnCl2 via cannula. The reaction
mixture was stirred for 3 h and the precipitate removed by filtration
through Celite. The solution was then added to 0.120 g (0.5 mmol)
of KCpFe(CN)2(CO) while stirring at ambient temperature for 1 h.
The resulting yellow solution was filtered a second time and
evaporated to dryness under vacuum to yield a yellow powder of
complex 7.
Synthesis of [CpFe(PPh3)(µ-CN)2ZnI(THF)]2, 8. A slurry of
0.067 g (0.21 mmol) of ZnI2 in 20 mL of CH3CN was transferred
slowly to a Schlenk flask containing 0.1 g (0.21 mmol) of KCpFe-
(CN)2(PPh3) in 10 mL of CH3CN via cannula. The solution was
stirred for 1 h at which point a yellow solution with white precipitate
was observed. The solvent was removed under vacuum, and the
solid dissolved in THF. The KI precipitate was then removed by
filtration through a frit containing Celite. The solvent was evapo-
rated under vacuum to give the yellow solid 8 in 74% yield. Anal.
Experimental Section
Methods and Materials. All manuipulations were carried out
using standard Schlenk and drybox techniques under an atmosphere
of argon unless otherwise stated. Prior to their use, hexane, diethyl
ether, THF, benzene, and toluene were distilled over sodium
benzophenone. Methanol was dried by distillation over magnesium
turnings and iodine. Acetonitrile was dried by successive distillation
over CaH2 and P2O5, whereas dichloromethane was distilled over
P2O5. All phosphines were purchased from Strem. Zinc iodide and
all phenols were obtained from Aldrich. Zinc dust was purchased
from Baker Scientific, and nitrosyl tetrafluoroborate was obtained
from Lancaster Synthesis Inc. KCpFe(CN)2CO,13,14 KCpFe-
(CN)2PPh3,10 [KCpFe(CN)2]2(µ-Ph2P(CH2)3PPh2),10 and Zn(CH3-
CN)4(BF4)21 were prepared using the published procedures.
Physical Measurements. All vibrational studies were carried
out on a Mattson 6021 FTIR spectrometer, using a 0.1 mm CaF2
sealed cell. 1H and 31P NMR were recorded on a 300 MHz Varian
Unity Plus spectrometer (121.4 MHz 31P), with acetonitrile as the
solvent unless otherwise stated. Deuterated water was used as the
lock solvent, and all spectra were referenced to an 85% phosphoric
acid solution. All samples for solid-state 15N and 13C NMR
measurements were ground and packed into zirconium oxide rotors
and fitted with Kel-F caps, and spectra were recorded at room
temperature on a 7.05 T Bruker MSL 300 superconducting
spectrometer. The spin rate of the 7 mm MAS probe was controlled
by a Bruker controller. Glycine and adamantane were used as
external references for all 15N and 13C spectra, respectively. Cross
polarization techniques were employed for signal enhancement with
(20) Chisholm, M. H.; Navarro-Llobet, D.; Zhou, Z. Macromolecules 2002,
35, 6494-6504.
(21) Kubas, G. J. In Inorganic Syntheses; Shriver, D. F., Ed.; John Wiley
and Sons: New York, 1979; pp 19, 90-92.
7816 Inorganic Chemistry, Vol. 42, No. 24, 2003

Double Metal Cyanides of Iron and Zinc
Table 4. X-ray Crystallographic Data for Complexes 8, 10a,b, and 11
8
10a
10b
11
formula
fw
cryst system
space group
Z
C58H56Fe2I2N4O2P2Zn2
1399.32
monoclinic
P21/n
C115H95Fe4IN11P4Zn3‚2NC5H5AgI2
C70H60Ag2Fe2I4N8P2Zn2
2041.06
monoclinic
P21/c
C45H52Fe2I2N4O2P2Zn2
1239.15
orthorhombic
Pnma
2821.30
triclinic
P1h
4
2
4
6
a, Å
b, Å
c, Å
10.936(4)
15.982(6)
16.338(6)
14.240(4)
16.089(5)
17.644(5)
63.302(6)
69.002(6)
79.220(6)
3370.1(17)
1.907
13.415(4)
11.005(3)
25.062(7)
29.029(11)
18.924(7)
9.052(4)
R, deg
â, deg
γ, deg
V, Å3
90.362(8)
104.232(5)
2855.3(19)
1.627
3586.4(18)
1.890
4973(3)
2.077
d
calcd, g/cm3
temp, K
110(2)
0.710 73
2.509
1.641
20.84
110(2)
0.710 73
3.560
1.161
18.38
100(2)
0.710 73
3.399
0.824
2.29
293(2)
0.710 73
4.282
1.020
5.41
wavelength, Å
abs coeff, mm-1
goodness of fit on F2
R,a %
Rw,b %
51.87
45.00
3.89
12.76
a R ) Σ|Fo| - |Fc|/Σ|Fo|. b Rw ) {[Σw(Fo - Fc )2]/[Σw(Fo )2]}1/2
.
2
2
2
Calcd for C58H56Fe2Zn2P2N4I2O2 (Mr ) 1399.3): C, 49.78; H, 4.03;
N, 4.00. Found, C, 51.02; H, 4.17; N, 4.37.
solvent was removed under vacuum, and the solid redissolved in
THF leaving behind a precipitate which was separated by filtration
through a frit containing Celite. The solvent was evaporated under
vacuum to give a yellow powder of complex 12 in 64% yield.
Synthesis of [CpFe(PPh3)(µ-CN)2Zn(BF4)(CH3CN)]2, 9. A
solution of 0.1 g (0.25 mmol) of Zn(CH3CN)4(BF4)2 in 15 mL of
CH3CN was transferred to a Schlenk flask containing 0.118 g (0.25
mmol) of KCpFe(CN)2(PPh3) in 10 mL of CH3CN via cannula. A
color change from orange to yellow was seen as the reaction
proceeded for 1 h, with no appearance of a precipitate. The solution
was evaporated to dryness to yield the yellow solid 9 along with
the residual KBF4 byproduct.
Synthesis of [{CpFe(PPh3)(µ-CN)2Zn(NC5H5)}2{CpFe(PPh3)-
(µ-CN)2}2ZnI(NC5H5)][AgI2], 10a, and [CpFe(µ-CN)2(PPh3)Zn-
(NC5H5)2]2[AgI2]2, 10b. A slurry of 0.083 g (0.26 mmol) of ZnI2
in 15 mL of CH3CN was transferred via cannula to a Schlenk flask
containing 0.123 g (0.26 mmol) of KCpFe(CN)2(PPh3) in 15 mL
of CH3CN. The reaction mixture was stirred for 2 h at room
temperature. The KI precipitate was removed by filtration, and the
solvent was removed under vacuum. The solid was then dissolved
in 20 mL of THF and transferred to a Schlenk flask charged with
0.043 g (0.26 mmol) of AgOOCCH3 via cannula and stirred for 1
h. The solution was filtered through a frit containing Celite to
remove the AgI precipitate and evaporated to dryness under vacuum.
The yellow solid was then dissolved in pyridine to form the
polymeric complex 10a. Compound 10b was crystallized from the
same solution as 10a, with the only exception being the length of
time taken for crystal growth.
Synthesis of [CpFe(µ-CN)2ZnI(THF)]2(µ-dppp), 11. A slurry
of 0.076 g (0.24 mmol) of ZnI2 in 20 mL of CH3CN was transferred
to a Schlenk flask containing 0.1 g (0.12 mmol) of [KCpFe(CN)2]2-
(µ-dppp) in 10 mL of CH3CN via cannula. The reaction was stirred
at room temperature for 2 h to give a yellow solution along with a
precipitate. The solvent was removed under vacuum, and the solid
was dissolved in THF. The solution was filtered through a frit
containing Celite and the solvent removed under vacuum to give
the yellow solid of 11. X-ray-quality crystals of 11 were obtained
from THF/hexane at -20 °C. Anal. Calcd for C45H52Fe2I2N4O2P2-
Zn2 (Mr ) 1239.15): C, 43.62; H, 4.23; N, 4.52. Found: C, 43.87;
H, 4.31; N, 4.67.
Synthesis of [CpFe(µ-CN)2Zn(BF4)(THF)]2(µ-dppp), 12. A
solution of 0.45 g (1.1 mmol) of Zn(CH3CN)4(BF4)2 in 15 mL of
CH3CN was added dropwise to a slurry of [KCpFe(CN)2]2(µ-dppp)
in 10 mL of CH3CN via cannula. The reaction mixture was stirred
at room temperature for 4 h to afford a turbid yellow mixture. The
Synthesis of H[CpFe(PPh3)(CN)2], 13. An excess (0.7 mL) of
degassed concentrated HCl was added to a flask containing 0.2 g
(0.42 mmol) of KCpFe(CN)2(PPh3) in 10 mL of MeOH while
stirring at room temperature. Approximately 50 mL of degassed
H2O was added to the flask after 1 h resulting in the formation of
a yellow precipitate. The solution was decanted via cannula, and
the precipitate was washed repeatedly with H2O. The product was
then washed with MeOH and dried under vacuum to give a yellow
powder of 13 in 83% yield. Anal. Calcd for C25H21N2PFe (Mr )
436.27): C, 68.83; H, 4.85. Found: C, 68.24; H, 5.01.
Synthesis of [HCpFe(CN)2]2(µ-dppp), 14. An excess (1.0 mL)
of degassed concentrated HCl was added to a flask containing 0.4
g (0.48 mmol) of [KCpFe(CN)2]2(µ-dppp) in 10 mL of MeOH while
stirring at room temperature. To the resulting green solution
approximately 50 mL of H2O was added to afford a yellow
precipitate. The solution was decanted via cannula, and the
precipitate was washed repeatedly with H2O. The product was then
washed with Et2O and dried under vacuum to give a yellow powder
of 14 in 89% yield. Anal. Calcd for C41H38N4Fe2P2 (Mr ) 760.4):
C, 64.76; H, 5.04. Found: C, 65.01; H, 4.93.
Synthesis of [CpFe(PPh3)(µ-CN)2Zn(2,6-OC6H3(tert-butyl)2)-
(THF)]2, 15. A 0.035 g (0.09 mmol) amount of Zn[N(SiMe3)2]2 in
10 mL of THF was transferred to a Schlenk flask containing 0.04
g (0.09 mmol) of 13 in 10 mL of THF via cannula. The reaction
mixture was stirred at room temperature for 15 min. A solution of
0.019 g (0.09 mmol) of 2,6-di-tert-butylphenol in 5 mL of THF
was added via cannula to the reaction mixture, which was stirred
for 30 min. The yellow solution was evaporated to dryness under
vacuum to provide a yellow powder of complex 15.
Synthesis of [CpFe(µ-CN)2Zn(2,6-OC6H3(tert-butyl)2)(THF)]2-
(µ-dppp), 16. A 0.071 g (0.18 mmol) amount of Zn[N(SiMe3)2]2
in 10 mL of THF was transferred to a Schlenk flask containing
0.07 g (0.09 mmol) of 14 in 10 mL of THF via cannula. The
reaction mixture was stirred at room temperature for 15 min. A
solution of 0.037 g (0.18 mmol) of 2,6-di-tert-butylphenol in 5 mL
of THF was added via cannula to the reaction mixture, which was
stirred for 30 min. The yellow solution was evaporated to dryness
under vacuum to give a yellow powder of complex 16.
Inorganic Chemistry, Vol. 42, No. 24, 2003 7817

Darensbourg et al.
Polymerization Reactions. A sample of catalyst (generally
0.10-0.15 g) was dissolved in 20 mL of epoxide. The solution
was loaded via an injection port into a 300 mL stainless steel Parr
autoclave, which had previously been dried overnight under vacuum
at 80 °C. The reactor was pressurized to 52 bar with carbon dioxide
and heated to 80 °C. After a period of time (usually 24 h), the
reactor was cooled and opened and the products were isolated by
dissolution in CH2Cl2. The crude reaction mixture was initially
analyzed for polycarbonate and cyclic carbonate by infrared
spectroscopy in the ν(CO2) region. The copolymer was characterized
hybridization, with isotropic thermal parameters fixed at 1.2 or 1.5
times the value of the attached atom. All atom scattering factors
were obtained from the International Tables for X-ray Crystal-
lography, Vol. C.
Structure solutions for compounds 8, 10a,b, and 11 were obtained
using the following programs: data collection, SMART;22 cell
refinement, SAINT;23 data reduction, SAINT;23 structure solution,
SHELXTL-Plus;24 molecular graphics, SHELXTL-Plus.25
Acknowledgment. Financial support from the National
Science Foundation (Grant Nos. CHE 99-10342 and CHE
02-34860 and CHE 98-07975 for purchase of X-ray equip-
ment), the Robert A. Welch Foundation, and the Texas
Advanced Research Technology Program (Grant No. 0390-
1999) is greatly appreciated.
1
by H NMR spectroscopy, where protons adjacent to carbonate
linkages afford a signal at 4.6 ppm, while the polyethers appear at
3.5 ppm.
X-ray Crystallography. Yellow single crystals of compounds
8, 10a,b, and 11 were coated with Paratone and mounted on a glass
fiber using Apiezon grease at room temperature. The crystals were
then placed on a Bruker Smart 1000 diffractometer equipped with
a CCD detector in a nitrogen cold stream maintained at 110 K.
More than a hemisphere of data was collected on each crystal over
three batches of exposures using Mo KR radiation (λ ) 0.710 73
Å). A fourth set of data was measured and compared to the initial
set to monitor and correct for decay, which was negligible in all
cases. Details of crystal data collected on compounds 8, 10a,b, and
11 are provided in Table 4.
Supporting Information Available: Complete details of the
X-ray diffraction study of complexes 8, 10a,b, and 11 (CIF). This
material is available free of charge via the Internet at
http://pubs.acs.org.
IC0347900
(22) SMART 1000 CCD; Bruker Analytical X-ray Systems: Madison, WI,
1999.
(23) SAINT-Plus, version 6.02; Bruker Analytical X-ray Systems: Madison,
Systematic absences and intensity statistics were used in space
group determination. All structures were solved using direct
methods. Anisotropic structure refinements were achieved using
full-matrix least-squares techniques on all non-hydrogen atoms. All
hydrogen atoms were placed in idealized positions based on
WI. 1999.
(24) Sheldrick, G. SHELXS-86 Program for Crystal Structure Solution;
Institut fur Anorganische Chemie der Universitat: Gottingen, Ger-
many, 1986.
(25) SHELXTL, version 5.0; Bruker Analytical X-ray Systems: Madison,
WI, 1999.
7818 Inorganic Chemistry, Vol. 42, No. 24, 2003