An exploration of the coupling reactions of epoxides and carbon dioxide catalyzed by tetramethyltetraazaannulene chromium(III) derivatives: Formation of copolymers versus cyclic carbonates

Article abstract of DOI:10.1021/ic801728e

The high catalytic activity of a tetramethyltetraazaannulene (tmtaa) chromium complex toward the copolymerization of cyclohexene oxide and carbon dioxide to discriminatively provide poly(cyclohexylene carbonate) has directed further studies into the capabilities of the catalyst system. Various [PPN]X (PPN⁺ = bis(triphenylphosphoranylidene)-ammonium) cocatalysts, where X = Cl, N₃, Br, CN, and OBzF₅, in the presence of (tmtaa)CrCl were examined for catalytic reactivity and selectivity for polycarbonate formation, achieving turnover frequencies of 1500 h^-1 at 80°C in the case of PPNCI. The catalyst system was examined under varied pressures and found to be active even at 1 bar of CO₂ pressure. In addition to cyclohexene oxide, the (tmtaa)CrCl complex was investigated for catalytic activity toward the coupling of carbon dioxide with propylene oxide, isobutylene oxide, 1,2-epoxyhexane, styrene oxide, and 4-vinyl cyclohexene oxide. Activation energies were found for the copolymerization reaction between cyclohexene oxide and carbon dioxide utilizing the tetramethyltetraazaannulene catalyst system to be 67.1 ± 4.2 kJ·mol^-1 and 65.2±2.5 kJ·mol^-1 in neat epoxide and with methylene chloride cosolvent, respectively, upon monitoring these processes by in situ infrared spectroscopy. Supplementary to the studies involving (tmtaa)CrCl, electronic effects at the metal center on catalytic activity were examined through derivatization of the tmtaa ligand, resulting in increased activity as electron-donating substituents were added.

Full text of DOI:10.1021/ic801728e

Inorg. Chem. 2008, 47, 11868-11878
An Exploration of the Coupling Reactions of Epoxides and Carbon
Dioxide Catalyzed by Tetramethyltetraazaannulene Chromium(III)
Derivatives: Formation of Copolymers versus Cyclic Carbonates
Donald J. Darensbourg* and Shawn B. Fitch
Department of Chemistry, Texas A&M UniVersity, College Station, Texas 77843
Received September 8, 2008
The high catalytic activity of a tetramethyltetraazaannulene (tmtaa) chromium complex toward the copolymerization
of cyclohexene oxide and carbon dioxide to discriminatively provide poly(cyclohexylene carbonate) has directed
further studies into the capabilities of the catalyst system. Various [PPN]X (PPN+ ) bis(triphenylphosphoranylidene)-
ammonium) cocatalysts, where X ) Cl, N3, Br, CN, and OBzF5, in the presence of (tmtaa)CrCl were examined for
catalytic reactivity and selectivity for polycarbonate formation, achieving turnover frequencies of 1500 h-1 at 80 °C
in the case of PPNCl. The catalyst system was examined under varied pressures and found to be active even at
1 bar of CO2 pressure. In addition to cyclohexene oxide, the (tmtaa)CrCl complex was investigated for catalytic
activity toward the coupling of carbon dioxide with propylene oxide, isobutylene oxide, 1,2-epoxyhexane, styrene
oxide, and 4-vinyl cyclohexene oxide. Activation energies were found for the copolymerization reaction between
cyclohexene oxide and carbon dioxide utilizing the tetramethyltetraazaannulene catalyst system to be 67.1 ( 4.2
kJ · mol-1 and 65.2 ( 2.5 kJ· mol-1 in neat epoxide and with methylene chloride cosolvent, respectively, upon
monitoring these processes by in situ infrared spectroscopy. Supplementary to the studies involving (tmtaa)CrCl,
electronic effects at the metal center on catalytic activity were examined through derivatization of the tmtaa ligand,
resulting in increased activity as electron-donating substituents were added.
Introduction
benign route toward the production of polycarbonates.
Currently, polycarbonates are produced industrially by the
interfacial reaction between a caustic solution containing a
diol, typically bisphenol A, and phosgene in a chlorinated
solvent or the neat reaction between a diol and diphenyl
carbonate, where the diphenyl carbonate is typically syn-
thesized using phosgene. The resultant polycarbonate pos-
sesses excellent marks for impact strength, optical clarity,
durability, and processability and has found useful applica-
tions in optics, electronics, sheeting, and the medical and
automotive industries, among others.2 This process poses
several environmental concerns. Carcinogenic chlorinated
solvents are used in the biphasic reaction, while phosgene
gained infamy in World War I when it was employed as a
chemical weapon. Coming into a more recent spotlight,
bisphenol A is currently under scrutiny for leaching from
poly(bisphenol A) carbonate products and causing hormone
In recent years, carbon dioxide has received much attention
in both the scientific community and the general public. With
atmospheric concentrations of CO2 on the rise and the
impending threat of “global climate change”, many scientists
are looking for ways to reduce CO2 emissions. Others are
concentrating their efforts on capture and sequestering or
utilizing this greenhouse gas for industrial or health and
safety benefits. Carbon dioxide has found growing applica-
tion in refrigerants, water treatment, dry-cleaning, the food
and agriculture industry, and chemical synthesis, either as a
solvent or reagent or for chemical separation.1
For over a decade, our research efforts have focused on
the utilization of carbon dioxide for a more environmentally
* Author to whom correspondence should be addressed. Fax: (979) 845-
0158. E-mail: djdarens@mail.chem.tamu.edu.
(1) (a) De Simone, J. Greenhouse Gas Control Technologies, Proceedings
of the International Conference on Greenhouse Gas Control Technolo-
gies, 6th, Kyoto, Japan, Oct. 1-4, 2002; Elsevier Ltd.: Oxford, U.K.,
2003; p 33. (b) Vansant, J. In RecoVery and Utilization of Carbon
Dioxide; Aresta, M., Ed.; Klumer Academic Publishers: The Nether-
lands, 2003; pp 3-50.
(2) Advances in Polycarbonates. Proceedings of Symposium of the
American Chemical Society, March 2003, Washington, DC; Brunelle,
D. J., Korn, M. R., Eds.; American Chemical Society: Washington,
DC, 2005; ACS Symposium Series 898, pp281.
11868 Inorganic Chemistry, Vol. 47, No. 24, 2008
10.1021/ic801728e CCC: $40.75  2008 American Chemical Society
Published on Web 11/13/2008

Coupling Reactions of Epoxides and Carbon Dioxide
disruption and reproductive damage to infants.3 The afore-
mentioned, more environmentally benign route for the
production of polycarbonates under investigation in our
laboratories, shown in eq 1, involves the alternative coupling
of epoxides to carbon dioxide, facilitated by a metal catalyst.
This subject has been reviewed comprehensively4a,b and
includes studies involving both transition and main group
metal catalysts.4c,d
Figure 1. Skeletal representation and crystallographic structure of the
(tmtaa)CrCl catalyst for the copolymerization of epoxides and carbon
dioxide.
Scheme 1
The catalyst system presented in this correspondence
utilizes a well-studied tetramethyltetraazaannulene (tmtaa)
ligand
architecture.5
5,7,12,14-Tetramethyldibenzo-
[b,i][1,4,8,11]tetraazacyclotetradecine was first synthesized
by Jaeger6 and later modified by Goedken and co-workers,7
involving the nickel-template condensation of acetylacetone
and o-phenylenediamine. The resultant macrocycle contains
a four-nitrogen donor core similar to that of porphyrin and,
like porphyrin, can bind a plethora of transition metals, both
early and late.5 Despite their similarities, tmtaa has a pore
size ∼0.1 Å smaller than that of porphyrin. Steric interactions
between the methyl groups and aromatic hydrogens cause
the ligand to adopt a saddle shape, which, concomitantly,
redirects the bonding orbitals of the donor plane slightly
upward. This pair of ligand attributes results in a bound metal
atom residing above the donor plane. The degree of deviation
is largely dependent upon the size of the metal ion, leading
to the largest deviations with early transition metals and little
to no deviation with later metals. The nonplanarity of the
tmtaa ligand also grants considerable solubility in common
organic solvents without the need for derivation, a major
drawback with porphyrin complexes. Another advantage of
the tmtaa ligand is its relative low cost compared to that of
porphyrins, with the tmtaa ligand being about 10-fold less
expensive. The complexes of interest to our group incorporate
a chromium(III) metal center, as an expansion of previous
and ongoing work with the chromium(III) salen systems. The
(tmtaa)CrCl complex (Figure 1), first reported by Cotton and
co-workers,8 has previously been communicated by us to
be a highly active and robust catalyst for the copolymeri-
zation of cyclohexene oxide and carbon dioxide.9 Further-
more, it is assumed to operate in the presence of cocatalysts
such as PPNX (PPN+ ) (Ph3P)2N+) by a similar mechanism
to that of its (salen)CrX counterparts (Scheme 1).
The scope of this presentation is to report the further
investigation and optimization of the (tmtaa)CrCl system as
it pertains to ligand derivatization, the nature of the cocata-
lyst, and dependence on temperature and pressure, and to
employ different epoxides with the hope of accessing an
assortment of polycarbonates and ascertaining their inherent
properties. Comprehensive kinetic studies of the copolym-
erization reaction performed both with a cosolvent and under
solventless conditions as monitored by in situ high-pressure
infrared spectroscopy are also included in this study.
Experimental Section
Reagents and Methods. All chemicals and solvents were
purified using the appropriate methods immediately prior to use
unless otherwise noted. 2,4-Pentanedione, o-phenylenediamine, 4,5-
diamino-o-xylene, 4,5-dichloro-o-phenylenediamine, nickel acetate
tetrahydrate, chromium(III) chloride, ammonium hexafluorophos-
phonate, bis(triphenylphosphoranylidene)ammonium chloride, so-
dium azide, potassium bromide, sodium cyanide, and sodium
pentafluorobenzoate were purchased from Aldrich, TCI, Alfa Aesar,
Strem, and Acros chemical companies and used without further
purification. Unless otherwise specified, all manipulations were
carried out on a double-manifold Schlenk vacuum line under an
(3) Mullin, R.; Morrisey, S. Chem. Eng. News 2008, 28, 11.
(4) (a) Darensbourg, D. J. Chem. ReV. 2007, 107, 2388–2410. (b) Coates,
G. W.; Moore, D. R. Angew. Chem., Int. Ed. 2004, 43, 6618. (c) Xiao,
Y.; Wang, Z.; Ding, K. Chem.sEur. J. 2005, 11, 3668–3678. (d) Xiao,
Y.; Wang, Z; Ding, K. Macromolecules 2006, 39, 128–137.
(5) (a) Cotton, F. A.; Czuchajowska, J. Polyhedron 1990, 9, 1553–2566.
(b) Mountford, P. Chem. Soc. ReV. 1998, 27, 105–116.
(6) Jaeger, E. G. Z. Anorg. Chem. 1969, 364, 177–191.
(7) Goedken, V. L.; Molin-Case, J. A.; Whang, Y.-A. J. Chem. Soc., Chem.
Commun. 1973, 337–338.
(8) Cotton, F. A.; Czuchajowska, J.; Falvello, L. R.; Feng, X. Inorg. Chim.
Acta 1990, 172, 135–136.
(9) Darensbourg, D. J.; Fitch, S. B. Inorg. Chem. 2007, 46, 5474–5476.
1
atmosphere of argon or in an argon-filled glovebox. H and 13C
NMR spectra were acquired on a Varian Inova 300 MHz super-
Inorganic Chemistry, Vol. 47, No. 24, 2008 11869

Darensbourg and Fitch
conducting NMR spectrometer. Infrared spectra were recorded using
a Bruker Tensor 27 FTIR spectrometer.
activated mixture was then dissolved in 20 mL of neat epoxide
and injected via the injection port into the autoclave that had been
heated to the appropriate reaction temperature. A 128-scan back-
ground spectrum was collected followed by charging of the
autoclave with CO2 to the desired pressure. Concurrently with the
autoclave being charged with CO2, data collection started with a
single 128-scan spectrum taken every 3 min during the course of
the reaction. Profiles of the absorbance at 1750 cm-1 (polycarbon-
ate) and ∼1808 cm-1 (cyclic carbonate) versus time were recorded
after baseline correction. After cooling and venting the autoclave,
the polymer was extracted with dichloromethane and the solution
allowed to evaporate. The unpurified mixture was then redissolved
in a minimal amount of dichloromethane and precipitated with a 1
M methanol solution of hydrochloric acid. Upon settling, the
supernatant solution containing the catalyst and cyclic carbonate
was decanted and discarded. The purified polymer that remained
was dried in vacuo at 60 °C overnight. The highest catalytic
activities were obtained from the crystalline catalyst. All other
activities reported herein are for comparative purposes only.
For the copolymerizations conducted for the purposes of co-
catalyst comparison, the same procedure outlined above was used.
The various PPNX salts used include azide, chloride, cyanide,
bromide, and pentafluorobenzoate. In each run, 50 mg of catalyst
with either 1 or 2 equiv of PPNX cocatalyst were used along with
20 mL of cyclohexene oxide and 500 psi of carbon dioxide,
maintained at 80 °C throughout the reaction.
Investigations to determine how chloro and methyl substituents
affect the electronic nature, and ultimately catalyst activity, of the
ligand system were carried out at 80 °C with the catalyst, 2 equiv
of PPNN3, 20 mL of cyclohexene oxide, and 500 psi of carbon
dioxide. The amount of catalyst was adjusted to maintain a constant
monomer-to-initiator ratio (M/I ) 1700); therefore, 57 mg and 66
mg of Cr(omtaa)Cl and Cr(tmtaaCl4)Cl were used, respectively.
The copolymerizations done for the pressure dependence studies
were conducted at 80 °C using 50 mg of (tmtaa)CrCl catalyst, 2
equiv of PPNCl, 20 mL of cyclohexene oxide, and the desired
pressure of carbon dioxide. The different epoxides attempted for
the copolymerization reaction were treated in the same manner as
previously prescribed, involving 50 mg of (tmtaa)CrCl catalyst, 2
equiv of PPNCl as a cocatalyst, 20 mL of epoxide, and 500 psi of
carbon dioxide. The reaction temperature varied depending on the
epoxide used.
Cosolvent copolymerizations were conducted using the same
procedure outlined above with the exception that the reaction
mixture was comprised of 75 mg of catalyst, 2 equiv of PPNN3
cocatalyst, 10 mL of epoxide, and 10 mL of dichloromethane in
order to keep all reaction products in solution. The reactions were
carried out at 500 psi of carbon dioxide and a range of temperatures.
Reaction rates were monitored by the appearance of the carbonate
bands in the infrared for the copolymer and cyclic carbonate.
Copolymerizations conducted for the kinetic study of (tmtaa)CrCl
utilizing an anionic cocatalyst in neat epoxide were carried out at
the desired temperature using 50 mg of catalyst, 2 equiv of PPNN3,
20 mL of cyclohexene oxide, and 500 psi of carbon dioxide. A
best fit line was applied to the first few minutes of each copolymer
and cyclic carbonate profiles of each polymerization and used to
determine their respective initial rates. These initial rates were used
to calculate the energies of activation for polycarbonate and cyclic
carbonate formation.
Synthesis of 5,7,12,14-Tetramethyldibenzo[b,i][1,4,8,11]-
tetraazacyclotetradecine. The ligand synthesis used was as previ-
ously reported by Goedken and Weiss10 and is as follows. A total
of 2 equiv of o-phenylenediamine was stirred with 1 equiv of
Ni(OAc)2 ·4H2O in refluxing methanol for 5 min, resulting in a
dark blue solution. An aqua-green solution was obtained upon the
addition of 2 equiv of 2,4-pentanedione. The reaction was refluxed
for 3 days, cooled at -4 °C, and filtered. The purple microcrystalline
Ni(tmtaa) was washed with methanol, dried in vacuo, and weighed.
This complex was then suspended in absolute ethanol, and
anhydrous HCl was bubbled through the slurry overnight. The
reaction flask was cooled and the white precipitate filtered and
collected. This solid, a tetrachloronickelate salt of the ligand, was
dissolved in water, and 3 equiv of ammonium hexafluorophospho-
nate (based on Ni(tmtaa)) was added to precipitate the ligand as a
bis(hexafluorophosphonate) salt. The neutral ligand was obtained
by the dropwise addition of triethylamine to a methanol solution
-
of the PF6 salt until the golden yellow product no longer
precipitated from solution. The neutral ligand, H2tmtaa, was filtered,
dried in the air, and weighed (36.7% overall yield).
Synthesis of 5,7,12,14-Tetramethylbis-4,5-dimethylbenzo-
[b,i][1,4,8,11]tetraazacyclo-tetradecine
(H2omtaa)
and
5,7,12,14-Tetramethylbis-4,5-dichlorobenzo[b,i][1,4,8,11]-tetra-
azacyclotetradecine (H2tmtaaCl4). The syntheses of H2omtaa
(octamethyltetraazaannulene) and H2tmtaaCl4 were adapted from
literature procedures11 and followed the procedure outlined above
with the appropriate substituted o-phenylenediamines.
Synthesis of (tmtaa)CrCl (1), (omtaa)CrCl (2), and
(tmtaaCl4)CrCl (3). Catalysts were synthesized following the
literature procedure for (tmtaa)CrCl reported by Cotton and co-
workers, whereby the appropriate free ligand and anhydrous CrCl3
were refluxed in benzene along with 2 equiv of triethylamine for 3
days.8 The dark burgundy solution was filtered and the solvent
removed in vacuo. The resultant material was triturated with
hexanes to extract any unreacted ligand. The product is somewhat
moist, air-sensitive, and typically synthesized prior to use as a
catalyst. Isolated yields for the synthesis of (tmtaa)CrCl, (omtaa)CrCl,
and (tmtaaCl4)CrCl were 62%, 84%, and 77%, respectively, and
are largely dependent on the quantity of benzene used to keep the
product in solution. Anal. calcd for C26H30N4CrCl·2H2O: C, 59.82;
H, 6.56; N, 10.73. Found: C, 59.88; H, 6.66; N, 10.19. Anal. calcd
for C22H18N4CrCl5 ·2H2O: C, 43.77; H, 3.67; N, 9.28. Found: C,
44.15; H, 4.45; N, 8.78.
Synthesis of PPNX (X ) N3, Br, CN, or OBzF5). The syntheses
were slight modifications of those previously reported in the
literature.12 PPNCl was stirred overnight with a slight excess of
NaX in absolute ethanol. The solution was filtered and the filtrate
dried in vacuo. The salt was recrystallized from dichloromethane
and ether, filtered, and dried in vacuo.
Copolymerization of Epoxides and Carbon Dioxide. High-
pressure copolymerizations were carried out using a stainless steel
Parr autoclave modified with a SiComp window to allow for
attenuated total reflectance spectroscopy using infrared radiation
on an ASI ReactIR 1000 in situ probe. For each copolymerization
reaction, the desired amount of catalyst and the appropriate
equivalent(s) of PPNX cocatalyst were dissolved in dichloromethane
and allowed to stir for 30 min before being dried in vacuo. This
X-Ray Structural Studies. A Leica microscope, equipped with
a polarizing filter, was used to identify suitable crystals from a
sample of crystals from the same habit. The representative crystal
was coated in a cryogenic protectant, such as paratone, and affixed
(10) Goedken, V. L.; Weiss, M. C. Inorg. Synth. 1980, 20, 115.
(11) Place, D. A.; Ferrara, G. P.; Harland, J. J.; Dabrowiak, J. C.
J. Heterocycl. Chem. 1980, 17, 439–443.
(12) L’Epplatenier, F. A.; Pugin, A. HelV. Chim. Acta 1975, 58, 917–929.
11870 Inorganic Chemistry, Vol. 47, No. 24, 2008

Coupling Reactions of Epoxides and Carbon Dioxide
Table 1. Crystallographic Data for Complex 2 and H2tmtaaCl4
(omtaa)CrCl (2)
H2tmtaaCl4
empirical formula
fw, g/mol
temp (K)
wavelength
cryst syst
space group
a (Å)
C26H30ClCrN4
485.99
293(2)
C22H20Cl4N4
482.22
110(2)
0.71073
monoclinic
P21/n
9.644(4)
7.846(4)
28.364(13)
90
0.71073
triclinic
j
P1
8.974(11)
12.227(15)
13.150(16)
79.57(2)
75.63(2)
76.14(2)
1346(3)
2
b (Å)
c (Å)
R (deg)
ꢁ (deg)
98.908(4)
90
2120.1(17)
4
γ (deg)
cell volume (Å3)
Z
density (calcd)
1.199
0.543
1.511
0.577
abs coeff (mm-1
)
obsd no. of reflns
no. of unique reflns (I > 2σ)
GoF
9863
4526
1.000
8.00
18086
3615
1.002
8.79
22.31
R,a % [I > 2σ]
Rw,a % [I > 2σ]
11.79
Figure 2. X-ray crystal structures of H2tmtaa, H2omtaa, and H2tmtaaCl4,
respectively. Thermal ellipsoids are shown at the 50% probability level with
hydrogens omitted for clarity with the exception of the amine hydrogens.
a R ) ∑|Fo| - |Fc|/∑|Fo|. Rw ) {[∑w(Fo - Fc )2]/[∑w(Fo )2]}1/2
.
2
2
2
to a nylon sample loop attached to a copper mounting pin. The
mounted crystals were then placed in a cold nitrogen stream
(Oxford) maintained at 110 K on either a Bruker SMART 1000,
GADDS, or APEX 2 three-circle goniometer.
Jaeger6 and later modified by Goedken and co-workers.7 The
neutral H2tmtaa ligand was isolated following removal of
the dianionic tmtaa from Ni(tmtaa) in absolute ethanol by
The X-ray data were collected on either a Bruker CCD, GADDS,
or APEX 2 diffractometer and covered more than a hemisphere of
reciprocal space by a combination of three (or nine in the case of
the GADDS) sets of exposures; each exposure set had a different
ꢀ angle for the crystal orientation, and each exposure covered 0.3°
in ω. Crystal data and details on collection parameters are given in
Table 1. The crystal-to-detector distance was 5.0 cm for all crystals.
Crystal decay was monitored by repeating the data collection for
50 initial frames at the end of the data set and analyzing the
duplicate reflections and was found to be negligible. The space
group was determined on the basis of systematic absences and
intensity statistics.13 The structure was solved by direct methods
and refined by full-matrix least-squares on F2. All non-H atoms
were refined with anisotropic displacement parameters. All H atoms
attached to C and N atoms were placed in idealized positions and
refined using a riding model with aromatic C-H ) 0.93 Å, methyl
C-H ) 0.96 Å, and amine N-H ) 0.86 Å and with fixed isotropic
displacement parameters equal to 1.2 (1.5 for methyl H atoms) times
the equivalent isotropic displacement parameter of the atom to
which they are attached. The methyl groups were allowed to rotate
about their local 3-fold axis during refinement.
-
anhydrous HCl, precipitation as the PF6 salt affording
[H4tmtaa][PF6]2, and neutralization with triethylamine. The
chromium(III) derivative was prepared following the litera-
ture procedure reported by Cotton and co-workers from
H2tmtaa and anhydrous chromium(III) chloride in refluxing
benzene in the presence of 2 equiv of triethylamine.8 Other
tetraazaannulene ligands were similarly prepared from the
appropriately substituted o-phenylenediamine and their cor-
responding Cr(III) derivatives synthesized in an analogous
manner to that described for (tmtaa)CrCl. Although most of
the studies reported at this time involved the (tmtaa)CrCl
complex, other derivatives examined for catalytic activity
include complexes 2 and 3, that is, (omtaa)CrCl and
(tmtaaCl4)CrCl, where the crystal structures of H2omtaa and
H2tmtaaCl4 are shown in Figure 2.
The X-ray structures of H2tmtaa and H2omtaa have been
reported elsewhere.7,17 In the course of our studies, we have
structurally characterized H2tmtaaCl4. These three structures
of the protonated ligands exhibit only minimal differences
in bond distances and bond angles, with the largest distinction
being the degree of distortion inherent to the saddle shape.
In order to quantify the magnitude of planar distortion, the
angles between phenylenediamine moieties were measured
and found to be 133.5°, 136.2°, and 140.4° for H2tmtaa,
H2tmtaaCl4, and H2omtaa, respectively. No obvious trend
could be seen from these deviations, which could merely be
due to crystal packing. The chromium(III) complex derived
from H2omtaa, complex 2, has been fully characterized by
X-ray crystallography. A thermal ellipsoid representation of
For all structures, the following were used: data reduction,
SAINTPLUS (Bruker13); program used to solve structures, SHELXS
(Sheldrick14); program used to refine structures, SHELXL-97
(Sheldrick15); molecular graphics and preparation of material for
publication, SHELXTL-Plus version 5.0 (Bruker16) and XSEED.
Results and Discussion
The tetramethyltetraazaannulene (tmtaa) ligand was syn-
thesized by the nickel template condensation of o-phenylene-
diamine with pentane-2,4-dione as described initially by
(13) SAINT-Plus, version 6.02; Bruker: Madison, WI, 1999.
(14) Sheldrick, G. SHELXS-86; Institut fur Anorganische Chemie der
Uniseritat Gottingen: Gottingen, Germany, 1986.
(15) Sheldrick, G. SHELXL-97; Institut fur Anorganische Chemie der
Uniseritat Gottingen: Gottingen, Germany, 1997.
(17) (a) Alcock, N. W.; Cannadine, J. C.; Errington, W.; Moore, P.;
Wallbridge, M. G. H. Acta Crystallogr., Sect. C: Cryst. Struct.
Commun. 1994, 50, 2037–2039. (b) Buzatu, D. A.; Nolan, S. P.;
Stevens, E. D. Acta Crystallogr., Sect. C: Cryst. Struct. Commun. 1995,
51, 1855–1857.
(16) SHELXTL, version 5.0; Bruker: Madison, WI, 1999.
Inorganic Chemistry, Vol. 47, No. 24, 2008 11871

Darensbourg and Fitch
Figure 3. Thermal ellipsoid plot of 2 at 50% probability. Hydrogen atoms
omitted for clarity.
Table 2. Selected Bond Distances and Angles for (tmtaa)CrCl and
(omtaa)CrCl
(tmtaa)CrCla
(omtaa)CrCl
Cr-N1 (Å)
1.9881(18)
1.9745(17)
1.9900(18)
1.9716(17)
2.2607(7)
92.98(7)
81.76(7)
92.96(7)
81.78(7)
0.4235
1.947(6)
2.000(6)
1.995(6)
1.957(7)
2.244(3)
91.8(3)
82.8(3)
91.8(3)
83.2(2)
0.4209
Figure 4. Reaction profiles indicating copolymer and cyclic carbonate
formation with time for the copolymerization of cyclohexene oxide and
CO2 in the presence of (tmtaa)CrCl and 35 bar of carbon dioxide at 80 °C
at various quantities of added PPNCl.
Cr-N2 (Å)
Cr-N3 (Å)
Cr-N4 (Å)
Cr-Cl (Å)
N1-Cr-N2 (deg)
N2-Cr-N3 (deg)
N3-Cr-N4 (deg)
N1-Cr-N4 (deg)
Cr-N4 (Å)
chosen due to their ease of being prepared anhydrously, and
the weakly interacting PPN cation provides an essentially
free anion for interaction with the chromium center and ring-
opening of the epoxide monomer. In our preliminary report
covering the (tmtaa)CrCl catalyst utilizing PPNN3 as a
cocatalyst, we observed maximum catalytic activity when
employing at least 2 equiv of cocatalyst.9 The same was
found to be true for PPNCl. This is illustrated in Figure 4
by the copolymer peak profiles over time for the copolym-
erization of cyclohexene oxide and carbon dioxide in the
presence of (tmtaa)CrCl and the addition of 1, 2, and 4 equiv
of PPNCl.
a Data taken from structure determination in ref .
complex 2 is shown in Figure 3, where the chromium center
resides 0.4209 above the nitrogen donor plane, along with
the atom labeling scheme. Table 2 lists comparative selected
bond distances and bond angles for the closely related
(tmtaa)CrCl and (omtaa)CrCl derivatives. As can be seen
from Table 2, there exist only minor structural differences
between these two complexes.
It should be recalled upon titrating (tmtaa)CrCl with up
to 8 equiv of azide ions in methylene chloride that the
infrared bands assignable to metal-bound azides reached a
maximum absorbance at 2 equiv of azide ions, with a further
addition of azide ions remaining free in solution, as revealed
In a previous publication, we have reported the kinetic
parameters for the copolymerization of cyclohexene oxide
and propylene oxide with carbon dioxide employing (salen)-
CrCl as a catalyst.18 In a later report, we compared the use
of various cocatalysts (N-heterocyclic amines, phosphines,
and salts containing anionic initiators) for enhancing the
effectiveness of the (salen)CrCl catalyst system and clearly
established that greater catalytic activity is noted when
employing salts possessing nucleophilic anions.19 Hence, for
the (tmtaa)CrCl catalytic studies, only anionic initiators were
utilized. Despite its sensitivity to aerobic environments, the
catalyst is very robust under polymerization conditions and
does not appear to degrade over time, as evidenced by the
constant rate of polymer formation observed from peak
profiles obtained from in situ infrared monitoring of the
copolymerization process.
9
by an infrared band at 2005 cm-1. The (tmtaa)Cr(N3)2
-
anionic complex exhibits a strong asymmetric stretching
vibration of NNN at 2043 cm-1, with a shoulder at 2057
cm-1. Similar (salen)chromium(III) complexes, including the
(salen)CrCl2- species, have been fully characterized by X-ray
crystallography.20 Hence, it is assumed that (tmtaa)CrCl in
-
the presence of excess chloride ions exists as (tmtaa)CrCl2 ,
which in the presence of the epoxide monomer take part in
the equilibrium process depicted in eq 2.
Copolymerization Reactivity Studies Utilizing
Various PPNX Salts as Initiators. In efforts aimed at
developing an optimal catalytic system, several anionic
initiators were examined for their effectiveness as cocatalysts
for the copolymerization of cyclohexene oxide and carbon
dioxide in the presence of (tmtaa)CrCl. For this study, the
PPN(bis(triphenylphosphoranylidene)ammonium) salts were
As is apparent from the reaction profiles in Figure 4, the
catalytic activity for the production of poly(cyclohexylene
carbonate) dramatically increased when the concentration of
PPNCl was increased from 1 to 2 equiv, with only a
negligible change upon further increasing it to 4 equiv. The
slight initiation period observed for the copolymerization
reaction employing 4 equiv of PPNCl might be attributed to
(18) (a) Darensbourg, D. J.; Yarbrough, J. C. J. Am. Chem. Soc. 2002,
124, 6335–6342. (b) Darensbourg, D. J.; Yarbrough, J. C.; Ortiz, C. G.;
Fang, C. C. J. Am. Chem. Soc. 2003, 125, 7586–7591.
(19) Darensbourg, D. J.; Mackiewicz, R. M. J. Am. Chem. Soc. 2005, 127,
14026–14038.
(20) Darensbourg, D. J.; Moncada, A. I. Inorg. Chem. 2008, in press.
11872 Inorganic Chemistry, Vol. 47, No. 24, 2008

Coupling Reactions of Epoxides and Carbon Dioxide
In general, for runs utilizing chloride or azide initiators,
13-15 g (∼50% conversion) of purified copolymer was
isolated from a 1.0 h reaction period. The molecular weight
and polydispersity of the copolymers in THF were measured
by gel permeation chromatography. Under these reaction
conditions, the typical Mn value was found to be 23 000 with
a polydispersity index (PDI) of 1.07. A more detailed
presentation of Mn as a function of the number of equivalents
of initiator may be found in our earlier report on this process.9
Copolymerization Optimization through Catalyst
Derivatization. Unlike chromium(III) salen complexes,
which require the addition of substituents on the phenolate
moieties to achieve sufficient solubility, tmtaa complexes do
not require such derivatization. These added ligand modifica-
tions, aside from altering solubility, also changed the
electronic structure of the salen ligand, which concomitantly
affected the copolymerization activity of the catalyst. For
the salen ligand, it was found that the addition of electron-
donating groups to the phenolate moieties and the ligand
backbone increased the catalyst’s activity toward the co-
polymerization of cyclohexene oxide and carbon dioxide.22
In order to further optimize the chromium(III) tmtaa system,
ligands were synthesized with either electron-withdrawing
chloro or electron-donating methyl groups on the phenylene-
diamine portions of the ligand framework. The expectation
was for these substituents to affect the electron density around
the chromium center, thereby influencing the activity of the
catalyst.
Figure 5. Reaction profiles for the copolymer formation with time for the
copolymerization of cyclohexene oxide and CO2 in the presence of
(tmtaa)CrCl and 35 bar of carbon dioxide at 80 °C employing various PPNX
salts as cocatalysts.
Table 3. Turnover Data for the Copolymerization of Cyclohexene
Oxide and CO2 Utilizing (tmtaa)CrCl as Catalyst and Either 1 or 2
equiv of Various PPN Salt Cocatalystsa
TOF (time (h))
PPNX
1 equiv
2 equiv
Cl
N3
Br
CN
OBzF5
1017b (0.5)
606 (1.55)
542b (0.67)
302 (0.5)
1478b (0.5)
1482b (0.5)
795 (0.67)
1161b (0.5)
656 (1.0)
415 (1.5)
a Copolymerization reactions were conducted with 50 mg of (tmtaa)CrCl
as a catalyst at 80 °C and under 500 psi of CO2 pressure for the duration
indicated. b Runs were conducted using crystalline catalyst.
As might be anticipated on the basis of our earlier studies
involving related (salen)CrCl catalysts, the TOF values for
the copolymerization of cyclohexene oxide and carbon
dioxide with complexes (tmtaa)CrCl (1), (omtaa)CrCl (2),
and (tmtaaCl4)CrCl (3) under comparable reaction conditions
at 80 °C increase in the order 3 < 1 < 2. That is, the catalytic
activity increases as the nucleophilicity of the chromium(III)
center increases, with TOFs at 80 °C being 1290 h-1 (2) >
943 h-1 (1) > 401 h-1 (3). Nevertheless, multiple runs
employing various reaction conditions utilizing complexes
1 and 2 as catalysts in the presence of different anion
initiators reveal only small differences in catalytic activity
between these two chromium(III) derivatives. Hence, because
of the greater ease of synthesizing and purifying complex 1,
we have employed this complex in our more comprehensive
studies of this catalytic system.
Epoxides and Carbon Dioxide Coupling Reactions
as a Function of Carbon Dioxide Pressure. We next
explored the effect of carbon dioxide pressure on the rate of
copolymer formation as well as the extent of CO2 incorpora-
tion, that is, the percentage of ether linkages in thus formed
polycarbonate. It was noted in previous reports that the high
epoxide affinity of our early generation zinc phenoxide
catalysts required high carbon dioxide pressures to counteract
the propensity for polyether formation over that of polycar-
bonate. In contrast, the chromium(III)-based salen catalysts
were shown to be very ineffective at homopolymerizing
competitive metal binding between the epoxide monomer
and chloride ion. The turnover frequencies (TOF; mol
epoxide consumed/mol catalyst · h) when using 1 to 2 equiv
of PPNCl were 1017 and 1478 h-1, respectively. Consistent
with our initial report utilizing PPNN3 as the initiator, the
production of cyclic carbonate is minimal throughout the
reaction until the reaction mixture becomes highly viscous
and the epoxide monomer has been sufficiently depleted. At
this point, the copolymerization process becomes limited by
mass transfer and backbiting of the polymer chain to provide
cyclic carbonate becomes more favorable.
Figure 5 illustrates the reaction profiles for copolymer
formation as a function of various PPNX salts present in 2
equiv relative to (tmtaa)CrCl. Table 3 lists the turnover
frequencies for this copolymerization process in the presence
of both 1 and 2 equiv of these PPNX salts. Comparatively,
it is shown in Table 3 that the chloride and azide anions
gave the highest catalytic activity, followed closely by
cyanide, while pentafluorobenzoate and the bromide ion
displayed the lowest activity of the initiators investigated
herein. The pentafluorobenzoate anion has been previously
reported to be a very efficient cocatalyst for cobalt(III) salen
catalysts for the copolymerization of propylene oxide and
carbon dioxide.21
(21) (a) Cohen, C. T.; Chu., T.; Coates, C. W. J. Am. Chem. Soc. 2005,
127, 10869–10878. (b) Cohen, C. T.; Coates, G. W. J. Polym. Sci.,
Part A 2006, 44, 5182–5191. (c) Cohen, C. T.; Thomas, C. M.; Peretti,
K. L.; Lobkovsky, E. B.; Coates, G. W. Dalton Trans 2006, 237.
(22) Darensbourg, D. J.; Mackiewicz, R. M.; Rodgers, J. L.; Fang, C. C.;
Billodeaux, D. R.; Reibenspies, J. H. Inorg. Chem. 2004, 43, 6024–
6034.
Inorganic Chemistry, Vol. 47, No. 24, 2008 11873

Darensbourg and Fitch
Figure 6. Effect of CO2 pressure on the rate of polycarbonate formation
from the coupling of CO2 and cyclohexene oxide using (tmtaa)CrCl and 2
equiv of PPNCl as a cocatalyst at 80 °C.
epoxides into polyether and, hence, did not require high
concentrations of carbon dioxide to produce polycarbonates
with greater than 95% carbonate linkages. A similar situation
was obvious when utilizing the (tmtaa)CrCl/PPNX catalyst
system. That is, although the rate of copolymerization of
cyclohexene oxide and CO2 varied significantly with CO2
pressure, at 1 atm of CO2 pressure, the copolymer produced
1
contained 97% carbonate linkages as evident by H NMR.
Figure 6 displays the reaction profiles for the copolym-
erization of cyclohexene oxide and CO2 at 80 °C over a range
of CO2 pressures. The concentration of CO2 dissolved in the
cyclohexene oxide medium at atmospheric CO2 pressure was
rapidly depleted, thereby inhibiting continuation of the
copolymerization process. Obviously, a maintained [CO2]
under 1 atm of carbon dioxide pressure would yield an
increase in copolymer production. Similar to what was noted
in the case of the (salen)CrX catalysts, as the CO2 pressure
was decreased, the production of the cyclic carbonate
byproduct increased.23 The most notable instance of signifi-
cant cyclohexylene carbonate formation was when the initial
CO2 pressure was 1 atm. That is, at the higher CO2 pressures,
the production of cyclic carbonate was negligible.
In Situ Kinetic Studies of the Cyclohexene Oxide/
Carbon Dioxide Coupling Reaction. The coupling reaction
of cyclohexene oxide and carbon dioxide as catalyzed by
the (tmtaa)CrCl/PPNN3 catalyst system was monitored by
in situ infrared spectroscopy in order to clearly define the
kinetic parameters for this process. Since the reaction
mixtures for polymerization reactions conducted in neat
epoxide become more viscous and governed by mass transfer
as the reaction progresses, accurate kinetic data for the
process over extended time periods are impossible to obtain.
Hence, in order to keep all components of the reaction in
solution and to thwart any viscosity changes, instead of
employing neat cyclohexene oxide as a solvent, a 1:1 mixture
of cyclohexene oxide and dichloromethane was employed
in these studies initially. Because no appreciable rate
Figure 7. Kinetic traces recorded for the formation of (a) poly(cyclohexyl-
ene carbonate) and (b) cyclohexyl carbonate as a function of the temperature
in methylene chloride. Reaction conditions: (tmtaa)CrCl (1.74 × 10-4 mol)
and 2 equiv of PPNN3 in 10 mL of cyclohexene oxide and 10 mL of
methylene chloride and a CO2 pressure of 34.5 bar.
enhancement was observed in excess of 2 equiv of cocatalyst,
all polymerization runs were performed employing 75 mg
of (tmtaa)CrCl (1.74 × 10-4 moles) and 2 equiv of PPNN3
in 10 mL each of cyclohexene oxide and methylene chloride.
These reactions were conducted over a 30° temperature
range, each process resulting in an asymptotic rate profile
for both copolymer and cyclic carbonate formation (Figure
7). In all cases, the introduction of the cosolvent, methylene
chloride, drastically increased the time needed for the
complete conversion of epoxide to products. Observed
reaction rate constants at each temperature were obtained
for both copolymer and cyclic carbonate formation by fitting
the data in plots of ln[A∞ - At] versus time, where A∞ and
At correspond to the infrared absorbances for the copolymer
(1750 cm-1) and cyclic carbonate (1808 cm-1) at time )
infinity and time ) t, respectively (Figures 8 and 9). The
respective rate constants are listed in Table 4.
Eyring plots of the kinetic data provided in Table 4 are
shown in Figure 10, with the calculated activation parameters
listed in Table 5. As anticipated from the similarities in the
reaction profiles provided in Figure 7, the ∆G‡ values for
copolymer and cyclic carbonate formation in methylene
chloride do not differ significantly at 80 °C, that is, 115.6
and 117.6 kJ ·mol-1, respectively. Although, the enthalpy of
activation for the production of the copolymer (62.2
kJ · mol-1) is less than that for cyclic carbonate production
(23) Darensbourg, D. J.; Mackiewicz, R. M.; Billodeaux, D. R. Organo-
metallics 2005, 24, 144–148.
11874 Inorganic Chemistry, Vol. 47, No. 24, 2008

Coupling Reactions of Epoxides and Carbon Dioxide
Figure 10. Eyring plot for the formation of polymer and cyclic carbonate.
Figure 8. Plots of ln[A∞ - At] vs time as a function of time, where A∞ and
At are the infrared absorbances for the polycarbonate (1750 cm-1) at time
) infinity and time ) t, respectively. R2 values at the various temperatures
(°C) are 70 (0.9990), 80 (0.9995), 90 (0.9998), and 100 (0.9973).
Table 5. Activation Parameters for the Coupling Reactions of
Cyclohexene Oxide and Carbon Dioxide in Methylene Chloridea
copolymer
cyclic carbonate
∆H‡ (kJ·mol-1
∆S‡ (J·mol-1 ·K-1
∆G‡ (kJ·mol-1
Ea (kJ·mol-1
)
62.2 ( 2.5
-151.3 ( 6.9
115.6
77.3 ( 5.7
-114.2 ( 16.0
117.6
)
)
)
65.2 ( 2.5
80.3 ( 2.5
a ∆G‡ value calculated at the typical reaction temperature of 80 °C.
under solVentless conditions. As mentioned earlier, because
of the increase in viscosity during copolymer production in
a neat carbon dioxide swollen cyclohexene oxide medium,
we are only able to quantitatively follow the reaction’s
progress during the early stages. Figure 11 depicts the
reaction profiles for copolymer formation for processes
carried out over the temperature range between 60 and 100
°C. The reactions were monitored via in situ infrared
spectroscopy employing 50 mg (1.16 × 10-4 mol) of
(tmtaa)CrCl and 2 equiv of PPNN3 in 20 mL of cyclohexene
oxide at a CO2 pressure of 34.5 bar. Initial rates for
copolymer and cyclic carbonate were determined over the
first few minutes of the processes. These rate data are
tabulated in Table 6.
Figure 9. Plots of ln[A∞ - At] vs time as a function of time, where A∞ and
At are the infrared absorbances for the cyclic carbonate (1808 cm-1) at
time ) infinity and time ) t, respectively. R2 values at the various
temperatures (°C) are 70 (0.9980), 80 (0.9988), 90 (0.9995), and 100
(0.9992).
Table 4. Temperature-Dependent Rate Constants for the Coupling of
Cyclohexene Oxide and Carbon Dioxide in Methylene Chloridea
Arrhenius plots for both copolymer and cyclic carbonate
formation obtained from the initial rate data listed in Table
6 are provided in Figure 12. The calculated energies of
k (M-1 sec-1) × 103
temp. (°C)
copolymer
cyclic carbonate
70
80
90
3.45
8.44b
12.9
1.37
4.08b
7.31
100
21.0
13.6
a Reaction carried out in 10 mL of methylene chloride, 1.74 × 10-4
mol of (tmtaa)CrCl and 2 equiv of PPNN3 in 10 mL of cyclohexene oxide
and a CO2 pressure of 34.5 bar. b Average value of two runs.
(77.3 kJ · mol-1), as has been previously reported for chro-
mium(III) and zinc(II) catalyst systems, it is offset by a more
negative ∆S‡ value.18,24
Although the kinetic study presented above determined
under homogeneous solution conditions has proven to be
insightful, our focus on a more environmentally benign route
for the production of polycarbonates dictates the necessity
of performing a comparable, albeit limited, investigation
Figure 11. Effect of temperature on the rate of polycarbonate formation
from the coupling of CO2 and cyclohexene oxide using the (tmtaa)CrCl
and 2 equiv of PPNN3 catalyst system at a CO2 pressure of 34.5 bar in
pure cyclohexene oxide.
(24) 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.
Inorganic Chemistry, Vol. 47, No. 24, 2008 11875

Darensbourg and Fitch
Table 6. Temperature-Dependent Initial Rates for the Coupling of
Cyclohexene Oxide and Carbon Dioxide in the Absence of a Cosolventa
initial rate (M/sec) × 105
temp. (°C)
copolymer
cyclic carbonate
60
65
70
75
80
90
100
4.75
10.1
11.2
18.
25.6
36.3
77.9
0.121
0.394
0.754
0.731
1.50
1.48
9.21
a Reactions carried out in 20 mL of cyclohexene oxide, 1.16 × 10-4
Figure 13. Time-dependent reaction profiles for the solventless copolym-
erization of cyclohexene oxide (20 mL) and CO2 (34.5 bar) in the presence
of 0.116 mmol of (tmtaa)CrCl and 3 equiv of PPNN3 at 80 °C. Note
enhancement of the cyclic carbonate production at 0.7 h after copolymer
formation has reached a plateau.
mol of (tmtaa)CrCl and 2 equiv of PPNN3 at a CO2 pressure of 34.5 bar.
Figure 14. Epoxides surveyed for copolymer production catalyzed by
(tmtaa)CrCl/PPNN3.
that higher temperatures generally lead to enhanced produc-
tion of cyclic carbonates, these experiments were performed
starting at temperatures close to ambient, with subsequent
ramping up of the temperatures as needed to observe reaction
progress. Copolymerization reactions were carried out in a
stainless steel reactor and were monitored by in situ infrared
spectroscopy. The epoxides examined were propylene oxide
(a), styrene oxide (b), 1,2-epoxyhexane (c), isobutylene oxide
(d), and 4-vinylcyclohexene oxide (e), as depicted in Figure
14.
The most widely investigated monomer for the copolym-
erization with carbon dioxide studied to date other than
cyclohexene oxide is propylene oxide. The challenge in
polymerizing aliphatic epoxides such as ethylene or propyl-
ene oxide is the ease at which these comonomers provide
cyclic carbonates as opposed to polycarbonates. Nevertheless,
there are several reports for the selective production of
poly(propylene carbonate) from propylene oxide and CO2
in the presence of (salen)MX complexes (M ) Cr25 or Co26).
In particular, the (salen)CoX complexes in the association
with anionic cocatalysts have been shown to provide
copolymers with stereo- and regioselectivity. In this study,
Figure 12. Arrhenius plots for polymer (R2 ) 0.9747) and cyclic carbonate
(R2 ) 0.9018) formation from polymerizations carried out in neat cyclo-
hexene oxide. Conditions: 1.16 × 10-4mol of (tmtaa)CrCl and 2 equiv of
PPNN3 in 20 mL of cyclohexene oxide under a CO2 pressure of 34.5 bar.
activation for the two processes are 67.1 ( 4.2 and 91.2 (
10.5 kJ · mol-1, respectively. Comparison of the Ea values
for copolymer production in a cosolvent (methylene chlo-
ride), and in the early stages of the reaction performed under
solventless conditions, indicate little difference for the two
processes. On the other hand, the Ea for cyclic carbonate
production is larger under the reaction conditions involving
pure cyclohexene oxide. However, because of the low level
of cyclic carbonate production in this instance, the Ea is
subject to significant uncertainty, that is, an error of (10.5
kJ · mol-1. Nevertheless, a higher value for Ea is anticipated
on the basis of the low rate of cyclic carbonate formation
observed in the absence of a cosolvent. This is explained on
the basis of the backbiting process, which leads to cyclic
carbonate formation being inhibited in pure cyclohexene
oxide. As previously noted, once the reaction medium
becomes viscous due to copolymer formation, where chain
growth is greatly retarded, the rate of cyclic carbonate
production dramatically increases.9 This is clearly illustrated
in Figure 13 for a solventless copolymerization process
carried out under typical reaction conditions.
(25) (a) Darensbourg, D. J.; Phelps, A. L. Inorg. Chem. 2005, 44, 4622–
4629. (b) Eberhardt, R.; Allmendinger, M.; Rieger, B. Macromol.
Rapid Commun. 2003, 24, 194–196. (c) Chisholm, M. H.; Zhou, Z.
J. Am. Chem. Soc. 2004, 126, 11030–11039.
(26) (a) Lu, X.-B.; Wang, Y. Angew. Chem., Int. Ed. 2004, 43, 3574–
3577. (b) Paddock, R. L.; Nguyen, S. T. Macromolecules 2005, 38,
6251–6253. (c) Qin, Z.; Thomas, C. M.; Lee, S.; Coates, G. W. Angew.
Chem., Int. Ed. 2003, 42, 5484–5487. (d) Cohen, C. T.; Chu, T.;
Coates, G. W. J. Am. Chem. Soc. 2005, 127, 10869–10878. (e) Cohen,
C. T.; Coates, G. W. J. Polym. Sci., Part A 2006, 44, 5182–5191. (f)
Nakano, K.; Kamada, T.; Nozaki, K. Angew. Chem., Int. Ed. 2006,
45, 7274–7277. (g) Noh, E. K.; Na, S. J.; S, S.; Kim, S.-W.; Lee,
B. Y. J. Am. Chem. Soc. 2007, 129, 8082–8083. (h) Lee, B. Y.; Kwon,
H. Y.; Lee, S. Y.; Na, S. J.; Han, S.-i.; Yun, H.; Lee, H.; Park, Y.-W.
J. Am. Chem. Soc. 2005, 127, 3031–3037.
Copolymerization of Other Epoxides. We have surveyed
several other epoxides for their selectivity to afford the
copolymer versus cyclic carbonate in the presence of the
(tmtaa)CrCl/PPNN3 catalyst system. These studies were
conducted under typical reaction conditions, that is, 50 mg
of (tmtaa)CrCl, 2 equiv of PPNN3, and 34.5 bar of CO2 in
the absence of a cosolvent. Because it is well-established
11876 Inorganic Chemistry, Vol. 47, No. 24, 2008

Coupling Reactions of Epoxides and Carbon Dioxide
the copolymerization of propylene oxide and carbon dioxide
using complex 1 and 2 equiv of PPNN3 at 40 °C yielded a
21% conversion to the product in 1 h, with 36% of the
product being the copolymer. Additionally, a 60 °C run
carried out for 3 h provided 49% conversion where 15% of
the product was poly(propylene carbonate). Hence, (tmtaa)CrCl
in the presence of PPNN3 is not a particularly effective
catalyst system for this monomer, that is, less effective than
the (salen)MX derivatives of chromium(III) or cobalt(III).
In an effort to examine an epoxide monomer with two
substituents on an ipso-carbon center, the copolymerization
of isobutylene oxide (d) with carbon dioxide was investi-
gated. Our hope was that (tmtaa)CrCl, with its metal center
being more accessible, would allow for the interaction with
this sterically encumbered monomer and concomitant ring-
opening. However, after some time at 40 °C, only trace
quantities of cyclic carbonate were observed, which increased
in its rate of formation as the temperature was raised to 60
°C and then to 80 °C. At no time was there spectral evidence
for the formation of polycarbonate. Similar observations were
made for reactions performed with 1,2-epoxyhexane (c) and
carbon dioxide, where only cyclic carbonate was produced.
It should be recalled that in the patent literature Jacobsen
and co-workers have reported that (salen)CrCl in the absence
of a cocatalyst was effective at copolymerizing 1,2-epoxy-
hexane and CO2 to polycarbonate.27
Polystyrene has become a widely used commodity ther-
moplastic as it is colorless, transparent, rigid, economical,
and easy to cast. In an effort to build upon the same aromatic
functionality of polystyrene while possibly synthesizing a
polymer with desirable properties, styrene oxide was sub-
jected to copolymerization with carbon dioxide using the
(tmtaa)CrCl complex as a catalyst. As mentioned vide supra
pertaining to aliphatic epoxides, the ease of cyclic carbonate
formation poses a challenge to affording copolymer. In earlier
attempts in our laboratory, the copolymerization of styrene
oxide (b) and CO2 utilizing (salen)CrN3 and PPNN3 catalysts
resulted in trace quantities of copolymer and moderate
amounts of cyclic styrene carbonate.28 In this study, the
reaction was carried out starting at a temperature of 40 °C,
under which conditions a small yield of cyclic carbonate was
observed. Further increases in temperature to 80 °C and then
to 100 °C resulted in increases in cyclic carbonate formation,
but no poly(styrene carbonate) was observed via infrared
spectroscopy.
Figure 15. In situ infrared profiles of polycarbonate and cyclic carbonate
formation for the copolymerization of CO2 (34.5 bar) and 4-vinyl cyclo-
hexene oxide (20 mL or M/I ) 1300) utilizing 50 mg of (tmtaa)CrCl (0.077
mmol %) and 2 equiv of PPNCl as a cocatalyst.
of an alicyclic epoxide similar to cyclohexene oxide with
the advantage of a pendant vinyl functionality that can be
post-polymerization-modified. Indeed, studies have shown
it to be possible to prepare block copolymers of poly(cy-
clohexylene carbonate) and poly(4-vinylcyclohexylene car-
bonate),30 as well as intramolecular cross-linked terpolymers
of cyclohexene oxide, 4-vinylcyclohexene oxide, and carbon
dioxide.31 We have previously employed a similar strategy
for intermolecularly cross-linking a copolymer prepared from
[2-(3,4-epoxycyclohexyl)ethyl]trimethoxysilane and CO2 via
silsesquioxane units.32 Figure 15 represents the reaction
profiles for the formation of the copolymer and cyclic
carbonate from the coupling of 4-vinylcyclohexene oxide and
carbon dioxide as catalyzed by (tmtaa)CrCl and 2 equiv of
PPNCl at 80 °C. The copolymerization reaction proceeded
at a slower rate, that is, ∼3 times slower, than the corre-
sponding process with the cyclohexene oxide monomer.
Nevertheless, the process was highly selective for copolymer
formation, providing a polycarbonate with a narrow poly-
dispersity of 1.12.
Summary and Conclusions
In part, our interest in exploring tetraazaannulene chro-
mium(III) derivatives as catalysts for the copolymerization
of cyclohexene oxide and carbon dioxide was to compare
their activity with the closely related salicylaldimine and
porphyrin analogs. Although differences in reaction condi-
tions make it difficult to quantitatively contrast the catalytic
activities involving these metal complexes, it is apparent that,
under optimal conditions, the (salen)CrX and (tmtaa)CrX
metal complexes in the presence of anionic cocatalysts
display similar activities. For example, (salen)CrX complexes
exhibit a TOF of ∼1150 h-1 at 80 °C and provide high
molecular weight copolymers (50 000) with low PDIs
Although the cyclohexene oxide monomer has been
invaluable for understanding the process of epoxide/carbon
dioxide coupling, the inherent properties of the resulting
copolymer, poly(cyclohexylene carbonate), make it currently
commercially unattractive.29 A closely related monomer,
4-vinylcyclohexene oxide (e), offers the selective reactivity
(27) Jacobsen, E. N.; Tokunaga, M.; Larrow, J. F. PCT Int. Appl. 2000,
WO 2000009463.
(28) Mackiewicz, R. M. Structural and mechanistic studies into the
copolymerization of carbon dioxide and epoxides catalyzed by
chromium salen complexes. http://handle.tamu.edu/1969.1/3849 (ac-
cessed Oct. 2008).
(29) Koning, C.; Wildeson, J. R.; Parton, R.; Plum, B.; Steeman, P.;
Darensbourg, D. J. Polymer 2001, 42, 3995–4004.
(30) Hsu, T.-J.; Tan, C.-S. Polymer 2002, 43, 4535–4543.
(31) Cherian, A. E.; Sun, F. C.; Sheiko, S. S.; Coates, G. W. J. Am. Chem.
Soc. 2007, 129, 11350–11351.
(32) Darensbourg, D. J.; Rodgers, J. R.; Fang, C. C. Inorg. Chem. 2003,
42, 4498–4500.
Inorganic Chemistry, Vol. 47, No. 24, 2008 11877

Darensbourg and Fitch
(1.13).23 Under identical reaction conditions the (tmtaa)CrX
complexes are somewhat more active with TOF of ∼1500
h-1 at 80 °C yielding a copolymer of molecular weight
30,000 with a PDI of 1.07. On the other hand, the complex
containing the more expensive N4-2 ligand, (porphyrin)CrX,
afforded a TOF ≈ 150 h-1 at 110 °C, yielding a low
molecular weight copolymer (Mn ≈ 3500) with a PDI of
1.19.33 Hence, the (tmtaa)CrX complex is slightly more
active than the correspondent complex containing the
ubiquitous salicylaldimine ligand and is considerably more
active than analogous chromium derivatives containing the
porphyrin ligand. This reactivity trend parallels the expected
electron-donating ability of the ligand set about the chro-
mium(III) center, where tmtaa > salen . porphyrin.
Consistent with this interpretation, the addition of electron-
withdrawing chloride substituents to the tmtaa ligand retards
the catalytic activity of the (tmtaaCl4)CrCl derivative
significantly.
An important observation resulting from this investiga-
tion is that the selectivity for formation of the copolymer
versus cyclic carbonate is highly dependent on whether
cosolvents are used. That is, a comparison of product
selectivity under solventless and cosolvent conditions
clearly shows that in the absence of a cosolvent the rate
of the backbiting process for cyclic carbonate production,
either at the metal center or subsequent to chain dissocia-
tion, is greatly retarded. Our current thoughts on this
behavior are that enhanced cyclic carbonate formation is
the result of copolymer chain dissociation from the metal
center followed by facile cyclization.26f,g,34,35
Figure 16. The H2tmtaa ligand containing an aromatic strapping unit on
the underside of the bowl-shaped ligand.37
words, two polymer chains originating from a single metal
center.26d,36 Because it is relatively easy to hinder one side
of the tetraazaannulene ligand with a strap (see Figure 16),11
the resultant metal complex should allow small anions to
bind to both axial sites but inhibit monomer enchainment
on the strapped side. Hence, a study of the type that this
baseline report describes utilizing electronically similar
complexes with and without the strap should address this
question in a very definitive manner. These studies are
currently being pursued in our laboratories.
Finally, these (tmtaa)chromium(III) derivatives, like other
closely related complexes, are rather disappointing as
catalysts for the copolymerization of aliphatic epoxides and
carbon dioxide. However, they were shown to be very
effective at copolymerizing other alicyclic epoxides than
cyclohexene oxide, such as 4-vinylcyclohexene oxide.
The mechanistic aspects of the copolymerization reactions
of epoxides and carbon dioxide employing (salen)CrX and
those presented herein for (tmtaa)CrX correspond intimately
to one another. This in turn allows for an opportunity with
this tetraazaannulene chromium complex to examine one of
the unresolved issues with regard to the behavior of this
general class of metal complexes. That is, because the
observed molecular weights of the thus formed copolymers
are generally significantly less than predicted on the basis
of monomer/initiator, this has been proposed to result from
monomer enchainment above and below the approximately
Acknowledgment. We gratefully acknowledge the finan-
cial support from the National Science Foundation (CHE 05-
43133) and the Robert A. Welch Foundation (A-0923).
Supporting Information Available: X-ray crystallographic files
in CIF format for the structural determination of complex 2 and
the H2tmtaaCl4 ligand. This material is available free of charge via
the Internet at http://pubs.acs.org.
-2
-2
planar arrangement of N2O2 or N4 ligands, in other
(33) Mang, S.; Cooper, A. I.; Colelough, M. E.; Chauhan, N.; Holmes,
A. B. Macromolecules 2000, 33, 303–308.
(34) Darensbourg, D. J.; Bottarelli, P.; Andreatta, J. R. Macromolecules
2007, 40, 7727–7729.
(35) Min, J. K.; Seong, J. E.; Na, S. J.; Lee, B. Y. Angew. Chem., Int. Ed.
2008, 47, 7306–7309.
IC801728E
(36) Aida, T.; Inoue, S. Acc. Chem. Res. 1996, 29, 39–48.
(37) Unpublished X-ray structure from our laboratories.
11878 Inorganic Chemistry, Vol. 47, No. 24, 2008