Triphenylphosphane-modified cobalt catalysts for the selective carbonylation of ethyl diazoacetate

Article abstract of DOI:10.1021/om100480v

The triphenylphosphane-substituted carbonyl cobalt complexes Co ₂(CO)₇(PPh₃), Co₂(CO) ₆(CHCO₂Et)(PPh₃), and [Co(CO) ₃(PPh₃)₂][Co(CO)₄] were found to be more effective precatalysts in the carbonylation of ethyl diazoacetate under atmospheric pressure of carbon monoxide at 10 °C in dichloromethane solution than the parent Co₂(CO)₈ and Co₂(CO) ₇(CHCO₂Et) complexes. The highly reactive (ethoxycarbonyl)ketene is the primary product of the catalytic carbonylation, which dimerizes in the absence of a proper scavenger. In the presence of ethanol as the trapping reagent diethyl malonate is the final product of the carbonylation reaction. The formation of (ethoxycarbonyl)ketene using the catalyst precursor Co₂(CO)₇(PPh₃) occurs in a catalytic cycle, where Co₂(CO)₆(PPh₃) and Co₂(CO)₆(CHCO₂Et)(PPh₃) are the repeating species. The 16e species Co₂(CO)₆(PPh ₃) is involved in the deazotization of ethyl diazoacetate, and Co₂(CO)₆(CHCO₂Et)(PPh₃) leads to the (ethoxycarbonyl)ketene formation. In the absence of carbon monoxide or at low CO concentration the reaction of Co₂(CO)₆(CHCO ₂Et)(PPh₃) with ethyl diazoacetate is the source of Co₂(CO)₅(CHCO₂Et)₂(PPh₃), which is not an active catalyst for the carbonylation of ethyl diazoacetate. Using [Co(CO)₃(PPh₃)₂][Co(CO)₄] as the catalyst precursor, the intermediary formation of [Co(CO) ₃(PPh₃)₂][Co(CO)₃(O=C=CHCO ₂Et)] through radical pairs is assumed. Substituting PPh₃ in Co₂(CO)₇(PPh₃), Co₂(CO) ₆(CHCO₂Et)(PPh₃), and [Co(CO) ₃(PPh₃)₂][Co(CO)₄] by polymer-bound PPh₃ results in active and reusable catalysts for the selective carbonylation of ethyl diazoacetate in dichloromethane solution at 40 °C and 11 bar of pressure with up to 5.1 mol of product/((mol of catalyst) h) turnover frequency.

Full text of DOI:10.1021/om100480v

Organometallics 2010, 29, 3837–3851 3837
DOI: 10.1021/om100480v
Triphenylphosphane-Modified Cobalt Catalysts for the Selective
Carbonylation of Ethyl Diazoacetate
00
†
Neszta Ungvari, Eszter Fordos, Janos Balogh, Tamas Kegl, Laszlo Parkanyi, and
†
†
†
‡
ꢀ
€
ꢀ
Ferenc Ungvary*
ꢀ
ꢀ
ꢀ
ꢀ ꢀ ꢀ
,†
ꢀ
†
ꢀ
Department of Organic Chemistry, University of Pannonia, 8201 Veszprem, Egyetem u. 10,
Hungary, and ‡Institute of Structural Chemistry, Chemical Research Center,
Hungarian Academy of Sciences, P.O. Box 17, 1525 Budapest, Hungary
Received May 17, 2010
The triphenylphosphane-substituted carbonyl cobalt complexes Co2(CO)7(PPh3), Co2(CO)6-
(CHCO2Et)(PPh3), and [Co(CO)3(PPh3)2][Co(CO)4] were found to be more effective precatalysts
in the carbonylation of ethyl diazoacetate under atmospheric pressure of carbon monoxide at 10 °C in
dichloromethane solution than the parent Co2(CO)8 and Co2(CO)7(CHCO2Et) complexes. The
highly reactive (ethoxycarbonyl)ketene is the primary product of the catalytic carbonylation, which
dimerizes in the absence of a proper scavenger. In the presence of ethanol as the trapping reagent
diethyl malonate is the final product of the carbonylation reaction. The formation of (ethoxy-
carbonyl)ketene using the catalyst precursor Co2(CO)7(PPh3) occurs in a catalytic cycle, where
Co2(CO)6(PPh3) and Co2(CO)6(CHCO2Et)(PPh3) are the repeating species. The 16e species Co2-
(CO)6(PPh3) is involved in the deazotization of ethyl diazoacetate, and Co2(CO)6(CHCO2Et)(PPh3)
leads to the (ethoxycarbonyl)ketene formation. In the absence of carbon monoxide or at low CO
concentration the reaction of Co2(CO)6(CHCO2Et)(PPh3) with ethyl diazoacetate is the source of
Co2(CO)5(CHCO2Et)2(PPh3), which is not an active catalyst for the carbonylation of ethyl
diazoacetate. Using [Co(CO)3(PPh3)2][Co(CO)4] as the catalyst precursor, the intermediary forma-
tion of [Co(CO)3(PPh3)2][Co(CO)3(OdCdCHCO2Et)] through radical pairs is assumed. Substitut-
ing PPh3 in Co2(CO)7(PPh3), Co2(CO)6(CHCO2Et)(PPh3), and [Co(CO)3(PPh3)2][Co(CO)4] by
polymer-bound PPh3 results in active and reusable catalysts for the selective carbonylation of ethyl
diazoacetate in dichloromethane solution at 40 °C and 11 bar of pressure with up to 5.1 mol of
product/((mol of catalyst) h) turnover frequency.
1. Introduction
using simple one-pot procedures. The mechanism of the catalytic
carbonylation of ethyl diazoacetate in CH2Cl2 solution at 10 °C
in the presence of ethanol using Co2(CO)8 as the catalyst
precursor has been established. According to Scheme 1 the
formation of diethyl malonate occurs in parallel in two different
cycles, A and B. In the first cycle Co2(CO)7 and Co2(CO)7-
(CHCO2Et) are the working repeating species, whereas in the
second cycle the species are Co2(CO)6(CHCO2Et) and Co2-
(CO)6(CHCO2Et)2. On the basis of the experimentally deter-
mined rates of the steps in the suggested cycles and the
composition of the cobalt complexes in the reaction mixture
under various conditions, the contribution of cycles A and B to
the formation of the diethyl malonate product were calculated.
According to the calculations most of the diethyl malonate
product forms in cycle B at low (1 bar) carbon monoxide
pressure, but in cycle A formation occurs if the carbon monoxide
pressure is high (100 bar).7
The substitution of the diazo group in diazoalkanes by carbon
monoxide to obtain ketenes has great potential in synthetic
organic chemistry.1,2 Among the various transition-metal com-
plexes capable of mediating the ketene formation,3 Co2(CO)8
was found to be a highly active and selective catalyst for the
preparation of (trimethylsilyl)ketene by the carbonylation of
(trimethylsilyl)diazomethane.4 The very reactive (ethoxycar-
bonyl)ketene formed in the Co2(CO)8-catalyzed carbonylation
of ethyl diazoacetate can be trapped in situ by alcohols, phenol,
secondary amines, or imines to obtain the corresponding malo-
nic acid derivatives3,5 or β-lactams,6 respectively, in high yields
*To whom correspondence should be addressed. E-mail: ungvary@
almos.vein.hu.
(1) Tidwell, T. T. Ketenes, 2nd ed.; Wiley: Hoboken, NJ, 2006.
(2) Tidwell, T. T. Eur. J. Org. Chem. 2006, 563–576.
We found that triphenylphosphane-substituted cobalt
carbonyl complexes such as Co2(CO)7(PPh3)8-10 (1),
ꢀ
ꢀ
(3) Ungvari, N.; Ungvary, F. Carbonylation of diazoalkanes. In
ꢀ
Modern Carbonylation Methods; Kollar, L., Ed.; Wiley-VCH: Weinheim,
11
Co2(CO)6(PPh3)2 (2), Co2(CO)6(CHCO2Et)(PPh3)12 (3),
Germany, 2008; Chapter 8, pp 199-221.
ꢀ
ꢀ
ꢀ
(4) Ungvari, N.; Kegl, T.; Ungvary, F. J. Mol. Catal. A: Chem. 2004,
219, 7–11.
00
ꢀ
ꢀ
€
ꢀ
ꢀ
(7) Ungvari, N.; Fordos, E.; Kegl, T.; Ungvary, F. Inorg. Chim. Acta
(5) Tuba, R.; Ungvary, F. J. Mol. Catal. A: Chem. 2003, 203, 59–67.
00
€
ꢀ
(6) Fordos, E.; Tuba, R.; Parkanyi, L.; Kegl, T.; Ungvary, F. Eur. J.
ꢀ
ꢀ
ꢀ
2010, 363, 2016-2028.
ꢀ
Org. Chem. 2009, 1994–2002.
(8) Bor, G.; Marko, L. Chem. Ind. 1963, 912–913.
r
2010 American Chemical Society
Published on Web 08/03/2010
pubs.acs.org/Organometallics

ꢀ
Ungvari et al.
3838 Organometallics, Vol. 29, No. 17, 2010
Scheme 1. Mechanism of the Catalytic Carbonylation of Ethyl Diazoacetate in the Presence of Ethanol using Co2(CO)8 as the Catalyst
Precursor
Co2(CO)5(CHCO2Et)(PPh3)212 (4), and [Co(CO)3(PPh3)2]-
[Co(CO)4]11,13 (5) are all useful precursors for the catalytic
carbonylation of ethyl diazoacetate. Thus, using 2 mol % of
these complexes under 50 bar pressure of carbon monoxide at
45 °C in the presence of ethanol, up to 100% yield of diethyl
malonate was achieved in 24-150 h reaction time.12 With the
exception of 2 and 4 these complexes are active at lower
temperature (10 °C) and under atmospheric pressure of carbon
monoxide as well. Under the milder conditions they show even
higher catalytic activity than Co2(CO)8 in the carbonylation of
ethyl diazoacetate. On the basis of the results of infrared spectro-
scopic, preparative, and kinetic studies a tentative mechanism of
these catalytic reactions is now reported.
Catalytic carbonylation of ethyl diazoacetate in dichloro-
methane solution using the immobilized analogues of 5,10,14,15
1,10,14,15 and 3 on polymer-bound triphenylphosphane as the
catalysts give at 40 °C and 11 bar of CO pressure products
close to 100% yields free of cobalt with up to 5.1 mol/(mol h)
turnover frequency.
the interference method.18 1H, 13C, and 31P NMR spectra were
recorded at room temperature on a Bruker 400 MHz spectro-
meter using CDCl3 as the solvent for solution spectra. Cobalt
and phosphorus analyses were carried out by using an ICP-OES
apparatus (inductively coupled plasma optical emission spectro-
meter from Perkin-Elmer) from samples prepared with concen-
trated nitric acid by using the HPA-S apparatus (high-pressure
asher from Anton-Paar, Graz, Austria). Heptacarbonyl(triphenyl-
phosphane)dicobalt(0),10 octacarbonyldicobalt,19 the μ2-(ethoxyc-
arbonyl)carbene complexes 312 and 4,12 [Co(CO)3(PPh3)2]-
2010 cm-1
1890 cm-1
2009 cm-1
[Co(CO)4]11,13 (IR: εM
= 3556, and εM
= 5752
= 335,
cm2/mmol in CH2Cl2), Na[Co(CO)4]20,21 (IR: εM
1887 cm-1
1858 cm-1
εM
= 2995, and εM
= 1366 cm2/mmol in THF),
and [PPN][Co(CO)4]22 (PPN = [Ph3PdNdPPh3]þ) (IR:
1889 cm-1
εM
= 6100 cm2/mmol in CH2Cl2) were prepared by
standard literature procedures. Stock solutions (0.050 mol/dm3)
of [Co(EtOH)6][Co(CO)4]2 were prepared by dissolving Co2(CO)8
(257 mg, 0.75 mmol) in absolute EtOH (9.9 cm3) under argon at
2010 cm-1
1903 cm-1
room temperature (IR: εM
= 140, εM
= 4230 cm2/
mmol, and a strong shoulder at 1870 cm-1 in EtOH) and were used
immediately after preparation. Carbon monoxide (99.99% purity)
was obtained from Messer Hungary. 13CO (99% isotope purity)
was obtained from Sigma-Aldrich. Cobaltocene (99% purity) was
obtained from Strem. Polystyrenediphenylphosphane cross-linked
with 1% divinylbenzene (1.6 mmol P/g, 100-200 mesh particle
size) was obtained from Fluka, and polystyrenediphenylphos-
phane cross-linked with 1% divinylbenzene (∼2.0 mmol P/g,
100-200 mesh particle size) was obtained from Sigma-Aldrich.
2.2. Catalytic Carbonylations. The catalytic carbonylation
reactions of ethyl diazoacetate using soluble catalyst precursors
were performed in a thermostated glass reactor at atmospheric
pressure or in a glass insert of a thermostated autoclave at higher
carbon monoxide pressure using 0.02-0.20 mmol aliquots of
the catalyst precursor complexes in the form of freshly prepared
stock solutions. The reactions were started by adding neat ethyl
diazoacetate to the reaction mixture. The concentration of the
carbonylation products and that of other components in the
reaction mixture was determined at different time intervals from
withdrawn samples by quantitative infrared spectroscopy. In-
itial rates of diethyl malonate formation were calculated from
2. Experimental Section
2.1. General Procedures. Handling of the cobalt carbonyl
complexes was carried out under an atmosphere of dry P4O10
and deoxygenated (BTS contact, room temperature) argon or
carbon monoxide utilizing standard Schlenk techniques.16 Di-
chloromethane was dried and distilled under an atmosphere
of argon or carbon monoxide according to standard proce-
dures.17 IR spectra were recorded on Thermo Nicolet Avatar
330 and Bruker Tensor 27 FTIR spectrometers using 0.002 65,
0.007 65, 0.0218, or 0.0505 cm CaF2 liquid cells, calibrated by
ꢀ
ꢀ
(9) Szabo, P.; Fekete, L.; Bor, G.; Nagy-Magos, Z.; Marko, L.
J. Organomet. Chem. 1968, 12, 245–248.
(10) Comely, A. C.; Gibson, S. E.; Hales, N. J.; Johnstone, C.;
Stevenazzi, A. Org. Biomol. Chem. 2003, 1, 1959–1968.
(11) Hieber, W.; Freyer, W. Chem. Ber. 1958, 91, 1230–1234.
00
€
ꢀ
(12) Tuba, R.; Fordos, E.; Ungvary, F. J. Mol. Catal. A: Chem. 2005,
236, 113–118.
(13) Vohler, O. Chem. Ber. 1958, 91, 1235–1238.
the concentrations determined using the molar absorbancies of
=
1730cm-1
diethyl malonate in CH2Cl2 at 1730 and 1749 cm-1 (εM
ꢀ
(14) Comely, A. C.; Gibson (nee Thomas), S. E.; Hales, N. J. Chem.
1749 cm-1
666 cm2/mmol, εM
conversions.
= 579 cm2/mmol) between 5 and 15%
Commun. 1999, 2075–2076.
(15) Comely, A. C.; Gibson (nee Thomas), S. E.; Hales, N. J. Chem.
ꢀ
Commun. 2000, 305–306.
(16) Shriver, D. F. The Manipulation of Air-Sensitive Compounds;
Krieger: Malabar, FL, 1982.
(17) Armarego, W. L. F.; Perrin, D. D. Purification of Laboratory
Chemicals, 4th ed.; Butterworth-Heinemann: Oxford, U.K., 1996.
(18) Willard, H. H.; Merritt, L. L., Jr.; Dean, J. A.; Settle, F. A., Jr.
Instrumental Methods of Analysis, 6th ed.; Wadsworth: Belmont, CA, 1981;
p 206.
ꢀ
ꢀ
(19) Szabo, P.; Marko, L.; Bor, G. Chem. Technol. (Leipzig) 1961, 13,
549–550.
(20) Edgell, W. F.; Lyford, J., IV Inorg. Chem. 1970, 9, 1932–1933.
(21) Winzenburg, M. L. Ph.D. Thesis, University of Minnesota,
1979; Appendix A, p 152.
(22) Ruff, J. K.; Schlienz, W. J. Inorg. Synth. 1974, 15, 84–90.

Article
The catalytic carbonylation reactions of ethyl diazoacetate
Organometallics, Vol. 29, No. 17, 2010 3839
dichloromethane solutions in the Schlenk flask contained 0.149
mmol of Co2(CO)8 (according to quantitative infrared analysis)
not taken up by the polymer in the loading procedure. Drying
the cobalt-loaded polymer first in a stream of CO and then
under vacuum afforded 3.12 g of dark yellow-red beads as the
final product. Anal. Calcd (for a composition of 88% of
polymer-bound 5-type complex and 12% of polymer-bound 1-
type complex): Co, 7.86; P, 3.69. Found: Co, 7.9; P, 3.7. IR
with the heterogenized catalysts were performed in a glass insert
of a thermostated magnetically stirred autoclave between 5 and
50 bar of carbon monoxide pressure using 0.050 mmol portions
of the catalysts. Recycling experiments were performed using a
glass insert with a sintered glass frit (P3) which was closed on the
bottom of the frit by a Teflon stopper. After the reaction the
liquid product was removed from the glass insert under argon by
filtration through the frit. When the bottom of the filter tube was
closed by a Teflon stopper, new solvent and reactant were added
to the recycled catalyst and a new cycle was started by pressur-
izing with CO.
2.3. Photochemical Reactions. Photochemical experiments
were performed in a double-jacketed thermostated glass reactor
using a Pen-Ray light source emitting at 254 or 366 nm
(Ultraviolet Products Inc.) immersed into the reaction mixture
in a protecting quartz tube.
2.4. Synthesis of Co2(CO)5(CHCO2Et)2(PPh3) (6). To a so-
lution of Co2(CO)6(CHCO2Et)(PPh3) (635 mg, 1.0 mmol)
in dichloromethane (9.4 cm3) was added ethyl diazoacetate (0.106
cm3, 1.0 mmol) at 0 °C under argon. For 2 h argon was slowly
bubbled through the reaction mixture, and then a second portion
of ethyl diazoacetate (0.106 cm3, 1.0 mmol) was added, and the
argon purging was continued for another 3 h period. After the
reaction mixture was kept at 0 °C in the refrigerator overnight, the
solution was filtered and concentrated to a volume of 3 cm3 before
flash column chromatography on silica gel (1.5 cm ꢀ24 cm
column) with dichloromethane as the eluent. The first 180 cm3
fraction was discarded, and from the second 410 cm3 fraction the
solvent was removed under vacuum. Complex 6was obtained as an
(KBr): ν(CO) 2072 m, 1992 vs, br, 1880 vs, br cm-1
.
31P NMR
(solid state, δ): 65.0 ppm.
2.7. Preparation of Polystyrenediphenylphosphane (PSDP)-
Anchored Carbonyl Cobalt Catalyst from PSDP and Co2(CO)7-
(CHCO2Et). Following the loading procedure as in part 2.6 but
using a solution of Co2(CO)7(CHCO2Et) (3.712 mmol) in
dichloromethane (38 cm3) added in four installments to the
dichloromethane-swelled polystyrenediphenylphosphane (2.32
g, 3.712 mmol P) resulted in 3.81 g of dark purple-red beads after
drying. Anal. Calcd (for a polymer-bound 3-type complex): Co,
11.84; P, 3.11. Found: Co, 12.1; P, 3.0. IR (KBr): ν(CO) 2075 m,
2029 s, 2011 s, 1997 m, br, 1975 m, br cm-1 (complex 3 IR (KBr):
ν(CO) 2075 s, 2027 s, 2017 s, 2005 s, 1990 m, 1825 s, 1668 m
cm-1). 31P NMR (solid state, δ): 61.0 ppm.
2.8. Preparation of Polystyrenediphenylphosphane (PSDP)-
Anchored N2CHCO2Et. Using in the loading procedure poly-
styrenediphenylphosphane (1.03 g, 2.047 mmol P) and ethyl
diazoacetate (0.37 g, 3.24 mmol) gave recovered ethyl diazoa-
cetate in the filtrate (0.14 g, 1.23 mmol), and 1.26 g of yellow
beads in the filter tube after drying. Anal. Calcd (for a polymer-
bound Ph3PdNNdCHCO2Et-type phosphazane): P, 5.02. Found:
P, 4.9. IR (KBr): ν(CdO) 1710 s, ν(NdC) 1675 s, ν(PdN) 1520
cm-1 31P NMR (CH2Cl2-swelled beads, δ): 19.3 ppm.
.
orange oil (215 mg, 31% yield). TLC on silica gel/CH2Cl2 gave one
= 522 cm2/
2087 cm-1
2.9. 13CO Exchange Reactions. The 13CO exchange reactions
of [Co(CO)3(PPh3)2][Co(CO)4] (5), [Co(EtOH)6][Co(CO)4]2 (7),
Na[Co(CO)4] (8), and [PPN][Co(CO)4] (9) were performed with
0.03-0.05 mmol amounts of the complexes and 1.0 mmol of
13CO in a thermostated glass reactor at atmospheric pressure.
The change in the concentration of the complexes was followed
by quantitative infrared spectroscopy using a syringe pump to
circulate the reaction mixture between the reactor and the
infrared liquid cell.
spot with Rf = 0.32. IR (CH2Cl2): 2087 (εM
2041 cm-1
1975 cm-1
mmol), 2041 (εM
= 673 cm2/mmol), 1975 (εM
=
= 121 cm2/mmol), 1685
1723 cm-1
3009 cm2/mmol), 1723 (εM
1685 cm-1
(εM
= 440 cm2/mmol) cm-1. 1H NMR (CDCl3, δ/ppm):
7.58-7.37 (m, 15 H, C6H5), 4.15-3.29 (m, 4H, CH2), 2.13 (s, 2H,
CH), 1.11 (m, 6H, CH3). 13C NMR (CDCl3, δ/ppm): 198.8 and
198.6 (term COs), 181.1 and 180.3 (CO2Et), 134.4-128.0 (C6H5),
112.3 (CH), 60.5 and 59.5 (CH2), 14.1 (CH3) 5.2 (d, 2JCP = 16 Hz,
CH). 31P NMR (CDCl3, δ/ppm): 60.36.
2.10. Theoretical Methods. All of the minima and transition
states were computed with the hybrid HF/DFT method using
the combination of the three-parameter Becke exchange func-
tional with the Lee-Yang-Parr correlation functional known
as B3LYP.23 For cobalt atoms the Stevens-Basch-Krauss-
Jasien-Cundari basis set, which is valence triple-ζ for transition
metals, with the corresponding small-core pseudopotential24
was utilized. This basis set follows the contraction pattern of
(8s,8p,6d) f [4s,4p,3d] for cobalt. For the other atoms the
6-31G(d,p) basis set25 was used.
The calculations were carried out using PC-GAMESS 7.1.E/
Firefly software,26 which is partially based on the GAMESS-US
source code27 and the Gaussian 03 suite of programs.28 Local
minima were identified by the absence of negative eigenvalues in
the vibrational frequency analysis. The Hessian matrix of
transition state structures has only one negative eigenvalue.
Intrinsic reaction coordinate (IRC) analyses29 were used at the
same level of theory as the geometry optimizations in order to
confirm that the stationary points are smoothly connected to
each other. To estimate the effect of the solvent, single-point
2.5. Synthesis of EtO2CCH2Co(CO)3(PPh3). A solution of
Co2(CO)5(CHCO2Et)(PPh3)2 (434 mg, 0.5 mmol) in dichloro-
methane (4.7 cm3) was refluxed under argon for 15 min. The
brown precipitate (200 mg, 46 m/m%) was separated by filtra-
tion and washed with dichloromethane (3 ꢀ 2 cm3). IR (KBr):
1960 sh, 1943 vs, 1930 sh cm-1. IR (CH2Cl2): 1969.6 m, 1941 vs,
br, 1897 m cm-1. Flash column chromatography of the concen-
trated orange-red filtrate was performed on silica gel (1.0 cm ꢀ
24 cm column) with CH2Cl2 as the eluent. The first 60 cm3
fraction was discarded, and from the second 75 cm3 fraction the
solvent was removed under vacuum. The complex EtO2CCH2-
Co(CO)3(PPh3) was obtained as yellow crystals (196 mg, 77%
= 185 cm2/mmol),
2046 cm-1
yield). IR (CH2Cl2): 2046 (εM
1974 cm-1
1685 cm-1
1974 (εM
= 2500 cm2/mmol), and 1685 (εM
=
342 cm2/mmol) cm-1
.
2.6. Preparation of Polystyrenediphenylphosphane (PSDP)-
Anchored Carbonyl Cobalt Catalyst from PSDP and Co2(CO)8.
To polystyrenediphenylphosphane (2.32 g, 3.712 mmol P) in a
filter tube (P3 frit 2.0 cm diameter, 31 cm3 volume, fitted with
a gas inlet for CO and a silicon septum and connected to a
graduated Schlenk flask receiver of 100 cm3 capacity) under a
CO atmosphere was added dichloromethane (16 cm3) at room
temperature. A slight positive pressure difference between the
Schlenk flask and the filter (5-10 mmHg) kept the solvent in
the filter tube and served to stir the mixture. After 30 min, a solu-
tion of Co2(CO)8 (0.762 g, 2.228 mmol) in dichloromethane (8.0
cm3) was added by a syringe. When the pressure difference was
reversed after 60 min, the solution was collected in the Schlenk
flask and the dark brown polymer beads were washed five
times with 8 cm3 portions of dichloromethane. The combined
(23) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652.
(24) Stevens, W. J.; Basch, H.; Krauss, M. J. Chem. Phys. 1984, 81,
6026–6033.
(25) Hehre, W. J.; Ditchfield, R.; Pople, J. A. J. Chem. Phys. 1972, 56,
2257–2261.
(26) http://classic.chem.msu.su/gran/gamess/index.html.
(27) Schmidt, M. W.; Baldridge, K. K.; Boatz, J. A.; Elbert, S. T.;
Gordon, M. S.; Jensen, J. H.; Koseki, S.; Matsunaga, N.; Nguyen, K. A.;
Su, S.; Windus, T. L.; Dupuis, M; Montgomery, J. A. J. Comput. Chem.
1993, 14, 1347–1363.

ꢀ
Ungvari et al.
3840 Organometallics, Vol. 29, No. 17, 2010
Table 1. Initial Rate of Diethyl Malonate Formation (rDEM) in the Carbonylation of Ethyl Diazoacetate (EDA) ([EDA]0 = [EtOH]0 =
0.25 mol/dm3) at 10 °C in CH2Cl2 Solution under Atmospheric Pressure of Carbon Monoxide ([CO]0 = 0.00500-0.00510 mol/dm3) in the
Presence of Different Carbonyl Cobalt Complexes as the Precatalysts: Co2(CO)8, Co2(CO)7(PPh3) (1), Co2(CO)6(CHCO2Et)(PPh3) (3),
Co2(CO)5(CHCO2Et)(PPh3)2 (4), and [Co(CO)3(PPh3)2][Co(CO)4] (5)
entry [Co2(CO)8]0 (mol/dm3) [1]0 (mol/dm3) [3]0 (mol/dm3) [4]0 (mol/dm3) [5]0 (mol/dm3) [PPh3]0 (mol/dm3) 105rDEM (mol/(dm3 s))
1
2
3
4
5
6
7
0.050
0.50
4.95
0.050
0.050
0.48a
10.85
1.36b
3.00
0.050
0.050
0.050
0.050
0.500
0.11
8
9
0.0050
0.0050
0.0050
0.0050
0.0050
0.0050
0.0050
0.0050
0.0050
3.30
3.33c
3.28d
1.35e
0.67f
2.11
10
11
12
13
14
15
16
0.012
0.025
0.050
1.55
0.67
0.17g
a Under 3152 mmHg pressure of carbon monoxide, [CO]0 = 0.0288 mol/dm3. b Under 1544 mmHg pressure of carbon monoxide, [CO]0 = 0.01298
mol/dm3. c [EtOH]0 = 0.50 mol/dm3. d [EtOH]0 = 1.00 mol/dm3. e [EDA]0 = 0.10 mol/dm3. f [EDA]0 = 0.05 mol/dm3. g Under 10 bar pressure of carbon
monoxide, [CO]0 = 0.072 mol/dm3.
calculations on the gas-phase optimized structures were carried
out employing the polarized continuum model (PCM)30 with the
dielectric constant ε0 = 8.93 for dichloromethane.
higher than that observed with Co2(CO)8 (entry 1). In 3 h
reaction time complete conversion of the starting ethyl dia-
zoacetate (EDA) into diethyl malonate (DEM) occurs with
complex 1 but only 20% conversion with Co2(CO)8.7 The
infrared spectra of the reaction mixtures of the former show at
different conversions the presence of complex 1 at 2078, 2023,
3. Results
2078cm-1
2023cm-1
1818 cm-1
1988, 1955, and 1818 cm-1 (εM
= 2381, εM
= 609, and εM
=
=
,
Typical initial rates of diethyl malonate formation (rDEM
)
1988cm-1
1955cm-1
in the carbonylation reaction of ethyl diazoacetate (EDA)
according to eq 1 using different cobalt complexes as the
catalyst precursors are compiled in Table 1.
1147, εM
= 4205, εM
90 cm2/mmol in CH2Cl2)31 and a shoulder at 2033 cm-1
which is the most intense band of complex 3.12 When the CO
pressure is raised from atmospheric to 4 bar, the initial rate of
diethyl malonate formation (rDEM) is strongly reduced (entry
3), and the shoulder at 2033 cm-1 cannot be seen in the
infrared spectrum of the reaction mixture. According to the
X-ray structure of the crystalline Co2(CO)7(PPh3) (see Figure
S10 in the Supporting Information, CCDC756129) there is no
carbon monoxide ligand(s) in a bridging position. However,
two of the equatorial carbonyl ligands around the cobalt
atoms are tilted by 6-7° toward the neighboring cobalt
atoms.
N2CHCO2Et þ CO þ EtOH
EDA
catalyst
EtO2CCH2CO2Et þ N2
ð1Þ
CH2Cl2, 10 °C, 1 bar
DEM
The rDEM data in Table 1 show that using the PPh3-
containing carbonyl cobalt complexes Co2(CO)7(PPh3) (1),
Co2(CO)6(CHCO2Et)(PPh3) (3), Co2(CO)5(CHCO2Et)(PPh3)2
(4), and [Co(CO)3(PPh3)2][Co(CO)4] (5) as catalyst precur-
sors result in more active catalysts than Co2(CO)8.
With complex 3 as the catalyst precursor (entry 4), 22 times
higher initial rate of the carbonylation reaction of EDA was
observed than with Co2(CO)8. In 20 min 50% conversion
and in 180 min practically complete conversion of EDA into
DEM was observed. At 50% conversion the composition of
the cobalt complexes in the solution is 90% complex 3
Using Co2(CO)7(PPh3) (1) as the precatalyst, the initial
rate of diethyl malonate formation (rDEM) (entry 2) is 10 times
(28) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.;
Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.;
Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson,
G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.;
Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai,
H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken,
V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev,
O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.;
Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.;
Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas,
O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.;
Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov,
B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.;
Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.;
Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.;
Gonzalez, C.; Pople, J. A. Gaussian 03; Gaussian, Inc., Wallingford, CT,
2004.
2080 cm-1
2033 cm-1
1688 cm-1
2013 cm-1
1666 cm-1
(εM
2462, εM
= 1517, εM
= 2941, εM
= 174, and εM
=
=
1829 cm-1
= 597, εM
151 cm2/mmol in CH2Cl2) and 10% complex 1, according to
the infrared spectra of the reaction mixture. Raising the carbon
monoxide pressure to 2 bar strongly inhibits the rate of diethyl
malonate formation (entry 5). In experiments started under
atmospheric pressure of carbon monoxide new bands grow in
at 2087 and 2041 cm-1 at higher conversions. If complex 3 and
EDA are applied in a 1:1 to 1:2 molar ratio in the absence of
carbon monoxide and ethanol, the complex with the character-
istic absorptions at 2087 and 2041 cm-1 becomes the main
(29) Gonzalez, C.; Schlegel, H. B. J. Chem. Phys. 1989, 90, 2154–
2161.
(30) Miertus, S.; Scrocco, E.; Tomasi, J. Chem. Phys. 1981, 55, 117–
129.
(31) In CH2Cl2 solutions of Co2(CO)7(PPh3) a very weak absorption
in the bridging ν(CO) range at 1818 cm-1 is always present (see Figure S1
in the Supporting Information), which is, however, absent in hexane8,9
or THF10 solutions.

Article
Organometallics, Vol. 29, No. 17, 2010 3841
product of the reaction (eq 2). Flash column chromatography
on silica gel with CH2Cl2 at 4 °C affords the new complex in
pure form as Co2(CO)5(CHCO2Et)2(PPh3) (6) having absorp-
tions in CH2Cl2 solution at 2087 s, 2041 vs, 1979 vs, 1971 vs,
Co(CO)3(PPh3)34-36 in 77% isolated yield (based on the
CHCO2Et content of complex 4) having absorptions in
2046 cm-1
CH2Cl2 solution at 2046 (εM
= 185 cm2/mmol),
1974 cm-1
1974 (εM
(εM
= 2500 cm2/mmol), and 1685 cm-1
1685 cm-1
1723 w, and 1685 m cm-1
.
= 342 cm2/mmol). A single-crystal X-ray dif-
fraction study of our yellow solid product confirmed the
assumed structure (see Figure S11 in the Supporting Infor-
mation, CCDC 756130), which is basically identical with that
of EtO2CCH2Co(CO)3(PPh3). The only difference is that in
our low-temperature (133(2) K) structure the ethyl group is
not disordered. X-ray studies of close analogues of similar
[(alkoxycarbonyl)methyl]cobalt type complexes have also
been described.37
Co2ðCOÞ6ðCHCO2EtÞðPPh3Þ þ N2CHCO2Et
3
0 °C, 24 h
Co2ðCOÞ5ðCHCO2EtÞ2ðPPh3Þþ CO þ N2 ð2Þ
CH2Cl2, Ar
6
The carbonylation of ethyl diazoacetate with complex 6 as
the precatalyst applied in 0.005 mol/dm3 concentration under
the conditions shown in Table 1 results in <0.03 mol/(dm3 s)
initial rate of diethyl malonate formation.
Using complex 4 as the catalyst precursor, the initial rate
of diethyl malonate formation (entry 6) was only about 28%
of that observed with complex 3 (entry 4). In the infrared
spectrum at 15% conversion of EDA the composition of the
cobalt complexes in the solution is 30% of complex 3 and
The complex [Co(CO)3(PPh3)2][Co(CO)4] (5) under an
atmosphere of carbon monoxide shows a remarkable activity
in the selective carbonylation of ethyl diazoacetate. Even at
low concentrations (0.0050 mol/dm3) complex 5 (entry 8) is
more active than Co2(CO)8 (entry 1). Added PPh3 lowers the
observed initial rate of carbonylation (entries 13-15). The
initial rate of the carbonylation reaction is lower at 10 bar
pressure of CO (entry 16) than under atmospheric pressure.
The initial rate of diethyl malonate formation (rDEM) is first
order in EDA and zero order in EtOH concentration
(compare the rate in entry 8 with those in entries 9-12).
The infrared spectra of the reaction mixtures containing the
catalyst, EDA, and ethanol show bands at 2112, 2010, 1889,
1750, 1730, and 1692 cm-1 during the catalysis (see Figure 1
for an illustration). The ν(N2) and ν(CO) bands at 2112 and
1692 cm-1, respectively, belong to the starting ethyl diazoa-
70% of complex 4 on the basis of the intensity ratio of
the bands at 1829 cm-1 (εM
1829 cm-1
= 597 cm2/mmol) and
1786 cm-1
1786 cm-1 (εM
= 506 cm2/mmol). When this experi-
ment was repeated in the presence of a 10-fold molar excess
of added PPh3 (entry 7), only trace amounts of complex 3
were formed from complex 4 and the initial rate of DEM
formation dropped to about 4% of that without added PPh3
(entry 6). These results suggest that complex 3 rather than
complex 4 is responsible for the catalytic DEM formation
observed in entries 5 and 6, since these complexes are in
equilibrium32 according to eq 3. In addition the observed
three bands in the infrared spectrum at 1711, 1688, and 1534
cm-1 revealed the formation of the known phosphazane
2112cm-1
1692cm-1
cetate (EDA) (εM
= 800 cm2/mmol and εM
=
605 cm2/mmol), the ν(CO) bands at 2010 and 1889 cm-1
2010cm-1
(εM
1889cm-1
= 3556 cm2/mmol and εM
= 5752 cm2/
1711 cm-1
1688 cm-1
mmol) belong to the catalyst complex [Co(CO)3(PPh3)2]-
[Co(CO)4] (5), and the ν(CO) bands at 1750 and 1730
cm-1 belong to the product diethyl malonate (DEM)
Ph3PdNNdCHCO2Et33 (εM
308, and εM
= 340, εM
=
= 349 cm2/mmol) from ethyl diazoace-
1534 cm-1
1750 cm-1
1730 cm-1
tate and the excess PPh3.
Co2ðCOÞ5ðCHCO2EtÞðPPh3Þ2 þ CO
(εM
= 579 cm2/mmol and εM
= 666 cm2/
mmol). The intensity ratio of the bands at 2010 and 1889
cm-1 during the catalysis are, however, not the same as that
in the spectrum of the original [Co(CO)3(PPh3)2][Co(CO)4]
complex before participating in the carbonylation reaction.
While the intensity of the band at 2010 cm-1 is unchanged
during the catalytic carbonylation of ethyl diazoacetate, the
band at 1889 cm-1 shows a small but significant decrease of
intensity.
The change of the intensity ratio is much more pro-
nounced in experiments started in the absence of ethanol
(see Figure 2 for illustration). In Figure 2 we can see
not only the dramatic decrease of the band intensity at
1889 cm-1 but the appearance of two new bands at 1734 and
1637 cm-1. The band at 1734 cm-1 can be observed in
the case of all the other cobalt catalysts as well if the
carbonylation of ethyl diazoacetate is performed in the
4
Kð10 °C, CH2Cl2Þ ¼ 1:03 þ 0:03
R
Co ðCOÞ ðCHCO EtÞðPPh Þ
F
2
2
3
6
3
þ PPh3
ð3Þ
If a solution of complex 4 in CH2Cl2 without any other
added reagents is refluxed under argon within a few minutes,
it completely converts into a product which consists of a
brown precipitate of the composition [Co(CO)2(PPh3)]x
(46% m/m yield based on complex 4; IR (KBr) 1960 sh,
1943 vs, 1930 sh cm-1; IR (CH2Cl2) 1969.6 m, 1941 vs, br
1897 m cm-1) and an orange-red solution with the main
characteristic IR band at 1974 cm-1. Flash column chroma-
tography on silica gel with CH2Cl2 separates this component
as yellow crystals in pure form as the known EtO2CCH2-
absence of ethanol, and this turned out to be the character-
istic ν(CdO) band (εM
1734 cm-1
= 440 cm2/mmol) of the
known dimer of (ethoxycarbonyl)ketene, 3-(ethoxycar-
bonyl)-4-hydroxy-6-ethoxy-2H-pyran-2-one, which exists
00
ꢀ
€
ꢀ
ꢀ
(32) Ungvari, N.; Fordos, E.; Kegl, T.; Ungvary, F. Inorg. Chim. Acta
2009, 362, 1333–1342.
(33) Staudinger, H.; Meyer, J. Helv. Chim. Acta 1919, 2, 619–635.
ꢀ
(34) Galamb, V.; Palyi, G.; Cser, F.; Furmanova, M. G.; Struchkov,
ꢀ ꢀ
(37) (a) Szabo, M.; Szilagyi, R.; Bencze, L.; Boese., R.; Zucchi, C.;
ꢀ
Caglioti, L.; Palyi, G. Enantiomer 2000, 5, 549–559. (b) Zucchi, C.; Boese.,
Y. T. J. Organomet. Chem. 1981, 209, 183–195.
ꢀ
ꢀ
(35) Palyi, G.; Alberts, K.; Bartik, T.; Boese, R.; Frater, G.; Herberich,
T.; Herfurth, A.; Kriebel, C.; Sorkau, A.; Tschorner, C. M.; Zucchi, C.
Organometallics 1996, 15, 3253–3255.
ꢀ
ꢀ
R.; Alberts, K.; Herbrich, T.; Toth, G.; Bencze, L.; Palyi, G. Eur. J. Inorg.
Chem. 2001, 2297–2304. (c) Zucchi, C.; Tiddia, S.; Boese., R.; Tschoerner,
ꢀ
ꢀ
ꢀ
C. M.; Bencze, L.; Palyi, G. Chirality 2001, 13, 458–464. (d) Zucchi, C.;
(36) Bencze, L.; Szabo, M.; J. Szilagyi, R. K.; Boese, R.; Zucchi, C.;
ꢀ
ꢀ
ꢀ
Palyi, G. In Fundamentals of Life; Palyi, G., Zucchi, C., Caglioti, L., Eds.;
Elsevier: Paris, 2002; pp 451-471.
Turrini, D.; Boese., R.; Bencze, L.; Kurdi, R.; Caglioti, L.; Palyi, G. Can. J.
Chem. 2005, 83, 882–893.

ꢀ
Ungvari et al.
3842 Organometallics, Vol. 29, No. 17, 2010
Figure 1. Change of the infrared spectra between 2200 and 1560 cm-1 during the catalytic carbonylation of ethyl diazoacetate (EDA)
in the presence of ethanol and [Co(CO)3(PPh3)2][Co(CO)4] (5) as the catalyst precursor under atmospheric pressure of carbon
monoxide at 10 °C in CH2Cl2 solution using a 0.007 65 cm CaF2 liquid cell: (black line) reaction time 30 min; (violet line) reaction time
90 min. Initial concentrations: [EDA]0 = [EtOH]0 = 0.25 mol/dm3, [5]0 = 0.0050 mol/dm3, [CO]0 = 0.00506 mol/dm3.
Figure 2. Change of the infrared spectra between 2200 and 1500 cm-1 during the catalytic carbonylation of ethyl diazoacetate (EDA)
in the absence of ethanol with [Co(CO)3(PPh3)2][Co(CO)4] (5) as the catalyst precursor under atmospheric pressure of carbon
monoxide at 10 °C in CH2Cl2 solution using a 0.021 80 cm CaF2 liquid cell: (blue line) reaction time 18 min; (violet line) reaction time 66
min; (green line) reaction time 134 min; (black line) reaction time 265 min. Initial concentrations: [EDA]0 = 0.0182 mol/dm3, [5]0 =
0.0045 mol/dm3, [CO]0 = 0.00504 mol/dm3.
as a solvent-dependent equilibrium mixture of the keto and
enol forms (eq 4).38
of light. Especially [Co(CO)3(PPh3)2][Co(CO)4] (complex 5)
turned out to be extremely sensitive to light. Its electronic
absorption spectrum in CH2Cl2 shows an outer-sphere metal
to metal charge transfer or intervalence transfer absorption
band at 390 nm (εM390 nm = 2645 cm2/mmol). Irradiation of
0.005 M solutions of complex 5 in CH2Cl2 at 366 nm results
in rapid (<1 min at 10 °C) formation of Co2(CO)6(PPh3)2
(complex 2) in the form of a sparingly soluble red-brown
precipitate similar to that obtained in acetone solution.39 If
the irradiation was performed with a 1:1 molar mixture
of Co2(CO)8 and [Co(CO)3(PPh3)2][Co(CO)4], the rapid
(<1 min at 10 °C) and quantitative formation of Co2(CO)7-
(PPh3) (complex 1) was observed. Under thermal reaction
The growing intensity of the band at 1637 cm-1 and the
declining intensity of the band at 1889 cm-1 show a linear
correlation (see Figure 5 in the Discussion).
In order to obtain reproducible initial rates, it was neces-
sary to run the carbonylation experiments with the exclusion
(38) Staudinger, H.; Becker, H. Ber. Bunsen-Ges. Phys. Chem. 1927,
50, 1016–1024.
(39) Vogler, A.; Kunkely, H. Organometallics 1988, 7, 1449–1450.

Article
Organometallics, Vol. 29, No. 17, 2010 3843
Figure 3. Change of the infrared spectra of [Co(CO)3(PPh3)2][Co(CO)4] (5) in CH2Cl2 solution ([5]0 = 0.005 mol/dm3) between 2110
and 1700 cm-1 under an atmospheric pressure of 13CO at 10 °C using a 0.0218 cm CaF2 liquid cell: (blue line) reaction time 65 s; (violet
line) reaction time 1080 s; (green line) reaction time 2135 s; (black line) reaction time 7585 s.
conditions (40 °C, in the dark) the above transformations
need more than 24 h for completion.
[Co(CO)4]- shows no parallelism between the rates. With
respect to the rates of 13CO exchange, Na[Co(CO)4] shows
the highest activity, in accord with our earlier observation,40
and complex [Co(CO)3(PPh3)2][Co(CO)4] shows the lowest
activity. The rates of 13CO exchange with the complexes
[Co(EtOH)6][Co(CO)4]2 and [PPN][Co(CO)4] are between
those observed with [Co(CO)3(PPh3)2][Co(CO)4] and Na-
[Co(CO)4]. In the presence of 100 and 50 mol % cobaltocene
the complexes 5 and 7 (entries 1, 2 and entries 3, 5, res-
pectively) show little change in activities in the CO exchange
and in the ethyl diazoacetate carbonylation reactions. Com-
plexes 8 and 9, however, show strongly diminished activi-
ties in these reactions if 50 mol % cobaltocene is present
(entries 6-8).
Commercially available polystyrenediphenylphosphane
(PSDP) (1.6 mmol P/g; 100-200 mesh particle size) swollen
in CH2Cl2 was treated with solutions of Co2(CO)8 in CH2Cl2
at room temperature. A mixture of the polymer-bound
analogues of the complexes [Co(CO)3(PPh3)2][Co(CO)4] (5)
and Co2(CO)7(PPh3) (1) was obtained as red-brown beads
after drying. Infrared spectra in Nujol mull and KBr pellet
confirmed the presence of the polymer-bound complexes.
The ratio of the polymer-bound complexes in the loaded
polymer was calculated from the material balance (see the
Experimental Section), which was in accord with the result of
Co and P analyses. Using this material as the catalyst in the
carbonylation of ethyl diazoacetate in CH2Cl2 solution give
the best results at 40 °C under 11 bar of CO pressure (see
Table 3 for the results under various reaction conditions).
Complete conversion into diethyl malonate was achieved,
and no cobalt leaching from the catalyst into the solution of
the product was observed by infrared spectroscopy. Turn-
over frequencies up to 5.1 mol/(mol h) were achieved. The
catalyst can be recycled, but the activity decreases in the
consecutive cycles (see entries 12-16). According to ICP-
OES analyses of the catalyst after the fifth cycle the phos-
phorus content is practically unchanged but the cobalt
content drops to 95.8% of the original value. In the absence
Figure 3 shows that under an atmosphere of 13CO at 10 °C
in the dark the intensity of the 2010 cm-1 band of complex 5
remained practically unchanged for more than 2 h but the
intensity of the band at 1889 cm-1 decreased dramatically
and was shifted to lower wavenumbers. The initial rate of CO
exchange based on the decrease of intensity of the band at
1889 is 0.44 ꢀ 10-5 mol of CO/(dm3 s).
The solvent-free reaction product of the 13CO exchange
experiment stored in the refrigerator for 10 days shows,
according to its infrared spectrum, a complete scrambling
of 13CO not only in the anionic part of [Co(CO)3(PPh3)2]-
[Co(CO)4] but in the cationic part as well. High 13CO
enrichment (ca. 75%) in both parts of [Co(CO)3(PPh3)2][Co-
(CO)4] was achieved by preparing 13CO-enriched (ca. 75%)
Co2(CO)8 in CH2Cl2 solution at first and then adding 2 equiv
of PPh3 to form the disproportionation product. The CO
ligands of 13CO-enriched [Co(CO)3(PPh3)2][Co(CO)4] show
13C resonances at 214.10 (octet, 1J(13C-59Co) = 280.1 Hz)
and 193.64 ppm (triplet, 2J(31P-13C) = 24.9 Hz) and a 31
P
resonance at 54.79 ppm (quartet, 2J(31P-13C) = 12.5 Hz).
Repeating the carbonylation experiment of ethyl diazoa-
cetate in the absence of ethanol but under a 13CO atmo-
sphere (Figure 4) confirmed the relatively rapid CO
exchange in the anionic part of complex 5 shown in Figure 3
and the correlation shown in Figure 2. In comparison
to Figure 2 the results in Figure 4 show that under 13CO
the appearance of the 1734 cm-1 band of the ketene dimer
is unchanged but the band at 1637 cm-1 is shifted to
1606 cm-1
.
We found that not only [Co(CO)3(PPh3)2][Co(CO)4] but
also [Co(EtOH)6][Co(CO)4]2 and Na[Co(CO)4] are active
catalyst precursors. Their catalytic activity in the carbonyla-
tion of ethyl diazoacetate, however, is lower than that obser-
ved with [Co(CO)3(PPh3)2][Co(CO)4]. On the other hand,
the complex [PPN][Co(CO)4]22 (PPN = [Ph3PdNdPPh3]þ)
shows no catalytic activity in the carbonylation reaction at
all (Table 2). A comparison of the rates of carbonylation
(rDEM) with those of CO exchange (rCO exchange) measured
under a 13CO atmosphere for the four different complexes of
ꢀ
(40) Ungvary, F.; Wojcicki, A. J. Am. Chem. Soc. 1987, 109, 6848–
6849.

ꢀ
Ungvari et al.
3844 Organometallics, Vol. 29, No. 17, 2010
Figure 4. Change of the infrared spectra between 2200 and 1500 cm-1 during the catalytic carbonylation of ethyl diazoacetate (EDA)
in the absence of ethanol with [Co(CO)3(PPh3)2][Co(CO)4] (5) as the catalyst precursor under atmospheric pressure of 13CO at 10 °C in
CH2Cl2 solution using a 0.021 80 cm CaF2 liquid cell: (blue line) reaction time 5 min; (light violet line) reaction time 30 min; (dark blue
line) reaction time 60 min; (green line) reaction time 120 min; (dark violet line) reaction time 180 min; (red line) reaction time 240 min.
Initial concentrations: [EDA]0 = 0.0186 mol/dm3, [5]0 = 0.0043 mol/dm3, [13CO]0 = 0.00508 mol/dm3.
Table 2. Initial Rate of 13CO Exchange (rCO exchangea) and the Initial Rate of Diethyl Malonate Formation (rDEMb) in the Presence of
Different Complexes of Tetracarbonylcobaltate(-): [Co(CO)3(PPh3)2][Co(CO)4] (5), [Co(EtOH)6][Co(CO)4]2 (7), Na[Co(CO)4] (8), and
[PPN][Co(CO)4] (9) under Atmospheric Pressure (736-745 mmHg) of 13CO and CO, Respectively, in CH2Cl2 Solution at 10 °C
complex
entry
[Co(CO)4-]0 (mol/dm3)
102[CO]0c (mol/dm3)
105rCOexchangea (mol/(dm3 s))
105rDEMb (mol/(dm3 s))
5
1
2
3
4
5
6
7
8
9
0.005
0.005
0.008
0.005
0.050
0.005
0.005
0.005
0.005
0.504
0.508a
0.790a
0.802
0.790
0.760a
0.760
0.510a
0.510
3.47 (3.19)d
0.44 (0.47)d
2.34e (2.23)f
7
0.16e
1.70e (1.41)f
8
9
3.67g (0.04)f
1.50 (0.01)f
1.68g (0.10)f
<0.001
a The CO exchange reaction started under an atmosphere of 13CO at 743 mmHg total pressure. b Carbonylation reaction of ethyl diazoacetate (EDA)
started with [EDA]0 = [EtOH]0 = 0.25 mol/dm3 under an atmosphere of CO. c Calculated from the partial pressure of CO and the known solubility of
CO in CH2Cl241,42 expressed in mole fraction: xCO(10 °C, 1 bar of P(CO)) = 0.000 473 converted to [CO] = 0.007 489 mol/dm3. Solubility of CO in
EtOH43 expressed in mole fraction: xCO(10 °C, 1 bar of P(CO)) = 0.000 487 converted to [CO] = 0.008 438 mol/dm3. The solubility of CO in THF [CO]
(10 °C, 1 bar of P(CO)) = 0.0087 ( 0.0011) was extrapolated from published values at 30 and 25 °C.44 d In the presence of 100 mol % cobaltocene.
e Reactions performed in EtOH solutions. f In the presence of 50 mol % cobaltocene. g Reactions performed in THF solutions.
of ethanol nearly complete conversion into 3-(ethoxycarbonyl)-
4-hydroxy-6-ethoxy-2H-pyran-2-one (dimer of (ethoxycarbo-
nyl)ketene) was obtained (entry 11).
isolated complexes, we performed experiments with the
latter at 40 °C as well. The data in Table 5 show that the
complexes 3, 1, and 5 in 0.005 mol/dm3 concentration are up
to 2 orders of magnitude more active under these conditions
than the polymer-bound complexes applied in 0.05 mmol
quantities/cm3 in the reaction mixture. Using complex 6 as
the catalyst precursor under the same conditions, a lower
yield and turnover frequency was observed than with the
other complexes (cf. entry 3 with entries 1, 2, and 4). At the
end of 1 h reaction time the infrared spectrum of the reaction
mixture shows the almost complete conversion of complex 6
into complex 1.
Using CH2Cl2 solutions of the complex Co2(CO)7-
(CHCO2Et) for the cobalt loading onto PSDP, we obtained
the polymer-bound analogue of the complex Co2(CO)6-
(CHCO2Et)(PPh3) (3) as dark red beads. This catalyst shows
activity in the carbonylation of ethyl diazoacetate lower than
that observed by using the mixture of 88% PSDP-anchored
“[Co(CO)3(PPh3)2][Co(CO)4]” and 12% PSDP-anchored
“Co2(CO)7(PPh3)” under the same reaction conditions (cf.
the results in Table 4 with those in Table 3). However, despite
its lower activity, the polymer-bound analogue of the com-
plex Co2(CO)6(CHCO2Et)(PPh3) is also a suitable catalyst
for the carbonylation of ethyl diazoacetate.
The 31P resonances of dichloromethane-swelled polystyr-
enediphenylphosphane (PSDP) beads are at -7.2 (60%) and
In order to have a comparison of the activity of the
polymer-bound catalysts with those of the corresponding
(42) Magee, M. P.; Li, H.-Q.; Morgan, O.; Hersh, W. H. Dalton
Trans. 2003, 387–394.
(43) Cargill, R. W., Ed. IUPAC Solubility Data Series; Pergamon
Press: Oxford, U.K., 1990; Vol. 43 (Carbon Monoxide), p 166.
(44) Payne, M. W.; Leussing, D. L.; Shore, S. G. J. Am. Chem. Soc.
1987, 109, 617–618.
(41) Jonasson, A.; Persson, O.; Rasmussen, P. J. Chem. Eng. Data
2000, 45, 642–646.

Article
Organometallics, Vol. 29, No. 17, 2010 3845
The low activity of complex 6 as the catalyst precursor and
its slow conversion during the catalytic reaction into com-
plex 1 does not conform to a catalytic cycle in which complex
6 and Co2(CO)5(CHCO2Et)(PPh3) are the repeating species
(Scheme 3). Rather, the observed catalytic activity in this
case can be assigned to complex 1, formed from complex 6.
The facile decomposition of complex 4 in dichloromethane
solution at 40 °C gives two distinct products according to
eq 5. We assume that the source of the second hydrogen atom
in the (ethoxycarbonyl)methyl group of EtO2CCH2Co(CO)3-
(PPh3) is the solvent. The insoluble byproduct [Co(CO)2-
(PPh3)]x may be the same as [Co(CO)2(PPh3)]n obtained from
[Co(CO)3(PPh3)2][Co(CO)4] in refluxing acetone solutions.47
Co2ðCOÞ5ðCHCO2EtÞðPPh3Þ2
40 °C, 15 min EtO2CCH2CoðCOÞ ðPPh3Þ
3
Figure 5. Correlation of the infrared band intensities at 1889 and
1637 cm-1 in [Co(CO)3(L)]- (L = CHCO2Et, OdCdCHCO2Et)
of a reaction mixture with the initial concentrations [5]0 = 0.0045
mol/dm3, [N2CHCO2Et]0 = 0.0182 mol/dm3, and [CO]0 =
0.00504 mol/dm3 as measured at the beginning (O) and at 66 (9),
134 (þ), 193 (ꢀ) and 265 min (b) reaction time using a 0.0218 cm
CaF2 liquid cell. 5 denotes the complex [Co(CO)3(PPh3)2]-
[Co(CO)4].
CH2Cl2
þ ð1=xÞ½CoðCOÞ2ðPPh3Þꢁx
ð5Þ
Research on various carbonylcobalt cation-anion pairs
has revealed already a rich thermal chemistry and photo-
chemistry.48,49 We believe that the observed high activity of
[Co(CO)3(PPh3)2][Co(CO)4] (5) in catalyzing the carbonyla-
tion of ethyl diazoacetate opens an interesting new chapter in
the chemistry of this classical homonuclear ion pair.
-8.0 (40%) (lit.45 -5.4 ppm for PPh3). The dichloro-
methane-swelled beads of the PSDP-bound phosphazane
show a 31P resonance at 19.3 ppm (lit.46 22.7 ppm for
Ph3PdNNdCHCO2Et). These signals are completely absent
in the PSDP-bound cobalt catalysts being used in carbonyla-
tion experiments of ethyl diazoacetate.
The lower intensity of the 1889 cm-1 band of complex 5
during the catalytic carbonylation of ethyl diazoacetate
suggests that the tetracarbonylcobaltate(-) part of the com-
plex is involved in the deazotization and carbonylation of
ethyl diazoacetate. For the deazotization of diazoalkanes
metal complexes with a vacant coordination site were found
to be active.50 A vacant coordination site in [Co(CO)3-
(PPh3)2][Co(CO)4] may be formed in the tetracarbonylco-
baltate(-) part of the complex, which then transfers the
diazoalkane into a carbene ligand and dinitrogen. The reac-
tion of the carbene ligand with carbon monoxide can result in
the formation of a ketene ligand. The observed correlation of
the decreasing band intensity at 1889 cm-1 with the increas-
ing band intensity of a new absorption at 1637 cm-1 (see
Figure 5) might be an indication of intermediate carbene
and/or ketene complexes. The good correlation shown in
Figure 5 and the absence of new ν(CtO) bands in the vicinity
of 1889 cm-1 supports the idea that the ν(CtO) band
position of the carbonyl ligands of [Co(CO)3(L)]- are coin-
cidentally the same as those of [Co(CO)4]-. The computed
average C-O bond distances in [Co(CO)3(L)]- (1.1640 and
4. Discussion
The carbonylation of ethyl diazoacetate using complex 1 or
complex 3as the catalyst precursor can be explained on the basis
of the observed negative effect of CO and on the basis of the
observed cobalt complexes in the reaction mixture during
catalysis by a catalytic cycle in which Co2(CO)6(PPh3) and Co2-
(CO)6(CHCO2Et)(PPh3) are the repeating species (Scheme 2).
The suggested catalytic cycle shown in Scheme 2 can be put
into effect in two ways: either by adding complex 1 or by
adding complex 3 to the reaction mixture. The presence of
both complexes can be seen by infrared spectroscopy during
the catalytic reaction. We have already shown that the
(ethoxycarbonyl)carbene ligand in complex 3 can be driven
out by CO with the simultaneous quantitative formation of
complex 1 and diethyl malonate (if equimolar amounts of
ethanol were also present).12 The bottleneck of the cycle is
most probably the dediazotization of ethyl diazoacetate by
complex 1. The observed negative effect of carbon monoxide
on the rate of the reaction suggests that the actual reaction
partner of ethyl diazoacetate in the deazotization reaction is
the coordinative unsaturated Co2(CO)6(PPh3) formed in
undetectable low concentrations by reversible CO loss from
complex 1. Diethyl malonate is the product of the fast
reaction of the (ethoxycarbonyl)ketene intermediate with
ethanol. In the absence of ethanol the (ethoxycarbonyl)-
ketene dimerizes to 3-(ethoxycarbonyl)-4-hydroxy-6-ethoxy-
2H-pyran-2-one.38
˚
1.1643 A) are indeed very close to the C-O bond distances in
-1
[Co(CO)4]- (1.1690 A). The band at 1637 cm can be
˚
assigned either to the ν(CdO) of the (ethoxycarbonyl)-
carbene ligand in [Co(CO)3(CHCO2Et)]- or to the ester
ν(CdO) and the ketene ν(CdO) of the (ethoxycarbonyl)ketene
(47) Sacco, A. Gazz. Chim. Ital. 1963, 93, 542–548.
(48) (a) Atwood, J. D. Inorg. Chem. 1987, 26, 2918–2920. (b) Zhen,
Y. Q.; Feighery, W. G.; Atwood, J. D. J. Am. Chem. Soc. 1991, 113, 3616–
3618.
(49) (a) Bockman, T. M.; Kochi, J. K. J. Am. Chem. Soc. 1989, 111,
4669–4683. (b) Kochi, J. K.; Bockman, T. M. Adv. Organomet. Chem. 1991,
33, 51–124. (c) Wei., C. H.; Bockman, T. M.; Kochi, J. K. J. Organomet.
Chem. 1992, 428, 85–97. (d) Kochi, J. K. Adv. Phys. Org. Chem. 1994, 29,
185–272and references therein.
(45) Albright, T. A.; Freeman, W. J.; Schweizer, E. J. Org. Chem.
1976, 41, 2716–2720.
€
(46) Pedro, F. M.; Santos, A. M.; Baratta, W.; Kuhn, F. E. Organo-
(50) Doyle, M. P.; McKervey, M. A.; Ye, T. Modern Catalytic
Methods for Organic Synthesis with Diazo Compounds; Wiley: New York,
1998; pp 62-66.
metallics 2007, 26, 302–309.

ꢀ
Ungvari et al.
3846 Organometallics, Vol. 29, No. 17, 2010
Table 3. Yield of Diethyl Malonate (DEM) from Ethyl Diazoacetate (EDA) under Carbon Monoxide Pressure in the Presence of 75 mg
(0.05 mmol “Co2”) Polystyrenediphenylphosphane (PSDP)-Anchored Carbonyl Cobalt Catalyst (7.87 m/m % Co and 3.68 m/m % P,
Corresponding to a Mixture of 88% PSDP-Anchored “[Co(CO)3(PPh3)2][Co(CO)4]” and 12% PSDP-Anchored “Co2(CO)7(PPh3)”) in
1.0 cm3 Methylene Chloride Solution under Various Reaction Conditions
entry
T (° C)
Ptotal (bar)
[EDA]0 (mol/dm3)
[EtOH]0 (mol/dm3)
reacn time (h)
yield of DEMa (%)
TOFb (h-1
)
1
2
3
4
5
6
7
8
10
25
40
40
40
40
40
40
40
40
40
40
40
40
40
40
11
11
11
11
11
11
11
11
5
50
50
11
11
11
11
11
0.25
0.25
0.25
0.50
1.00
2.00
4.00
6.00
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.30
0.30
0.30
0.60
1.00
2.00
4.00
6.00
0.30
0.30
0.30
0.30
0.30
0.30
0.30
0.30
4
4
4
4
4
4
4
4
4
4
4
4
4
16
26
29
14
66
100
99
99
50
26
17
0.18
0.83
1.25
2.48
4.95
5.01
5.10
5.10
1.19
1.18
1.20c
1.23
1.21
0.31
0.16
0.11
9
95
10
11
12
13
14
15
16
94
96c
98
97d
99e
83f
61g
1730 cm-1
a Determined by quantitative infrared spectroscopy using the molar absorbances of DEM in CH2Cl2 at 1730 and 1749 cm-1 (εM
mmol, εM
= 666 cm2/
1749 cm-1
= 579 cm2/mmol). b In units of mol of product/((mol of catalyst) h). c 3-(Ethoxycarbonyl)-4-hydroxy-6-ethoxy-2H-pyran-2-one
1734 cm-1
(dimer of ethoxycarbonylketene), determined by quantitative infrared spectroscopy using the molar absorbance in CH2Cl2 at 1734 cm-1 (εM
=
422 cm2/mmol). The TOF was calculated for the monomer. d Catalyst of entry 12 recycled. e Catalyst of entry 13 recycled. f Catalyst of entry 14 recycled.
g Catalyst of entry 15 recycled. IPC-OES analysis of the recovered catalyst: 3.65 m/m % P and 7.54 m/m % Co.
Table 4. Yield of Diethyl Malonate (DEM) from Ethyl Diazoacetate (EDA) under Carbon Monoxide Pressure in the Presence of 51.3 mg
(0.05 mmol “Co2”) of Polystyrenediphenylphosphane (PSDP)-Anchored “Co2(CO)6(CHCO2Et)(PPh3)” (11.10 m/m % Co and 3.02 m/m
% P), in 1.0 cm3 Dichloromethane Solution under Various Reaction Conditions
entry
T (° C)
Ptotal (bar)
[EDA]0 (mol/dm3)
[EtOH]0 (mol/dm3)
reacn time (h)
yield of DEMa (%)
TOFb (h-1
)
1
2
3
4
5
6
25
40
40
40
40
40
11
11
11
11
11
11
0.25
0.25
0.25
1.00
2.00
4.00
0.30
0.30
0.30
1.20
2.00
4.00
24
4
24
24
24
24
14
23
92
95
53
29
0.03
0.29
0.20
0.79
0.88
0.97
1730 cm-1
a Determined by quantitative infrared spectroscopy using the molar absorbances of DEM in CH2Cl2 at 1730 and 1749 cm-1 (εM
mmol, εM
= 666 cm2/
1749 cm-1
= 579 cm2/mmol). b In units of mol of product/((mol of catalyst) h).
Table 5. Yield of Diethyl Malonate from Ethyl Diazoacetate (EDA) at 1 h Reaction Time in the Presence of Ethanol under 11 bar
Total Pressure of Carbon Monoxide at 40 °C in the Presence of Co2(CO)6(CHCO2Et)(PPh3) (3), Co2(CO)7(PPh3) (1),
Co2(CO)5(CHCO2Et)2(PPh3) (6), or [Co(CO)3(PPh3)2][Co(CO)4] (5) as the Catalyst Precursors in 0.005 mol/dm3 Initial
Concentration and the Calculated Turnover Frequency (TOF)
entry
catalyst
[EDA]0 (mol/dm3)
[EtOH]0 (mol/dm3)
[EDA]0/[catalyst]0
yield of DEMa (%)
TOFb (h-1
)
1
2
3
4
5
6
7
8
3
1
6
5
5
5
5
5
0.25
0.25
0.25
0.25
0.50
1.00
2.50
5.00
0.30
0.30
0.30
0.30
0.60
1.20
3.00
5.00
50
50
50
50
100
200
500
1000
76
70
10
29
19
19
19
10
60
35
4
15
19
38
95
101
1730 cm-1
a Determined by quantitative infrared spectroscopy using the molar absorbances of DEM in CH2Cl2 at 1730 and 1749 cm-1 (εM
mmol, εM
= 666 cm2/
1749 cm-1
= 579 cm2/mmol) after 1 h reaction time. b In units of mol of product/((mol of catalyst) h).
ligand in [Co(CO)3(OdCdCHCO2Et)]-. The shift of the
observed band at 1637 cm-1 to 1606 cm-1 in the experiment
performed under 13CO supports the assumption that we
see in the IR spectra an anionic ketene complex rather
than an anionic carbene complex. The existence of [Co-
(CO)3(CHCO2Et)]- and [Co(CO)3(OdCdCHCO2Et)]-
has not been described in the literature so far. The only
anionic cobalt carbonyl carbene complexes known at
present are those having an alkyl or an aryl isocyanide
ligand.51 These complexes were prepared by the reac-
tion of alkyl and aryl isocyanides with the highly re-
duced tricarbonylcobaltate(3-) anion. Examples of anio-
nic ketene complexes of some transition metals are also
known.52
(51) Ellis, J. E.; Barger, P. T.; Winzenburg, M. L.; Warnock, G. F.
J. Organomet. Chem. 1990, 383, 521–530.
(52) Geoffroy, G. L.; Bassner, S. L. Adv. Organomet. Chem. 1988, 28,
1–83.

Article
Scheme 2. Suggested Mechanism for Catalytic Diethyl
Organometallics, Vol. 29, No. 17, 2010 3847
of ethyl diazoacetate, different mechanisms might operate in
each case.
Malonate Formation using Co2(CO)7(PPh3) or
Co2(CO)6(CHCO2Et)(PPh3) as the Precatalyst for the
Carbonylation of Ethyl Diazoacetate
½CoðCOÞ3ðPPh3Þ2ꢁ½CoðCOÞ4ꢁ þ x13CO
13
h½CoðCOÞ3ðPPh3Þ2ꢁ½Coð COÞxðCOÞ4 - xꢁ þ xCO ð6Þ
It is known that the carbonyl ligands in Na[Co(CO)4] can
be substituted by 13CO, phosphites, phosphanes, and acti-
vated olefins in facile thermal and photochemical reac-
tions.40,53 According to theoretical calculations the thermal
formation of [Co(CO)3]- from [Co(CO)4]- by CO dissocia-
tion needs too much energy (42.4 kcal/mol54). The observa-
tion that the presence of catalytic amounts of Co2(CO)8 or
Co4(CO)12 are necessary for the 13CO scrambling in [Co-
(CO)4]- 55 suggests an energetically much favorable radical
pathway with the involvement of the tetracarbonylcobalt
radical in both 13CO scrambling and in the reaction with
ethyl diazoacetate. For the former the reactions in Scheme 5
and for the latter the reaction sequence in Scheme 6 seem to
be more favorable. Essential for keeping the reaction cata-
lytic in both cases are the facile one-electron exchange
reactions between the 17e radical species and the tetracar-
bonylcobaltate(-) ion in the final steps. Traces of oxygen can
serve for the initial formation of the tetracarbonylcobalt
radical by oxidation of Na[Co(CO)4] and of [PPN][Co-
(CO)4]. In accord with this idea is the experimental observa-
tion that there are strongly diminished rates of 13CO scram-
bling and ethyl diazoacetate carbonylation if the highly
reducing cobaltocene is present in the reaction mixture in
the case of Na[Co(CO)4] and of [PPN][Co(CO)4].
Scheme 3. Tentative Mechanism for the Carbonylation of
Ethyl Diazoacetate using Co2(CO)6(CHCO2Et)(PPh3) or
Co2(CO)5(CHCO2Et)2(PPh3) as the Catalyst Precursor
For the formation of [•Co(CO)4] the presence of Co2(CO)8
is not necessary in the case of complex 5. The transfer of an
electron from [Co(CO)4]- to [Co(CO)3(PPh3)2]þ in complex
5 under thermal or photochemical conditions results in the
formation of a radical pair according to eq 7. Hence, radical
pathways such as those in Schemes 5 and 6 may work by
starting with complex 5 alone in the absence of Co2(CO)8.
According to the correlation shown in Figure 5 a log
I0/I value of 0.0214 at 1637 cm-1 corresponds to a composi-
tion of [Co(CO)3(PPh3)2][Co(CO)3(L)] with 0.0045 mol/dm3
concentration. Taking into account the thickness of the
½CoðCOÞ3ðPPh3Þ2ꢁ½CoðCOÞ4ꢁ½•CoðCOÞ3ðPPh3Þ2ꢁ½•CoðCOÞ4ꢁ
ð7Þ
applied solution cell (0.0218 cm), a molar absorbance of
= 218 cm2/mmol can be calculated for the sup-
1637 cm-1
εM
A similar one-electron transfer may occur from [Co(CO)4]- to
[Co(EtOH)6]2þ in complex 7 as well. Interestingly, in contrast to
the observations with the complexes Na[Co(CO)4] and [PPN]-
[Co(CO)4] the presence of cobaltocene does not affect the cata-
lytic activities of complexes 5and 7in the CO exchange and in the
ethyl diazoacetate carbonylation reaction significantly.
4.1. Computational Results of the CO Exchange Reaction
of [Co(CO)4]- and •Co(CO)4. In order to check the likelihood
of the involvement of anionic and/or radical carbonyl cobalt
species in the 13CO scrambling and in the carbonylation
reactions, we performed theoretical calculations. The most
appealing difference in the structures of the anionic [Co-
(CO)4]- (10) and the radical •Co(CO)4 (10r) is the significant
difference in cobalt-carbon bond lengths (see Figure 6), as it
posed [Co(CO)3(PPh3)2][Co(CO)3(L)] at 1637 cm-1
.
Since the concentrations of CO and PPh3 have a negative
effect on the rate of the catalytic carbonylation of ethyl
diazoacetate, we assume the reversible formation of the
[Co(CO)3(PPh3)2][Co(CO)3] intermediate. Reaction of this
intermediate with ethyl diazoacetate might initiate the dea-
zotization of ethyl diazoacetate and the formation of the
corresponding complex carbenoid [Co(CO)3(PPh3)2][Co(CO)3-
(CHCO2Et)], which upon reaction with CO (external or coor-
dinated) gives (ethoxycarbonyl)ketene, and a new cycle can start
according to Scheme 4.
The observed zero-order ethanol dependence of the diethyl
malonate formation suggests that the reaction of (ethoxy-
carbonyl)ketene with ethanol is faster than its formation in
the catalytic cycle. In the absence of ethanol (ethoxycarbonyl)-
ketene dimerizes into 3-(ethoxycarbonyl)-4-hydroxy-6-ethoxy-
2H-pyran-2-one.38
ꢀ
(53) Ungvary, F.; Gallucci, J.; Wojcicki, A. Organometallics 1991, 10,
3053–3062.
(54) Huo, C. F.; Li, Y. W.; Wu, G. S.; Beller, M.; Jiao, H. J. Phys.
Chem. A 2002, 106, 12161–12169.
(55) Fachinetti, G.; Funaioli, T. Angew. Chem., Int. Ed. Engl. 1992,
31, 1596–1599.
Although the CO exchange in the anionic part of complex
5 observed in the presence of 13CO according to eq 6 occurs
on a time scale similar to that for the catalytic carbonylation

ꢀ
Ungvari et al.
3848 Organometallics, Vol. 29, No. 17, 2010
Scheme 4. Suggested Mechanism for the Catalytic Diethyl Malonate Formation using [Co(CO)3(PPh3)2][Co(CO)4] (5) as the Precatalyst
for the Carbonylation of Ethyl Diazoacetate
found on the •Co(CO)3(N2CHCOOEt) potential energy
hypersurface, which describes the coordination of EDA on
the R-carbon to the cobalt center with simultaneous N2
extrusion. The characteristic single imaginary frequency
for 14rTS is 533i cm-1. IRC calculations revealed that
14rTS is connected with 13r, a diazo complex with an η1
coordination mode. The formation of 13r from 11r and EDA
is endothermic in terms of free energy by 6.7 kcal/mol (6.4
kcal/mol in gas phase). The free energy barrier for the N2
extrusion and carbene formation was found to be 15.6 kcal/
mol for the gas phase and 16.4 kcal/mol including sol-
vation effects. The reaction of 11r with EDA results in
carbene complex 15r possessing Cs symmetry with a reac-
tion free energy of -20.0 kcal/mol (-21.7 kcal/mol in gas
phase). The structures of 13r, 14rTS, and 15r are depicted in
Figure 7.
Scheme 5. Suggested Radical Pathway for the 13CO Exchange in
[Co(CO)4]-
˚
is remarkably short (1.769 A) in 10, whereas the distance of
˚
the longer Co-C bond in 10r is 1.858 A. The anionic
complex [Co(CO)4]- possesses Td symmetry, whereas Co-
•
(CO)4 has a trigonal-pyramidal C3v structure, in accord with
earlier theoretical results.54 The much higher bond dissocia-
tion free energy (ΔGd) for 10 than for 10r is in line with the
difference in Co-C bond lengths in the two complexes. For
[Co(CO)4]- ΔGd = 35.0 kcal/mol in the gas phase and
ΔGd = 32.0 kcal/mol when solvation effects are estimated,
considering dichloromethane as solvent. For the 17e com-
The carbene group in 15r couples with one of the CO
ligands connected to the cobalt center with shorter bond
˚
length (1.878 A). The intramolecular reaction takes place via
•
plex Co(CO)4 the corresponding gas-phase and solvated
transition state 16rTS with a free energy barrier of 15.5 kcal/
mol (16.9 kcal/mol in the gas phase). The characteristic
imaginary frequency for 16rTS is 424i cm-1. The CO-
carbene coupling results in the 15e species 17r with a reaction
free energy of -7.0 kcal/mol (-5.1 kcal/mol in the gas
phase). Complex 17r takes up one CO in an exothermic
reaction (the reaction free energy is predicted to be -12.9
kcal/mol in the dichloromethane phase and -14.5 kcal/mol
in the gas phase), resulting in the 17e complex 18r. The ketene
elimination step is predicted to be exothermic as well, with a
free energy of -3.2 kcal/mol (-2.0 kcal/mol in the gas
phase). This step proceeds in a fast reaction, as the free
energy barrier is only 5.1 kcal/mol (5.2 kcal/mol in the gas
phase). The characteristic imaginary frequency for the tran-
sition structure 19rTS describing the ketene dissociation is
338i cm-1. Intrinsic reaction coordinate (IRC) calculations
revealed that 19rTS indeed connects 18r with the dissociated
(ethoxycarbonyl)ketene along with the coordinatively unsa-
turated 11r, which is ready to take up EDA, thus closing the
catalytic cycle. The minima and the transition-state struc-
tures involved in the ketene formation and elimination step
are depicted in Figure 8.
dissociation free energies are 15.2 and 14.5 kcal/mol, respec-
tively. These results support the idea that the 15e •Co(CO)3
complex (11r) is involved in both CO exchange and diazo
activation.
However, for CO exchange of 10r an alternative reaction
route seems more favorable than the dissociative pathway.
CO addition and the dissociation of a carbonyl ligand can
take place following an SN2-like reaction mechanism via
transition structure 12rTS with D3h symmetry. The free
energy barrier for this process is 7.4 kcal/mol. Including
solvation effects, the activation free energy is slightly ele-
vated to 7.7 kcal/mol. The somewhat similar SN2-like transi-
tion structure 12TS was found for the CO exchange of 10 as
well; however, the free energy is much higher than that for
the radical pathway, namely 40.8 kcal/mol, and 43.6 kcal/
mol when solvation correction is applied. Instead of a
D3h transition state, 12TS possesses C2 symmetry with the
shortest Co-C bond as the rotation axis.
4.2. Computational Results of the Diazo Activation and
Ketene formation. We made several attempts to find a diazo
activation pathway with an SN2-like reaction mechanism
starting from 10r and ethyl diazoacetate (EDA), but no
transition state similar to 12rTS was found, suggesting the
dissociative mechanism to be the only viable pathway. Thus,
for the diazo activation step the transition state 14rTS was
The calculated free energy profile of the ketene formation
reaction catalyzed by •Co(CO)3 is shown in Figure 9.
The entire anionic pathway has been computed as well
in order to make a comparison with the radical ketene

Article
Organometallics, Vol. 29, No. 17, 2010 3849
Scheme 6. Suggested Radical Pathway for the Formation of [Co(CO)3(CHCO2Et)]- and [Co(CO)3(OdCdCHCO2Et)]-
Figure 8. Computed structures of the transition state for in-
tramolecular ketene formation 16rTS, the coordinative unsatu-
rated ketene complex 17r, the coordinative saturated ketene
complex 18r, and the transition state for ketene dissociation
Figure 6. Computed structures of [Co(CO)4]- (10), [Co(CO)3]-
(11), Co(CO)4 (10r), Co(CO)3 (11r), and the transition states
˚
12rTS and 12TS. Bond distances are given in A.
•
•
˚
19rTS. Bond distances are given in A.
The transition-state structure for the N2 extrusion 14TS
showssomesimilaritytothe radical14rTS, althoughtheC-N
bond as well as the Co-Ccarbonyl bonds are notably shorter in
the former. The characteristic single imaginary frequency for
14TS is 527i cm-1. According to the IRC calculations the
transition state 14TS is connected with diazo complex 13,
which possesses η2(C-N) coordination. This kind of diazo
coordination to transition metals is postulated,56 but no
example has been reported experimentally or theoretically,
to the best of our knowledge. It must be noted, however, that
the formation of 13 from 11 and EDA is endothermic by 10.3
kcal/mol (2.7 kcal/mol in the gas phase); therefore, both the
free energy of activation (17.9 kcal/mol with solvation correc-
tions and 12.6 kcal/mol in the gas phase) and the reaction free
energy (-24.6 kcal/mol with solvation corrections and -30.1
kcal/mol in the gas phase) for the formation of the unsatu-
rated carbene intermediate complex 15 are given with respect
to 11 and EDA.
The connection of the carbene carbon to the neighbo-
ring carbonyl carbon via transition state 16TS takes place
with a smaller barrier compared to that for the pathway
Figure 7. Computed structure of the diazo adduct 13r, transi-
•
tion state 14rTS, and carbene complex Co(CO)3(CHCOOEt)
˚
(15r). Bond distances are given in A.
formation mechanism. The computed structures and the
transition states obtained from 11 are depicted in Figure 10.
(56) Dartiguenave, M.; Menu, M. J.; Deydier, E.; Dartiguenave, Y.;
Siebald, H. Coord. Chem. Rev. 1998, 178-180, 623–663.

ꢀ
Ungvari et al.
3850 Organometallics, Vol. 29, No. 17, 2010
Figure 9. Calculated free energy profile of (ethoxycarbonyl)ketene formation catalyzed by •Co(CO)3.
The dissociation of (ethoxycarbonyl)ketene from 18 pro-
ceeds in a barrierless manner, providing tricarbonylcobal-
tate(-) (11) and closing the catalytic cycle. The free energy of
the dissociation is 11.3 kcal/mol including solvation correc-
tions and 18.1 kcal/mol in the gas phase.
The calculated free energy profile of the ketene formation
reaction catalyzed by [Co(CO)3]- is shown in Figure 11.
During the catalytic carbonylation of ethyl diazoacetate,
we found a correlation between the decreasing ν(CtO)
intensity of [Co(CO)4]- and the increasing intensity of a
new ν(CdO) band at 1637 cm-1 (see Figure 5). This new
band might belong to an anionic species such as either [Co-
(CO)3(CHCOOEt)]- (15) or [Co(CO)3(OdCdCHCOOEt)]-
(18), which are the anionic counterparts of 15r and 18r,
respectively, and may be formed by taking one electron from
[Co(CO)4]-. Ketene formation via the anionic carbene inter-
mediate cannot be excluded; the radical pathway seems to be
more plausible. Calculations suggest for both pathways that the
rate-determining step of the catalytic cycle is the activation of
ethyl diazoacetate; thus, the N2 extrusion step results in the
corresponding carbene complexes.
Mechanisms similar to those for the homogeneous cases
can be assumed for the corresponding polystyrene-sup-
ported catalysts as well. On the supported catalysts, how-
ever, the site of the reaction is confined to the solvent-swollen
polymer beads and restricted material transport may explain
the much lower activities compared to those for the homo-
geneous cases. In accord with this assumption the turnover
frequencies show no linear correlation with the initial con-
centration of ethyl diazoacetate above an initial EDA/cata-
lyst molar ratio of 20 (see Table 3, entries 3-8). Under
homogeneous reaction conditions complex 3 shows higher
catalytic activity than either complex 1 or complex 5
(compare the TOF values of entries 1, 2, and 4 in Table 5).
The order of activity is 3 > 1 > 5. The polymer-bound
analogues of 3 and 5 show the reverse order of activity under
the same reaction conditions. A reason for the reverse order
of activity is that the activation of the substrate occurs on
the [Co(CO)]4 part of the polymer-bound analogue of 5,
which is on the more accessible periphery of the catalyst
beads, whereas the substrate activation in the case of the
polymer-bound analogue of 3 occurs closer to the polymer
core.
Figure 10. Computed structures of the diazo adduct 13, the
transition state 14TS, the ionic carbene complex [Co(CO)3-
(CHCOOEt)]- (15), the transition state for intramolecular
ketene formation 16TS, the coordinative unsaturated ketene
complex 17, and the coordinative saturated ketene complex
[Co(CO)3(OdCdCHCOOEt)]- (18). Selected bond distances
˚
are given in A.
15r f 16rTS f 17r. The free energy of activation for this step
is 8.5 kcal/mol (8.3 kcal/mol in the gas phase); the free energy
of the reaction resulting in the coordinative unsaturated
ketene complex 17 is -9.5 kcal/mol (-9.0 kcal/mol in gas
phase). The characteristic single imaginary frequency for
16TS is 271i cm-1. Complex 17 takes up one external carbon
monoxide to form the saturated ketene complex 18.

Article
Organometallics, Vol. 29, No. 17, 2010 3851
Figure 11. Calculated free energy profile of (ethoxycarbonyl)ketene formation catalyzed by [Co(CO)3]-.
5. Conclusion
and the Hungarian Scientific Research Fund for financial
support under Grant No. OTKA NK 71906. We thank the
Supercomputer Center of the National Information Infra-
structure Development (NIIF) Program for support and
the Flextra Lab KFT for access to the Bruker Tensor 27
FTIR spectrometer. T.K. acknowledges the support of a
Bolyai Grant of the Hungarian Academy of Sciences and
the support of the Research Team on Innovation (SROP-
4.2.2/08/1/2008-2011) project. We thank the reviewers for
their valuable suggestions.
Modifying Co2(CO)8 and Co2(CO)7(CHCO2Et) by PPh3 in
the form of Co2(CO)7(PPh3), Co2(CO)6(CHCO2Et)(PPh3),
and [Co(CO)3(PPh3)2][Co(CO)4] results in catalysts with en-
hanced activity in the selective carbonylation of ethyl diazoa-
cetate to (ethoxycarbonyl)ketene under ambient conditions. In
the presence of ethanol the highly reactive (ethoxycarbonyl)-
ketene is trapped in situ and gives diethyl malonate as the final
carbonylation product. Substituting PPh3 in Co2(CO)7(PPh3),
Co2(CO)6(CHCO2Et)(PPh3), and [Co(CO)3(PPh3)2][Co(CO)4]
by polymer-bound PPh3 results in active and reusable catalysts
for the selective carbonylation of ethyl diazoacetate in dichloro-
methane solution at 40 °C and 11 bar pressure with up to 5.1 mol
of product/((mol of catalyst) h) turnover frequency.
Supporting Information Available: Figures giving solution IR
spectra of the complexes 1 and 5-9, solution and solid-state IR
spectra of [Co(CO)2(PPh3)]x, and microscopic views of the
PSDP-bound complexes and figures and CIF files giving
X-ray structures and crystallographic data for complex 1 and
EtO2CCH2Co(CO)3(PPh3). This material is available free of
charge via the Internet at http://pubs.acs.org.
Acknowledgment. We thank Profs. C. D. Hoff (Coral
ꢀ
Gables) and G. Palyi (Modena) for stimulating discussions