Spectroscopic detection and theoretical confirmation of the role of Cr 2(CO)5(C5R5)2 and ·Cr(CO)2(ketene)(C5R5) as intermediates in carbonylation of N=N=CHSiMe3 to O=C=CHSiMe3 by ·Cr(CO)3(C5R5) (R = H, CH3)

Article abstract of DOI:10.1021/ja075008o

Conversion of N=N=CHSiMe₃ to O=C=CHSiMe₃ by the radical complexes ·Cr(CO)₃C₅R₅ (R = H, CH₃) derived from dissociation of [Cr(CO)₃(C ₅R₅)]₂ have been investigated under CO, Ar, and N₂ atmospheres. Under an Ar or N₂ atmosphere the reaction is stoichiometric and produces the Cr≡Cr triply bonded complex [Cr(CO)₂(C₅R₅)]₂. Under a CO atmosphere regeneration of [Cr(CO)₃(C₅R₅)] ₂ (R = H, CH₃) occurs competitively and conversion of diazo to ketene occurs catalytically as well as stoichiometrically. Two key intermediates in the reaction, ·Cr(CO)₂(ketene)(C ₅R₅) and Cr₂(CO)₅(C ₅R₅)₂ have been detected spectroscopically. The complex ·Cr(¹³CO)₂(O=¹³C=CHSiMe ₃)(C₅Me₅) has been studied by electron spin resonance spectroscopy in toluene solution: g(iso) = 2.007; A(⁵³Cr) = 125 MHz; A(¹³CO) = 22.5 MHz; A(O=¹³C=CHSiMe₃) = 12.0 MHz. The complex Cr₂(CO)₅(C₅H ₅)₂, generated in situ, does not show a signal in its ¹H NMR and reacts relatively slowly with CO. It is proposed to be a ground-state triplet in keeping with predictions based on high level density functional theory (DFT) studies. Computed vibrational frequencies are also in good agreement with experimental data. The rates of CO loss from ³Cr₂(CO)₅-(C₅H₅) ₂ producing ¹[Cr(CO)₂(C₅H ₅)]₂ and CO addition to ³Cr₂(CO) ₅(C₅H₅)₂ producing ¹[Cr(CO)₃(C₅H₅)]₂ have been measured by kinetics and show ΔH^? ? 23 kcal mol^-1 for both processes. Enthalpies of reduction by Na/Hg under CO atmosphere of [Cr(CO)ⁿ(C₅H₅)]₂ (n = 2,3) have been measured by solution calorimetry and provide data for estimation of the Cr≡Cr bond strength in [Cr(CO)₂(C₅H ₅)]₂ as 72 kcal mol^-1. The complex [Cr(CO) ₂(C₅H₅)]₂ does not readily undergo ¹³CO exchange at room temperature or 50°C implying that ³Cr₂(CO)₅(C₅H₅) ₂ is not readily accessed from the thermodynamically stable complex [Cr(CO)₂(C₅H₅)]₂. A detailed mechanism for metalloradical based conversion of diazo and CO to ketene and N₂ is proposed on the basis of a combination of experimental and theoretical data.

Full text of DOI:10.1021/ja075008o

Published on Web 10/26/2007
Spectroscopic Detection and Theoretical Confirmation of the
Role of Cr2(CO)5(C5R5)2 and ‚Cr(CO)2(ketene)(C5R5) as
Intermediates in Carbonylation of NdNdCHSiMe3 to
OdCdCHSiMe3 by ‚Cr(CO)3(C5R5) (R ) H, CH3)
George C. Fortman,† Tama´s Ke´gl,*,‡ Qian-Shu Li,*,§ Xiuhui Zhang,§
Henry F. Schaefer III,| Yaoming Xie,| R. Bruce King,*,| Joshua Telser,*, and
Carl D. Hoff*,†
Contribution from the Department of Chemistry, UniVersity of Miami, Coral Gables, Florida,
33126, Department of Organic Chemistry, UniVersity of Pannonia, Veszpre´m, Hungary, The
Institute for Chemical Physics, Beijing Institute of Technology, Beijing, 100081 People’s
Republic of China, Department of Chemistry and Center for Computational Chemistry,
UniVersity of Georgia, Athens, Georgia 30606, and Department of Biological, Chemical and
Physical Sciences, RooseVelt UniVersity, Chicago, Illinois 60605
Received July 6, 2007; E-mail: c.hoff@miami.edu
Abstract: Conversion of NdNdCHSiMe3 to OdCdCHSiMe3 by the radical complexes ‚Cr(CO)3C5R5 (R )
H, CH3) derived from dissociation of [Cr(CO)3(C5R5)]2 have been investigated under CO, Ar, and N2
atmospheres. Under an Ar or N2 atmosphere the reaction is stoichiometric and produces the CrtCr triply
bonded complex [Cr(CO)2(C5R5)]2. Under a CO atmosphere regeneration of [Cr(CO)3(C5R5)]2 (R ) H, CH3)
occurs competitively and conversion of diazo to ketene occurs catalytically as well as stoichiometrically.
Two key intermediates in the reaction, ‚Cr(CO)2(ketene)(C5R5) and Cr2(CO)5(C5R5)2 have been detected
spectroscopically. The complex ‚Cr(13CO)2(O)13CdCHSiMe3)(C5Me5) has been studied by electron spin
resonance spectroscopy in toluene solution: g(iso) ) 2.007; A(53Cr) ) 125 MHz; A(13CO) ) 22.5 MHz;
A(Od13CdCHSiMe3) ) 12.0 MHz. The complex Cr2(CO)5(C5H5)2, generated in situ, does not show a signal
1
in its H NMR and reacts relatively slowly with CO. It is proposed to be a ground-state triplet in keeping
with predictions based on high level density functional theory (DFT) studies. Computed vibrational
frequencies are also in good agreement with experimental data. The rates of CO loss from 3Cr2(CO)5-
(C5H5)2 producing 1[Cr(CO)2(C5H5)]2 and CO addition to 3Cr2(CO)5(C5H5)2 producing 1[Cr(CO)3(C5H5)]2 have
been measured by kinetics and show ∆H‡ = 23 kcal mol-1 for both processes. Enthalpies of reduction by
Na/Hg under CO atmosphere of [Cr(CO)n(C5H5)]2 (n ) 2,3) have been measured by solution calorimetry
and provide data for estimation of the CrtCr bond strength in [Cr(CO)2(C5H5)]2 as 72 kcal mol-1. The
complex [Cr(CO)2(C5H5)]2 does not readily undergo 13CO exchange at room temperature or 50 °C implying
that 3Cr2(CO)5(C5H5)2 is not readily accessed from the thermodynamically stable complex [Cr(CO)2(C5H5)]2.
A detailed mechanism for metalloradical based conversion of diazo and CO to ketene and N2 is proposed
on the basis of a combination of experimental and theoretical data.
1. Introduction
methylsilyl)diazomethane is catalyzed by Co2(CO)8 and selec-
tively produces (trimethylsilyl)ketene:
The generation of ketenes (R2CdCdO) is important because
of their prominent role in organic synthesis.1 In this connection
diazoalkanes (R2CdNdN) can serve as ketene sources by their
thermal or photochemical reaction with CO. Ungva´ry and
co-workers2 have recently shown that carbonylation of (tri-
Co2(CO)8
NdNdCHSiMe3 + CO
8 OdCdCHSiMe3 + N2 (1)
Such methods provide an attractive route to the in situ generation
of ketenes. However the mechanism of the metal-catalyzed
conversion of diazoalkanes to ketenes is not fully understood
even though the coordination chemistry of both diazoalkanes
and ketenes has been extensively investigated.3
A number of fascinating aspects of ketene coordination
chemistry have been discovered. Grotjahn and co-workers4 have
demonstrated coordination of ketenes by both C,C- and C,O-
† University of Miami.
‡ University of Pannonia.
§ Beijing Institute of Technology.
| University of Georgia.
Roosevelt University.
(1) Tidwell, T. T. Ketenes, Wiley, New York, 1995.
(2) (a) Ungva´ri, N.; Ke´gl, T.; Ungva´ry, F. J. Mol. Catal. A: Chem. 2004, 219,
7. (b) Ke´gl, T.; Ungva´ry, F. J. Organomet. Chem. 2007, 692, 1825. (c)
Tuba, R.; Ungva´ry, F. J. Mol. Catal. A: Chem. 2003, 203, 59. (d) Fo¨rdo¨s,
E.; Ungva´ri, N.; Ke´gl, T.; Ungva´ry F. J. Eur. J. Inorg. Chem. 2006, 1875.
(3) Zollinger, H. Diazo Chemistry; VCH: Weinheim, Germany, 1995.
9
14388
J. AM. CHEM. SOC. 2007, 129, 14388-14400
10.1021/ja075008o CCC: $37.00 © 2007 American Chemical Society

Confirmation of Carbonylation Intermediates
A R T I C L E S
bonding modes to trans-ClIr[P(iPr)3]2. However, related Ir(I)
complexes containing chelating phosphines were also shown
to split bound ketene ligands to CO and carbene units in
reversible reactions.5 Hermann and co-workers6 reported the
synthesis and structures of η2-ketene complexes by slow high-
pressure carbonylation of metallocarbenes as shown in eq 2:
2. Experimental
2.1. General Procedures. Unless stated otherwise, all operations
were performed in a Vacuum Atmospheres glovebox under an
atmosphere of purified argon or utilizing standard Schlenk tube
techniques under argon. Toluene and THF were purified by distillation
under argon from sodium benzophenone ketyl into flame dried
glassware. [Cr(CO)3(C5R5)]2 was prepared as described in earlier
publications.7 Solutions of NdNdCHSiMe3 (2.0 M in hexanes) were
obtained from Sigma-Aldrich and used without further purification.
FTIR data were obtained on a Perkin-Elmer system 2000 spectrometer.
FTIR spectra for kinetic measurements were taken using a closed system
FTIR cell that has been described in previous publications.22 NMR
spectra were obtained on a Bruker AVANCE 500 MHz spectrometer.
Electron spin resonance (ESR) spectra were taken with a Bruker EMX
ESR Spectrometer.
Qualitative Studies of Reaction of [Cr(CO)3(C5R5)]2 (R ) H, Me)
and NNCHSiMe3. Qualitative studies of reactions of [Cr(CO)3(C5R5)]2
(R ) H, CH3) and N2CHSiMe3 were performed using standard Schlenk
techniques. Reaction products were analyzed in terms of the bands
shown in the Supporting Information Table ST-1 for the known
complexes utilized. Approximate percent conversions were made on
the basis of band-shape analysis using standard procedures.
Kinetic Studies of Reaction of [Cr(CO)3(C5R5)]2 (R ) H, Me)
and N2CHSiMe3 under Low Pressures of Ar, CO, and N2. In a
typical procedure, a solution of 0.1360 g [Cr(CO)3(C5Me5)]2 in 15 mL
of toluene was prepared in a Schlenk tube under argon to make a 0.0167
M solution (0.0334 M solution in radical). The Cr radical solution was
then stirred for about 20 min to allow the compound to dissolve
completely. In the glovebox the solution was loaded into a Hamilton
gastight syringe fitted with a syringe filter. Outside the glovebox the
[Cr(CO)3(C5Me5)]2 solution was filtered into the closed system IR cell
via syringe. The cell had been previously placed under an atmosphere
of 9 psi of CO. The sample was shaken vigorously to ensure that the
maximum amount of CO was dissolved into the solution. The IR cell
was then attached to a constant temperature bath and allowed to reach
thermal equilibrium at 298 K. The initial IR spectrum was then taken.
A Hamilton gastight syringe was filled with 0.125 mL of a 2.0 M
solution of N2CHSiMe3 in hexanes. Simultaneously the diazo solution
was added and the timer was started. The sample was then shaken
vigorously to ensure thorough mixing. IR spectra were taken on an
average of every 60 s. Similar procedures were used to study the
reaction of [Cr(CO)3(C5R5)]2 (R ) H, Me) and N2CHSiMe3 over a range
of temperatures (T ) 278-308 K) and pressures of Ar, CO, and N2 (p
) 1-4.5 atm).
Kinetic Studies of Reaction of [Cr(CO)3(C5Me5)]2 and N2CHSiMe3
at 12.7 atm CO. In a typical procedure, a solution of 0.1936 g [Cr-
(CO)3(C5Me5)]2 in 20 mL toluene was prepared in a Schlenk tube under
argon to make a 0.0179 M solution. The solution was then mixed for
about 20 min to allow the compound to dissolve. In the glovebox the
solution was loaded into a Hamilton gastight syringe fitted with a
syringe filter. Outside the glovebox the [Cr(CO)3(C5Me5)]2 solution was
filtered into the closed system IR cell via syringe using Schlenk
techniques under Ar. The cell was then connected to a tank of CO. A
Hamilton gastight syringe was used to add 0.180 mL of a 2.0 M solution
of N2CHSiMe3 in hexanes to the cell after which the cell was
immediately pressurized to a total of 173 psi. The reaction was
thermostated at 290 K. IR spectra were taken as the reaction proceeded
over the next week. Slow build up of the ‚Cr(CO)2(ketene)C5Me5 was
monitored by IR peaks at 1672, 1908, and 1984 cm-1. Other side-
products that were seen during the course of the reaction include a
slow buildup of both Cr(CO)6 (1981 cm-1) and an unknown organic
carbonyl near 1700 cm-1 which overlapped the broad band at 1670
cm-1 peak of the ketene complex.
(η5-C5H5)Mn(CO)2(dC(C6H5)2) CO8
(η5-C5H5)Mn(CO)2(η2-OdCdC(C6H5)2) (2)
The stability of the η2-ketene bonding mode was indicated by
the ability of such metal complexes to withstand high pressures
of CO without undergoing ligand displacement. Related η2-
ketene complexes have also been prepared by the reaction of
photogenerated (η5-C5H5)Mn(CO)2(THF), (η5-C5R′5)Mn(CO)3,
and diazoalkanes. This surprising reaction was proposed to occur
by direct transfer of a carbene to a bound carbonyl ligand. This
work reports investigation of reaction 1 using [Cr(CO)3(C5R5)]2
(R ) H, CH3) instead of Co2(CO)8. Compared to Co, the Cr
complexes show a greater tendency to dissociate and yield
reactive monomeric 17e- radical species by homolytic cleavage
of the weak Cr-Cr bond.7 An additional difference in the Cr
system is the presence of the highly stable CrtCr complexes
[Cr(CO)2(C5R5)]2 for which there is no known stable counterpart
in cobalt carbonyl chemistry. Interestingly, the complex Cr2-
(CO)5(C5H5)2, which is the analogue of the often proposed
intermediate Co2(CO)7, does not appear to have been detected.
In recent years methods based on density functional theory
(DFT) have had great impact on organotransition metal
chemistry.8-16 For example, recent work on metal-metal
bonded complexes of first row transition metals has yielded
quantitative predictions regarding structure and stability.17-20
A recent DFT study of the diazo to ketene conversion21 is of
direct relevance to this work. This paper reports the convergence
of theory and experiment to study the mechanism of activation
of a diazo compound by concerted attack of two metal radicals.
(4) (a) Grotjahn, D. B.; Collins, L. S. B.; Wolpert, M.; Bikzhanova, G. A.;
Lo, H. C.; Combs, D.; Hubbard, J. L. J. Am. Chem. Soc. 2001, 123, 8260.
(b) Grotjahn, D. B.; Lo, H. C. J. Am. Chem. Soc. 1996, 118, 2097. (c)
Urtel, H.; Bikzhanova, G. A.; Grotjahn, D. B.; Hofmann, P. Organometallics
2001, 20, 3938.
(5) Grotjahn, D. B.; Bikzhanova, G. A.; Collins, L. S. B.; Concolino, T.; Lam,
K. C.; Rheingold, A. L. J. Am. Chem. Soc. 2000, 122, 5222.
(6) (a) Herrmann, W. A.; Plank, J. Angew. Chem., Int. Ed. Engl. 1978, 17,
525. (b) Herrmann, W. A.; Plank, J.; Ziegler, M. L.; Weidenhammer, K. J.
Am. Chem. Soc. 1979, 101, 3133.
(7) (a) Hoff, C. D. Coord. Chem. ReV. 2000, 206, 451. (b) Watkins, W. C.;
Jaeger, T.; Kidd, C. E.; Fortier, S.; Baird, M. C.; Kiss, G.; Roper, G. C.;
Hoff, C. D. J. Am. Chem. Soc. 1992, 114, 907. (c) McLain, S. J. J. Am.
Chem. Soc. 1988, 110, 643. (d) Adams, R. D.; Collins, D. E.; Cotton, F.
A. J. Am. Chem. Soc. 1974, 96, 749. (e) The value of 2.215 Å is the average
of the two reported Cr-Cr, bond lengths of 2.200 and 2.230 Å reported
for [Cr(CO)2Cp]2: Curtis, M. D.; Butler, W. M. J. Organomet. Chem. 1978,
155, 131.
(8) Ehlers, A. W.; Frenking, G. J. Am. Chem. Soc. 1994, 116, 1514.
(9) Delly, B.; Wrinn, M.; Lu¨thi, H. P. J. Chem. Phys. 1994, 100, 5785.
(10) Li, J.; Schreckenbach, G.; Ziegler, T. J. Am. Chem. Soc. 1995, 117, 486.
(11) Jonas, V.; Thiel, W. J. Chem. Phys. 1995, 102, 8474.
(12) Barckholtz, T. A.; Bursten, B. E. J. Am. Chem. Soc. 1998, 120, 1926.
(13) Niu, S.; Hall, M. B. Chem. ReV. 2000, 100, 353.
(14) Macchi, P.; Sironi, A. Coord. Chem. ReV. 2003, 238, 383.
(15) Carreon, J.-L.; Harvey, J. N. Phys. Chem. Chem. Phys. 2006, 8, 93.
(16) Bu¨hl, M.; Kabrede, H. J. Chem. Theory Comput. 2006, 2, 1282.
(17) Schaefer, H. F.; III; King, R. B. Pure Appl. Chem. 2001, 73, 1059.
(18) King, R. B.; Xie, Y.; Schaefer, H. F.; Richardson, N.; Li, S. Inorg. Chim.
Acta 2005, 358, 1442.
(19) Wang, H.; Xie, Y.; King, R. B.; Schaefer, H. F., III. Inorg. Chem. 2006,
45, 10849.
(20) Li, Q.-S.; Zhang, X.; Xie, Y.; King, R. B.; Schaefer, H. F. J. Am. Chem.
Soc. 2007, 129, 3433.
(21) Seres, B.; Fo¨rdo¨s, E.; Ungva´ri, N.; Fortman, G. C.; Hoff, C. D.; Ungva´ry
F.; Ke´gl, T.Unpublished work.
(22) Capps, K. B.; Bauer, A.; Sukcaroenphon, K.; Hoff, C. D. Inorg. Chem.
1999, 38, 6206.
9
J. AM. CHEM. SOC. VOL. 129, NO. 46, 2007 14389

A R T I C L E S
Fortman et al.
ESR Measurements of the Reaction of [Cr(CO)3(C5R5)]2 (R )
H, Me) and Co2(CO)8. In a representative experiment a 0.0162 M
[Cr(CO)3(C5R5)]2 (R ) H, Me) solution was prepared under Schlenk
techniques. The [Cr(CO)3(C5R5)]2 solution (1.5 mL) was syringe filtered
into a Pyrex NMR tube.23 The end of the NMR tube was sealed and
taken for ESR analysis. To the tube 0.5 equiv of N2CH(SiMe3) (12 µL
of 2.0 M in hexanes) was added. The tube was inverted several times
to ensure proper mixing. The ESR spectrum was then taken. The
reaction of [Cr(CO)3(C5Me5)]2 with diazoalkane leads to an intense
signal at g ) 2.007. The reaction of [Cr(CO)3(C5H5)]2 with diazoalkane
leads to a small signal at g ) 2.005. No signal was detected for the
reaction of Co2(CO)8 with the diazoalkane. Experimental conditions:
T, 298 K; microwave power, 0.630 mW; microwave frequency, 9.840
GHz; modulation amplitude, 1.37 G.
Preparations of 13CO Labeled [Cr(CO)3(C5R5)]2 (R ) H, Me).
In the glovebox 0.2146 g of [Cr(CO)3(C5Me5)]2 was dissolved in 25
mL of toluene (0.016 M) with stirring for approximately 20 min. Using
Schlenk techniques the solution was then syringe filtered into the closed
system IR cell which in turn was pressurized to 230 psi with 13CO (99
atom % obtained from ISOTEC). CO exchange was allowed to occur
overnight. The next morning the 13CO in the cell was evacuated and
replaced by 193 psi of fresh 13CO. FTIR of the sample lead to the
conclusion that this yielded approximately 96% [Cr(13CO)3(C5Me5)]2
with the remaining 4% primarily consisting of [Cr(13CO)2(12CO)(C5-
Me5)]2.
ESR of [Cr(13CO)3(C5R5)]2 (R ) H, Me). In a representative
experiment 1.5 mL of a 0.2146 M [Cr(13CO)3(C5Me5)]2 solution
prepared as described above was syringe filtered into the a 5 mm pyrex
NMR/ESR tube fitted with a screw cap and Teflon-coated silicone
septum. The NMR/ESR tube had been purged with argon prior to
addition. An ESR spectrum run prior to injection of N2CHSiMe3 showed
no significant signal in keeping with the previously reported observation
that in solution ‚Cr(CO)3(C5Me5) does not produce an observable ESR
spectrum.24 To the ESR tube 3.3 equiv of N2CHSiMe3 (20 µL of 2.0
M in hexanes) was added. The tube was inverted several times to ensure
proper mixing, and the ESR spectrum was taken. As in the experiments
with natural isotopic abundance complexes, the reaction of [Cr(13CO)3-
(C5Me5)]2 with diazoalkane led to an intense signal at g ) 2.007 and
the reaction of [Cr(13CO)3(C5H5)]2 with diazo led to a small signal at
g ) 2.005. An additional small signal seen in the spectrum for [Cr-
(13CO)3C5H5]2 was attributed to an unknown organic radical. Experi-
mental conditions: T, 298 K; microwave power, 0.630 mW; microwave
frequency, 9.807 GHz; modulation amplitude, 1.00 G.
NMR Studies of the Reaction of [Cr(CO)3(C5H5)]2 and N2CHSiMe3.
In a typical experiment 0.0321 g of [Cr(CO)3(C5H5])2 was dissolved
in 5 mL C6D6. One milliliter was then syringe filtered into an argon
purged 5 mm Pyrex NMR tube fitted with a screw cap and septum.
The sample was placed in the 500 MHz Bruker AVANCE 500 MHz
NMR spectrometer. It was left in the NMR at 288 K for about 15 min
to reach thermal equilibrium. The tube was then quickly ejected, loaded
with 0.5 equiv of N2CHSiMe3 (12.0 µL of 2 M in hexanes), mixed,
and placed back in the NMR. Spectra were run starting at about 240 s
and successively run roughly every 80 s. In the Cp region (3.0-6.0
ppm) the only significant peak detected was a signal at ∼4.21 ppm
attributed to [Cr(CO)2C5H5]2 which grew steadily during the course of
the reaction. The reaction was confirmed by the FTIR spectra that were
taken following the conclusion of the NMR experiment. The contents
of the final solution included OCCHSiMe3 (2107 cm-1), traces of
unreacted [Cr(CO)3(C5H5)]2 (2009, 1946, 1920 cm-1), and [Cr(CO)2-
(C5H5])2 (1901, 1878 cm-1). NMR spectra were referenced to C6HD5
in the C6D6 at 7.15 ppm.
Calvet Calorimetric Measurement of Heat of Reaction of [Cr2-
(CO)2(C5H5)]2 and [Cr(CO)3)(C5H5)]2 with 1% Na/Hg. These experi-
ments represent a modification of experimental procedures used in a
previous report on enthalpies of reduction of organometallic com-
pounds.25 In the glovebox 0.5 mL of 1% Na/Hg was loaded into the
cell of the Setaram-C-80 Calvet microcalorimeter under an atmosphere
of CO. Subsequently added to the cell was 4.5 mL of THF that had
been shaken and stored in the glovebox over the 1% Na/Hg and under
a CO atmosphere. The solid-containing compartment of the calorimeter
was loaded with 0.0132 g of [Cr(CO)2(C5H5)]2 which had been recently
recrystallized from heptane and methylene chloride. The calorimeter
cell was assembled and sealed under a CO atmosphere and taken from
the glovebox and loaded into the calorimeter. After reaching temperature
equilibration the reaction was initiated. The thermogram indicated a
rapid reaction which returned cleanly to baseline with no thermal signal
indicative of any substantial secondary reactions occurring.
Following return to baseline, the cell was taken back into the
glovebox. A sample analyzed by FTIR spectroscopy showed the
presence of Na+[Cr(CO)3(C5H5)]- as the sole organometallic product.
The calorimetric signal was integrated and the average value for
reduction of solid [Cr(CO)2(C5H5)]2 to yield a THF solution of
Na+[Cr(CO)3(C5H5)]- was -94.6 ( 0.75 kcal mol-1. The above
procedure was also followed for the reaction of [Cr(CO)3(C5H5)]2 with
1% Na/Hg. An average value of -80.7 ( 1.3 kcal mol-1 was obtained
for the reduction of this solid to Na+[Cr(CO)3(C5H5)]- in THF solution.
Spectral data for the formation of Na+[Cr(CO)3(C5H5)]- were identical
for reductions starting from solid [Cr(CO)2(C5H5)]2 and [Cr(CO)3-
(C5H5)]2. Heats of solution for both complexes are assumed to be very
similar. Spectral data for all relevant compounds are listed in the
Supporting Information Table ST-1.
CO Exchange of [Cr(CO)2(C5H5)]2 with 13CO. Under Schlenk
techniques 0.055 mmol of [Cr(CO)2(C5H5)]2 was dissolved in 10 mL
of toluene. Under argon the solution was then transferred to the closed
system FTIR cell via Hamilton gastight syringe. An initial FTIR
spectrum was taken. The cell was then pressurized to 112 psi of 13CO
and thermostated at 295 K for 1 h. No change was observed. In addition
the temperature was increased from 295 to 310 K for an hour and finally
to 323 K for 2 h. No significant change was seen. The sample was left
at room temperature overnight and again no significant change was
seen.
2.2. Theoretical Methods. Two DFT methods were used in this
study. The first functional was the B3LYP method, which is a hybrid
HF/DFT method using the combination of the three-parameter Becke
functional (B3) with the Lee-Yang-Parr (LYP) generalized gradient
correlation functional.26,27 The other DFT method used in the present
paper was BP86, which combines Becke’s 1988 exchange functional
(B) with Perdew’s 1986 gradient corrected correlation functional method
(P86).28,29 It has been noted elsewhere that the BP86 method may be
somewhat more reliable than B3LYP for the type of organometallic
systems considered in this paper.30-32 In the present study, the B3LYP
(25) (a) Data obtained here for reduction of Cr2(CO)6(C5H5)2 using Na/Hg in
THF under CO was found to be similar to previous measurements of the
same reaction under Ar. Initial attempts at measuring the enthalpy of
reduction utilized Na+benzophenone- as has been described previously (ref
25b). During the course of these investigations it was found that because
of pinacol coupling-type reactions that occur in some systems involving
Na+benzophenone- the Na/Hg reactions are more accurate. The purpose
of the current work was to generate reliable enthalpies of reduction for
Cr2(CO)4(C5H5)2. To the authors’ knowledge the reaction itself has not been
previously reported. Reduction of Cr2(CO)n(C5H5)2 (n ) 2,3) was measured
by both methods Na+benzophenone- and Na/Hg reduction. The differences
between enthalpies of reduction (for n ) 2, 3) were similar. Owing to
pinacol coupling-like reactions only the Na/Hg measurements are considered
reliable and reported here. Additional work on enthalpies of reduction are
in progress and will be reported separately. (b) Kiss, G.; Nolan, S. P.; Hoff,
C. D. Inorg. Chim. Acta 1994, 227, 285.
(23) Initial studies were made using quartz ESR tubes. It was found that a 5
mm screw cap NMR tube fitted with a teflon coated silicone septum showed
no significant signal that overlapped with the signal for the Cr complex
for these concentrated solutions.
(24) Morton, J. R.; Preston, K. F.; Cooley, N. A.; Baird, M. C.; Krusic, P. J.;
McLain, S. J. J. Chem. Soc., Faraday Trans. 1 1987, 12, 3535.
(26) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
(27) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785.
(28) Becke, A. D. Phys. ReV. A 1988, 38, 3098.
(29) Perdew, J. P. Phys. ReV. B 1986, 33, 8822.
(30) Furche, F.; Perdew, J. P. J. Chem. Phys. 2006, 124, 044103.
9
14390 J. AM. CHEM. SOC. VOL. 129, NO. 46, 2007

Confirmation of Carbonylation Intermediates
A R T I C L E S
The observed reactivity during “stage one”39 corresponds
roughly to the approximate stoichiometry shown in eq 4 below:
and BP86 methods agree with each other fairly well in predicting the
structural characteristics of the Cp2Cr2(CO)n (the common abbreviations
Cp ) C5H5 and Cp* ) C5Me5 will be used interchangeably) derivatives
of interest. Although both the B3LYP and BP86 results are shown in
the figures and tables, unless specifically noted, only the BP86 results
(geometries, energies, and vibrational frequencies) are discussed in the
text.
All calculations were performed using the double-ú plus polarization
(DZP) basis sets. The DZP basis sets used for carbon and oxygen add
one set of pure spherical harmonic d functions with orbital exponents
Rd(C) ) 0.75 and Rd(O) ) 0.85 to the standard Huzinaga-Dunning
contracted DZ sets33,34 and are designated (9s5p1d/4s2p1d). For
hydrogen, a set of p polarization functions Rp(H) ) 0.75 is added to
the Huzinaga-Dunning DZ set. The loosely contracted DZP basis set
for chromium is the Wachters primitive set35 augmented by two sets
of p functions and a set of d functions, contracted following Hood,
Pitzer, and Schaefer,36 designated (14s11p6d/10s8p3d).
The geometries of all structures were fully optimized using the DZP
B3LYP and DZP BP86 methods, and the vibrational frequencies were
determined by evaluating analytically the second derivatives of the
energy with respect to the nuclear coordinates. The corresponding
infrared intensities were also evaluated analytically. All of the computa-
tions were carried out with the Gaussian 03 program,37 exercising the
fine grid option (75 radial shells, 302 angular points) for evaluating
integrals numerically, while the tight (10-8 hartree) designation is the
default for the self-consistent field (SCF) convergence.
2 •Cr(CO)3(C5Me5)
1/2[Cr(CO)2(C5Me5)]2 + N2
1/2OdCdCHSiMe3
+
f
+
NdNdCHSiMe2
1/2‚Cr(CO)2(ketene)(C5Me5) +
1/4Cr2(CO)5(C5Me5)2 + 5/4CO (4)
During this phase of the reaction, as shown in Figure 1, most
of the chromium radical reacts and produces ketene and the
CrtCr triple-bonded product Cr2(CO)4(C5Me5)2 as well as
intermediate products. Infrared bands assigned to the two
proposed intermediate complexes ‚Cr(CO)2(ketene)(C5Me5) and
Cr2(CO)5(C5Me5)2 are observed to increase during this first stage
of the reaction
During the second stage of the reaction, bands assigned to
the two intermediate complexes are observed to decay according
to the approximate stoichiometry shown in
1/2‚Cr(CO)2(ketene)(C5Me5)
1/2[Cr(CO)2(C5Me5)]2
+
f
+
1/4Cr2(CO)5(C5Me5)2 + 1/4CO 1/2OdCdCHSiMe3
(5)
3. Results and Discussion
Spectroscopic data during the later stage of reaction is shown
in Supporting Information Figure S1. During this time the bands
assigned to intermediates have reached a maximum and begin
to decay liberating additional ketene and produce additional
triple-bonded product. The clean nature of the reaction under
these conditions is shown by the presence of an isosbestic point
as shown in Supporting Information Figure S2. The fact that a
significant amount of free ketene is liberated during this phase,
in spite of low levels of free diazo present in solution, indicates
that at least one of the intermediates contains a bound form of
ketene or a precursor to it. Reactions performed under an N2
atmosphere were studied in an attempt to see if either rate or
product distribution differed significantly from reactions under
Ar atmosphere. No real differences were observed implying that
reversible dissociation of N2 was not a rate determining step in
the overall mechanism.
3.1.2. Reaction of 2 •Cr(CO)3(C5Me5) and NNCHSiMe3
under CO. Reaction of 2‚Cr(CO)3(C5Me5) and NdNd
CHSiMe3 under CO at 1.6 atm CO pressure showed several
distinctive differences compared to reaction under argon. The
first is a markedly slower rate of reaction. The second is that
less [Cr(CO)2(C5Me5)]2 is produced and less ‚Cr(CO)3(C5Me5)
is consumed during ketene production. This implies that it is
CO gas that is utilized to produce ketene rather than the
organometallic complex being decarbonylated by the diazoal-
kane. In addition, the level of detectable intermediates under
CO at a 2/1 Cr/diazoalkane mole ratio is lower than what is
observed under argon. Finally, as shown in Figure 2 the reaction
shows a characteristic “S” shape curve for production of ketene,
consistent with an induction period followed by a relatively rapid
3.1. Qualitative and Kinetic Studies. Reaction 3 presented
a number of surprises as well as difficulties in its study:
NdNdCHSiMe3 + CO + [Cr(CO)3(C5R5)]2 f products (3)
The yield and nature of the products were found to depend on
the concentration of virtually every reactant and product in the
equation.38 Furthermore, the precursor complexes [Cr(CO)3-
(C5R5)]2 (R ) H, CH3) dissociate to radicals to varying degrees
depending on the temperature and R group. The third-order rate
laws often found7 for ‚Cr(CO)3(C5Me5) also implicate a strong
dependence on the absolute concentration of the metal. This
section summarizes key aspects of this reaction studied under
conditions designed to simplify and isolate given types of
reactivity. The focus is on the reactions that occur at relatively
high concentrations38 (typically from 1 to 20 mM) of metal
complex and ratios of Cr/diazo of 2/1 and 1/1, under CO, Ar,
and N2 atmospheres.
3.1.1. Reaction of 2‚Cr(CO)3(C5Me5) and NNCHSiMe3
under Ar or N2. With a 2/1 Cr/diazoalkane mole ratio under
Ar or N2, the reaction is best viewed as occurring in two stages.
(31) Wang, H. Y.; Xie, Y.; King, R. B.; Schaefer, H. F. J. Am. Chem. Soc.
2005, 127, 11646.
(32) Wang, H. Y.; Xie, Y.; King, R. B.; Schaefer, H. F. J. Am. Chem. Soc.
2006, 128, 11376.
(33) Dunning, T. H. J. Chem. Phys. 1970, 53, 2823.
(34) Huzinaga, S. J. Chem. Phys. 1965, 42, 1293.
(35) Wachters, A. J. H. J. Chem. Phys. 1970, 52, 1033.
(36) Hood, D. M.; Pitzer, R. M.; Schaefer, H. F. J. Chem. Phys. 1979, 71, 705.
(37) Frisch, M. J.; et al. Gaussian 03, revision C 02; Gaussian, Inc.: Wallingford,
CT, 2004.
(38) Several intermediate complexes are either detected or proposed to be present
in reactions studied at relatively low concentrations of diazoalkane. Attempts
to extend these studies to regimes of high diazoalkane concentration
produced unknown products. It seems plausible that these resulted from
radical intermediates already containing a bound diazoalkane or ketene
moiety reacting further with additional equivalents of free diazoalkane to
produce oligomeric products. Catalytic conversion of diazoalkane to ketene
could be achieved by repeated injections of small amounts of diazoalkane.
However “flooding” the system with a high concentration of diazoalkane
was found to produce side products of unknown and presumably oligomeric
nature.
(39) Separation of the reaction into two phases is somewhat artificial and is
meant to show the qualitative nature of the two periods. In the first phase
intermediates build up and in the second they decay. In terms of reaction
kinetics there is of course one continuous process and, particularly in the
middle where the “two phases” overlap, there is no clear distinction.
Nevertheless the very beginning and ending parts of the reaction approach
the limiting stoichiometries discussed.
9
J. AM. CHEM. SOC. VOL. 129, NO. 46, 2007 14391

A R T I C L E S
Fortman et al.
Figure 1. FTIR spectral data during initial stage of reaction of ‚Cr(CO)3(C5Me5) (∼0.033 M) with NdNdCHSiMe3(∼0.0165 M) at 25 °C in toluene under
an argon atmosphere. Bands assigned to int0 ) Cr2(CO)5(C5Me5)2 are at 1957, 1894, and 1799 cm-1 and bands assigned to int1 ) ‚Cr(CO)2(ketene)(C5Me5)
occur at 1984, 1908, and 1672 cm-1. In the text and figures the common abbreviations Cp ) C5H5 and Cp* ) C5Me5 are used interchangeably. For complete
band assignments see Table ST-1.
Figure 2. Absorbance versus time plots for key species in reaction of ‚Cr(CO)3(C5Me5) (∼0.033 M) with NdNdCHSiMe3 (∼0.0165 M) at 25 °C in
toluene under Ar and CO at 1.6 atm of pressure. Absorbance bands used for quantification of key species: N2CHSiMe3 ) 2066 cm-1; OCCHSiMe3 ) 2107
cm-1; ‚Cr(CO)3(C5Me5) )1994 cm-1: b, N2CHSiMe3 (CO); O, N2CHSiMe3 (Ar); 9, OCCHSiMe3 (CO); 0, OCCHSiMe3 (CO); 2, ‚Cr(CO)3(C5Me5)
(CO); 4, ‚Cr(CO)3(C5Me5) (Ar).
reaction. The “S” shape is more pronounced under Ar atmo-
sphere compared to CO.
3.1.3. Reaction of ‚Cr(CO)3(C5Me5) and ‚Cr(13CO)3-
(C5Me5) with NNCHSiMe3, FTIR and ESR Studies. Reaction
of ‚Cr(CO)3(C5Me5) with NdNdCHSiMe3 in a 1/1 ratio was
found to show significant differences from those studied at a
2/1 molar ratio. This is presumably because the reductive
elimination step is dinuclear and the presence of excess diazo
in solution generates ‚Cr(CO)2(ketene)(C5Me5) more quickly
than it is consumed. In addition it was found that under a CO
9
14392 J. AM. CHEM. SOC. VOL. 129, NO. 46, 2007

Confirmation of Carbonylation Intermediates
A R T I C L E S
Figure 3. Experimental and simulated room-temperature X-band (9.807 GHz) ESR spectra for ‚Cr(13CO)2(η2-Od13CdCHSiMe3)(C5Me5). Simulated data
are fit assuming 95% isotopic substitution in 13C and natural abundance in 53Cr and the following parameters: g(iso) ) 2.007; A(53Cr) ) 42.5 MHz; A(13C
a)2 ) 22.5 MHz; A(13C b)1 ) 12.0 MHz; 3 MHz Gaussian line widths. Experimental and simulated ESR spectra for natural isotopic abundance
‚Cr(CO)2(OdCdCHSiMe3)(C5Me5) are shown in Figure S3 and exhibit hyperfine splitting only from 53Cr.
atmosphere40 the intermediate complex Cr2(CO)5(C5Me5)2 is
carbonylated but ‚Cr(CO)2(ketene)(C5Me5) reacts only extremely
slowly with CO. The net result is that significant quantities of
the radical ketene complex buildup. This increase can be seen
in both the IR and ESR spectra.
FTIR studies of the reaction of ‚Cr(CO)3(C5Me5) with Nd
NdCHSiMe3 at a 1/1 molar ratio were studied under CO
pressures of 10-700 psi. Under higher pressures the reaction
is slow and may take up to a week to go to completion at room
temperature. The principal product under these conditions is
the intermediate complex formulated as ‚Cr(CO)2(ketene)(C5-
Me5). Attempts to obtain crystals of this highly air-sensitive
complex were unsuccessful despite repeated attempts. During
these long reactions under high CO pressure, slow buildup of
Cr(CO)6, identified by its characteristic IR spectrum, and an
unknown organic carbonyl compound were observed to be
formed, presumably from41
reactions of ‚Cr(CO)3(C5R5) (R ) H, CH3) with NNCHSiMe3
by ESR spectroscopy showed the presence of radical species
that build up and decay in rough accord with the FTIR spec-
troscopic studies discussed above. Representative spectroscopic
data illustrating the buildup and slow decay of this complex as
well as the spectrum of the natural isotope abundance species
are shown in Supporting Information Figure S3.
Reaction studies using analogous ESR techniques to those
shown for ‚Cr(CO)3(C5Me5) were also performed for reactions
of ‚Cr(CO)3(C5H5) and Co2(CO)8. As shown in Supporting
Information Figure S4 the overall level of radical concentration
is approximately 10 times lower for the Cp compared to the
Cp* chromium systems. The reaction of Co2(CO)8 with diazo
showed no detectable signal above background noise. Radical
species in the cobalt system were estimated to be present at
levels at least 6 orders of magnitude lower than for the
chromium systems under comparable conditions.
The 13CO labeled radical ‚Cr(13CO)3(C5Me5) complex was
prepared and allowed to react with NdNdCHSiMe3 to prepare
in situ the isotopically labeled complex ‚Cr(13CO)2(η2-Od13Cd
CHSiMe3)(C5Me5). ESR spectra for this complex and a com-
‚Cr(CO)2(ketene)(C5Me5) + nCO f
Cr(CO)6 + “R-C(dO)(C5Me5)” (6)
The formulation of the intermediate as ‚Cr(CO)2(ketene)(C5-
Me5) is based on its vibrational spectrum (Table ST-1) and
observed release of ketene upon reaction with ‚Cr(CO)3(C5Me5)
or Cr2(CO)5Cp*2 as well as ESR spectroscopy. Study of the
(41) (a) Production of Mo(CO)6 from RMo(CO)3C5R5 under CO pressure has
been shown to yield R-C(dO)-C5R5 as described in ref 41b. The complex
‚Cr(CO)2(η2-ketene)(C5Me5) is resistant to carbonylation. It seems possible
that this is due to an unfavorable pre-equilibrium to generate ‚Cr(CO)3-
(η1-ketene)(C5Me5). Once formed, this radical complex, rather than liberate
ketene may form a C-C bonded complex in which the (η1-ketene) moiety
migrates to the cyclopentadienyl ring and then adds further CO yielding
ultimately Cr(CO)6 and some complex organic carbonyl. The growth of a
broad band near 1700 cm-1 in these experiments is in keeping with that
hypothesis but no attempt was made to further characterize this side product.
(b) Nolan, S. P.; de la Vega, R. L.; Mukerjee, S. K.; Hoff, C. D. Inorg.
Chem. 1986, 25, 1160.
(40) In addition to external gas pressure utilized in the reaction, the shape of
the reaction container and how it is stirred can also play a role in terms of
how quickly any CO released in solution can escape to the gas phase. In
the narrow confines of a sealed ESR tube, for example, even reactions run
initially under argon may soon buildup significant CO concentrations
serving to quench the reaction as well as alter the product distribution. For
that reason ESR studies described are primarily qualitative in nature.
9
J. AM. CHEM. SOC. VOL. 129, NO. 46, 2007 14393

A R T I C L E S
Fortman et al.
Figure 4. Stock solution of [Cr(CO)3Cp]2 (0.0163 M) in toluene under argon prior to addition of NNCHSiMe3 (0.0108 M) in a 1.5/1 molar ratio at 25 °C
(red spectrum), t ≈ 30 s following addition of NNCHSiMe3 (blue spectrum) and t ≈ 60 s (green spectrum). Little reaction occurs during the initial induction
period as shown in the blue spectrum. The green spectrum shows rapid growth in free ketene, nearly complete reaction of initial diazo and [Cr(CO)3Cp]2,
minor buildup of [Cr(CO)2Cp]2, and the presence of an intermediate complex. For band positions see Table ST-1.
puter simulation are shown in Figure 3. Both display a coupling
to 13CO not present in the natural abundance spectrum shown
in Supporting Information Figure S3. The spectrum of the
13CO labeled complex consists essentially of a triplet of doublets
centered at g ) 2.007, with low-intensity satellites due to 53Cr
significant Cr(I) character of ‚Cr(CO)2(η2-OdCdCHSiMe3)(C5-
Me5). The complex ‚Cr(CO)3(C5H5) also exhibited hyperfine
coupling from the carbonyl ligands (detected in natural isotopic
abundance) with aiso(13C) ) 25 MHz,43b quite close to the value
observed here (we are unaware of any relevant data for the
ketene-based 13C hyperfine coupling).
3
(I ) /2, 9.5% abundant). The largest hyperfine coupling is to
the metal, A(53Cr) ) 42.5 MHz, with additional hyperfine
splitting assigned as 22.5 MHz for the two equivalent 13CO
ligands and 12.0 MHz for η2-Od13CdCHSiMe3. No hyperfine
An interesting difference, however, is that the symmetrical
parent radical, as well as related derivatives ‚Cr(CO)3(C5R5)
(R ) H, CH3, Ph) are known to display no solution-phase ESR
spectra at room temperature.42 However, for species of lower
effective symmetry such as ‚Cr(CO)2(L)(C5Me5),43 ‚Cr(CO)2-
(RC≡CR)(C6Me6)+,44 and ‚Cr(CO)2(η3-tris(pyrazolylborato)-
(PMe3)45 solution-phase spectra are reported even at ambient
temperature. These arguments support assignment of the radical
species in solution as ‚Cr(CO)2(η2-OdCdCHSiMe3)(C5Me5),
which likewise has lower effective symmetry. Computed SOMO
and spin density data are shown in Supporting Information
Figure S5 and also support a highly delocalized radical which
is predominantly metal based. Similar conclusions44 were made
for the cationic acetylene complexes ‚Cr(CO)2(RCtCR)(C6-
Me6)+ which are probably closest in nature to the proposed
intermediate ketene radical complex reported here.
Reaction of ‚Cr(13CO)3(C5Me5) and NdNdCHSiMe3 was
also investigated by FTIR spectroscopy as shown in Supporting
Information Figure S6. A shift in the assigned vibrational bands
of all assigned bands was observed (see Supporting Information
Table ST-1). However of principal interest was the shift in
ν(CO) of the assigned band of the coordinated ketene from 1672
to 1637 cm-1 for ‚Cr(13CO)2(η2Od13CdCHSiMe3)(C5Me5).
1
splitting is assignable to H, as indicated by the simple ESR
spectra of the natural abundance spectrum shown in Supporting
Information Figure S3.
It cannot be excluded that low levels of other radical species
may be present in solution, but the dominant pattern seen in a
number of experiments was reproducible and the fit of computed
and experimental data are good. The growth and decay of the
peaks assigned to this intermediate as followed by ESR were
also observed in FTIR studies done in parallel reactions.
These ESR results can be compared with those for other
organochromium radicals.42,43 The value for isotropic 53Cr
hyperfine coupling, 42.5 MHz, is comparable, but larger than
that seen for ‚Cr(CO)3(C5H5) (aiso(53Cr) ) 23 MHz, based on
measurements at low temperatures in doped single crystals of
the diamagnetic host, CpMn(CO)3),43b which indicates the
(42) (a) Rieger, P. H. Coord. Chem. ReV. 1994, 135/136, 203. (b) Hoobler, R.
J.; Hutton, M. A.; Dillard, M. M.; Castellani, M. P.; Rheingold, A. L.;
Rieger, A. L.; Rieger, P. H.; Richards, T. C.; Geiger, W. E. Organometallics
1993, 12, 116. (c) Fortier, S.; Baird, M. C.; Preston, K. F.; Morton, J. R.;
Ziegler, T.; Jaeger, T. J.; Watkins, C.; MacNeil, J. H.; Watson, K. A.;
Hensel, K.; Page, Y. L.; Charland, J. P.; Williams, A. J. J. Am. Chem. Soc.
1991, 113, 542.
(43) (a) Cooley, N. A.; Baird, M. C.; Morton, J. F.; Preston, K. F.; Page, Y. L.
J. Magn. Res. 1988, 76, 325. (b) Morton, J. R.; Preston, K. F.; Cooley, N.
A.; Baird, M. C.; Krusic, P. J.; McLain, S. L. J. Chem. Soc., Faraday
Trans. 1 1987, 83, 3535.
(44) Connelly, N. G.; Orpen, A. G.; Rieger, A. L.; Rieger, P. H.; Scott, C. J.;
Rosair, G. M. Chem. Commun. 1992, 1293.
(45) MacNeil, J. H.; Watkins, W. C.; Baird, M. C.; Preston, K. F. Organome-
tallics 1992, 11, 2761.
9
14394 J. AM. CHEM. SOC. VOL. 129, NO. 46, 2007

Confirmation of Carbonylation Intermediates
A R T I C L E S
This is strong evidence that loss of N2 occurs previous to the
formation of this intermediate.
∆S† ) 8 cal mol-1 K-1 for reaction 9. The product branching
ratio between reactions 9 and 10 was observed to depend on
CO pressure, but for a fixed CO pressure the selectivity showed
little temperature variation among experiments done at 15, 25,
and 35 °C. This implies a similar enthalpy of activation for both
loss of CO as well as addition of CO to Cp2Cr2(CO)5.
3.1.4. Reaction of [Cr(CO)3(C5H5)]2 and NNCHSiMe3.
Reaction of [Cr(CO)3Cp]2 with NdNdCHSiMe3 in a 1.5/1
stoichiometric ratio {a 3/1 ratio based on the monomeric radical
‚Cr(CO)3(C5H5)} under argon was found to occur more rapidly
for the C5H5 than the C5Me5 system. That rapid occurence is
despite the fact that lower concentrations of radical species are
present in the Cp system. Production of ketene was nearly
complete following a short induction period as shown in Figure
4. The initial organometallic products are a previously unknown
intermediate complex with major bands near 1931 and 1833
cm-1 and a minor band near 2006 cm-1 as well as small
quantities of the known triple-bonded complex [Cr(CO)2Cp]2.
As discussed later, the ν(CO) frequencies are in good agreement
with formulation of the previously unreported product as Cp2-
Cr2(CO)5 on the basis of high level density functional theory.
Thus the first stage of the reaction of [Cr(CO)3Cp]2 with Nd
NdCHSiMe3 is considered to occur in a manner similar to that
for 2‚Cr(CO)3(C5Me5) but with only trace amounts of the bound
ketene radical and a much greater amount of Cp2Cr2(CO)5.
The relatively slow rate of reactions 9 and 10, as well as the
DFT calculations discussed below, suggested that Cp2Cr2(CO)5
might be a ground-state triplet, and as such not display a
detectable NMR signal (nor ambient temperature ESR). The
reaction of [Cr(CO)3Cp]2 with NdNdCHSiMe3 was therefore
studied by NMR spectroscopy in C6D6 at 15 °C. The first NMR
spectrum was taken approximately 240 s after initiating reaction.
At that time, complete conversion of the diazo to ketene had
occurred, in keeping with earlier FTIR studies. A very small
broad peak due to [Cr(CO)3Cp]2 was detected near 5.3 ppm
and underwent only minor changes during the reaction. The only
major change seen in the NMR spectra was the appearance “out
of nowhere” of a steadily increasing signal near 4.2 ppm due
to [Cr(CO)2Cp]2 as shown in Supporting Information Figure
S10. These studies strongly support the formulation of the
3
intermediate as Cp2Cr2(CO)5.
[Cr(CO)3Cp]2
1/2[Cr(CO)2(C5H5)]2 +
OdCdCHSiMe3
f +
3.1.5. Calorimetric Study of Enthalpies of Reduction of
[Cr(CO)2Cp]2 and [Cr(CO)3Cp]2 by Na/Hg under CO. To
make an estimate of the CrtCr bond strength in [Cr(CO)2Cp]2,
+
the energies of reactions 11 [∆E ) -94.9 ( 0.8 kcal mol-1
]
NdNdCHSiMe3
trace‚Cr(CO)2(ketene)(C5H5) +
1/2Cp2Cr2(CO)5
and 12 [∆E ) -80.7 ( 1.3 kcal mol-1] were measured by
(7)
solution calorimetry in THF solution.
A very small band near 1700 cm-1 can be detected and is
assigned to small amounts of ‚Cr(CO)2(ketene)(C5H5) by
comparison with the band near 1672 cm-1 assigned to ‚Cr(CO)2-
(ketene)(C5Me5). In addition, ESR measurements for the Cp
system show the presence of a rapidly decaying radical species
in low concentration as discussed above. Reactions done with
a greater excess of [Cr(CO)3Cp]2 such that the Cr/diazo mole
ratio was 3/1 were found to produce the bands assigned to Cp2-
Cr2(CO)5 more rapidly with the near absence of initially
produced [Cr(CO)2Cp]2.
[Cr(CO)2Cp]2(s) + Na/Hg + 2CO(g) f
2NaCrCp(CO)3(soln) (11)
[Cr(CO)3Cp]2(s) + Na/Hg f 2NaCrCp(CO)3(soln) (12)
The difference between reactions 11 and 12 yield a value for
reaction 13 of ∆E ) -14.2 ( 2.1 kcal mol-1 since all other
terms cancel.
[Cr(CO)2Cp]2(s) + 2CO(g) f [Cr(CO)3Cp]2(s) (13)
The second stage of reaction under an argon atmosphere as
shown in Supporting Information Figure S7 was initially
interpreted as a disproportionation reaction:
The computed energy change is based on reactions of the solids
as shown in eq 13. Conversion of the value for ∆Esolids to ∆Hsolns
however results in no significant change in the value of -14.2
kcal mol-1 for the enthalpy of reaction 13 with all species in
solution.46 Previously we have reported47 measurements of the
2Cp2Cr2(CO)5 f [Cr(CO)3Cp]2 + [Cr(CO)2Cp]2 (8)
However decay of Cp2Cr2(CO)5 is best described by
enthalpy of reaction 14 ∆H ) -40.3 kcal mol-1
:
Cp2Cr2(CO)5 f [Cr(CO)2Cp]2 + CO
Cp2Cr2(CO)5 + CO f [Cr(CO)3Cp]2
(9)
[Mo(CO)2Cp]2(toluene soln) + 2CO(g) f
(10)
[Mo(CO)3Cp]2(toluene soln) (14)
The product ratio observed in decomposition of Cp2Cr2(CO)5
was found to depend on CO pressure. At high CO pressures
predominantly [Cr(CO)3Cp]2 is produced, and at very low CO
pressures [Cr(CO)2Cp]2 is the major product as shown in
Supporting Information Figure S8.
Kinetic studies of the rate of decay of Cp2Cr2(CO)5 as a
function of temperature and CO pressure were performed. First-
order plots of the rate of decay of Cp2Cr2(CO)5 under argon as
a function of temperature, producing primarily [Cr(CO)2Cp]2
are shown in Supporting Information Figure S9, and lead to
estimated activation parameters of ∆H† ) 23 kcal mol-1 and
The considerably more exothermic reaction of the MotMo
complex as shown in reaction 14 is in keeping with its generally
(46) The value for the enthalpy of reaction 13 as measured from solid to solid
should require little adjustment for reaction in toluene solution. Measure-
ment of the enthalpy of solution of [Cr(CO)3Cp]2 presents problems owing
to its dissociation to radicals. However experiments with the corresponding
molybdenum complexes (which undergo no such dissociation) suggest that
the enthalpy of solution of [Cr(CO)3Cp]2 in toluene is approximately 1.2
kcal mol-1 more endothermic than for [Cr(CO)2Cp]2. However, the actual
calorimetric measurements are for ∆E and not ∆H in the closed system.
Since ∆H ) ∆E + ∆PV, correction for the two moles of gas absorbed in
the reaction (RT∆n of ∼1.2 kcal mol-1) will essentially cancel the difference
in the estimated enthalpies of solution.
(47) Nolan, S. P.; de la Vega, R. L.; Hoff, C. D. Inorg. Chem. 1986, 25, 4446.
9
J. AM. CHEM. SOC. VOL. 129, NO. 46, 2007 14395

A R T I C L E S
Fortman et al.
Figure 5. Proposed reaction scheme for reaction of ‚Cr(CO)3(C5Me5).
much greater reactivity compared to the relatively inert CrtCr
complex.48 The difference between the enthalpies of reactions
13 and 14 (ca. -26 kcal mol-1) implies that while the Cr-Cr
single bond is much weaker than the Mo-Mo single bond, that
situation does not extend to the triple-bonded complexes.49
In fact, use of a value of Cr-CO ) 37 kcal mol-1 50 leads to
any formation of Cp2Cr2(CO)4(13CO) from [Cr(CO)2Cp]2 by
direct addition under these conditions. Since carbonylation of
[Cr(CO)2Cp]2 by 2CO is known to occur51 albeit slowly, and
since 3Cp2Cr2(CO)5 (see theoretical section below) adds CO at
roughly the same rate as it loses it, we conclude that the addition
of CO to1[Cr(CO)2Cp]2 is slow to the combination of both spin
state and geometric changes occurring in reaction 15.52
3.1.7. Proposed Reaction Pathways for Carbonylation of
NdNdCHSiMe3 by ‚Cr(CO)3(C5R5) R ) H, CH3. A general
mechanism in keeping with experimental studies described
above and computational results described later is presented in
Figure 5. The first step, 5i, is consistent with both the inhibiting
effect of CO as well as the observed induction period and “S”
shaped curve. This step consists of the slow loss of CO to form
the radical diazo complex ‚Cr(CO)2(NdNdCHSiMe3)(C5R5) in
a thermodynamically unfavored associative ligand displacement
reaction. In step 5ii this intermediate complex combines with a
second metal radical to form an initial adduct (the structure of
which is discussed in later theoretical sections) which goes on
to produce ‚Cr(CO)2(ketene)(C5R5) and ‚Cr(CO)2(N2)(C5R5) by
cleavage of the C-N bond. One of the products of this reaction,
‚Cr(CO)2(ketene)(C5R5) is detectable by ESR spectroscopy;
however, the second product of step 5ii, ‚Cr(CO)2(N2)(C5R5)
is undetected and presumed to be highly reactive owing to the
expected weak nature of the Cr-N2 bond.53 Reaction 5iii
involves associative displacement of N2 from ‚Cr(CO)2(N2)-
(C5R5) by free NdNdCHSiMe3 to reform ‚Cr(CO)2(NdNd
CHSiMe3)(C5R5). This is part of a reaction cycle (shown in red
in Figure 5) which accounts for the “S” shaped nature of the
reaction.
an estimate of the CrtCr bond strength of 74 kcal mol-1
slightly larger than the previous reported estimate of the Mot
Mo bond strength of 69 kcal mol-1
,
.
3.1.6. Attempted 13CO Exchange in [Cr(CO)2Cp]2. Inves-
tigation as to the reversibility of reaction 9 showed that CO did
not readily add at pressures up to ∼10 atm at room temperature.
Wrighton and co-workers have reported slow carbonylation at
pressures in excess of 100 atm.51 That reaction involves the
addition of 2 mol of CO presumably through intermediate Cp2-
Cr2(CO)5. To test this possibility, we studied reaction 15
[Cr(CO)2Cp]2 + 13CO f Cp2Cr2(CO)3(13CO) + CO (15)
At 13CO pressure of ∼10 atm and at temperatures up to
∼50 °C no real sign of incorporation of 13CO label in
[Cr(CO)2Cp]2 was observed. This was taken to indicate little if
(48) Curtis, M. D. Polyhedron 1987, 6, 759.
(49) The Cr-Cr single bond strength is on the order of 15 kcal mol-1 and the
Mo-Mo bond strength is on the order of 32.5 kcal mol-1. This difference
(17.5 kcal mol-1) alone would make the enthalpy of eq 14 exothermic by
that amount. The Mo-CO bond strength is typically viewed as being ∼4
kcal mol-1 greater than the Cr-CO bond strength. Since two CO bonds
are transferred that would make an estimate of ∆H = -17.5 + 2 × -4 )
-25.5 kcal mol-1. This is close to the measured difference and implies
that the CrtCr and MotMo bond strengths are roughly comparable in
terms of the reaction: Cp(CO)2MtM(CO)2Cp f 2MCp(CO)2 as a
definition of the bond dissociation energy (BDE) reaction. It is apparently
the case that the weak nature of the Cr-Cr single bond does not carry
over to the triple bond, at least in comparison to corresponding Mo
complexes.
(52) (a) Theoretical treatment of the role of spin-state changes in organometallic
reactions for mononuclear species is still evolving (see ref 52b). The authors
are not aware of theoretical analysis of concomitant spin state/geometric
changes being reported for metal-metal bonded systems similar to this
one. (b) Carreo´n-Macedo, J. L.; Harvey, J. N. J. Am. Chem. Soc. 2004,
126, 5789.
(50) Lewis, K. E.; Golden, D. M.; Smith, G. P. J. Am. Chem. Soc. 1984, 106,
3905.
(51) (a) Ginley, D. S.; Bock C. R.; Wrighton, M. S. Inorg. Chim. Acta 1977,
85. (b) Abrahamson, H. B.; Palazotto, M. C.; Reichel, C. L.; Wrighton, M.
S. J. Am. Chem. Soc. 1979, 101, 4123.
9
14396 J. AM. CHEM. SOC. VOL. 129, NO. 46, 2007

Confirmation of Carbonylation Intermediates
A R T I C L E S
Table 1. Computed Total Energiesa (kcal mol-1) for
Atom-Balanced Systems Relative to Cp2Cr2(CO)6 (6S-1) and
Cr-Cr Distances by B3LYP and BP86 Methods
system
B3LYP
BP86
Cr−
Crb
2CpCr(CO)2 + 2CO
1Cp2Cr2(CO)4 + 2CO
1Cp2Cr2(CO)5 + CO
3Cp2Cr2(CO)5 + CO
2CpCr(CO)3
88.3
38.5
29.2
18.8
1.2
123.0
45.9
36.6
31.7
17.1
0
2.238, 2.239 (2.215)c
2.818, 2.690
2.704, 2.645
Cp2Cr2(CO)6
0
3.340, 3.246 (3.281)d
a Energies relative to Cp2Cr2(CO)6. Data computed by conversion of
calculated energies of individual species (in hartrees) by either B3LYP or
BP86 method using a computed value of -113.328658 (B3LYP) or
-113.327235 (BP86) hartrees for CO. Data for atom balanced systems were
then subtracted from computed energy of Cp2Cr2(CO)6 (6S-1) which was
-3156.327850(B3LYP) or -3156.738922(BP86) and converted from
hartrees to kcal mol-1 b Computed Cr-Cr bond distances (Å). B3LYP value
.
is listed first, followed by BP86 value, and then experimental value (if
available) in parenthesis. c See ref 7e. d See ref 7d.
Figure 6. Computed optimized geometries of the lowest energy states for
CpCr(CO)3, CpCr(CO)2, 1Cp2Cr2(CO)6 (6S-1), 3Cp2Cr2(CO)5 (5T-1), 1Cp2-
Cr2(CO)5 (5S-1), and 1Cp2Cr2(CO)4 (4S-1). The upper numbers (bold) were
obtained by the B3LYP method, while the lower numbers (italics) were
obtained by BP86.
The structures in Figure 6 and data in Table 1 were generated
utilizing two widely accepted theoretical approaches for orga-
nometallic complexes. The computed Cr-Cr bond length for
6S-1 (∼3.3 Å) and CrtCr bond length 4S-1 (∼2.3 Å), as shown
in Table 1, are in good agreement with existing experimental
data and give confidence that the calculated value for 5T-1
(∼2.7 Å) is reasonable for the CrdCr bond length. These data
are significant in that interconversion of the single, double and
triple bonded complexes by either addition or elimination of
CO involves not only spin changes but also significant changes
in complex geometrysmost notably metal-metal bond length.
The relatively slow addition of CO to 5T-1 producing 6S-1 and
elimination of CO from 5T-1 producing 4S-1 could possibly
involve not only spin change barriers, but also concomitant
geometric changes as well.
The undetected complex “‚Cr(CO)2(N2)(C5R5)”53 is proposed
to react rapidly with a number of additional reagents depending
critically on relative concentrations. Reaction 5iv results in
deactivation by reaction of ‚Cr(CO)2(N2)(C5R5) with CO to
regenerate ‚Cr(CO)3(C5R5). Reaction 5v is the reaction of ‚Cr-
(CO)2(N2)(C5R5) with ‚Cr(CO)2(ketene)(C5R5) (the two im-
mediate products of reaction 5ii) to form final products. Reaction
5vi, radical combination of ‚Cr(CO)2(N2)(C5R5) with ‚Cr(CO)3-
(C5R5), is proposed as a major route leading ultimately to Cr2-
(CO)5(C5R5)2. Once formed, Cr2(CO)5(C5R5)2 can either elimi-
nate or add CO as shown in steps 5viii and 5ix. Finally,
dinuclear elimination of ketene from ‚Cr(CO)2(ketene)(C5R5)
and ‚Cr(CO)3(C5R5) is postulated in step 5vii.
3.2. Computational Results. 3.2.1. Structure of Mono-
nuclear Complexes. Full analysis of ground-state and excited-
state structures of the complete set of mononuclear CpCr(CO)n
(n ) 3, 2, 1) and dinuclear Cp2Cr(CO)n (n ) 6, 5, 4, 3, 2, 1)
complexes by computational techniques similar to those previ-
ously reported have been performed and will be described in
detail in a separate publication.54 The focus of this section is
on the structure, spin state, and energetics of those species
relevant to experimental kinetic and thermodynamic data in this
paper. Optimized structures computed for selected lowest energy
species are shown in Figure 6; energetic data are summarized
in Table 1.
Calculated data on reaction energies taken from Table 1
provide some insight into reaction pathways, as well as the rare
opportunity to compare computed and experimental data on
metal-metal bond strengths. The bond dissociation energy of
Cp2Cr2(CO)6 to 2‚Cr(CO)3Cp is relatively well-known: on the
order of 15 kcal mol-1 for both the Cp and Cp* systems.7 The
computed values of 1.2 (B3LYP) and 17.1 (BP86) kcal mol-1
differ from each other by a significant amount. The BP86 value
is in reasonable accord with experiment. The computed energies
of conversion of Cp2Cr2(CO)6 to Cp2Cr2(CO)4 + 2CO of 38.5
(B3LYP) and 45.9 (BP86) kcal mol-1 are considerably higher
than the experimental measurement reported in this paper of
∼14 kcal mol-1. In this case the B3LYP is closer to the
experimental result. While solvation energies can complicate
the comparison of gas-phase theoretical values with solution-
phase measurements, in this case solvation energies for Cp2-
Cr2(CO)6 and Cp2Cr2(CO)4 are likely comparable. Therefore it
seems that the strength of the CrtCr triple bond may be
somewhat underestimated. The nature of semibridging carbonyl
bonding in these complexes (in the authors’ opinion) makes
theoretical treatment in complexes such as Cp2Cr2(CO)4 par-
ticularly challenging.
(53) (a) The labile nature of the Cr-N2 compared to the Cr-CO bond is
illustrated by the fact that the complex Cr(CO)5(N2) is stable enough to be
detected by fast infrared spectroscopy but decays with a first-order rate
constant of 1.7 s-1 (see ref 53b). This is presumably a dissociative
displacement of bound N2. In the system studied here, the Cr-N2 absolute
bond strength might be somewhat stronger, but associative displacement
in the 17 e- radical would also make it more labile to ligand displacement
by any stronger ligand. The LnM-CO bond is typically 15-20 kcal mol-1
stronger than the LnM-N2 bond for low valent complexes (see ref 53c). In
addition, 18 e- complexes Cr(CO)2(N2)(arene) have been photochemically
prepared (see ref 53d). These observations all point to expected thermo-
chemical stability for ‚Cr(CO)2(N2)C5Me5 but rapid kinetic lability. Despite
the reasonable nature of its postulation as a reactive intermediate in this
process, there is to the author’s knowledge no experimental data regarding
‚Cr(CO)2(N2)C5Me5. (b) Church, S. P.; Grevels, F. W.; Hermann, H.;
Schaffner, K. Inorg. Chem. 1984, 23, 3830. (c) Hoff, C. D. Prog. Inorg.
Chem. 1992, 40, 503. (d) Strohmeier, W.; Hellmann, H. Chem. Ber. 1964,
97, 1877.
An additional area of mechanistic interest is the energy of
loss of CO from ‚Cr(CO)3Cp to form ‚Cr(CO)2Cp + CO since
generation of this reactive intermediate 15e- complex could
conceivably play a role in diazo activation. Calculating this
energy from the data in Table 1 yields a value of 43.6 (B3LYP)
or 53.0 (BP86) kcal mol-1. This indicates that the binding of
(54) Li, Q. S.; Zhang, X.; Xie, Y.; Schaefer, H. F., III; King, R. B. Unpublished
work.
9
J. AM. CHEM. SOC. VOL. 129, NO. 46, 2007 14397

A R T I C L E S
Fortman et al.
Figure 7. Drawing of the calculated SOMO for (a) ‚Cr(CO)2(NNCH2)(C5H5) and (b) ‚Cr(CO)3(C5H5).
CO in ‚Cr(CO)3Cp may be even stronger than that in Cr(CO)6
for which an experimental value of 37 kcal mol-1 has been
reported.55
3.2.2. Calculation of Transition-State Structures. In place
of the experimentally studied NNCHSiMe3 the computationally
simpler molecule NNCH2 was used to calculate structures and
activation energies for key intermediates using the SDD basis
set with the corresponding effective core potential for chromium
and the 6-311G(d,p) basis set for all the other atoms.59 The first
step in Figure 5, responsible for the induction period in the
reaction and also its inhibition by CO, is substitution of a CO
ligand by NdNdCHSiMe3 to form ‚Cr(CO)2(NNCHSiMe3)-
(C5R5). A molecular orbital calculation of the SOMO of a
simplified model of this radical complex is shown in Figure 7.
The net reaction free energy of this elementary step is 3.6 kcal
Of key importance to this paper was the calculation that the
most stable state of Cp2Cr2(CO)5 is a triplet (5T-1) which is
calculated to lie 10.4 (B3LYP) or 4.9 (BP86) kcal mol-1 lower
than the lowest lying singlet state (5S-1). These data are in
complete accord with experimental data which showed no NMR
signal detectable for the intermediate complex formulated as
Cp2Cr2(CO)5 based on its observed disproportionation to Cp2-
Cr2(CO)4 and Cp2Cr2(CO)6. Another key feature is the accept-
able agreement between experimental and theoretical values for
ν(CO) frequencies utilizing the BP86 method, which in our
experience appears to be more reliable for calculating ν(CO)
than the B3LYP method for first row transition-metal carbonyl
derivatives.56,57 The computed infrared frequencies (ν(CO) in
cm-1 followed by intensity in parentheses) for 4S-1 of 1876
(0), 1890 (1332), 1904 (1146), and 1940 (0) are in good
agreement with our measured values in toluene solution (this
report) of 1878 and 1901 cm-1. The intermediate from the
reaction of Cp2Cr2(CO)6 and NdNdCHSiMe3 suggested above
to be Cp2Cr2(CO)5 exhibits a strong band at 1922 cm-1, a
mol-1
.
Significant spin density appears resident on the terminal C
of the diazo ligand of this radical. Addition of an H atom to
the proposed intermediate shown in Figure 7a would yield a
methyldiazenido complex Cr(CO)2(NNCH3)(C5H5) as shown in
‚Cr(CO)2(NNCH2)(C5H5) + ‚H f
Cr(CO)2(NNCH3)(C5H5) (16)
Analogous diazenido complexes of Mo and W are known.60,61
Formation of the diazenido complex, isoelectronic with the
highly stable nitrosyl complex Cr(CO)2(NO)(C5H5) would be
expected to be highly favorable. Such a radical combination
would be expected to be rapid62 and the structure of a key
reaction intermediate during step 5ii of Figure 5 is proposed to
be the addition of ‚Cr(CO)3C5H5 rather than ‚H to yield an
intermediate complex (C5H5)(CO)2Cr-NtN-CH2-Cr(C5H5)-
(CO)3). A computed structure for this proposed intermediate is
shown in Figure 8. The enthalpy of addition of NO to ‚Cr-
(CO)3C5Me5 has been measured and indicates NO acts as a three
electron donor. The Cr-NO bond strength is 33.2 ( 1.8 kcal
mol-1 stronger than the Cr-CO bond. This driving force, plus
the formation of the Cr-C bond (expected to be weak) would
serve to stabilize an intermediate adduct such as that shown in
medium band at 1832 cm-1, and a weak band near 2006 cm-1
.
These experimental values are consistent with the calculated
values for the lowest energy isomer of Cp2Cr2(CO)5 (5T-1) of
1842 (510), 1863 (101), 1910 (917), 1925 (1828), and 1961
(9) cm-1. Thus the experimentally observed medium band at
1832 cm-1 may correspond to the unresolved calculated ν(CO)
frequencies at 1842 and 1863 cm-1 for 5T-1. Similarly the
observed strong band at 1922 cm-1 may correspond to the
unresolved strong calculated ν(CO) frequencies at 1910 and
1925 cm-1 for 5T-1. The weak band near 2006 cm-1 may
correspond to the weak band computed to occur at 1961 cm-1
.
Finally it should be mentioned that the calculated Cr-Cr
2
distance in 5T-1 of 2.690 Å suggesting a σ + /2 π CrdCr
double bond for triplet Cp2Cr2(CO)5 is similar to the known58
FedFe double bond in the stable triplet Cp2Fe2(µ-CO)3, which
was also found to not display a measurable NMR signal.
(59) (a) Calculations of transition-state structures were performed using
analogous computational methods to those for stable molecules and as
described in references 57b and 57c below. (b) Dolg, M.; Stoll, H.; Preuss,
H.; Pitzer, R. M. J. Phys. Chem. 1993, 97, 5852. (c) Krishnan, R.; Binkley,
J. S.; Seeger, R.; Pople, J. A. J. Chem. Phys. 1980, 72, 650.
(60) Herrmann, W. A. Angew. Chem., Int. Ed. Engl. 1975, 14, 355.
(61) Herrmann, W. A.; Biersack, H. Chem. Ber. 1977, 110, 896.
(62) In, addition to rapid radical-radical combination reactions, studies of
ultrafast disproportionation reactions of the photochemically generated
radical ‚W(CO)3C5H5 have recently been reported: Kling, M. F.; Cahoon,
J. F.; Glascoe, E. A.; Shanoski, J. E.; Harris, C. B. J. Am. Chem. Soc.
2004, 126, 11414.
(55) This difference could be attributed to competition by coordinated CO for
electron density from the central metal atom via M f CO π* back-bonding.
Thus in Cr(CO)6 there are six CO groups competing for this electron density,
whereas in ‚Cr(CO)3Cp there are only three. Backbonding would be
expected to be stronger on this basis.
(56) Jonas, V.; Thiel, W. J. Phys. Chem. 1995, 102, 8474.
(57) Silaghi-Dumitrescu, I.; Bitterwolf, T. E.; King, R. B. J. Am. Chem. Soc.
2006, 128, 5342.
(58) Blaha, J. P.; Bursten, B. E.; Dewan, J. C.; Frankel, R. B.; Randolph, C. L.;
Wilson, B. A.; Wrighton, M. S. J. Am. Chem. Soc. 1985, 107, 4561.
9
14398 J. AM. CHEM. SOC. VOL. 129, NO. 46, 2007

Confirmation of Carbonylation Intermediates
A R T I C L E S
Two routes involving migration of the terminal CH2 of the
diazo group were also examined. The most natural way to form
ketene complexes seemed to proceed via a movement of the
CH2 group toward the carbon atom of one of the CO ligands.
This process, however, taking place through the transition state
TS2 (its characteristic imaginary frequency is 397i cm-1) is
predicted to be disfavored compared to the other competitive
pathways owing its relatively high free energy barrier of 33.6
kcal mol-1
.
Alternatively, the electron-rich terminal methylene group of
‚Cr(CO)2(NNCH2)(C5H5) may be attacked by external carbon
monoxide resulting in ketene and ‚Cr(CO)2(N2)(C5H5) in a
highly exothermic process with a reaction free energy of -30.4
kcal mol-1. This concerted reaction proceeds via the transition
state TS3 (its single imaginary frequency is 346i cm-1) with a
free energy of activation of 16.8 kcal mol-1, representing a
viable pathway for ketene formation with a lower activation
barrier, than for the analogous uncatalytic reaction.21 The
transition states involving methylene migration are depicted in
Figure 10.
Figure 8. Calculated structure of (C5H5)(CO)2Cr-NtN-CH2-Cr(C5H5)-
(CO)3).
Despite the fact that the pathway for direct addition of CO
via transition state TS3 is computed to be low in energy, it is
viewed by the authors as probably not occurring in the system
studied here. For the system studied, associative displacement
of the bound diazo by CO (the reverse of 5i in Figure 5) is
competitive and probably has a lower activation energy than
does addition of CO to the bound diazo radical as shown in
TS3. Nevertheless TS3 presents a viable alternative to dinuclear
production. In the absence of alternative reaction channels it is
expected to be competitive. Thus it is of fundamental interest
and could conceivably play a role in reactions of other
complexes or on surfaces.
Figure 9. Computed transition state for step 5ii in the mechanism in Figure
5.
Figure 8. The related compounds M(CO)2(NNCH3)(C5H5) (M
) Mo, W) are known.63
The intermediate complex shown in Figure 8 is poised for
formation of a C-C bond by migration of the Cr-C bond to a
coordinated terminal CO. A second critical step is proposed to
be migration of an alkyl radical in the bridging dinuclear
intermediate, shown in Figure 8, from chromium to a carbonyl
group in a manner analogous to CO insertion. As this occurs,
the driving force to form ketene and N2 provides the needed
energy to split the complex to the two intermediate radicals
‚Cr(CO)2(ketene)(C5R5) and ‚Cr(CO)2(N2)(C5R5) as shown in
step 5ii of Figure 5. The computed transition state is shown in
Figure 9. The reaction is exothermic with a free-energy
difference of -9.3 kcal mol-1 and proceeds with a free-energy
barrier of 21.3 kcal mol-1. Harmonic vibrational frequency
calculation confirmed that TS1 is indeed a transition state
containing a single imaginary frequency (132i cm-1).
One possible reaction of ‚Cr(CO)2(ketene)(C5R5) and ‚Cr-
(CO)2(N2)(C5R5) is proposed to be recombination to reform a
Cr-Cr bond and eliminate N2, ketene, and form Cr2(CO)4Cp2.
However, as shown in Figure 9, the geometry of cleaving the
C-N bond from an intermediate such as that shown in Figure
8 appears to present an initially unfavorable steric orientation
to radical combination to form a Cr-Cr bond along this
trajectory. Therefore it is proposed (especially for the experi-
mentally studied complex which has a pendant SiMe3 on the
diazo carbon) that the radicals predominantly diffuse into
solution rather than directly form N2, ketene, and Cr2(CO)4Cp2
in a single step. This is illustrated in Figure 9 and is also in
accord with spectroscopic data which shows the buildup of other
intermediate complexes rather than direct formation of products.
Finally, computation of the minimum energy configuration
of the intermediate complex formulated as ‚Cr(CO)2(η2ketene)-
(C5R5) yielded two tautomeric structures: a C,C bonded and a
C,O bonded form as shown in Figure 11.
The computed energies of the C,C and C,O tautomers shown
in Figure 11 are quite close, with the C,O tautomer being lower
in energy by 1.3 kcal mol-1. The CpMn(CO)2(ketene) com-
plexes prepared by Herrmann and co-workers6 exhibit the C,C
bonded form of the ketene. Work of Grotjahn has shown there
is a fine balance that governs bonding in the C,C ketene, C,O
ketene, and carbene + CO and other bonding motifs available
in the R2CdCdO system. Infrared and ESR data do not allow
for conclusive64 assignment of which of the two binding modes
shown in Figure 11 best fits experimental data. The relatively
strong analogy to the manganese carbonyl complex suggests
that the C,C form is the more likely structure, as well as the
most likely transition state leading to the bound form of ketene
as shown in Figure 9 suggests the C,C tautomer. At this point,
no firm conclusion can be made regarding whether ‚Cr(CO)2-
(64) Grotjahn and co-workers (see ref 4) have shown that for coordination to
the IrCl(PiPr3)2 fragment Ph2CdCdO coordinates in an η2-C,O form with
νCO ) 1632 cm-1, but that Ph(H)CdCdO coordinates in an η2-C,C form
with νCO ) 1752 cm-1. The Cr radical complex reported here has νCO )
1672 cm-1 which is somewhat closer to the C,O tautomer for Ir. However
the electronic differences between the Cr and Ir complexes are large enough
that no conclusion can be made at this time. Both the η2-C,O form and
η2-C,C form have two approximately symmetry equivalent CO ligands.
The ESR data showing hyperfine splitting from two equivalent 13C nuclei
derived from 13CO cannot distinguish between the two forms. Use of 13C
labeled diazoalkane might be helpful in identifying possible coordination
by this carbon (in the η2-C,C form), but is synthetically beyond the scope
of the present study.
(63) Hillhouse, G. L.; Haymore, B. L.; Herman, W. A. Inorg. Chem. 1979, 18,
2423.
9
J. AM. CHEM. SOC. VOL. 129, NO. 46, 2007 14399

A R T I C L E S
Fortman et al.
Figure 10. Computed transition states for the internal CH2 migration (TS2) and for the intermolecular CH2 migration step (TS3) resulting in ‚Cr(CO)2-
(N2)(C5H5) and ketene.
Figure 11. Calculated minimum energy structures for C,C and C,O bonded tautomers of ‚Cr(CO)2(η2-OdCdCH2)(C5H5). Selected bond lengths are given
in angstrom.
(η2-ketene)(C5H5) is a C,C or C,O bonded tautomer. Additional
experimental and computational work on this and related radical
diazo alkane and ketene complexes is planned.
that for Co2(CO)8 studied by Ungvary and co-workers.2 Even
in the well-developed field of organometallic chemistry of
transition metals such as chromium and cobalt, interesting new
complexes, reaction mechanisms, and theoretical understanding
remain a source of surprise.
4. Conclusion
Combinations of theoretical and experimental approaches are
used to propose a radical-based complex reaction mechanism
for the seemingly simple conversion of a diazoalkane and CO
to the corresponding ketene and N2. Two key paramagnetic
complexes were observed spectroscopically (triplet not actually
observed directly) and supported by DFT studies, namely3
(R5C5)2Cr2(CO)5 and ‚Cr(CO)2(η2-OdCdCHSiMe3)(C5R5). The
rate determining step is attributed to formation of the undetected
radical diazoalkane complex ‚Cr(CO)2(η1-NdNdCHSiMe3)-
(C5R5) which is then trapped by a second mole of ‚Cr(CO)3-
(C5R5) to form (C5H5)(CO)2Cr-NtN-C(H)(SiMe3)-Cr(C5H5)-
(CO)3) and subsequently leads to products as outlined in Figure
5. Mechanisms cannot be proven; however, the reasonable
accord between theory and experiment presented here supports
a concerted mechanism of this type and represents our best
understanding of the complicated sequence of reactions in this
system. The mechanism of carbonylation of diazoalkanes by
Cr2(CO)6(C5R5)2 reported here appears to be quite different from
Acknowledgment. The authors dedicate this paper to Profes-
sor Ferenc Ungva´ry of the University of Pannonia, Hungary.
Support of this work by the National Science Foundation Grants
CHE-0615743 (C.D.H.) and CHE-0209857 (R.B.K.), the Hun-
garian Scientific Research Fund Grant No. OTKA F046959
(T.K.) and the 111 Project B07012 (Q.L.) in China is gratefully
acknowledged.
Supporting Information Available: Figures S1-S10 and
Table ST-1, including FTIR spectra as a function of time, ESR
1
spectra, H NMR spectra, kinetic plots and computed SOMO
and spin densities for the tautomers of the proposed intermediate
[•Cr(CO)2(OdC)CH2)Cp]; FTIR data for selected compounds
and intermediates in toluene; complete ref 37. This material is
available free of charge via the Internet at http://pubs.acs.org.
JA075008O
9
14400 J. AM. CHEM. SOC. VOL. 129, NO. 46, 2007