Mechanistic and synthetic studies of the addition of alkyl aldehydes to vinylsilanes catalyzed by Co(I) complexes

Article abstract of DOI:10.1021/ja980610n

The mechanistic details of the cobalt-catalyzed intermolecular hydroacylation reaction have been investigated using kinetic, spectroscopic, and crystallographic methods. The Co(I) bisolefin complex 1, [C₅MesCo(C₂H₃SiMe₃)₂], was shown to catalyze the addition of a series of alkyl aldehydes (2a-1) to vinylsilanes to give the corresponding ketones with exclusive anti-Markovnikov selectivity under mild conditions. The catalytic cycle exhibits two resting states, complex 1 and a bisalkyl carbonyl complex, [C₅Me₅Co(CO)(R)(R')], 4a-1 which are in equilibrium. Kinetic investigations along with low-temperature NMR spectroscopy establish a sensitive balance between resting states during catalysis which is strongly dependent on substrate structure. The turnover-limiting step was established as the reductive elimination of ketone from intermediate 4. Using ferrocenecarboxaldehyde (Fc-C(O)H), 21, as substrate, the intermediate 41 [C₅Me₅Co(CO)(Fc)(CH₂CH₂SiMe₃)] was isolated at low temperatures and characterized by X-ray crystallography. Complex 41 was used to study the carbon-carbon bond-forming step directly by thermolysis in the presence of a trapping ligand L (P(OMe)₃, PMe₃). Kinetic analysis showed competitive ligand dependent and ligand independent pathways for ketone formation. Deuterium scrambling, isomerization of aldehydes prior to ketone formation, and production of isomeric ketones in certain cases establish that complex isomerization processes occur prior to productive ketone elimination from 4. A detailed mechanism accounting for all observations is proposed. Catalyst deactivation was shown to involve primarily decarbonylation to yield [C₅Me₅Co(CO)]₂ and [C₅Me₅Co(CO)(C₂H₃SiMe₃)]. When excess aldehyde is present, catalytic aldehyde dimerization occurs to give esters.

Full text of DOI:10.1021/ja980610n

J. Am. Chem. Soc. 1998, 120, 6965-6979
6965
Mechanistic and Synthetic Studies of the Addition of Alkyl
Aldehydes to Vinylsilanes Catalyzed by Co(I) Complexes
Christian P. Lenges, Peter S. White, and Maurice Brookhart*
Contribution from the Department of Chemistry, UniVersity of North Carolina at Chapel Hill,
Chapel Hill, North Carolina 27599-3290
ReceiVed February 23, 1998
Abstract: The mechanistic details of the cobalt-catalyzed intermolecular hydroacylation reaction have been
investigated using kinetic, spectroscopic, and crystallographic methods. The Co(I) bisolefin complex 1, [C5-
Me5Co(C2H3SiMe3)2], was shown to catalyze the addition of a series of alkyl aldehydes (2a-l) to vinylsilanes
to give the corresponding ketones with exclusive anti-Markovnikov selectivity under mild conditions. The
catalytic cycle exhibits two resting states, complex 1 and a bisalkyl carbonyl complex, [C5Me5Co(CO)(R)-
(R′)], 4a-l which are in equilibrium. Kinetic investigations along with low-temperature NMR spectroscopy
establish a sensitive balance between resting states during catalysis which is strongly dependent on substrate
structure. The turnover-limiting step was established as the reductive elimination of ketone from intermediate
4. Using ferrocenecarboxaldehyde (Fc-C(O)H), 2l, as substrate, the intermediate 4l [C5Me5Co(CO)(Fc)(CH2-
CH2SiMe3)] was isolated at low temperatures and characterized by X-ray crystallography. Complex 4l was
used to study the carbon-carbon bond-forming step directly by thermolysis in the presence of a trapping
ligand L (P(OMe)3, PMe3). Kinetic analysis showed competitive ligand dependent and ligand independent
pathways for ketone formation. Deuterium scrambling, isomerization of aldehydes prior to ketone formation,
and production of isomeric ketones in certain cases establish that complex isomerization processes occur prior
to productive ketone elimination from 4. A detailed mechanism accounting for all observations is proposed.
Catalyst deactivation was shown to involve primarily decarbonylation to yield [C5Me5Co(CO)]2 and [C5Me5-
Co(CO)(C2H3SiMe3)]. When excess aldehyde is present, catalytic aldehyde dimerization occurs to give esters.
Introduction
ate cyclopentanones using a cationic Rh(I) catalyst described
by Bosnich and co-workers.28-33 The mechanism of intramo-
The selective activation of a C-H bond by a transition metal
complex followed by the functionalization of the activated
substrate has been an area of intensive research. One of the
few catalytic processes reported which successfully combines
C-H bond activation with C-C bond formation1-10 is the
hydroacylation11-27 of olefins to generate ketones.
(13) Watanabe, Y.; Tsuji, Y.; Kondo, T.; Akazome, M. J. Org. Chem.
1990, 55, 1286-1291.
(14) Sneeden, R. P. A.; Denise, B.; Isnard, P. J. Organomet. Chem. 1982,
240, 285-288.
(15) Marder, T. B.; Roe, C. D.; Milstein, D. Organometallics 1988, 7,
1451-1453.
(16) Sakai, K.; Suemune, H.; Ichikawa, Y.; Xie, Z.-F. Chem. Pharm.
Bull. 1987, 35, 1812-1816.
(17) Sakai, K.; Kawahara, T.; Suemune, H. Chem. Pharm. Bull. 1986,
34, 550-557.
(18) Sakai, K.; Oda, O. Tetrahedron Lett. 1972, 4375-4376.
(19) Sakai, K.; Ide, J.; Oda, O.; Nakamura, N. Tetrahedron Lett. 1972,
1287-1290.
The most thoroughly studied example of this reaction has
been the intramolecular hydroacylation of 4-pentenals to gener-
(20) Miller, R. G.; Campbell, R. E.; Lochow, C. F.; Vora, K. P. J. Am.
Chem. Soc. 1980, 102, 5824-5830.
(1) Kim, Y. H.; Kang, J.-B.; Lim, Y.-G. J. Chem. Soc., Chem. Commun.
1994, 2267.
(21) Miller, R. G.; Lochow, C. F.; Krishnakant, P. V. J. Organomet.
Chem. 1980, 192, 257-264.
(2) (a) Jones, W. D.; Duttweiler, R. P.; Feher, F. J. J.; Hessell, E. T.
New J. Chem. 1989, 13, 725-736. (b) Jones, W. D.; Foster, G. P.; Putinas,
J. M. J. Am. Chem. Soc. 1987, 109, 5047-5048.
(3) Tanaka, M.; Sakakura, T. Pure Appl. Chem. 1990, 62, 1147-1150.
(4) Shul′pin, G. B.; Shilov, A. E. Chem. ReV. 1997, 97, 2879-2932.
(5) Jordan, R.; Taylor, D. F. J. Am. Chem. Soc. 1989, 111, 778.
(6) Murai, S.; Chatani, N.; Ie, Y.; Kakiuchi, F. J. Org. Chem. 1997, 62,
2604-2610.
(7) Murai, S.; Kakiuchi, F.; Tanaka, Y.; Kamatani, A.; Sonoda, M.;
Chatani, N. Bull. Chem. Soc. Jpn. 1995, 68, 62-83.
(8) Murai, S.; Kakiuchi, F.; Sekine, S.; Tanaka, Y.; Kamatani, A.; Sonoda,
M.; Chatani, N. Nature (London) 1993, 366, 529.
(9) Murai, S.; Chatani, N.; Yamada, A.; Kakiuchi, N.; Fujii, N. Chem.
Lett. 1997, 425.
(22) Miller, R. G.; Lochow, C. F. J. Am. Chem. Soc. 1976, 98, 1281-
1283.
(23) Larock, R. C.; Potter, G. F.; Oertle, K. J. Am. Chem. Soc. 1980,
102, 190-197.
(24) James, B. R.; Young, C. G. J. Chem. Soc., Chem. Commun. 1983,
1215-1216.
(25) Gable, K. P.; Benz, G. A. Tetrahedron Lett. 1991, 32, 3473-3476.
(26) Vinogradov, M. G.; Tuzikov, A. B.; Nikishin, G. I. IzV. Akad. Nauk
SSSR, Ser. Khim. 1985, 11, 2557-2562.
(27) Vinogradov, M. G.; Tuzikov, A. B.; Nikishin, G. I. IzV. Akad. Nauk
SSSR, Ser. Khim. 1983, 7, 1686-1687.
(28) Bosnich, B.; Barnhart, R. W. Organometallics 1995, 14, 4343-
4348.
(10) Trost, B. M.; Imi, K.; Davies, I. W. J. Am. Chem. Soc. 1995, 117,
5371.
(11) Watanabe, W. H.; Conlon, L. E. J. Am. Chem. Soc. 1957, 79, 2828.
(12) Watanabe, Y.; Tsuji, Y.; Kondo, T. Tetrahedron Lett. 1987, 28,
6229-6230.
(29) Bosnich, B.; Fairlie, D. P. Organometallics 1988, 7, 936-945.
(30) Bosnich, B.; Fairlie, D. P. Organometallics 1988, 7, 946-954.
(31) Bosnich, B.; Whelan, J.; Bergens, S. H.; Noheda, P.; Wang, X.;
Barnhart, R. W. Tetrahedron 1994, 50, 4335-4346.
(32) Bosnich, B.; Wang, X. Organometallics 1994, 13, 4131-4133.
S0002-7863(98)00610-6 CCC: $15.00 © 1998 American Chemical Society
Published on Web 07/07/1998

6966 J. Am. Chem. Soc., Vol. 120, No. 28, 1998
Lenges et al.
lecular hydroacylation was investigated in detail for this system
using labeling studies and product analysis. However, this
catalyst system is quite specific for intramolecular hydroacy-
lation and in this case for the formation of cyclopentanones.
The intermolecular version of this reaction has been reported
for only a few examples with, in general, reduced activity.11-15,34
Marder, Milstein, and Roe15 reported a system which involves
use of the indenylrhodium ethylene complex [(C8H7)Rh(C2H4)2]
as a catalyst for the addition of benzaldehyde to ethylene under
quite severe conditions (1000 psi of C2H4, 100 °C, 12 h). About
5 TO were observed in the conversion of benzaldehyde to
propiophenone with this system. Recently Jun et al. have
reported a procedure using a Rh(I)-phosphine catalyst in
combination with 2-aminopyridines to effect intermolecular
hydroacylations. The intermediate formation of a pyridineimine
is proposed to facilitate the catalysis; however, mechanistic
details are scarce.34
Due to the increased reactivity toward dissociative olefin loss,
we were able to investigate the reactivity of the accessible 16-
electron cobalt(I) system, [C5Me5Co(C2H3SiMe3)], toward other
C-H bonds. Recently we reported that Co(I) complex 1
exhibits H/D exchange in C6D6 at room temperature via a Co-
(III) intermediate produced from oxidative addition of benzene
to a Co(I) species (eq 1).39
We have reported the intermolecular hydroacylation of
vinylsilanes by aromatic aldehydes using a cobalt(I) bisolefin
complex as catalyst.35
In this paper we report the activity of 1 as a catalyst for the
hydroacylation of vinyl silanes using a variety of aliphatic
aldehydes (eq 2). Significant differences are observed in the
Key mechanistic features of this catalysis using cobalt complex
1 [C5Me5Co(C2H3SiMe3)2] were investigated. Complex 1 is
the resting state throughout catalysis; a first-order dependence
of the turnover process on aldehyde concentration and an inverse
first-order dependence on olefin concentration were discovered.
Crucial for these catalytic hydroacylations was the lability of
the bulky trimethylsilyl-substituted olefin.36-38 For example,
1 reacts readily with P(OMe)3 below room temperature to
generate the monophosphite adduct.39 The corresponding
bisethylene complex, [C5Me5Co(C2H4)2], shows the same
reactivity in the course of 1 h at 60 °C,40 and attempts to carry
out hydroacylations of ethylene or strongly binding olefins such
as styrene were unsuccessful.
hydroacylation of aliphatic aldehydes in comparison to aromatic
aldehydes. On the basis of a kinetic analysis as well as in situ
NMR spectroscopic studies and X-ray analysis of an intermedi-
ate during catalysis, the catalyst resting states have been
identified. Kinetic results in combination with labeling experi-
ments allow formulation of a mechanism for this complex
transformation which can explain deactivation routes as well
as the formation of isomeric products from certain substrates
in this catalytic process.
Results and Discussion
A. Synthesis and Structure of [C5Me5Co(C2H3SiMe3)2],
1. The cobalt(I) complex 1 was prepared in a two-step synthesis
(eq 3) starting from CoCl2 using a modified procedure based
on the reported synthesis of the parent Co(I)-bisethylene
complex [C5Me5Co(C2H4)2].41,42
(33) (a) Bosnich, B.; Whelan, J.; Bergens, S. H.; Noheda, P.; Wang, X.;
Barnhart, R. W. J. Am. Chem. Soc. 1994, 116, 1821-1830. (b) Bosnich,
B.; Barnhart, R. W.; McMorran, D. A. Inorg. Chim. Acta 1997, 263, 1-7.
(34) Jun, C.-H.; Lee, H.; Hong, J.-B. J. Org. Chem. 1997, 62, 1200.
Aminopyridines are used in this Rh catalyst system as cocatalysts for
intermolecular hydroacylation.
(35) Lenges, C. P.; Brookhart, M. J. Am. Chem. Soc. 1997, 119, 3165.
(36) This notion is supported by the observation of Spencer37 that the
corresponding bispropene complex [C5Me5Co(C2H3Me)2] is only stable in
the presence of excess propene in solution and highly reactive toward ligand
substitution reactions: on the other hand, the parent bisethylene Co(I)
complex [C5Me5Co(C2H4)2] is more resistant toward ligand substitution.
(37) Spencer, J. L.; Beevor, R. G.; Frith, S. A. J. Organomet. Chem.
1981, 221, C25-C27.
Suitable crystals for an X-ray structure determination were
grown from an acetone solution of 1 at -78 °C. As shown in
(38) (a) Boennemann, H. Angew. Chem., Int. Ed. Engl. 1985, 24, 248.
(b) Benn, R.; Cibura, K.; Hofmann, P.; Jonas, K.; Rufinska, A. Organo-
metallics 1985, 4, 2214-2221.
(39) Lenges, C. P.; Brookhart, M.; Grant, B. E. J. Organomet. Chem.
1997, 528, 199-203.
(40) Brookhart, M.; Grant, B.; Volpe, A. Organometallics 1992, 11,
3920-3922.
(41) Koelle, U.; Fuss, B.; Belting, M.; Raabe, E. Organometallics 1986,
5, 980-987.
(42) Koelle, U.; Khouzami, F.; Fuss, B. Angew. Chem., Int. Ed. Engl.
1982, 21, 131-132.

Addition of Alkyl Aldehydes to Vinylsilanes
J. Am. Chem. Soc., Vol. 120, No. 28, 1998 6967
Table 1. Hydroacylation of Isovaleraldehyde with
Vinyltrimethylsilane. Substrate Dependence of TOF and Ratio of
Co Complexes 1 to 4c during Catalysis
TOF concn of 4c
entry aldehyde olefin ratio A:O ratio 4c:1 [TO/h]
[mol/L]
1
2
3
4
5
6
7
8
9
40
40
30
40
30
20
20
20
20
20
20
10
20
20
20
40
30
20
20
30
40
50
70
40
2
2
1.5
1
1
1
1
0.75
0.5
0.4
0.29
0.25
3.3
3.4
2.7
1.8
2.0
2.3
2.1
1.5
1.1
1.0
0.4
0.5
3.75
3.75
3.35
3.00
3.00
2.80
2.95
2.60
2.50
1.70
1.25
1.20
0.0162
0.0163
0.0154
0.0136
0.0141
0.0147
0.0143
0.0127
0.0111
0.0106
0.0060
0.0070
10
11
12
Figure 1. ORTEP view of [C5Me5Co(C2H3SiMe3)2] (1). All atoms
are drawn at 50% probability ellipsoids. Selected bond distances (Å)
and angles (deg): C1-C2 1.404(4); Co-C1 2.002(3); Co-C2 2.058-
(3); C2-Si1 1.865(3); C3-C4 1.414(4); Co-C3 2.018(3); Co-C4
2.065(3); C4-Si2 1.855(3); C1-Co-C2 40.43(12); C1-Co-C3 97.71-
(12); C1-Co-C4 95.20(12); C2-Co-C4 115.55(11); Si1-C2-C1
128.11(15).
well as the ratio of aldehyde to olefin. In addition to persistence
of 1 during catalysis, 1H NMR monitoring reveals the formation
of a second Co-containing species which exhibits a ratio of Cp*
to SiMe3 groups of 1:1. Additional resonances in the aliphatic
region are observed but are quite complex. This intermediate
is assigned as the bisalkyl carbonyl cobalt(III) complex, 4c, (eq
5); a more rigorous proof of its structure is given below.
the ORTEP diagram43 (Figure 1), the arrangement of the
coordinated trimethylvinylsilane ligands generates a C2-sym-
1
metric structure which is consistent with the simple H NMR
spectrum of 1.
The carbon-carbon double bond length of the coordinated
olefin of 1.40 Å (C1-C2) is characteristic for a low-valent olefin
complex in which back-bonding from filled metal d-orbitals is
dominant in olefin binding.44-47 The Co-C olefin distances
are very similar; however, the length of Co-C2, the carbon
bearing the trimethylsilyl group, is slightly elongated, reflecting
the steric influence of this substituent. The angles C1-Co-
C3 (97.7°) and C2-Co-C4 (115.6°) indicate a twist in the
coordinated olefin in support of this assignment. Other isomers
with alternative orientations of the coordinated olefins are not
observed.48
The ratio of bisolefin complex 1 to the intermediate species
4c is dependent on the ratio of aldehyde to olefin substrates. In
a reaction with an aldehyde-to-olefin ratio of 1:1 and a 5 mol
% catalyst load, the ratio of complex 4c to 1 is 2:1 in the initial
phase of catalysis. Increasing aldehyde concentration increases
the ratio of 4c to 1, while increasing the olefin concentration
decreases this ratio. The equilibrium between the two Co-
containing species 1 and 4c is rapidly established with respect
to catalytic ketone formation. For example, addition of isov-
aleraldehyde to a d6-acetone solution of 1 and trimethylvinyl-
silane at room temperature establishes the ratio of compounds
1 and 4c in the course of minutes. Upon raising the temperature
to 35 °C, catalysis with a turnover frequency of about 3 TO/h
is observed and the ratio of 1 to 4c is maintained. Table 1
summarizes the ratios of 4c to 1 observed for various ratios of
isovaleraldehyde to trimethylvinylsilane and at various absolute
concentrations of these substrates. Initial turnover frequencies
(TOFs) under these conditions are also listed. Several conclu-
sions can be drawn from these data. Increasing the concentra-
tion of olefin in the reaction mixture results in decreased TOFs
for ketone formation and a change in the ratio of 4c to 1 which
favors 1. On the other hand, increasing aldehyde concentration
increases the ratio of 4c to 1 in favor of 4c and increases the
TOF. Addition of excess trimethylvinylsilane to a reaction
mixture with a substrate ratio of 1:1 during catalysis and after
10% ketone formation changes the ratio of 4c to 1 in favor of
1 and decreases the TOF. Even though a substrate dependent
B. Hydroacylation of Isovaleraldehyde. Mechanistic
Considerations. Complex 1 is an active catalyst for the
hydroacylation of trimethylvinylsilane using a variety of simple
alkyl aldehydes. Detailed kinetic studies of the turnover process
have been carried out with isovaleraldehyde as a representative
aldehyde. These studies are described in this section, followed
in the next section by a survey of the reactivity of other alkyl
1
aldehydes. A series of H NMR experiments were performed
in which the catalytic addition of isovaleraldehyde to trimeth-
ylvinylsilane to form ketone 3c (eq 4) was followed with time.
Complex 1 was employed as catalyst (1.27 × 10-5 M in d6-
acetone), and the ratio of substrate to catalyst was varied as
(43) For complete structural data see the Supporting Information.
(44) Schmid, G.; Kilanowski, B.; Boese, R.; Blaeser, D. Chem. Ber. 1993,
126, 899-906.
(47) Kang, Y.; Kim, K.-i.; Kang, S.-O.; Ko, J.; Heaton, B. T.; Barkley,
J. V. J. Organomet. Chem. 1997, 532, 79-82.
(48) For example, in the corresponding bisstyrene Co(I) complex, [C5-
Me5Co(C2H3C6H5)2], several isomers are observed in which the substituent
on the olefin is directed either toward the other olefin substituent or toward
the Cp* ligand. Brookhart, M.; Lenges, C. P. Unpublished results.
(45) Klein, H. F.; Lull, G.; Rodenhauser, B.; Cordier, G.; Paulus, H. Z.
Naturforsch. 1988, 43b, 1256-1262.
(46) Lehmkuhl, H.; Naeser, J.; Mehler, G.; Keil, T.; Danowski, F.; Benn,
R.; Mynott, R.; Schroth, G.; Krueger, C.; Betz, P. Chem. Ber. 1991, 124,
441-452.

6968 J. Am. Chem. Soc., Vol. 120, No. 28, 1998
Lenges et al.
Table 2. Substrate Dependence of Hydroacylation of
Scheme 1. Mechanism for Intermolecular Hydroacylation of
Trimethylvinylsilane with Isovaleraldehyde Catalyzed by 1
Triphenylvinylsilane with Isovaleraldehydea
equivalents of
entry
aldehyde
olefin
ratio A:O
TOF [TO/h]
1
2
3
4
5
6
7
8
9
20
30
40
20
20
20
20
10
20
30
20
20
20
20
40
50
70
50
50
50
1
1.5
2
4.0
4.2
4.2
3.9
3.9
4.2
4.2
4.0
4.2
3.8
1
0.5
0.4
0.29
0.2
0.4
0.6
10
a Aldehyde: isovaleraldehyde. Olefin: triphenylvinylsilane. 1.27 ×
10-5 mol of 1 in 0.6 mL of d6-acetone at 35 °C, TOF over at least
50% conversion. Internal reference: ferrocene.
initial ratio of 4c to 1 is reproducibly established initially and
persists throughout ca. 80% of catalysis, a change in this ratio
is apparant at high conversions. Using excess olefin, complex
1 is regenerated as the major species after nearly all aldehyde
has been consumed and only olefin and ketone remain in
solution. If aldehyde is the substrate in excess, hydroacylation
terminates prior to complete consumption of olefin and deac-
tivation and side products are observed, which will be discussed
in more detail later.
formation of one major new species, the mixed olefin complex
6, is observed by 1H NMR spectroscopy.49 Besides complex 6
and a small amount of remaining 1, two new species with Cp*
resonances at 1.38 and 1.42 ppm are observed. When all volatile
materials in this reaction mixture are removed in vacuo,
readdition of d6-acetone allows the NMR spectroscopic char-
acterization of what we believe are two isomers of a bis-
(triphenylvinylsilane)-Co(I) complex 7. Solutions prepared in
this manner feature the same characteristics in hydroacylation
catalysis as the use of 1 in the presence of excess triphenylvi-
nylsilane.
The TOF is not dependent on absolute concentrations of
substrates. In a series of experiments in which the ratio of
aldehyde to olefin was constant at 1:1 and the amount of catalyst
was varied from 5 to 3.3 and to 2.5 mol %, essentially complete
conversion was observed in each case with a TOF of 3.0 TO/h
at 35 °C. The initial TOF observed for a series of different
substrate concentrations is roughly linearly dependent on the
concentration of 4c (see the Supporting Information for the plot).
This indicates that 4c is intimately involved in the catalytic cycle
and not a result of a nonproductive side reaction.
Reaction of d1-Isovaleraldehyde. To test reversibility of
the oxidative addition of aldehyde, d1-isovaleraldehyde, (CH3)2-
CHCH2C(O)D, was examined. A reaction with 5 mol %
catalyst 1 and a d1-aldehyde-to-olefin ratio of 1:1 was followed
by 1H NMR spectroscopy at 35 °C. The aldehyde region
showed no 1H resonance after mixing of substrate and catalyst
in d6-acetone. After 2 h and 17 min, 20% of the aldehyde was
converted to ketone, which corresponds to a TOF of 2.3 TO/h
for catalytic ketone formation. At this point in the catalysis a
9.7 ppm triplet corresponding to h1-isovaleraldehyde was
observed. About 5% of all aldehyde present after 2 h is
protioaldehyde. Analysis of the rate of appearance of protio-
aldehyde with time indicates that intermediate(s) 5c and 4c (see
Scheme 1) return to complex 1 and aldehyde ca. 1/4 of the time
Triphenylvinylsilane reacts with isovaleraldehyde in acetone
in the presence of catalytic amounts of 1 to give the â-SiPh3-
substituted ketone (eq 7).
A series of kinetic experiments was performed in which the
substrate concentration of either olefin or aldehyde was varied
and the initial turnover frequency for ketone formation was
determined. Table 2 lists these results. First, the concentration
of aldehyde was increased with constant olefin concentration
(5 mol % catalyst load with respect to olefin, entries 1-4). In
all cases initial formation of ketone is observed with similar
TOFs; however, after 50-70% conversion considerable decrease
in activity is observed and only about 70% of the olefin is
converted to ketone. Deactivation of the catalyst occurs.
Reactions in the presence of excess olefin (entries 5-10) are
well-behaved and result in complete conversion of aldehyde.
Throughout the range of substrate concentrations discussed here
3
versus partitioning to product ca. /4 of the time. Analysis of
the reaction by 13C NMR spectroscopy after complete ketone
formation indicates that deuterium is selectively incorporated
at the â-carbon of the ketone.
The mechanistic scheme presented in Scheme 1 is consistent
with the data presented thus far.
Hydroacylation of Isovaleraldehyde with Triphenylvinyl-
silane. Intermolecular hydroacylation is also observed using
triphenylvinylsilane. Several mechanistic features are consider-
ably different from the trimethylvinylsilane case since the even
bulkier triphenylvinylsilane is less strongly coordinated to the
Co(I) center. Independent synthesis of a Co(I)-bis(triphenylvi-
nylsilane) complex by sodium-amalgam reduction of the Co-
(II) precursor was not successful. However, if complex 1 is
mixed with 16 equiv of triphenylvinylsilane in d6-acetone, the
(49) A characteristic of 6 is one Cp* resonance at 1.52 and one -SiMe3
resonance at -0.21; several olefinic resonances in the region of 2.5-0.3
ppm support this assignment.

Addition of Alkyl Aldehydes to Vinylsilanes
J. Am. Chem. Soc., Vol. 120, No. 28, 1998 6969
Table 3. Hydroacylation of Trimethylvinylsilane with Alkyl
Aldehydes
there is essentially no rate dependence on either aldehyde or
olefin concentration; a constant turnover frequency of 4.0 [TO/
h] at 35 °C for catalytic hydroacylation is observed.
Investigating this catalytic process by 1H NMR spectroscopy
furnishes additional information. Addition of isovaleraldehyde
to a mixture of excess triphenylvinylsilane (e.g., 20 equiv) and
1 in d6-acetone (vide supra) results in the immediate formation
of a single new species with a Cp* resonance at 1.50 ppm. This
material accounts for all Co-containing species. In the TMS
region only one singlet is observed which corresponds to
uncoordinated trimethylvinylsilane along with the appropriate
resonances in the olefinic region. Catalytic hydroacylation is
not significant at this stage. In the concentration ranges of
substrate investigated, the formation of one species with identical
spectroscopic features was observed at all times. As in the case
of trimethylvinylsilane, the new species observed is formulated
as the Co(III)-bisalkyl-CO complex 9c (see eq 7) on the basis
of NMR and IR characterization (Experimental Section, Table
5). Following catalytic ketone formation at 35 °C by 1H NMR
spectroscopy shows that 9c is the resting state throughout
catalysis.
Reaction of d1-Isovaleraldehyde. The hydroacylation of
triphenylvinylsilane by d1-isovaleraldehyde in the presence of
catalyst 1 (5 mol %, 35 °C, d6-acetone) was examined. Again,
immediately upon mixing both substrates, the formation of the
bisalkyl Co(III) species 9c is observed exclusively. An initial
TOF of 3.75 TO/h for the ketone formation is observed. Again,
H/D exchange and formation of h1-isovaleraldehyde occur
during catalysis. An initial TOF for the appearance of protio-
aldehyde of 3.4 [TO/h] was measured. This indicates that
monodeuterated intermediate 9c partitions about equally to
product and back to unlabeled aldehyde. Consistent with these
observations, the extent of H/D exchange reaches a steady state
condition after about 3 h, whereas only about 55% of ketone
has been formed (5 mol % catalyst, 1:1 ratio of substrates).
Deuterium incorporation occurs selectively at the carbon â to
the carbonyl group.
a [1] Initial stage of catalysis in a reaction with 5 mol % catalyst
1
load and an aldehyde to olefin ratio of 1:1, by H NMR at 35 °C in
d6-acetone, 1.27 × 10-5 mol of 1. [2] Scaled up reactions in a total
volume of 5 mL of acetone and 1.27 × 10-4 mol of 1 (0.05 g) for 24
h at 45 °C or 48 h at 25 °C. [3] Two diastereomeric complexes of 4d
and 4e are observed. [4] The actual intermediate observed during
catalysis is 4b vide infra.
from these experiments and remains constant during initial
stages of catalysis (1-4 h on average). These two species
account for all Co-containing material. Deviations at later times
are due to deactivation processes. Results are listed in Table
3. Conversions to product were based on NMR analysis.
Organic products were isolated and fully characterized.
Aldehydes with one or two substituents at Câ (entries b-e)
show similar turnover frequencies and similar ratios of 4 to 1,
suggesting that rates of reductive eliminations from all these
intermediates (4b-e) must be quite similar. In the case of the
neopentyl aldehyde (entry f), with three Câ substituents, the
increased bulk at Câ results in a somewhat higher TOF, yet a
significantly smaller ratio of 4f to 1 suggests that the lower
percentage of 4f present is compensated for by an increased
rate of reductive elimination from the more crowded 4f.
Consistent with this observation is the fact that cyclohexane-
carbaldehyde (entry h), with two CR substituents, shows a
reduced ratio of 4h to 1 yet TOFs similar to those of aldehydes
in entries a-e. For the highly crowded pivaldehyde system
(entry i) no intermediate 4i could be detected and the TOF was
very small; product corresponding to only 2 TO could be
detected after 24 h. While this system was not amenable to
mechanistic investigation, it seems probable that the turnover-
limiting step in the catalytic cycle may be altered. A reasonable
proposal is that the initial oxidative addition of the aldehyde to
the Co(I) center is turnover-limiting.
C. Hydroacylation of Various Alkyl Aldehydes. Cobalt-
(I) olefin complex 1 was also used in the catalytic hydroacylation
of trimethylvinylsilane with a series of alkyl aldehydes. The
efficiency of the catalyst is good. Essentially quantitative
conversion with 1 mol % catalyst load and a ratio of olefin to
aldehyde of 1:1 is observed. Catalytic reactions are also
effective in neat trimethylvinylsilane. Catalyst loading can be
reduced to 0.5 mol % with still nearly quantitative conversion
for certain substrates (Table 3, entries a-c and e-f). Reactions
employing a series of alkyl aldehydes (Table 3) were run under
comparable conditions with increased catalyst load (5 mol %)
either at 45 °C for 12 h or at room temperature for 48 h in
acetone, benzene, or THF. Turnover frequencies in the initial
stages of catalysis were obtained in reactions using 5 mol %
catalyst load and an aldehyde-to-olefin ratio of 1:1 in d6-acetone
at 35 °C. Exclusive formation of the linear ketone with respect
to olefin addition is observed in all cases. Reactions using 5
mol % catalyst load (1.27 × 10-5 mol) and an aldehyde-to-
olefin ratio of 1:1 were more closely investigated by 1H NMR
spectroscopy at 35 °C in 0.6 mL of d6-acetone for the same
series of alkyl aldehydes. TOFs were extracted from the initially
linear TON vs time plots. In all cases investigated, an
intermediate Co complex was observed with spectroscopic
features (Cp*:SiMe3 of 1:1) analogous to those of the bisalkyl-
CO-cobalt(III) complex in the isovaleraldehyde case, complex
4c. The ratio of this intermediate 4 to 1 could be extracted
The importance of steric effects observed in this system is
also illustrated in the reactivity of different alkyl aldehydes with

6970 J. Am. Chem. Soc., Vol. 120, No. 28, 1998
Lenges et al.
triphenylvinylsilane. As discussed above, reactivity with isov-
aleraldehyde or cyclohexanecarbaldehyde is excellent and good
turnovers are observed (>200 TO). On the other hand, if
n-butyraldehyde (or propionaldehyde) is reacted in a 1:1 mixture
with triphenylvinylsilane (5 mol % catalyst) only about 7
turnovers to form ketone are observed and rapid deactivation
terminates catalysis early.
This Co(III)-CO-bisalkyl complex is extremly sensitive, and
storage of the d6-acetone solution of 4a at -78 °C results in
considerable decomposition after 24 h. Additional support for
the formation of an intermediate with the structure of 4a was
gained from its reaction with P(OMe)3 at -15 °C. Addition of
P(OMe)3 (10 equiv) to the solution of 4a with remaining 1 (5%)
in d6-acetone resulted in slow formation of ketone and the
bisphosphite complex [Cp*Co(P(OMe)3)2] (10) as the major
species (95%) and the monophosphite-olefin adduct [Cp*CoP-
(OMe)3(C2H3SiMe3)] (11) as a minor species (5%).
D. Spectroscopic Characterization of the Intermediates
Formed during Catalytic Hydroacylation. A general method
has been developed for generating intermediates 4 in high
concentration at low temperature for spectroscopic observation.
In a series of 1H NMR experiments an excess of 10-20 equiv
of the appropriate alkyl aldehyde was added to a d6-acetone
solution of complex 1 at -30 °C; no reaction is observed at
this temperature. Controlled and slow warming of the NMR
probe results in the formation of the intermediates 4 observed
during catalysis. In addition to this, 1 equiv of uncoordinated
trimethylvinylsilane is observed in solution, indicating that olefin
substitution has occurred and 1 equiv of aldehyde has been
incorporated into the new Co complex. With careful temper-
ature adjustments these intermediates 4 can essentially be
quantitatively generated without significant side reactions in
most cases (for exceptions see below). The d6-acetone solution
has changed color at this stage from the red of 1 to yellow for
4. Even at -10 °C slow hydroacylation is observed which
eventually consumes all uncoordinated olefin to generate ketone.
However, lifetimes were sufficiently long at -15 °C to allow
1
characterization of these intermediates by H and 13C NMR
Reaction of 1 at these temperatures with excess P(OMe)3
results in the formation of only the monophosphite adduct [C5-
Me5Co(P(OMe)3)(C2H3SiMe3)] (11). Displacement of the
second CH2dCHSiMe3 to give the bisphosphite adduct does
not occur on these time scales.39
spectroscopy. Since aliphatic aldehydes are used in excess and
slow ketone formation is also observed, the new, complex
1
resonances in the H NMR spectrum which correspond to this
intermediate were difficult to assign. Significant is the observa-
1
tion that the region downfield of ca. 2.2 ppm in the H NMR
spectrum contains only resonances for uncoordinated olefin and
excess aldehyde.50 After olefin consumption is observed at these
temperatures, further reactivity in the presence of excess
aldehyde results in side reactions and deactivation of the catalyst
(aldehyde dimerization to esters is observed, vide infra) which
complicate characterization of complexes 4.
In the case of propionaldehyde we were able to isolate and
characterize intermediate 4a by NMR spectroscopy. Complex
1 was treated with excess propionaldehyde at -15 °C in d6-
acetone, and after 2 h the intermediate 4a was generated (95%).
All volatile materials were removed in vacuo at 0 °C from the
NMR tube followed by readdition of d6-acetone solvent through
vacuum transfer. Characteristic NMR data for 4a show singlets
at 1.6 ppm (Cp*) and at -0.04 ppm (-SiMe3) in the ratio of
15:9 and a methyl triplet at 0.91 ppm along with multiplets at
1.22, 1.33, 0.81, 0.62, and 0.90 ppm which correspond to the
diastereotopic proton resonances of the aliphatic groups. The
13C shifts for the two carbons bound to cobalt appear as
broadened resonances at 14.2 and 15.3 ppm; the carbonyl carbon
is not observed.51,52 IR spectroscopy shows one band at 1970
If intermediate 4a were to be a simple aldehyde-olefin
adduct, the displacement of aldehyde would be facile and the
monophosphite adduct would be the expected major product.
The formation of the bisphosphite adduct further supports the
formulation of 4a as the bisalkyl-CO-Co(III) complex since
ketone formation removes all ligands at cobalt and allows direct
conversion to the bisphosphite complex.53 The minor olefin-
phosphite adduct is generated from residual 1.
To provide additional evidence for and to simplify the NMR
assignment of a Co(III) dialkyl carbonyl complex, a system with
higher symmetry was desired. Therefore, the aldehyde
Me3Si-CH2-CH2-C(O)H 2k54 was prepared and reacted in
excess (10 equiv) with 1 at -15 °C in d6-acetone. Over the
course of 2 h the formation of only one new species was
observed which featured 1H NMR resonances for Cp* and
-SiMe3 groups in a ratio of 1:2 as expected at 1.61 (Cp*, 15,
s) and -0.04 ppm (-SiMe3, 18, s). Furthermore, a simplified
1H NMR spectrum with three resonances centered at 1.33 (m,
2), 0.77 (m, 2), and 0.63 ppm (m, 4) for the eight -CH2CH2-
protons was observed, confirming the structural assignment for
4k. Convincing as well is the simple 13C spectrum with
cm-1
.
(50) Addition of trimethylvinylsilane to a reaction mixture prepared in
this fashion allows entry into a typical catalytic hydroacylation cycle and
regeneration of the equilibrium mixture of 1 and 4.
7
(51) The quadrupolar broadening effect of the 59Co nucleus (I ) /2) is
attributed in part to this effect. See: Farrar, T. C.; Becker, E. D. Pulse and
Fourier Transform NMR; Academic Press: New York, 1971; p 61.
(52) In an analogous experiment using CH3(CH2)313C(O)H, an intermedi-
ate of type 4 was generated in a low-temperature NMR experiment after
addition of excess [13C]pentanal to a d6-acetone solution of 1. NMR
characteristics of the solutions obtained are quite complex, but 13C{1H}
NMR analysis of the resulting reaction mixture showed a broad resonance
at 210 ppm (ν1/2 = 45 Hz) which corresponds to a Co-CO.
(53) The formation of a Co(III) species of type [C5Me5Co(R)(-C(O)R)-
P(OMe)3] was not observed in this reaction, vide infra.
(54) Sarkar, T. K.; Gosh, S. K.; Satapathi, T. K. Tetrahedron 1990, 46,
1885-1898.

Addition of Alkyl Aldehydes to Vinylsilanes
J. Am. Chem. Soc., Vol. 120, No. 28, 1998 6971
resonances for Cp* (8.33 and 97.98 ppm) and -SiMe3 (-1.85
ppm) and two resonances at 15.5 (broad) and 22.81 ppm for
the CH2 groups; the carbonyl resonance was not observed (νCO
) 1969 cm-1).
catalyst and a substrate ratio of ferrocenecarbaldehyde to
trimethylvinylsilane of 1:1, the formation of a new species with
a ratio of Cp* to -SiMe3 groups of 1:1 was observed together
with complex 1. However, the hydroacylation using this
substrate is considerably slower than for alkyl aldehydes;
comparable turnover frequencies are observed only at temper-
atures of 50 °C to generate ketone 3l (eq 12). Catalytic solutions
prepared in this manner contain 95% 4l with 5% of 1 remaining
throughout the reaction.
A direct synthesis of 4l was now successful. Three equiva-
lents of ferrocenecarbaldehyde were added to an acetone solution
of 1 at 0 °C. After 4 h, the reaction mixture was cooled to
-79 °C and yellow crystalline material was obtained. 1H NMR
analysis (-15 °C, d6-acetone) of this material indicates the
presence of a substituted ferrocene unit (3.96 (s, 5, C5H5); 3.51
(m, 1, C5H3H); 3.87 (m, 1, C5H3H); 4.09 (m, 2, C5H2H2)) and
a Cp* resonance (1.52 ppm, s, 15H) along with a -SiMe3
resonance (0.09 ppm, s, 9H). Several more complex multiplet
patterns are observed in the alkyl region integrating for 4
protons. The 13C NMR spectrum features five resonances for
the Co-substituted Cp ligand; significant is the resonance of
the Co-bound carbon at 98.6 ppm. With a concentrated sample
a resonance for the CO-carbon could be observed (broad signal,
ν1/2 ≈ 37 Hz, at 208 ppm). The IR spectrum shows a CO stretch
at 1983 cm-1 (compare with 4a: νCO ) 1970 cm-1). This
indicates that the ferrocenyl substituent is effective in removing
electron density from cobalt compared to the bisalkyl complexes
and explains the slower reductive elimination of 4l.
Warming the solution results in ketone formation and
decomposition products.55 This reaction was repeated under
catalytic conditions (5 mol % 1, ratio of aldehyde to trimeth-
ylvinylsilane of 1:1), and the same intermediate, 4k, together
with 1 was observed during catalytic ketone formation. Com-
plex 4k was also prepared directly and characterized by low-
temperature NMR spectroscopy.
For essentially all aldehydes listed in Table 3 the formation
of the intermediate Co(III) species 4 was observed. Even though
the ratio of 1 to 4 is quite substrate dependent during catalysis
with the turnover frequency for ketone formation similar in the
listed series of aldehydes, the low-temperature NMR techniques
described above allowed the characterization of these complexes
by in situ NMR spectroscopy under conditions where side
reactions and deactivation processes are slow. A trend in the
IR data for complexes 4 substantiates the proposed difference
in reactivity of complexes 4 in ketone formation. Complexes
formed from linear aldehyes (4a,b,k) have a similar νCO at 1970
cm-1. Complexes 4f and 4h which build up only to a small
extend during catalysis show a νCO at 1962 cm-1, indicating a
more electron rich metal center. NMR and IR data for additional
selected examples are summarized in the Experimental Section.56
The low-temperature generation of the intermediates with
aldehydes 2e and 2d of Table 3 results in the observation of
two diastereomeric Co(III) bisalkyl CO complexes for both 4e
and 4d. In the reaction of excess racemic 3-phenylbutanal (entry
d) with 1 at -15 °C, two diastereomeric Co(III) complexes,
4d, were generated in a 1.8:1 ratio which remains constant at
-15 °C. With this substrate a corresponding observation is
made during catalysis; the ratio of diastereomers remains
constant at 1.8:1 throughout catalysis.
E. Isolation, Characterization, and Reactivity of an
Intermediate Co(III) [C5Me5Co(R)(R′)(CO)] Complex. Iso-
lation of species of type 4 on a preparative scale using alkyl
aldehydes was not successful. For example, 1 was reacted in
acetone with excess propanal (10 equiv) at -20 °C and slowly
warmed to 0 °C. After removal of all volatiles a yellow oil
was isolated which showed the desired compound as the major
species by NMR analysis (90%). This compound is thermally
quite sensitive, and attempts to crystallize it or purify it by
chromatography were unsuccessful. With the goal of isolating
a Co(III) intermediate in this catalytic process ferrocenecarbal-
dehyde, 2l, was investigated. In a reaction using 10 mol %
It was possible to characterize complex 4l by X-ray crystal-
lography. The ORTEP diagram is shown in Figure 2. Selected
bond lengths and angles are summarized in Table 4. This
complex is clearly a [C5Me5Co(R)(R′)(CO)] type species with
a direct Co-ferrocene bond and provides further proof that all
of the intermediates 4 are complexes of this type. The
arrangement of ligands generates an octahedral coordination
environment at cobalt. In addition to facial coordination of the
Cp* spectator ligand, angles between the Co and each of the
two alkyl (ferrocenyl) and the CO ligands are very close to 90°.
The Co-C1, Co-C2, and Co-C24 bond distances reflect the
nature of the attached carbon in terms of hybridization. The
large angles C2-C3-Si and Co-C2-C3 reflect the steric impact
of the bulky trimethylsilyl substituent in the anti-periplanar
configuration with respect to C2-C3, supporting again the effect
of the -SiMe3 group in directing the exclusive anti-Markovni-
kov addition reaction. The inter-carbon distances in the Co-
coordinated Cp ring are quite similar; however, a considerable
impact on the Fe-C5H4 distances is observed. The C24 to Fe
(55) The stability of aldehydes with â-silyl substituents has been
discussed. See: Hoveyda, A. H.; Young, D. G. J.; Hale, M. R. Tetrahedron
Lett. 1996, 37, 827-830.
(56) In comparison to other Co(III)-CO complexes the series of
complexes of type 4 show relatively low CO bands. See for example: Dahl,
L. F.; Nagaki, D. A.; Olson, W. L. Organometallics 1986, 5, 630-634.

6972 J. Am. Chem. Soc., Vol. 120, No. 28, 1998
Lenges et al.
Scheme 2. Reaction of 4l with P(OMe)3 in d6-Acetone at
35 °C
Figure 2. ORTEP view of 4l. All atoms are drawn at 50% probability
ellipsoids.
pathway. This was also repeated with triphenylvinylsilane
which gave a rate constant of 4 × 10-5 s-1 (10 equiv) for this
process.
Table 4. Selected Bond Lengths and Bond Angles for 4l (Å and
deg)
C1-O1
Co-C1
Co-C2
1.143(5) C3-Si
1.876(5) Fe-C24
2.091(3)
1.726(4) Fe-C21 1.026(4) C21-C25 1.413(6)
2.034(3) Fe-C23 2.036(3) C24-C25 1.430(5)
Co-C24 1.962(3) Fe-C25 2.043(3) C24-C23 1.418(5)
C2-C3
1.516(5) Fe-C22 2.029(4) C21-C22 1.407(6)
C11-Fe-C23
C12-Fe-C24
C13-Fe-C25
C14-Fe-C21
C15-Fe-C22
C1-Co-C2
108.19(17)
109.85(15)
109.44(16)
106.20(17)
105.17(17)
90.35(16)
90.38(16)
178.1(3)
C22-C21-C25
C21-C22-C23
C22-C23-C24
C23-C24-C25
C21-C25-C24
Co-C2-C3
108.1(3)
107.2(3)
109.8(3)
105.5(3)
109.4(3)
116.34(24)
113.2(3)
125.8(3)
89.39(14)
C1-Co-C24
Co-C1-O
Si-C3-C2
Co-C24-C23
C2-Co-C24
To further characterize the reductive elimination step from
4l, a series of reactions with excess P(OMe)3 was performed
1
and followed by H NMR spectroscopy in d6-acetone solution
distance is elongated with respect to the four remaining carbons
in the top ring, with the shortest distances being C22-Fe and
C21-Fe. Essentially Fe has moved from the centroid of the
top ring away from C24.43
(Scheme 2). The formation of the ferrocene-coupled ketone is
observed along with the clean formation of the known bispho-
sphite complex [Cp*Co(P(OMe)3)2] 10.58 Pseudo-first-order
rate constants were determined for a series of reactions with
different P(OMe)3 concentrations. A linear increase in reductive
elimination rate was observed with increasing phosphite con-
centration.59 A general rate law of the type rate ) (ko + k1-
[P(OMe)3]) [4l] is observed which indicates that two pathways
are viable for reductive elimination of ketone, a ligand
independent pathway and a ligand accelerated pathway in which
phosphite assists ketone formation. A plot of the data appears
in the Supporting Information. The value of ko determined from
the intercept matches the rate observed in the reaction of 4l
with excess trimethylvinylsilane (ko ) 0.3 × 10-4 s-1), and a
second-order rate constant of k1 ) 3.2 × 10-4 s-1 (L/mol). At
0.4 M phosphite the rate of ligand-assisted ketone formation is
ca. 6.5 times the rate of formation via the unassisted pathway.
The presumed 18-electron Co(III) phosphite acyl alkyl complex
(assigned structure 12, vide infra) however has not been
observed under these conditions.60
The availability of compound 4l now allows direct study of
the C-C bond-forming step in catalytic hydroacylation. Inser-
tion of CO into either Co-C bond generates an alkyl acyl
intermediate which can undergo reductive elimination to form
the corresponding ketone. In a series of kinetic experiments,
compound 4l was treated at 35 °C in d6-acetone with excess
trimethylvinylsilane. The bisolefin complex 1 was formed
together with the corresponding ferrocene-coupled ketone, 3l.
The pseudo-first-order rate constant for this process was
determined by monitoring the disappearing Cp* resonance of
4l. Essentially no significant variation in rate was observed
for different olefin concentrations (kobs ) 2 × 10-5 s-1, 18
equiv; 2 × 10-5 s-1, 25 equiv; 4 × 10-5 s-1, 30 equiv; 3 ×
10-5 s-1, 50 equiv). This indicates that reductive elimination
is not assisted by olefin under these conditions and occurs from
a 16-electron, possibly solvated intermediate or, perhaps more
likely, from an 18-electron η2-acyl intermediate. In an ad-
ditional reaction, 4l was treated with 13 equiv of triethylvinyl-
silane in d6-acetone and a pseudo-first-order rate (kobs ) 4 ×
10-5 s-1) was observed for ketone formation.57 The observed
pseudo-first-order rate constant is similar to that of the series
with trimethylvinylsilane, supporting the substrate independent
(58) Werner, H.; Heiser, B.; Klingert, B.; Doefel, R. J. Organomet. Chem.
1982, 240, 179-190.
(59) The following pseudo-first-order rate constants were observed for
the corresponding P(OMe)3 concentration: 1.03 × 10-4 s-1, 11 equiv; 1.18
× 10-4 s-1, 16 equiv; 1.42 × 10-4 s-1, 18 equiv; 1.52 × 10-4 s-1, 24
equiv; 2.04 × 10-4 s-1, 32 equiv; 2.91 × 10-4 s-1, 44 equiv; 3.52 × 10-4
s-1, 56 equiv.
(60) Kinetics are consistent with either rapid preequilibrium between 4l
and 12 favoring 4l followed by reductive elimination or rate-determining
trapping of 5l followed by rapid reductive elimination.
(57) In the reation of 4l with triethylvinylsilane the formation of the
bisolefin complex [C5Me5Co(C2H3SiEt3)2] is observed.

Addition of Alkyl Aldehydes to Vinylsilanes
J. Am. Chem. Soc., Vol. 120, No. 28, 1998 6973
We have investigated this reaction in more detail at lower
temperatures using PMe3, a more electron rich ligand which is
expected to more effectively stabilize an intermediate of type
12. Compound 4l in d6-acetone was treated with 20 equiv of
PMe3 at 0 °C, and the slow formation of a new Co complex
was observed. At 10 °C compound 13 was generated during
the course of 2 h. Compound 13 exhibits a 1:1:1 ratio for Cp*
(1.54 ppm, d, 0.9 Hz) to -SiMe3 (0.09 ppm, s) to PMe3 (0.90
ppm, d, 9 Hz) groups. A series of complex resonances for the
aliphatic hydrogens is observed which are shifted from the
resonances for 4l. Particularly informative is a multiplet at 3.21
ppm (ddd, 15.9, 11.9, 3.7 Hz, 1 proton), which is nearly 1 ppm
downfield from a corresponding resonance of 4l. In the
ferrocene region five signals are observed for the substituted
Cp unit.61 Significant in the 13C spectrum is the Co-acyl
resonance at 270 ppm (d, 22.6 Hz) and a resonance at 49.9 ppm
(d, 7 Hz) for the R-carbon of the Co-acyl, supporting the
proposed regiochemistry of CO insertion. Even though com-
pound 13 can be generated to nearly 90% in the reaction mixture
at 10 °C, conversion to the bisphosphine complex [C5Me5Co-
(PMe3)2]58 is eventually observed along with the formation of
ketone 3l.62
Scheme 3. Formation of Acetone through Carbonylation of
a Cobalt Alkyl Complex
Scheme 4. Decomposition of 4l
These trapping reactions show that the reductive elimination is
more facile from an 18-electron system like 12 or 13 as opposed
to a 16-electron Co(III)-alkyl-acyl species in which η2-acyl
coordination may assist reductive elimination as in 5l.63
not excluded.67,68 Such assistance seems likely in view of the
above results. A significant feature of these reductive elimina-
tions is that the eliminated ketone is capable of stabilizing an
unsaturated species such as [C5Me5Co] or [C5Me5Co(L)]. The
reductive elimination of ketone to generate a metal-ketone
complex has been directly observed by Suggs et al. in a Rh
system.69
Compound 4l is stable at 25 °C for short times in solution
but decomposes over the course of several hours. For example,
after 24 h at 22 °C in d6-acetone the olefin CO complex 14 and
the carbonyl-bridged dimer 15 are formed in a ratio of about
1:1 in addition to small amounts of unassigned organometallic
products. Ferrocene, free trimethylvinylsilane, and a small
amount of ketone 3l (2%) are also present. The proposed
decomposition pathway is shown in Scheme 4. The presence
of excess trimethylvinylsilane in the catalytic system (and in
the thermolyses described in eq 15) retards formation of 14 and
15, thereby preventing catalyst deactivation.
Bergman et al. have discussed a closely related reductive
elimination reaction in which acetone was generated from a
cobalt(III) complex. [C5H5Co(PPh3)Me2] was treated with CO
(1-5 atm), and besides the formation of acetone, the cobalt
complexes [C5H5Co(CO)2] and [C5H5Co(PPh3)(CO)] were
observed. The expected intermediacy of [C5H5Co(CO)Me2],
4m, a Co(III) complex analogous to the complexes of type 4,
was confirmed by in situ NMR spectroscopy (1 atm of CO,
THF, 4m builds up to 8%, IR 1998 cm-1); see Scheme 3.64-66
The authors favored a transition state for reductive elimination
which resembles a 16-electron species, possibly assisted by an
η2-acyl interaction; however, the assistance of excess CO was
(61) NMR data for 13 (d6-acetone, -15 °C, 300 MHz): 1H NMR δ
1.54 (d, 0.9 Hz, 15H, C5Me5), 0.90 (d, 9.0 Hz, 9H, -PMe3), 0.09 (s, 9H,
-SiMe3), 0.77-0.94 (m, 2H, -CH2-), 1.82 (m, 1H, -CH2-), 3.21 (ddd,
15.9, 11.9, 3.7 Hz, 1H, -CH2-), 3.33, 4.03, 4.12, 4.19 (m, 1H, -C5H4),
3.99 (s, 5H, C5H5); 13C NMR δ 270.8 (d, 22.6 Hz, Co-C(O)-), 49.9 (6.8
Hz, -C(O)-CH2-), 12.7 (-CH2SiMe3), -1.9 (-SiMe3), 66.1, 66.5, 78.8
(-CC5H4), 79.9 (d, 6 Hz, -CC4H5), 67.9 (C5H5), 15.2 (d, 30 Hz, PMe3),
10.2 (C5Me5), 95.7 (d, 3 Hz, (C5Me5).
F. Evidence for the Formation of Isomeric Ketones and
Aldehydes during Hydroacylation. Catalytic Hydroacylation
Using Butyraldehydes. An additional feature in intermolecular
hydroacylation catalyzed by Co(I) complexes is illustrated in
the reactivity of the isomeric butyraldehydes (n-butyraldehyde
(62) Repeated attempts to isolate 13 were not successful; decomposition
to ketone 3l and unassigned organometallic products are observed.
(63) Bunel, E. E.; Sheridan, R. E.; Roe, D. C. J. Am. Chem. Soc. 1994,
116, 1163-1164.
(67) In a subsequent study Bergman et al. investigated a similar Co-
(III)-metallacyclopentanone system. Even at high temperatures the forma-
tion of cyclobutanone was not observed. See: (a) Bergman, R. G.; Theopold,
K. H. Organometallics 1982, 1, 1571-1579. (b) Bergman, R. G.; Theopold,
K. H.; Becker, P. N. J. Am. Chem. Soc. 1982, 104, 5250-5252.
(68) Maitlis, P. M.; Saez, I. M., Fanizzi, F. P.; Sunley, G. J. J. Organomet.
Chem. 1987, 330, C27-C30.
(64) Bergman, R. G.; Evitt, E. R. J. Am. Chem. Soc. 1980, 102, 7003-
7011.
(65) Bergman, R. G. Acc. Chem. Res. 1980, 13, 113-120.
(66) Bergman, R. G.; Bryndza, H. E.; Evitt, E. R. J. Am. Chem. Soc.
1980, 102, 4948-4951.
(69) Suggs, J. W.; Wovkulich, M. J.; Cox, S. D. Organometallics 1985,
4, 1101-1107.

6974 J. Am. Chem. Soc., Vol. 120, No. 28, 1998
Lenges et al.
that of the n-butyraldehyde. In the reaction of isobutyraldehyde
with olefin only trace amounts of n-butyraldehyde are generated
during catalysis; the isoaldehyde is converted to the mixture of
isomeric ketones without considerable formation of n-butyral-
dehyde. The formation of small amounts of propene in the
reactions of the isomeric butyraldehydes is also observed with
ongoing catalysis; a trace amount of a third aldehyde is observed
which we attribute to 2k, Me3SiCH2CH2C(O)H. When either
butyraldehyde is treated with 1 in a low-temperature NMR
experiment, the same intermediate is formed in each case which
we assign to 4b.70 Characteristic for 4b is a Cp*resonance at
1.62 ppm and a -SiMe3 resonance at -0.05 ppm in a ratio of
15:9 along with a νCO at 1970 cm-1
.
To support this assignment, we compare 4b with 4h, which
serves as a model for a Co(III) complex bound to a secondary
alkyl group as in 4g, [C5Me5Co(CO)(C2H4SiMe3)(CHMe2)].
Complex 4h, the Co-bisalkyl complex, which originates from
the reaction of cyclohexanecarbaldehyde with 1, has a Cp*
resonance at 1.52 ppm and a νCO at 1962 cm-1. Additional
evidence was gained in an in situ IR experiment. Addition of
10 equiv of n-butyraldehyde to a toluene solution of 1 generated
a yellow solution with a strong IR band at 1970 cm-1; repeating
this experiment with 10 equiv of isobutyraldehyde generated
an essentially identical IR spectrum (νCO ) 1970 cm-1).
Addition of 10 equiv cyclohexanecarbaldehyde, 2h, to a toluene
Figure 3. Conversion vs time plot for the hydroacylation of n-
butyraldehyde.
and isobutyraldehyde, entries b and g, Table 3) with trimeth-
ylvinylsilane in the presence of 1. In a typical reaction, a 1:1
mixture of olefin and aldehyde is treated at 35 °C in acetone
with 5 mol % catalyst. Analysis of the reaction mixture after
10 h indicates that all aldehyde has been consumed (in either
case this is observed) and that two isomeric ketones 3b and 3g
are formed as sole organic products (eq 16).
solution of 1 resulted in a species with νCO at 1962 cm-1
.
As an additional experiment designed to probe isomerization,
we have examined hydroacylation of 2,2-dideuteriopropional-
2
dehyde. Following the catalysis in h6-acetone by H NMR
spectroscopy shows clearly that deuterium is incorporated into
both Câ and CR sites of the ketone, which illustrates isomer-
ization also occurs in the case of propionaldehyde (eq 17).71
For example, reaction of n-butyraldehyde with trimethylvi-
nylsilane and 5 mol % 1 in d6-acetone at 35 °C results in 99%
conversion to ketone, of which 67% is linear (n, 3b) and 32%
is branched (iso, 3g). However, in a reaction of n-butyraldehyde
with olefin at 25 °C after 2 days, 35% of aldehyde is converted
to ketones, of which 58% corresponds to the branched ketone
3g. This indicates that at lower temperatures isomerization is
more extensive. In a reaction of isobutyraldehyde under the
same conditions, 43% of the linear ketone (3b) is observed
together with 56% of the branched product (3g). In a reaction
of a 1:1 mixture of both aldehydes (5 mol % catalyst), complete
conversion is also observed to a mixture of ketones (iso, 48%;
n, 51%). A decrease in catalyst load to 3 or 1 mol % gives a
similar result: complete conversion and 43% isopropyl vs 56%
On the basis of these observations and our kinetic investiga-
tions, we can propose an extended mechanistic scheme which
accounts for the isomerized products and interconversion of
isomeric aldehydes. This is illustrated in Scheme 5 for the
butyraldehydes.
The consequence of the complete reversibility of all elemen-
tary steps in this process with the exception of the C-C bond-
forming step is demonstrated in the formation of the mixture
of isomeric ketones in the reaction of butyraldehyde. A Co-
(III) acyl olefin hydride species is proposed which is responsible
for isomerization through insertion of propene with different
regiochemistry after rotation.47 This scheme also suggests that
reductive elimination can occur from any of the four alkyl-
1
n-propyl ketone. Following this latter reaction by H NMR
spectroscopy in d6-acetone at 35 °C shows that n-butyraldehyde
is converted more rapidly to product than isobutyraldehyde (3.2
TO/h vs 2.4 TO/h). The catalytic conversion of n-butyraldehyde
was followed by 1H NMR spectroscopy and the individual TOFs
for ketone formation extracted. Remarkably, during catalysis
isobutyraldehyde is formed as evidenced by the appearance of
a doublet at 9.62 ppm. This indicates that not only do
isomerized products form but the isomerization process actually
converts n-butyraldehyde to isobutyraldehyde during catalysis.
The time course of the reaction is shown in Figure 3. As is
evident, after conversion of ca. 16 equiv (out of 20), the ratio
of isobutyraldehyde to n-butyraldehyde is about 1:1, and at
longer times the concentration of the isobutyraldehyde exceeds
(70) In a reaction of 1 with 30 equiv of isobutyraldehyde in d6-acetone
at -15 °C the formation of one major intermediate assigned as 4b was
observed (Cp*: 1.62 ppm); at low temperature a second species builds up
to about 5% with a Cp* resonance at 1.52 ppm and a -SiMe3 resonance at
-0.02 ppm (ratio 15:9). This intermediate is depleted quickly after warming
to 0 °C, and only 4b in addition to 1 remains; this species is tentatively
assigned as 4g.
(71) n-Hexanal was also reacted with trimethylvinylsilane and catalyst
1. A mixture of two hydroacylation products is observed, the linear addition
product (n-C5H11C(O)C2H4SiMe3) and the methyl-branched addition product
(C3H7CH(Me)C(O)C2H4SiMe3); other products are not observed. During
catalysis the formation of isomeric hexanals is observed as well.

Addition of Alkyl Aldehydes to Vinylsilanes
J. Am. Chem. Soc., Vol. 120, No. 28, 1998 6975
Scheme 5. Catalytic Hydroacylation of Butyraldehydes.
Formation of Isomeric Ketones 3b and 3g
(B) which is responsible for product formation (E and F).
Isomerization to a rhodacyclopentanone complex (C) takes place
as indicated by the labeling experiment, but reductive elimina-
tion to generate a cyclobutanone is not observed. On the basis
of the results of additional labeling studies, a rhodacyclopen-
tane-CO complex (D) was proposed but not detected.
F. Deactivation Processes. Catalyst deactivation and side
reactions were investigated in more detail. In the presence of
excess aldehyde, hydroacylation occurs and, depending on the
structure of the aldehyde, olefin is converted to ketone in yields
of 80-100%. The color of the reaction mixture changes during
this process from orange-yellow to green. Analysis of the
reaction mixture by NMR and IR spectroscopy reveals the
formation of two CO-containing Co species which were
mentioned above and have been described as products of
decarbonylation reactions for aromatic aldehydes,35 the olefin
carbonyl complex [C5Me5Co(CO)(C2H3SiMe3)] 14 and the
carbonyl-bridged dimer [C5Me5Co(CO)]2 15. Both species are
observed as final products during the decomposition of complex
4l at room temperature as well. The ratio of 14 to 15 depends
on substrate ratios used and on the nature of the aldehyde. This
is illustrated in the case of 3-phenylbutanal 3d (entry d, Table
3). Besides the cobalt-carbonyl complexes 14 (90%) and 15
(10%), the expected stoichiometric amount of isopropylbenzene
is also observed after catalysis has stopped. These results are
confirmed in a reaction of 1 with 15 equiv of 3-phenylbutanal
at -15 °C. As described above the formation of the diastere-
omeric mixture of complexes 4d is observed. If this sample is
warmed to 25 °C, the ketone 3d is formed along with the
conversion of all organometallic material into complexes 14 and
15 now in a ratio of 1:1.
Scheme 6. Bosnich Mechanism for Intramolecular
Hydroacylation of 4-Pentenals Catalyzed by Cationic Rh(I)
Complexes28-33
In the case of other aldehydes summarized in Table 3,
deactivation reactions are more complex. Complexes 14 and
15 are observed; however, 1H NMR integration shows that the
material accounted for does not correspond to the amount of
catalyst 1 introduced. Additional decomposition products were
observed but could not be assigned.
Tishenko-Type Dimerization of Aldehyde. In the presence
of excess aldehyde, an additional catalytic process is observed:
dimerization of two molecules of aliphatic aldehyde to generate
the corresponding alkyl ester. This type of intermolecular
aldehyde disproportionation has a classic analogue in the base-
catalyzed Tishenko reaction. Transition metal catalyzed ver-
sions of this reaction have also appeared.72-76 For example, in
a reaction of isovaleraldehyde (40 equiv) with trimethylvinyl-
silane (20 equiv) with 1 the formation of ketone 3c is observed
acyl Co(III) intermediates since all isomers must be populated
during catalysis (Scheme 5).
Bosnich28-33 et al. have investigated the intramolecular
hydroacylation of 4-pentenals using labeling studies. Their
proposed mechanism is summarized in Scheme 6. A cationic
Rh(I) bisphosphine system is applied as a catalyst; no discrete
intermediates on the catalytic cycle build up in sufficient
concentrations to be detected. Deuterium-labeling studies
established the isomerization pathways also found in the Co
system investigated in this study. Scrambling the deuterium
label from the C1-position to generate aldehyde G confirms that
the aldehyde activation/olefin insertion sequence is faster than
reductive elimination to generate ketone. The proposed domi-
nant acyl alkyl Rh complex is a rhodacyclohexanone species
(72) March, J. AdVanced Organic Chemistry, 3rd ed.; John Wiley &
Sons: New York, 1985; p 1346.
(73) SanFilippo, J.; Villacorta, G. M. J. Org. Chem. 1983, 48, 1151-
1155.
(74) Takegami, Y.; Watanabe, Y.; Mitsudo, T.; Yamashita, M. Bull.
Chem. Soc. Jpn. 1976, 49, 3597-3600.
(75) Yamamoto, A.; Koshiro, Y.; Horino, H.; Ito, T. Bull. Chem. Soc.
Jpn. 1982, 55, 504-512.
(76) This reaction is observed with substrates without acidic R-protons
to avoid competitive enolate-type transformations. The Tishenko analogue
reaction reported here is curiously exclusive for alkyl aldehydes. Aromatic
aldehydes show regular decarbonylation chemistry to 14 and 15 in the
presence of excess aldehyde.

6976 J. Am. Chem. Soc., Vol. 120, No. 28, 1998
Lenges et al.
exclusively in the initial phase of catalysis. After 70%
conversion to ketone, Tishenko-type ester begins to appear.
During this process the resting states 1 and 4c greatly diminish
in intensity; uncharacterized Co species are formed. After
complete conversion of aldehyde, all olefin had been converted
to ketone and excess aldehyde was completely accounted for
by ester formation.
Scheme 7. Resting States during Catalytic Hydroacylation
Using 1 as Catalyst Precursor
This reaction was also observed in the low-temperature NMR
experiment described above used to generate intermediate 4c.
Even at low temperatures the formation of ketone was followed
by the formation of ester. Careful adjustment of temperature
was necessary to generate 4c without ester and ketone formation.
Depending on the aldehyde structure it was possible to cleanly
generate intermediates of type 4 without contamination by ester
(Table 3, entries a, b, d, f, and k). In other cases this was not
possible (Table 3, entries c and h). For example, in the case of
cyclohexanecarbaldehyde (entry h), intermediate 4h was gener-
ated at -12 °C in d6-acetone from 1 and 20 equiv of aldehyde.
After 2 h complex 4h was the major organometallic species
observed (80%) with 1 as the minor species (20%); no other
cobalt-containing complexes were observed at this stage (based
on C5Me5 and SiMe3 resonances). Besides small amounts of
free trimethylvinylsilane and ketone 4h, 60% of the cyclohex-
anecarbaldehyde had been converted to the ester. Isobutyral-
dehyde behaves in a similar manner. In the presence of excess
olefin clean hydroacylation to ketones 3g and 3b is observed.
On the other hand, in a reaction of 30 equiv of isobutyraldehyde
2g with 1 at 0 °C in d6-acetone, stoichiometric formation of 3g
was observed while formation of ester consumed all aldehyde.77
The presence of esters added to the difficulties of characterizing
intermediates 4 by NMR spectroscopy in these reaction mix-
tures.
G. Summary. The intermolecular hydroacylation of vinyl-
silanes with a variety of alkyl aldehydes using precatalyst 1
exhibits good turnover numbers under quite mild conditions.
In a kinetic and spectroscopic study it was discovered that a
sensitive interplay of binding affinity of olefin and reactivity
of aldehyde determines the nature of the catalyst resting state(s)
throughout catalysis. In our previous investigation of the
hydroacylation of aromatic aldehydes we found that 1 was the
sole resting state in catalysis;35 in the reaction of alkyl aldehydes
with complex 1, an equilibrium between 1 and 4, a Co(III)
complex of the type [C5Me5Co(CO)(C2H4SiR3)(R′)], is estab-
lished during catalysis. Using the more bulky olefin, vinyl-
triphenylsilane, the bisolefin Co(I) species becomes unfavored
and the Co(III)-bisalkyl-CO complex is observed as the sole
resting state (Scheme 7).
been pointed out that acyl alkyl reductive elimination is
especially facile,28,64,69,78a,b which might suggest that the hy-
droacylation of olefins is a specific example for a successful
combination of C-H bond activation and C-C bond formation;
the same catalyst system (complex 1) will reversibly activate
aromatic C-H bonds (eq 1), but olefin-arene coupling is not
observed (eq 1).78c The C-H bond activation reaction, especially
of aldehydes, has been well documented using Rh79-82 or Ir83
complexes. However, cobalt has been widely excluded from
bond activation studies,39,84,85 and the inability of the first-row
3
cobalt system [C5H5Co(CO)] to participate in such Csp -H
activation reactions has been discussed recently.86,87 The system
investigated here suggests that the oxidative addition of an
aldehydic C-H bond to [C5Me5Co(L)] is a facile process with
a barrier well below the reductive elimination barriers discussed
as the turnover-limiting step in hydroacylation catalysis.
Experimental Section
General Considerations. All manipulations of air- and/or water-
sensitive compounds were performed using standard high-vacuum or
Schlenk techniques. Argon and nitrogen were purified by passage
through columns of BASF R3-11 catalyst (Chemalog) and 4 Å
molecular sieves. Solid organometallic compounds were transferred
in an argon-filled Vacuum Atmospheres drybox and, unless stated
otherwise, were stored in the drybox at -20 °C. All coupling constants
are reported in hertz. 1H and 13C NMR chemical shifts were referenced
to residual 1H NMR signals and to the 13C NMR signals of the
deuterated solvents, respectively. NMR probe temperatures were
measured using an anhydrous methanol sample. Elemental analyses
were performed by Atlantic Microlabs, Inc., of Norcross, GA.
Materials. Toluene, pentane, tetrahydrofuran, and diethyl ether were
distilled under nitrogen from sodium-benzophenone ketyl. d6-Benzene
was dried over potassium, vacuum-transferred, degassed by repeated
freeze-pump-thaw cycles and stored in the drybox. d6-Acetone was
dried over CaH2, vacuum-transferred, degassed by repeated freeze-
pump-thaw cycles, and stored in the drybox. [C5Me5CoCl]2 was
(78) (a) Puddephatt, R. J.; Brown, M. P.; Upton, C. E. E.; Lavington, S.
W. J. Chem. Soc., Dalton Trans. 1974, 1613-1618. (b) See also: Graham,
W. A. G.; Gosh, C. K. J. Am. Chem. Soc. 1989, 111, 375-376. (c) Bergman,
R. G.; McGhee, W. D. J. Am. Chem. Soc. 1988, 110, 4246-4262.
(79) Suggs, W. J. J. Am. Chem. Soc. 1978, 100, 640-641.
(80) Milstein, D. Acc. Chem. Res. 1984, 17, 221-226.
(81) Bergman, R. G.; Luecke, H. F. J. Am. Chem. Soc. 1997, 119, 11538.
(82) Milstein, D. J. Chem. Soc., Chem. Commun. 1982, 1357-1358.
(83) Rauchfuss, T. B. J. Am. Chem. Soc. 1979, 101, 1045-1047.
(84) Wadepohl, H.; Borchert, T.; Buechner, H.; Pritzkow, H. Chem. Ber.
1993, 126, 1615-1618.
(85) Wadepohl, H.; Borchert, T.; Pritzkow, H. J. Chem. Soc., Chem.
Commun. 1995, 1447-1448.
(86) Siegbahn, P. E. M. J. Am. Chem. Soc. 1996, 118, 1487-1496.
(87) Bergman, R. G.; Moore, B. C.; Bengali, A. A. J. Am. Chem. Soc.
1995, 117, 3879-3880.
Further, it was established that the reductive elimination to
generate the C-C bond is the turnover-limiting step in this
process; as a consequence of this, isomerization processes can
occur prior to reductive elimination which not only generate a
mixture of isomeric ketones but also isomerize aldehydes during
catalysis, e.g., n-butyraldehyde into isobutyraldehyde. It has
(77) Turnover numbers on the order of 50-100 are observed for
aldehydes such as isobutyraldehyde or cyclohexanecarbaldehyde at 10 °C
with turnover frequencies of ∼4 TO/h under these conditions. 4 persists to
90% conversion of aldehyde to ester; then deactivation is observed, and
besides decarbonylation product 15 further unassigned organometallic
complexes were observed.

Addition of Alkyl Aldehydes to Vinylsilanes
J. Am. Chem. Soc., Vol. 120, No. 28, 1998 6977
Table 5. NMR and IR Spectroscopic Data for Compounds 4a-l and 9ca
1H NMR, δ [ppm]
13C NMR, δ [ppm]
-R
IR [cm-1
]
compound C5Me5 -SiMe3
-R
C5Me5 -SiMe3
(C7H8, 20 °)
4a
1.60
-0.09 0.91 (t, 3, J ) 7.5 Hz, -CH3),
8.8, 97.8 -1.80 14.2,, 15.3, 23.2 (-CH2-),
19.3 (-CH3, q, 113 Hz)
1970.3
1.22, 1.34, 0.81, 0.90 (m, 1, -CHH),
0.62 (m, 2, -CH2-)
4b
4c
1.62
1.61
-0.05 na
8.8, 97.1 -1.78 14.6, 15.4, 23.2, 25.4, 29.0
8.8, 97.5 -1.95 14.1, 17.3, 25.7 (-CH2-),
34.5 (-CHMe2), 27.6,
1970.0
1968.2
-0.05 0.52, 0.65, 0.91, 1.28-1.71,
2.31 (m, 13H)
33.7 (-CHMe2)
4k
4f
1.61
1.61
-0.04 1.33, 0.77 m, 2H)
8.3, 98.0 -1.85 15.5 (Co-CH2), 22.81 (-CH2-Si)
1969.1
1961.8
(18H)
0.63 (m, 4H)
-0.04 0.44, 1.43 (d, 1, 10 Hz), 0.68,
8.8, 97.8 -1.55 15.4, 24.0, 33.1, 34.3, 40.9
1.25 (m, 1H), 0.9-1.1 (m, 2H),
0.92 (s, 9H)
4h
4d
1.51
1.62
1.58
1.52
-0.01 na
8.2, 97.5 -1.72 na
1962.1
1968.1
-0.03 na
8.5, 97.2 -1.70 na
-0.03
8.7, 97.4 -1.70
4l
0.09 1.50, 1.05, 1.82, 2.46 (m, 1, -CHH-),
8.9, 98.1 -1.83 20.0, 26.2, (Co-CH2), 67.7,
67.8, 73.6, 77.4, 98.6, (C5H4-),
68.7, (C5H5)
1983.3
1966.0
3.51, 3.87, 4.09, 4.09 (m, 1, C5H4-),
3.96 (s, 5, C5H5)
na
9c
1.57
8.8, 97.8
15.4, 18.8 (Co-CH2), 26.0 (-CH2-Si),
34.2 (-CHMe2), 27.5, 33.8 (-CHMe2);
aromatics na
a na ) not assigned.
Table 6. Crystallographic Data and Refinement Details for 1 and
4l
prepared according to the procedure of Koelle.41 Alkyl aldehydes were
distilled twice prior to use and vacuum-transferred, degassed by repeated
freeze-pump-thaw cycles, and stored in the drybox. Trimethylvi-
nylsilane (Aldrich) contained traces of tetrahydrofuran (∼5%) which
were removed by stirring over ZrCl4 followed by vacuum transfer.
Triphenylvinylsilane (Celeste) was used as received without additional
purification. Ferrocenecarbaldehyde (Aldrich) was used after recrys-
tallization from acetone. The alkyl aldehydes used require utmost
purity. If the corresponding carboxylic acid is added even in stoichio-
metric amounts, catalytic activity is extensively reduced and precipita-
tion is observed. The aldehyde Me3SiCH2CH2C(O)H was prepared
following a literature procedure54 and stored at -30 °C, slow
decomposition to unassigned products is observed.88 d1-Isovaleraldehyde
was prepared from d2-isoamyl alcohol by perruthenate (TPAP) oxidation
(99%).89,90 d2-Propanal was prepared from a pyridine-catalyzed reaction
of propanal in d2-water.
[C5Me5Co(C2H3SiMe3)2] 1. The synthesis of the parent ethylene
complex [C5Me5Co(C2H4)2] has been reported in the literature. The
trimethylvinylsilane complex 1 described here was prepared in a
modification of that synthesis. [Cp*CoCl]2 (2 g, 4.4 × 10-3 mol) was
dissolved in 40 mL of tetrahydrofuran. Trimethylvinylsilane (4.4 g,
0.044 mol) was added via syringe, and the mixture was cannula
transferred to a previously prepared amalgam mixture covered by 30
mL of tetrahydrofuran (1.0 g of Na, 100 g, 7.4 mL of Hg, 1 mol %).
The amalgam reaction mixture was stirred for 1 h at room temperature.
A color change from dark brown to dark red was observed. After
filtration by cannula from the amalgam, the solvent was removed in
vacuo, leaving a red solid, which was dissolved in pentane, filtered,
and placed at -78 °C. Red crystals were isolated (1.3 g, 3.3 × 10-3
mol, 75%), and the compound was stored at -30 °C under an argon
atmosphere. Crystals for X-ray analysis were grown from an acetone
solution of 1 at -79 °C. NMR spectroscopic data for 1: 1H (400 MHz,
d6-acetone, 20 °C) δ 1.54 (s, 15, C5Me5), 0.09 (s, 18, SiMe3), 0.28 (d,
2, J ) 14.6 Hz, CH2 syn), 1.82 (d, 2, J ) 11.6 Hz, CH2 anti), 0.87 (dd,
2, J ) 14.6 Hz, J ) 11.6 Hz, CHSi syn); 13C{1H} (75 MHz, C6D6, 20
°C) δ 2.43 (SiMe3), 9.67 (C5Me5), 93.5 (C5Me5), 43.7 (CHSi), 47.4
(CH2).
1
4l
molecular formula
FW
CoSi2C20H39
394.63
CoFeSiOC26H37
508.44
crystal dimensions (mm) 0.25 × 0.25 × 0.04 0.40 × 0.20 × 0.20
no. of reflections for cell 6261
determination
8192
a (A)
b (A)
c (A)
17.7949(10)
10.2972(5)
12.9968(7)
19.050(10)
91.433(1)
2548.74(23)
P21/n
4
0.86
1075.62
-100
8970
14.0181(7)
18.9075(10)
106.128(1)
4530.9(4)
P21/c
8
0.86
â (deg)
V (A3)
space group
Z
µ (Mo, KR, mm-1
F(000)
1716.77
-100
T(°C)
no. of reflections
21778
no. of unique reflections 7900
3610
3019
no. of observations
6261
(I > 2.5σ(I))
T
min, Tmax
0.761, 0.888
0.048, 0.043
0.070
0.695, 1.000
0.039, 0.048
0.050
R, Rw
R (all data)
GOF
2.13
1.65
sealed with a septum and removed from the drybox. After cooling to
0 °C, the corresponding amount of aldehyde was added by syringe.
Shaking resulted in a homogeneous reaction mixture which was
transferred to the temperature equilibrated probe of a Bruker AMX
1
300. Ketone formation was followed by H NMR spectroscopy over
time using an implemented kinetic routine. As internal standard during
catalysis, either added ferrocene or the residual amount of tetrahydro-
furan (2%)35 present in the commercial vinyltrimethylsilane was used.
Turnover frequencies are extracted from the slope of plots of turnover
number (based on integration) over time for initial, linear segments
(linear regression). At these low concentrations, the cobalt species are
sensitive to impurities and accurate integration of resonances for 1 and
4c in the presence of large excesses of substrates is difficult. Thus,
substantial error limits apply to the TOFs in Table 1.
Turnover Frequency Determinations. A 5 mm NMR tube was
charged with 0.005 g (1.27 × 10-5mol) of 1. d6-Acetone (0.6 mL)
was placed on top, and the appropriate amount of vinyltrimethylsilane
was added via syringe by weight in the drybox. The NMR tube was
Reaction of 1 with Excess Aldehyde. To a solution of 1 in acetone
at -5 °C (e.g., 0.1 g (2.5 × 10-4 mol) in 15 mL) was added by syringe
the appropriate alkyl aldehyde (15-20 equiv). This mixture was stirred
for several hours (1-4 h), keeping the solution at 0 °C. A characteristic
color change from orange to yellow was observed and was used to
estimate the reaction time. Removal of volatiles at low temperatures
(88) Hoyeyda, A. H.; Young, D. G. J.; Hale, M. R. Tetrahedron Lett.
1996, 37, 827-830.
(89) Griffith, W. P.; Ley, S. V.; Whitcombe, G. P.; White, A. D. J. Chem.
Soc., Chem. Commun. 1987, 1625-1627.
(90) Griffith, W. P.; Ley, S. V. Aldrichim. Acta 1990, 23, 12-22.

6978 J. Am. Chem. Soc., Vol. 120, No. 28, 1998
Lenges et al.
Table 7. Spectroscopic and Analytical Data for Ketones Formed in Catalytic Hydroacylation
furnishes a yellow solid material (or an oil in some cases). Charac-
terization by 1H NMR spectroscopy reveals the formation of complexes
4a,k,l and is in agreement with the data found for the intermediates
observed during catalysis or in direct NMR experiments as described.
Additional resonances due to decomposition products or remaining 1
are often observed as well. Any attempt to purify the product by
chromatography (-60 °C, alumina, pentane) or recrystallization resulted
in extensive decomposition. Therefore, compounds 4a,k were char-
acterized by NMR and IR spectroscopy at this stage. Recrystallization
was successful in the case of 4l, and isolation on a larger scale was
possible. Table 5 summarizes characteristic NMR and IR data for
compounds 4 and 9c. Complexes 4a,k,l were isolated in the described
manner, 4b,d and 9c were characterized under catalytic conditions at
-20 °C, 4c,f,h were generated in a reaction of excess aldehyde with 1
at low temperatures (vide supra) and characterized at -20 °C. All
NMR data were obtained in d6-acetone. IR spectra were obtained at
20 °C after addition of excess aldehyde to a toluene solution of 1
immediately after mixing.
[C5Me5Co(CO)(C2H4SiMe3)(C5H4FeC5H5)] 4l. Complex 1 (0.2 g,
5 × 10-4 mol) and ferrocenecarbaldehyde (0.32 g, 15 × 10-4 mol)
were combined in a drybox, and the flask was sealed and attached to
a Schlenk line. After cooling to 0 °C, 15 mL of acetone was slowly
added while the Schlenk flask was moved to ensure complete
dissolution. The homogeneous solution was kept at 0 °C for 5-6 h,
during which time the color changed to yellow. The reaction mixture
was cannula filtered and the Schlenk flask placed at -78 °C for 24 h.
Yellow crystalline material was isolated after the mother liquor was
decanted (0.2 g, 77%). Compound 4l is soluble in organic solvents
such as pentane, ether, or toluene without decomposition; however,
slow decay is observed in the course of hours at room temperature.
The complex was stored at -20 °C under an Ar atmosphere. The
material obtained this way was suitable for an X-ray structure
determination. In repeated syntheses the coprecipitation of remaining
1 is frequently observed, and recrystallization results in largely reduced
yields. NMR spectroscopic data for 4l: 1H (300 MHZ, d6-acetone,
-10 °C) δ 1.52 (s, 15, C5Me5), 0.09 (s, 9, SiMe3), 1.05 (m, 1, CHH-
Si), 1.50 (m, 1, CHH-Si) 1.81 (m, 1, Co-CHH), 2.46 (m, 1, Co-
CHH), 3.51 (m, 1, C5H3H), 3.87 (m, 1, C5H3H), 4.09 (m, 2, C5H2H2),
3.96 (s, 5, C5H5); 13C{1H} (d6-acetone, -15oC) δ -1.83 (SiMe3), 8.91
(C5Me5), 98.05 (C5Me5), 19.98, 26.18 (-CH2-), 67.70, 67.82, 73.61,
77.35, 98.62 (C5H4-Co), 68.69 (C5H5); Anal. Calcd for C26H37-
CoFeOSi: C, 61.42; H, 7.33. Found: C, 60.92; H, 7.36; MS/EI 508
(m+); IR (toluene, 20 °C) 1983 cm-1
.
[C5Me5Co(CO)(C2H3SiMe3)] 14. This compound was synthesized
independently in a reaction of 1 in pentane at 0 °C with CO (1.1 equiv).
A solution of 0.2 g (5 × 10-4 mol) 1 in pentane (15 mL) was cooled
to 0 °C, and 12.5 mL of CO gas was added slowly with vigorous
stirring. Stirring was continued for 2 h at 0 °C. A green oil was
isolated after filtration and evaporation of all volatiles. This compound
is quite sensitive and is stored as a frozen C6D6 solution. The
spectroscopic features of this complex are identical with the data for
complex 14 characterized in situ after hydroacylation catalysis. NMR
spectroscopic data for 14: 1H (300 MHZ, C6D6, 20 °C) δ d 1.52 (s,
15, C5Me5), 0.22 (s, 9, SiMe3), 2.08 (d, 10.8 Hz, 1, CH2 syn), 2.44 (d,
14.0 Hz, 1, CH2 anti), 1.02 (dd, 14.0, 10.8 Hz, 1, CHSi syn); 13C{1H}
(C6D6, 20 °C) δ d 0.32 (SiMe3), 9.82 (C5Me5), 95.0 (C5Me5), 43.4
(CHSi), 44.2 (CH2); IR (C6D6, 20 °C) 1944 cm-1
.
[C5Me5Co(µ-CO)]2 15. This decarbonylation product of catalytic
hydroacylation was isolated for example from a reaction using excess
propanal and vinyltrimethylsilane; identical results were obtained in a

Addition of Alkyl Aldehydes to Vinylsilanes
J. Am. Chem. Soc., Vol. 120, No. 28, 1998 6979
reaction of other aldehydes with vinyltrimethylsilane. Complex 1, (0.5
g, 1.27 × 10-3 mol) was dissolved in acetone (15 mL) with 20 equiv
of vinyltrimethylsilane present (2.54 g, 0.025 mol). The orange solution
was stirred, and 40 equiv of propanal (2.94 g, 0.05 mol) was added.
This mixture was kept at 40 °C for 24 h, during which time the mixture
changed color from yellow to green. The solution was cooled to -78
°C for 24 h. A dark-green, crystalline material was isolated after
filtration (0.15 g, 53% for 100% decarbonylation to 15 (vide supra)):
1H NMR (300 MHz, 20 °C, C6D6) δ 1.41 (C5Me5, s); 13C NMR (20
°C, C6D6) δ 94.7 (C5Me5), 8.6 (C5Me5); MS (EI) m/z 444 (M+), 135
aldehyde to olefin was in general 1:1 with a catalyst load of 2-5 mol
%; in selected cases less catalyst was used (1 or 0.5 mol %). The
flask was sealed with the Teflon screw-top and stirred for a period of
48 h at 25 °C or for 24 h at 45 °C. Conversion was based on 1H NMR
analysis of a sample of the reaction mixture after the reaction was
stopped and opened to air. These reaction mixtures were filtered
through silica or Celite plugs to separate the catalyst, and all volatiles
were removed by aspirator vacuum. In selected examples (3a) the
material isolated at this stage was pure by NMR analysis and elemental
analysis. However, in most cases purification was accomplished by
column chromatography on silica eluting with pentane/ether. Organic
(base, C5Me5); IR (hexane, 20 °C) 1758 cm-1 91-93
.
1
products (ketones 3a-l; 8a-c,h) were characterized by H and 13C-
X-ray Structure Determination of 1 and 4l. Data were collected
on a Siemens SMART diffractometer, using the ω scans. The structures
were solved by direct methods. Refinement was by full-matrix least
squares with weights based on counter statistics. Hydrogen atoms were
included in the final refinement using a riding model with thermal
parameters derived from the atom to which they are bonded. Crystal
data and experimental conditions are given in Table 6. All computa-
tions were performed using the NRCVAX suite of programs.61,70
General Procedure for Catalytic Hydroacylation. In a drybox
under an argon atmosphere, a flask with a stir bar and with a Teflon
screw cap (Kontes) was charged with catalyst 1. A catalyst load of
0.05 g (1.27 × 10-4 mol) was usually used. The catalyst was dissolved
in 2-3.5 mL of acetone to form a homogeneous solution. The olefin
was added first, and then the appropriate amount of aldehyde was added.
The reaction mixture was in all cases homogeneous. The ratio of
{1H} NMR spectroscopy, elemental analysis, or mass spectroscopy
(Table 7).
Acknowledgment is made to the National Institutes of Health
(Grant GM 28938) for financial support. C.P.L. thanks the Fonds
der Chemischen Industrie, Germany, for a Kekule´ fellowship.
We are indebted to Dan Tempel for providing a sample of [13C]-
pentanal. For additional crystallographic details, contact P.S.W.
Supporting Information Available: Tables of crystal data,
structure solution and refinement, atomic coordinates, bond
lengths and angles, and anisotropic thermal parameters for 1
and 4l, kinetic plots for ligand trapping reactions of 4l with
P(OMe)3 and olefins, and plots for the substrate dependence of
hydroacylation of isovaleraldehyde (24 pages, print/PDF). See
any current masthead page for ordering information and Web
access instructions.
(91) Brintzinger, H. H.; Lee, W.-S. J. Organomet. Chem. 1981, 209,
401-406.
(92) Dahl, L. F.; Ginsburg, R. E.; Cirjak, L. M. Inorg. Chem. 1982, 21,
940-957.
(93) Dahl, L. F.; Ginsburg, R. E.; Cirjak, L. M. J. Chem. Soc., Chem.
Commun. 1979, 468-472.
JA980610N