Nonhydride mechanism of metal-catalyzed hydrosilylation

Article abstract of DOI:10.1021/ja111748u

A 1:1:1 reaction between complex (Tp)(ArN=)Mo(H)(PMe₃) (3), silane PhSiD₃, and carbonyl substrate established that hydrosilylation catalyzed by 3 is not accompanied by deuterium incorporation into the hydride position of the catalyst, thus ruling out the conventional hydride mechanism based on carbonyl insertion into the M-H bond. An analogous result was observed for the catalysis by (O=)(PhMe₂SiO)Re(PPh ₃)₂(I)(H) and (Ph₃PCuH)₆.

Full text of DOI:10.1021/ja111748u

COMMUNICATION
pubs.acs.org/JACS
Nonhydride Mechanism of Metal-Catalyzed Hydrosilylation
Oleg G. Shirobokov,† Lyudmila G. Kuzmina,‡ and Georgii I. Nikonov*,†
†Chemistry Department, Brock University, 500 Glenridge Avenue, St. Catharines, Ontario L2S 3A1, Canada
‡Kurnakov Institute of General and Inorganic Chemistry, 31 Leninsky prospect, Moscow 119991, Russia
S Supporting Information
b
Scheme 1. Preparation and Structure of (Tp)(ArNd)Mo
(H)(PMe3) (3)
ABSTRACT: A 1:1:1 reaction between complex (Tp)-
(ArNd)Mo(H)(PMe3) (3), silane PhSiD3, and carbonyl
substrate established that hydrosilylation catalyzed by 3 is
not accompanied by deuterium incorporation into the
hydride position of the catalyst, thus ruling out the con-
ventional hydride mechanism based on carbonyl insertion
into the MÀH bond. An analogous result was observed
for the catalysis by (Od)(PhMe2SiO)Re(PPh3)2(I)(H)
and (Ph3PCuH)6.
etal-catalyzed hydrosilylation of carbonyls has recently
M
received renewed attention as a convenient reduction
method.1,2 Much current effort is aimed at replacing heavy late-
transition-metal complexes for less expensive and more environ-
mentally benign early metals, such as Ti,3 Zr,4 and Mo,5,6 or first
d series metals,2d,7 particularly Fe8 and Cu.9 The mechanistic
proposals offered for these metal catalysts share a common
theme: the formation of a metal hydride intermediate upon the
addition of silane followed by carbonyl insertion into the MÀH
bond to give an alkoxide (the hydride mechanism).2e Indeed, some
metal hydride complexes have been obtained or observed under
the hydrosilylation conditions and proved to turn over.6,7b,10 But
are they the actual catalysts, and what is the true role of the
hydride ligand? Here we disclose a simple labeling experiment
which allows one to probe the hydride mechanism. We apply
this technique to several prominent recent examples of Mo, Re,
and Cu catalysis and show that the reaction of metal hydrides
with carbonyls is not necessarily part of the actual catalysis
reaction.
ΔHq = 81.8 ( 2.1 kJ/mol, ΔSq = 32.4 ( 7.0 J/(mol K)],
3
suggesting dissociation of one of the pyrazolyl “legs”. In contrast
to complex 2, which undergoes a slow but catalytically unim-
portant reaction with PhSiH3, 3 is nonreactive toward silane.
However, it slowly reacts with benzaldehyde to give the insertion
product (Tp)(ArN)Mo(OCH2Ph)(PMe3) (4). Unlike the cor-
responding reaction of 2, this insertion is dissociative
[k(328.7 K) = (1.18 ( 0.02) Â 10À1 MÀ1
s
À1, ΔHq = 107.0
3
( 7.0 kJ/mol, ΔSq = 61.2 ( 21.4 J/(mol K)], but the reaction
3
rate does not depend on the concentration of PMe3,11 suggesting
that the rate-determining step is the dissociation of one of the
pyrazolyl ligands. The next step, the reaction of 4 with phenylsi-
lane to regenerate the starting hydride 3, also shows a positive
entropy of activation [k(295.1 K) = (2.35 ( 0.02) Â 10À4 sÀ1
,
ΔHq = 96.6 ( 1.8 kJ/mol, ΔSq = 12.5 ( 6.2 J/(mol K)],
3
indicating a significant dissociative (presumably of the Tp ligand)
character of the rate-limiting step.
Our previous mechanistic studies on Mo-catalyzed hydrosil-
ylation suggested that complex (ArNd)(Me3P)3Mo(Cl)(H)
(1) reacts by a phosphine dissociation mechanism, whereas the
related complex Cp(ArNd)Mo(H)(PMe3) (2) reacts by an un-
expected associative mechanism.6 Wishing to explore this mech-
anistic dichotomy further, we turned to the isolobal system
Tp(ArNd)Mo(H)(PMe3) (3; Tp = tris(pyrazolyl) borate).
Complex 3 was prepared by reacting 1 with TpK and was studied
by NMR and X-ray diffraction (Scheme 1). 3 was found to
catalyze the hydrosilylation even more effectively than 2
(Table 1). Given the increased bulkiness and diminished donor
ability of the Tp relative to the Cp ligand, the increased catalytic
efficiency of 3 was surprising and prompted us to carry out a
mechanistic study.
With these stoichiometric reactions in hand, a clear mechan-
istic picture emerged that the catalytic reaction proceeds via the
silane heterolytic cleavage on the MÀOR bond,6 with all steps
being activated by dissociation of the Tp. However, this hydride
mechanism is not compatible with the very poor reactivity of 3
toward ketones. Their reactions were found to occur only upon
heating at 50 °C for several days, whereas catalytic reactions are
already fast at room temperature (Table 1, entries 2 and 3).
This controversy was eventually solved by the remarkable
observation that a stoichiometric (1:1:1) reaction of PhCHO
with PhSiD3 and 3 results in immediate formation of PhCHDO-
SiD2Ph (>95% by NMR), keeping the complex 3 unchanged.
First, by EXSY NMR we found an exchange process in the Tp
Received: December 30, 2010
Published: April 08, 2011
ligand of 3. This exchange is dissociative [k(295.1 K) = 1.14 sÀ1
,
r
2011 American Chemical Society
6487
dx.doi.org/10.1021/ja111748u J. Am. Chem. Soc. 2011, 133, 6487–6489
|

Journal of the American Chemical Society
COMMUNICATION
Table 1. Catalytic Hydrosilylation Mediated by 3 (5 mol %, in C6D6)
entry
1
substrate
silane
conditions
0.5 day, rt
conversion of substrate (%)
100
product (yield)
PhCH2OSiH2Ph (38%),
PhC(O)H
PhSiH3
(PhCH2O)2SiHPh (62%)
MePhHCOSiH2Ph (85%),
(MePhHCO)2SiHPh (15%)
cyclohexyl-OSiH2Ph (85%),
(cyclohexyl-O)2SiHPh (15%)
PhCH2OSiHMePh (30%)
MePhHCOSiHMePh (100%)
cyclohexyl-OSiHMePh (96%),
(cyclohexyl-O)2SiMePh (4%)
cyclohexyl-OSiMe2Ph (11%)
2
3
PhC(O)Me
PhSiH3
PhSiH3
1.5 days, rt
53 min, rt
100
100
cyclohexanone
4
5
6
PhC(O)H
PhMeSiH2
PhMeSiH2
PhMeSiH2
1.5 days, 50 °C
2.5 days, 50 °C
1 day, rt
30
100
100
PhC(O)Me
cyclohexanone
7
8
cyclohexanone
PhCN
PhMe2SiH
PhSiH3
1.5 days, 50 °C
3 days, 50 °C
11
20
PhCH=NSiH2Ph (17%), (PhCH=N)2SiHPh (3%)
1
Close monitoring of catalytic reactions by H and 31P NMR
revealed the presence of 3 in the reaction mixture. An analogous
result was obtained for the stoichiometric reaction of cyclohex-
anone with PhSiD3 and 3, in which the MoÀH functionality
remained unlabeled (>95%). Finally, we carried out the hydro-
silylation of cyclohexanone by PhSiD3 in the presence of 5 mol %
3 and observed the retention of the MoÀH group at the end of
catalysis. These observations unequivocally establish that carbo-
nyl insertion into the MoÀH bond of 3 is not a catalytically
relevant event.
complex 7 is able to turn over, kinetic modeling of the potential
hydride mechanism suggested that it is a minor process. We reacted
7 with PhHCdO and DSiEt3 in the 1:1:1 ratio and initially
observed formation of the insertion product (Od)Re(PPh3)2-
(OCH2Ph)Cl2, which then slowly reacts with the silane to give
PhH2CÀOSiEt3 and PhHDCÀOSiEt3 in a 7:1 ratio, consistent
with the occurrence of a minor hydride mechanism. We also
checked whether the H/D scrambling in the products could come
from a ReH/SiD exchange. To this end, complex 6 was reacted
with DSiEt3. No exchange was observed at room temperature, and
only a minor exchange was observed after 1 day at 70 °C. However,
we were surprised to notice that (Od)Re(PPh3)2Cl3 appears to be
a poorer catalyst than hydride 7 (Table 1 in ref 10b), which is at
odds with an ionic mechanism of hydrosilylation that avoids the
formation of the latter. Given the fact that kinetic modeling of the
hydride mechanism accounts for merely 20% of hydrosilylation,10b
these observations suggest that 7 can be the true catalyst operating
by a nonhydride mechanism.
Lipshutz et al. found that copper hydrides, such as the Stryker
reagent (Ph3PCuH)6, catalyze hydrosilylation of carbonyls.17
Several copper salts were found to mediate this reaction in the
presence of alkoxide,18 with the accepted mechanistic proposal
being that an intermediate copper hydride is generated by the
reaction of a copper alkoxide and silane followed by carbonyl
insertion into the CuÀH bond.17,18c,19,20 However, while alde-
hydes were indeed shown to insert into the CuÀH bond of
(Ph3PCuH)6, ketones do not react with copper hydrides,17b,18c
despite the efficiency of Cu-catalyzed hydrosilylation of ketones!
To verify the mechanism, we first showed that Ph3PCuOBut,
prepared in situ from CuCl, Ph3P, and NaOBut, does react with
HSiMe2Ph to give (Ph3PCuH)6 in 46% isolated yield. However,
when this hydride reacted21 with stoichiometric amounts of benz-
aldehyde and DSiMe2Ph, the exclusive formation of PhHDCÀ
OSiMe2Ph was observed, without any deuterium incorporation
into the hydride position, thus allowing us to rule out the hydride
mechanism.
Kinetic studies of the latter reaction showed that reaction rates
increase with increasing concentration of cyclohexanone but are
insensitive to variation of the silane concentration.11 Therefore,
we conclude that the rate-determining step must be the activa-
tion of carbonyl. As 3 is nonreactive to both the silane and
ketone, the only mechanistic alternative we are left with is that
the metal center activates the carbonyl as a Lewis acid.
Lewis acid catalysis has been previously invoked for the
hydrosilylation by B(C6F5)3,12 AlCl3, Ti(OPri)4, TiCl4, ZrCl4,
and HfCl4,1e but surprisingly it has not been considered for other
metal complexes. To check the scope of this reactivity, we rein-
vestigated some recent prominent examples of catalytic hydro-
silylation.
Toste et al. reported the hydrosilylation of carbonyls by
(Od)2Re(PPh3)2I10a (5), and the related O2MoCl2-catalyzed
reactions were disclosed by Royo et al.13 The Toste group
provided mechanistic evidence that the reaction occurs via SiÀH
addition across the RedO bond to give the hydride (Od)
(PhMe2SiO)Re(PPh3)2(I)(H) (6), followed by carbonyl inser-
tion into the ReÀH bond.14 We studied the 1:1:1 reaction of 6
with benzaldehyde and DSiMe2Ph. At 50% conversion, no
deuterium scrambling into the ReÀH position was observed
(integrates as 1H),15 and the methylene part of the hydrosilyla-
tion product16 integrates as 1H, which contradicts the conclusion
by Toste et al. that silane addition to 5 is fast and that the rate-
determining step is the aldehyde insertion into the ReÀH bond
of 6. Another indication that the hydride mechanism is not the
dominant reaction pathway comes from the observation that the
1:1:1 reaction of 6 with benzaldehyde and DSiMe2Ph takes about
1 h, whereas a 1:1 reaction of 6 with benzaldehyde was complete
only overnight.
One limitation for the proposed labeling test comes from the
possibility of fast H/D exchange.19a This is the case for our pre-
viously studied system (ArNd)(Me3P)3Mo(Cl)(H) (1), which
was suggested to catalyze hydrosilylation of carbonyls by the
hydride mechanism.6a 1 undergoes H/D exchange with silane at
a rate comparable with the addition of benzaldehyde, so that the
1:1:1 reaction of 1 with PhHCdO and D3SiPh results in deu-
terium scrambling in the hydride position and the product.22
In a related catalysis by monooxo Re complexes, Abu-Omar et al.
observed that (Od)Re(PPh3)2Cl3 reacts with HSiEt3 to give the
hydride derivative (Od)Re(PPh3)2(H)Cl2 (7).10b Although
6488
dx.doi.org/10.1021/ja111748u |J. Am. Chem. Soc. 2011, 133, 6487–6489

Journal of the American Chemical Society
COMMUNICATION
In conclusion, we have developed a simple stoichiometric
labeling experiment allowing one to probe the hydride mechan-
ism of hydrosilylation, which is important for the rational design
of new catalysts. The true role of the hydride ligand in this
catalysis remains unclear. One possibility is that it may provide
enough space for carbonyl η1-coordination to metal in the Lewis
acid catalysis.
Vyboishchikov, S. F.; Nikonov, G. I. J. Am. Chem. Soc. 2008, 130, 3732.
(c) Shaikh, N. S.; Junge, K.; Beller, M. Org. Lett. 2007, 9, 5429. (d)
Nishiyama, H.; Furuta, A. Chem. Commun. 2007, 760. (e) Tondreau,
A. M.; Lobkovsky, E.; Chirik, P. J. Org. Lett. 2008, 10, 2789. (f) Shaikh,
N. S.; Enthaler, S.; Junge, K.; Beller, M. Angew. Chem. Int. Ed. 2008,
5, 5429. (g) Langlotz, B. K.; Wadepohl, H.; Gade, L. H. Angew. Chem.,
Int. Ed. 2008, 47, 4670. (h) Addis, D.; Shaikh, N.; Zhou, S.; Das, S.;
Junge, K.; Beller, M. Chem. Asian J. 2010, 5, 1687. (i) Kandepi, V. V. K.
M.; Cardoso, J. M. S.; Peris, E.; Royo, B. Organometallics 2010, 29, 2777.
(9) (a) Díez-Gonzꢀalez, S.; Nolan, S. P. Acc. Chem. Res. 2008, 41, 349.
(b) Deutsch, C.; Krause, N.; Lipshutz, B. H. Chem. Rev. 2008, 108, 2916.
(10) (a) Nolin, K. A.; Krumper, J. R.; Pluth, M. D.; Bergman, R. G.;
Toste, F. D. J. Am. Chem. Soc. 2007, 129, 14684. (b) Du, G. D.; Fanwick,
P. E.; Abu-Omar, M. M. J. Am. Chem. Soc. 2007, 129, 5180.
(11) See the Supporting Information for details.
’ ASSOCIATED CONTENT
S
Supporting Information. Experimental procedures, spec-
b
tral data, and details of kinetic studies; complete crystallographic
data (CIF) for the complexes reported in this paper. This material is
available free of charge via the Internet at http://pubs.acs.org
(12) (a) Rendler, S.; Oestreich, M. Angew. Chem., Int. Ed. 2008,
47, 5997. (b) Parks, D. J.; Blackwell, J. M.; Piers, W. E. J. Org. Chem.
2000, 65, 3090.
’ AUTHOR INFORMATION
(13) (a) Royo, B.; Rom~ao, C. C. J. Mol. Catal. A: Chem. 2005,
236, 107. (b) Reis, P. M.; Rom~ao, C. C.; Royo, B. Dalton Trans.
2006, 1842. (c) Fernandes, A. C.; Fernandes, R.; Rom~ao, C. C.; Royo,
B. Chem. Commun. 2005, 213.
Corresponding Author
gnikonov@brocku.ca
(14) For DFT calculations, see: (a) Costa, P. J.; Rom~ao, C. C.;
Fernandens, A. C.; Royo, B.; Reis, P.; Calhorda, M. J. Chem.—Eur. J.
2007, 13, 3934. (b) Drees, M.; Strassner, T. Inorg. Chem. 2007, 25,
10850. (c) Chung, L. W.; Lee, H. G.; Lin, Z.; Wu, Y.D. J. Org. Chem.
2006, 71, 6000.
(15) By the end of catalysis, about 15% decomposition of the catalyst
is observed and about 20% of the remaining rhenium hydride is
deuterated, presumably by the competing hydride mechanism. No silyl
hydride signal of HÀSiMe2Ph was observed throughout the catalysis,
suggesting that ReH/SiD exchange (ref 10a) is slower than hydro-
silylation.
’ ACKNOWLEDGMENT
This work was supported by Petroleum Research Fund (grant
to G.I.N.), administered by the American Chemical Society.
L.G.K. also thanks the Russian Foundation of Basic Research. We
thank Mr. Razvan Simionescu for the help with NMR and many
useful discussions.
’ REFERENCES
(1) For general reviews on hydrosilylation, see: Roy, A. K. Adv.
Organomet. Chem. 2008, 55, 1. (b) Gibson, S. E.; Rudd, M. Adv. Synth.
Catal. 2007, 349, 781. (c) Marciniec, B. Appl. Organomet. Chem. 2000,
14, 527.(d) Ojima, I. In The Chemistry of Organic Silicon Compounds;
Patai, S., Rappoport, Z., Eds.; Wiley: New York, 1989; Chapter 25.
(e) Marciniec, B. Comprehensive Handbook on Hydrosilylation; Perga-
mon: Oxford, 1992.
(2) (a) Carpentier, J.-F.; Bette, V. Curr. Org. Chem. 2002, 6, 913.
(b) Riant, O.; Mosterfai, N.; Courmarcel, J. Synthesis 2004, 18, 2943.
(c) Diez-Gondlez, S.; Nolan, S. P. Org. Prep. Proc. Int. 2007, 39, 523.
(d) Du, G. D.; Abu-Omar, M. M. Curr. Org. Chem. 2008, 12, 1185.
(e) Chakraborty, S.; Guan, D. Dalton Trans. 2010, 39, 7427.
(3) (a) Berc, S. C.; Kreutzer, K. A.; Buchwald, S. L. J. Am. Chem. Soc.
1991, 113, 5093. (b) Berc, S. C.; Buchwald, S. L. J. Org. Chem. 1992,
57, 3751. (c) Broene, R. D.; Buchwald, S. L. J. Am. Chem. Soc. 1993,
115, 12569. (d) Carter, M. B.; Schiøtt, B.; Gutiꢀerrez, A.; Buchwald, S. L.
J. Am. Chem. Soc. 1994, 116, 11667. (e) Halterman, R. L.; Ramsey, T. M.;
Chen, Z. J. Org. Chem. 1994, 59, 2642. (f) Harrod, J. F.; Xin, S. Can. J.
Chem. 1995, 73, 999. (g) Yun, J.; Buchwald, S. L. J. Am. Chem. Soc. 1999,
121, 5640.
(16) A mixture of isotopomers PhCH2OSiMe2Ph and PhCHDOSi-
13
29
Me2Ph is observed by 1H NMR. 1HÀ C HSQC, DEPT, and 1HÀ Si
HSQC experiments confirmed that isotopomers of the same hydro-
silylation product were formed.
(17) (a) Lipshutz, B. H.; Chrisman, W.; Noson, K. J. Organomet.
Chem. 2001, 624, 367. (b) Lipshutz, B. H.; Noson, K.; Chrisman, W.;
Lower, A. J. Am. Chem. Soc. 2003, 125, 8779. (c) Lipshutz, B. H.;
Frieman, B. A. Angew. Chem., Int. Ed. 2005, 44, 6345.
(18) (a) Brunner, H.; Miehling J. Organomet. Chem. 1984, 275, C17.
(b) Lipshutz, B. H.; Noson, K.; Chrisman, W. J. Am. Chem. Soc. 2001,
123, 12917. (c) Issenhuth, J.-T.; Dagorne, S.; Bellemin-Laponnaz, S.
Adv. Synth. Catal. 2006, 348, 1991.
(19) (a) Issenhuth, J.-T.; Notter, F.-P.; Dagorne, S.; Dedieu, A.;
Bellemin-Laponnaz, S. Eur. J. Inorg. Chem. 2010, 529. (b) Rendler, S.;
Oestreich, M. Angew. Chem., Int. Ed. 2007, 46, 498.
(20) Carbonyl insertion is believed to be stereodefining in asym-
metric versions (see ref 17b).
(21) No hydride signal is observed for the Lipshutz’s catalyst (R-3,
5-xyl-MeO-BIPHEP)CuH,17b making it a bad candidate for this
experiment.
(4) Yun, S. S.; Yong, S. Y.; Lee, S. Bull. Kor. Chem. Soc. 1997, 18,
(22) The Re complex 6 does undergo H/D exchange with PhMe2-
SiD,10a but slower than the 1:1:1 hydrosilylation, whereas complex 7 and
(Ph3PCuH)6 do not show any H/D exchange with PhSiD3 and
PhMe2SiD, respectively, during several days at room temperature.
1058.
(5) (a) Royo, B.; Rom~ao, C. C. J. Mol. Catal. A: Chem. 2005,
236, 107. (b) Reis, P. M.; Rom~ao, C. C.; Royo, B. Dalton Trans.
2006, 1842. (c) Ziegler, J. E.; Du, G. D.; Fanwick, P. E.; Abu-Omar,
M. M. Inorg. Chem. 2009, 48, 11290.
(6) (a) Peterson, E.; Khalimon, A. Y.; Simionescu, R.; Kuzmina,
L. G.; Howard, J. A. K.; Nikonov, G. I. J. Am. Chem. Soc. 2009, 131, 908.
(b) Shirobokov, O. G.; Simionescu, R.; Kuzmina, L. G.; Nikonov, G. I.
Chem. Commun. 2010, 46, 7831.
(7) (a) Fontaine, F.-G.; Nguyen, R.-V.; Zargarian, D. Can. J. Chem.
2003, 81, 1299. (b) Chakraborty, S.; Krause, J. A.; Guan, H. Organome-
tallics 2009, 28, 582. (c) Tran, B. L.; Pink, M.; Mindiola, D. J. Organo-
metallics 2009, 28, 2234.
(8) (a) Bart, S. C.; Lobkovsky, E.; Chirik, P. J. J. Am. Chem. Soc. 2004,
126, 13794. (b) Gutsulyak, D. V.; Kuzmina, L. G.; Howard, J. A. K.;
6489
dx.doi.org/10.1021/ja111748u |J. Am. Chem. Soc. 2011, 133, 6487–6489