Studies on the Mechanism of B(C6F5)3-Catalyzed Hydrosilation of Carbonyl Functions

Article abstract of DOI:10.1021/jo991828a

The strong organoborane Lewis acid B(C₆F₅)₃ catalyzes the hydrosilation (using R₃SiH) of aromatic and aliphatic carbonyl functions at convenient rates with loadings of 1-4%. For aldehydes and ketones, the product silyl ethers are isolated in 75-96% yield; for esters, the aldehydes produced upon workup of the silyl acetal products can be obtained in 45-70% yield. Extensive mechanistic studies point to an unusual silane activation mechanism rather than one involving borane activation of the carbonyl function. Quantitative kinetic studies show that the least basic substrates are hydrosilated at the fastest rates; furthermore, increased concentrations of substrate have an inhibitory effect on the observed reaction rate. Paradoxically, the most basic substrates are reduced selectively, albeit at a slower rate, in competition experiments. The borane thus must dissociate from the carbonyl to activate the silane via hydride abstraction; the incipient silylium species then coordinates the most basic function, which is selectively reduced by [HB(C₆F₅)₃]^-. In addition to the kinetic data, this mechanistic proposal is supported by a kinetic isotope effect of 1.4(5) for the hydrosilation of acetophenone, the observation that B(C₆F₅)₃ catalyzes H/D and H/H scrambling in silanes in the absence of substrate, computational investigations, the synthesis of models for proposed intermediates, and other isotope labeling and crossover experiments.

Full text of DOI:10.1021/jo991828a

3090
J . Org. Chem. 2000, 65, 3090-3098
Stu d ies on th e Mech a n ism of B(C6F 5)3-Ca ta lyzed Hyd r osila tion of
Ca r bon yl F u n ction s
Daniel J . Parks, J ames M. Blackwell, and Warren E. Piers*
Department of Chemistry, University of Calgary, 2500 University Drive N. W.,
Calgary, Alberta T2N 1N4, Canada
Received November 30, 1999
The strong organoborane Lewis acid B(C6F5)3 catalyzes the hydrosilation (using R3SiH) of aromatic
and aliphatic carbonyl functions at convenient rates with loadings of 1-4%. For aldehydes and
ketones, the product silyl ethers are isolated in 75-96% yield; for esters, the aldehydes produced
upon workup of the silyl acetal products can be obtained in 45-70% yield. Extensive mechanistic
studies point to an unusual silane activation mechanism rather than one involving borane activation
of the carbonyl function. Quantitative kinetic studies show that the least basic substrates are
hydrosilated at the fastest rates; furthermore, increased concentrations of substrate have an
inhibitory effect on the observed reaction rate. Paradoxically, the most basic substrates are reduced
selectively, albeit at a slower rate, in competition experiments. The borane thus must dissociate
from the carbonyl to activate the silane via hydride abstraction; the incipient silylium species then
coordinates the most basic function, which is selectively reduced by [HB(C6F5)3]-. In addition to
the kinetic data, this mechanistic proposal is supported by a kinetic isotope effect of 1.4(5) for the
hydrosilation of acetophenone, the observation that B(C6F5)3 catalyzes H/D and H/H scrambling in
silanes in the absence of substrate, computational investigations, the synthesis of models for
proposed intermediates, and other isotope labeling and crossover experiments.
In tr od u ction
mediate this reaction also,9 although not catalytically. In
these studies, BF3 activation of the carbonyl moiety is
the starting point for mechanistic postulates and is
included as a key step in the overall mechanistic scheme.
It was therefore surprising that, for the B(C6F5)3-medi-
ated hydrosilation of these substrates, our preliminary
mechanistic investigations strongly suggested that the
borane/carbonyl adducts were not directly involved in the
addition of Si-H across CdO. Since a detailed under-
standing of the intimate mechanism of such processes is
crucial for the design of better and/or stereoselective
catalysts, we have explored the mechanism of this
reaction in detail and report a full account of this study
herein.
The importance of Lewis acids in the catalysis of
organic transformations involving carbonyl functions
cannot be understated.1 While the diversity of Lewis acids
is extensive, boron-based reagents remain prominent as
a result of their high Lewis acid strength and ready
availability. While BF3 remains the quintessential boron
2
Lewis acid, the strong organometallic Lewis acid B(C6F5)3
has emerged in recent years as a viable alternative.3
Comparable in strength to BF3,4 it is considerably more
hydrolytically stable and has indeed been shown to be
active even under aqueous conditions.5 Given that it is
now commercially available, the number of applications
in organic synthesis is growing; indeed, its use (as well
as that of other pentafluorophenyl-substituted boranes)
has recently been reviewed.6 In general, for transforma-
tions involving the carbonyl function, it has been as-
sumed that B(C6F5)3 operates in a similar way to BF3,
serving to activate the carbonyl via coordination to one
of the oxygen lone pairs.7
Resu lts a n d Discu ssion
B(C6F 5)3-Ca t a lyzed Hyd r osila t ion of Ca r b on yl
F u n ction s. In the presence of 1-4 mol % B(C6F5)3,
addition of 1 equiv of R3SiH across the carbon-oxygen
double bond of a variety of aldehydes, ketones, and esters
is facile at room temperature, affording silyl ethers or
silyl acetals in good isolated yields. In addition to the
aryl-substituted compounds we initially employed,8 ali-
phatic substrates were also found to undergo reduction
under these conditions (Table 1). While aldehydes and
ketones are hydrosilated quite cleanly, the silyl acetal
products of ester reduction are susceptible to further
silation reactions10 under all conditions examined. These
Some years ago, we found that B(C6F5)3 is an effective
catalyst for the hydrosilation of a variety of aldehydes,
ketones, and esters.8 BF3‚OEt2 has been reported to
* Ph: 403-220-5746. FAX: 403-289-9488. Email: wpiers@ucalgary.ca.
(1) (a) Santelli, M.; Pons, J .-M. Lewis Acids and Selectivity in
Organic Synthesis: CRC Press: New York, 1996. (b) Lewis Acid
Chemistry: Yamamoto, H., Ed.; Oxford University Press: New York,
1999.
(2) Massey, A. G.; Park, A. J . J . Organomet. Chem. 1966, 5, 218.
(3) Piers, W. E.; Chivers, T. Chem. Soc. Rev. 1997, 345.
(4) Luo, L.; Marks, T. J . Top. Catalysis 1999, 7, 97.
(8) Parks, D. J .; Piers, W. E. J . Am. Chem. Soc. 1996, 118, 9440.
(9) (a) Fry, J . L.; Orfanopoulo, M.; Adlington, M. G.; Dittman, W.
R.; Silverman, S. B. J . Org. Chem. 1978, 43, 374. (b) Doyle, M. P.;
West, C. T.; Donnelly, S. J .; McOsker, C. C. J . Organomet. Chem. 1976,
117, 129 and other papers in this series.
(10) We (see ref 12) and others have observed B(C6F5)3-catalyzed
cleavage of ethers with silanes. See: Gevorgyan, V.; Liu, J .-X., Rubin,
M.; Benson, S.; Yamamoto, Y. Tetrahedron Lett. 1999, 40, 8919.
(5) Ishihara, K.; Hanaki, N.; Funahashi, M.; Miyata, M.; Yamamoto,
H. Bull. Chem. Soc. J pn. 1995, 68, 1721.
(6) Ishihara, K.; Yamamoto, H. Eur. J . Org. Chem. 1999, 527, 7.
(7) (a) Shambayati, S.; Crowe, W. E.; Schrieber, S. L. Angew. Chem.,
Int. Ed. Engl. 1990, 29, 256. (b) Shambayati, S.; Schrieber, S. L. In
Comprehensive Organic Synthesis; Trost, B. M., Fleming, I., Eds.;
Pergammon: Oxford, 1991; Vol. 1, pp 283-324.
10.1021/jo991828a CCC: $19.00 © 2000 American Chemical Society
Published on Web 04/20/2000

B(C6F5)3-Catalyzed Hydrosilation
J . Org. Chem., Vol. 65, No. 10, 2000 3091
Ta ble 1. B(C6F 5)3-Ca ta lyzed Hyd r osila tion of Ca r bon yl
F u n ction s Usin g P h 3SiH
TBAF) are in the range of 60-70%. Nonetheless, this
transformation (i.e., ester to aldehyde) is generally dif-
ficult to effect directly,11 and for simple systems this
represents an alternative worth considering.
Experiments aimed at exploring the scope of the
reaction showed that carbonyl hydrosilation occurs se-
lectively only if more basic functions are absent from the
substrate. Thus, while halogens, olefins, and internal
alkynes are tolerated, nitrile and alcohol functions gener-
ally are not. This observation led us to develop in detail
the utility of the B(C6F5)3/R3SiH system for the dehydro-
genative silation of alcohols as an alternative procedure
for generating silyl ethers.12 While the scope of alcohol
silation is quite wide, the lower inherent basicity of
carbonyl functions limits the carbonyl hydrosilation
reaction to somewhat less functionalized substrates.
Nonetheless, we have used this reaction to study the
mechanism by which the B(C6F5)3/R3SiH reagent operates
in some detail.
Mech a n ist ic St u d ies. Lewis acids are generally
thought to activate carbonyl functions through coordina-
tion of the CdO moiety, which serves to further polarize
the double bond and render the C atom even more
electrophilic than in the uncomplexed species. There are
some reports in the literature that suggest that Lewis
acid catalyzed addition of Si-H to CdO proceeds in this
fashion,13 but these are generally unsupported by quan-
titative mechanistic data. While B(C6F5)3 forms stable
adducts with carbonyl-containing substrates,14 we have
made several observations that suggest a less conven-
tional mechanism is operative for this system.
First, quantitative rate studies on the hydrosilation of
benzaldehyde, acetophenone, and ethylbenzoate revealed
that the order of reactivity was opposite to what would
be expected on the basis of substrate basicity. That is,
the least basic substrate toward B(C6F5)3 (ethyl benzoate)
was hydrosilated the fastest (TON ) 637 hr-1), followed
by acetophenone (TON ) 45 hr-1) and benzaldehyde
(TON ) 19 hr-1). If carbonyl adducts of B(C6F5)3 were
instrumental in bringing about hydrosilation, the least
basic substrate would be the least active, since there
would be a lower concentration of adduct present in
solution to react and the level of substrate activation
would be lower.
a
Conditions: A, 2% B(C6F5)3, benzene or toluene; B, 1%
B(C6F5)3, toluene, syringe pump addition of silane over 1 h.
Isolated yields. c Yields of PhCHO were lower as a result of loss
b
of product via evaporation and/or oxidation during workup; in situ
conversion to the 2,4-dinitrophenylhydrazone was done to gauge
the selectivity of the reaction for benzaldehyde more accurately.
Reaction via enol tautomer of substrate. e cis diastereomer
d
produced exclusively.
processes compete effectively with carbonyl hydrosilation
at later stages of the reaction and over-reduction products
are in evidence. Careful examination of the reaction of
ethylbenzoate with 1 equiv of silane and a catalytic
amount of B(C6F5)3 reveals a product mixture as shown
in eq 1. The silyl ether and siloxane products result from
Second, the observed rates of hydrosilation for a series
of para-substituted derivatives of acetophenone lead to
a similar conclusion. As shown in eq 2, the equilibrium
hydrosilation of the initial silyl acetal product; further
reduction of these secondary products ultimately yields
toluene and ethane, both of which are observed in trace
amounts. Thus, while the silyl acetal is the major
product, each of the other species in eq 1 is observed to
some extent as well. Extensive variation in reaction
conditions (temperature, solvent, and rate of silane
addition) and the silane employed did not lead to more
selective chemistry, so the yields as reported in Table 1
are optimized to the extent possible in our hands. Thus,
maximum isolated yields of the products of ester hydro-
silation (the analogous aldehydes upon workup with
constants for the para-X acetophenones (X ) H, CH3, Cl,
NO2) vary as one would expect on the basis of the donor
properties of the group. Thus, as the X group becomes
(11) Zakharkin, L. I.; Khorlina, I. M. Tetrahedron Lett. 1962, 14,
619.
(12) Blackwell, J . M.; Foster, K. L.; Beck, V. H.; Piers, W. E. J . Org.
Chem. 1999, 64, 4887.

3092 J . Org. Chem., Vol. 65, No. 10, 2000
Parks et al.
Sch em e 1
F igu r e 1. Plot of 1/kobs (× 10-4) vs acetophenone concentration
for the hydrosilation of acetophenone with Ph3SiH.
more electron-withdrawing and makes the carbonyl
group less basic, the equilibrium shifts away from the
adduct. The observed rate constants for hydrosilation of
these substrates (kobs ) 3.61(7) × 10-4 s-1 for X ) Me;
2.01(5) × 10-3 s-1 for X ) Cl; 1.35(3) × 10-2 s-1 for X )
NO2)8 indicate that the rate of hydrosilation increases
dramatically as X becomes more electron-withdrawing,
again showing that less basic substrates are reduced
more rapidly.
Third and perhaps most revealing, the rate of hydro-
silation is inhibited by increases in the substrate con-
centration. Thus, in the hydrosilation of acetophenone,
the observed rate constant is inversely proportional to
the concentration of acetophenone (Figure 1). Since
increases in the substrate concentration should push the
equilibrium toward the adduct, the inhibitory effect of
this on the overall rate of hydrosilation is perhaps the
clearest indication that the carbonyl adducts of B(C6F5)3
are not directly involved in the hydrosilation reaction.
dinated to R3Si+ occurs either from the hydridoborate
(path a) or via another equivalent of R3SiH (path b).
Three issues regarding this mechanistic proposal arise:
(1) What chemical evidence is there for silane activation
via hydride abstraction by B(C6F5)3? (2) How does sub-
strate interact with the resulting borane/silane complex?
(3) Which pathway for hydride delivery is operative in
this system? Each of these questions is dealt with in turn
in the following sections.
While these equilibria strongly favor the carbonyl
borane adducts, they are quickly established, and ex-
change processes between free and bound substrate are
rapid on the chemical time scale.14 In light of the above
kinetic data, we proposed a mechanism in which dis-
sociation of borane from the substrate is necessary in
order that B(C6F5)3 may activate the silane.8 The manner
in which we now believe this activation takes place is
outlined in Scheme 1. In this picture, the resulting
borane/silane complex then reacts with substrate, which
displaces the hydridoborate counterion from the incipient
silylium ion. At this point, reduction is consummated by
one of two pathways, each of which is consistent with
the observed kinetic behavior of the system. Transfer of
a hydride to the carbonyl carbon of the substrate coor-
Hyd r id e Abstr a ction fr om R3SiH by B(C6F 5)3. It
is well-known in the literature that B(C6F5)3 is a strong
enough Lewis acid to abstract alkide15 and hydride16 ions
from neutral organometallic precursors.3 Indeed, this
process is the key activating step for generation of
cationic olefin polymerization catalysts.17 Furthermore,
it has been shown that the tributylstannyl cation can be
(20) Typically, neutral, three-coordinate pentafluorophenyl borane
compounds exhibit a peak separation between the para and meta
fluorine resonances of >15 ppm. As the environment about boron
proceeds through neutral, four coordinate to anionic, four coordinate
geometries, this parameter gets smaller as the resonance for the para
fluorine shifts upfield. This was first observed by Horton and de With20a
and used to assess the extent of ion pairing between organometallic
cations and [MeB(C6F5)3]-. It is a sensitive tool for assessing the
coordination environment around boron in -C6F5-substituted
boranes.20b-d ( a) Horton, A. D.; de With, J . Organometallics 1997, 16,
5424. (b) Ko¨hler, K.; Piers, W. E.; Xin, S.; Feng, Y.; Bravakis, A. M.;
J arvis, A. P.; Collins, S.; Clegg, W.; Yap G. P. A.; Marder, T. B.
Organometallics 1998, 17, 3557. (c) Williams, V. C.; Dai, C.; Li, Z.;
Collins, S.; Piers, W. E.; Clegg, W. C.; Elsegood, M. R. J .; Marder, T.
B. Angew. Chem., Int. Ed. 1999, in press. (d) Ko¨hler, K.; Piers, W. E.
Can. J . Chem. 1998,76, 1249.
(21) (a) Lambert, J . B., Kania, L.; Zhang, S. Chem. Rev. 1995, 95,
1191. (b) Reed, C. A. Acc. Chem. Res. 1998, 31, 325. (c) Lambert, J . B.;
Zhang, S.; Stern, C. L.; Huffman, J . C. Science 1993, 260, 1917. (d)
Corey, J . Y. J . Am. Chem. Soc. 1975, 97, 3237.
(22) Silverstein, R. M.; Webster, F. X. Spectroscopic Identification
of Organic Compounds, 6th ed.; J ohn Wiley and Sons: New York, 1998;
p 163.
(13) (a) Ojima, I. In The Chemistry of Organic Silicon Compounds;
Patai, S., Rappoport, Z., Eds.; Wiley: New York, NY, 1989. (b)
Marciniec, B. Comprehensive Handbook on Hydrosilation; Pergam-
mon: New York, NY, 1992.
(14) Parks, D. J .; Piers, W. E.; Parvez, M.; Atencio, R.; Zaworotko,
M. J . Organometallics 1998, 17, 1369.
(15) Yang, X.; Stern, C. L.; Marks, T. J . J . Am. Chem. Soc. 1991,
113, 3623.
(16) Yang, X.; Stern, C. L.; Marks, T. J . Angew. Chem., Int. Ed. Engl.
1992, 31, 1375.
(17) (a) Brintzinger, H. H.; Fischer, D.; Mulhaupt, R.; Rieger, B.;
Waymouth, R. M. Angew. Chem., Int. Ed. Engl. 1995, 34, 1143. (b)
Bochmann, M. J . Chem. Soc., Dalton Trans. 1996, 255. (c) Marks, T.
J . Acc. Chem. Res. 1992, 25, 57.
(23) Chatgilialoglu, C.; Newcomb, M. Adv. Organomet. Chem. 1999,
44, 67.
(18) Lambert, J . B.; Kuhlmann, B. J . Chem. Soc., Chem. Commun.
1992, 931.
(19) Lambert, J . B.; Zhang, S.; Ciro, S. M. Organometallics 1994,
13, 2430.
(24) (a) Chojnowski, J .; Foruniak, W.; Stnczyk, W. J . Am. Chem.
Soc. 1987, 109, 7776. (b) Chojnowski, J .; Wilczek, L.; Fortuniak, W.;
J . Organomet. Chem. 1977, 135, 13. (c) Mayr, H.; Basso, N.; Hagen,
G. J . Am. Chem. Soc. 1992, 114, 3060.

B(C6F5)3-Catalyzed Hydrosilation
J . Org. Chem., Vol. 65, No. 10, 2000 3093
generated via hydride abstraction from Bu3SnH using
B(C6F5)3.18 By analogy to these phenomena, it is reason-
able to propose that borane is able to abstract hydride,
at least partially, from R3SiH.
Attempts to garner direct spectroscopic evidence for a
B(C6F5)3/silane interaction were only partially successful.
When samples of B(C6F5)3 and either Ph3SiH or Et3SiH
(1:1 in C6D6) are monitored by 1H NMR spectroscopy,
little perturbation in the chemical shift of the resonances
for the silane hydrogen are observed. This indicates that,
in the absence of a substrate, borane/silane adduct
formation is only partial and that its concentration is low
under these conditions (i.e., for eq 3, Keq , 1). The
resonance for the 29Si nucleus shifts from 0.5 ppm in free
Et3SiH to 1.2 ppm in the presence of 1 equiv of B(C6F5)3.
While the trend is in the expected direction,19 the change
is too small to be conclusive.
In an effort to drive the equilibrium to the right, the
19F NMR spectrum of B(C6F5)3 in neat Et3SiH was
recorded at -80 °C. The chemical shift of the para-
fluorines of the boron C6F5 rings are quite sensitive to
the coordination environment about boron.20 The value
of ∆(δm,p) for B(C6F5)3 in neat Et3SiH at room temperature
is 16.6 ppm (cf. a value of 18.3 ppm for B(C6F5)3 in
C6D620d); at -80 °C, the value falls to 11.1 ppm, which is
in the range expected for a weak Lewis base/B(C6F5)3
adduct.14 Nonetheless, even under these conditions, the
equilibrium likely favors the constituent reagents. Thus,
while hydride abstraction from silanes using the isoelec-
tronic trityl ion is irreversible (and the method of choice
for generating silylium-like ions in the condensed
phase19,21), when B(C6F5)3 is the abstracting agent, the
process is incomplete and rapidly reversible.
F igu r e 2. 300 MHz 1H NMR spectra (C6D6, room tempera-
ture) of (A) Et3SiH and (B) Et3SiH plus 10% B(C6F5)3 (10 min
of mixing).
Ta ble 2. AM1 Op tim ized P a r a m eter s for B(C6F 5)3/R3SiH
Ad d u cts
species
Ph3SiH
RB-H (Å) RSi-H (Å) qB (e)a qH (e)a qSi (e)a
1.4695
1.7119
-0.24
0.31 -0.23
0.00 -0.05
0.32 -0.24
-0.23
1.31
1.57
Ph3SiH-B(C6F5)3 1.3938
[HB(C6F5)3]-
Et3SiH-B(C6F5)3
Et3SiH
1.2164
1.3929
1.6672
1.4701
1.41
1.09
B(C6F5)3
0.50
∆Ph3SiHb
0.1774
0.1765
0.2424
0.1971
-0.19
0.01
-0.18 -0.01
0.26
0.32
∆Et3SiHb
a
b
qx represents the charge of atom x in terms of an electron.
∆
refers to the difference between the property of the adduct and
free species.
1
There is, however, convincing indirect H NMR spec-
trocopic evidence that the chemistry in eq 3 is occurring.
Although the chemical shift change for the silane hydro-
gen in Et3SiH is insignificant when B(C6F5)3 is added to
the sample, the J coupling between Si-H and the CH2
protons of the ethyl groups is lost within minutes of
borane contact with silane (Figure 2). Similarly, the 1J Si-H
coupling is no longer in evidence in these spectra. This
phenomenon is reminiscent of that observed in the 1H
NMR spectra of alcohols before and after the addition of
catalytic amounts of H+.22 In this instance, the borane
catalyzes the rapid exchange of Si-H protons by gener-
ating small amounts of “[Et3Si]+[HB(C6F5)3]-”. This is
also manifested by the observed rapid scrambling of
deuterium when a 1:1 mixture of Et3SiH and Ph3SiD is
treated with 10% B(C6F5)3 (eq 4). In the absence of ketone
AM1 level showed that the adducts of B(C6F5)3 with Ph3-
SiH and Et3SiH fall into energy minima; geometry
optimization in the absence of any restraints gave the
metrical parameters and Mulliken charges collected in
Table 2. For comparison, similar level calculations were
done on the free silanes, B(C6F5)3 and [HB(C6F5)3]-. As
expected on the basis of steric considerations, the B-H-
Si angles in these adducts are essentially 180°. The B-H
and Si-H distances are elongated relative to the values
calculated for the hydridoborate and silane species,
indicating incomplete transfer of the hydride from silane
to borane. Comparison of the atomic charges for Si and
B in the adducts, which increase and decrease relative
to the uncomplexed silane and borane, respectively,
attests to significant charge separation occurring as the
adduct forms. Higher level ab initio computations (MP2/
6-31G**) on the adduct formed from SiH4 and BH3 lead
to similar conclusions.
Finally, the magnitude of the kinetic isotope effect
observed in the B(C6F5)3 catalyzed addition of Ph3SiX (X
) H, D) to acetophenone is also consistent with hydride
abstraction. While the availability of free borane to a
large degree controls the overall rate of the reaction, in
the two reactions that are actually involved in the
hydrosilation, abstraction of the hydride from the silane
is likely rate-limiting. Thus, a primary KIE (kH/kD) of
Lewis bases (vide infra), the deuterium is essentially
equally distributed between the two silanes within
minutes of borane addition.
Further evidence for the borane/silane adduct was
obtained via computational studies. Calculations at the

3094 J . Org. Chem., Vol. 65, No. 10, 2000
Parks et al.
1.40(5) is observed for this reaction, as measured by the
experiment shown in eq 5. On the basis of zero point
× 10-4 s-1).27 The selectivities are even better in competi-
tion experiments involving the less basic substrate ethyl-
benzoate; in these reactions, no products of ester hydro-
silation are detected. This phenomenon is also observed
in mixed substrates such as that shown in entry 4, Table
1. In this ketoester, it is the more basic ketone function
which is exclusively reduced. The rate of the reaction is
consistent with the presence of a group with the basicity
of a ketone; when a further equivalent of silane is added
to the solution, the rate of silane consumption increases
dramatically as hydrosilation of the ester function ensues.
These observations imply that the relative strength of
the Lewis base plays a critical role in determining the
selectivity of the reaction as well as the overall rate. The
computational studies described above show that the
LUMO of R3Si-H-B(C6F5)3 lies along the Si-H-B axis,
with a large orbital lobe protruding from the silicon trans
to the Si-H bond (I). This would likely be the preferred
energy considerations, the maximum isotope effect to be
expected for Si-(H,D) cleavage is ∼3.8,23 so the observed
magnitude is significant. Furthermore, it is in the same
range measured for the reactions of carbocations with
various silanes. Isotope effects of 1.4-1.9 have been
reported for these reactions24 and interpreted in terms
of hydride transfer24c (eq 6) rather than initial electron
transfer followed by a hydrogen atom shift24a,b (eq 7) since
transfer of H• from silane generally leads to larger isotope
effects (>2).23 The magnitude of the isotope effect ob-
served in the present hydrosilation reaction therefore is
more consistent with direct hydride abstraction than with
a pathway involving initial electron transfer from the
silane to B(C6F5)3 followed by transfer of a hydrogen atom
(eq 8), which must be considered since it has recently
site of nucleophilic attack on the adduct and would
furthermore aid in displacement of the hydrido borate
from silicon.
The chemical viability of a ketone adduct of R3Si+ was
tested experimentally by generating the arene-stabilized
triethylsilylium ion in bromobenzene (using [Ph3C]+-
[B(C6F5)4]- according to the methodology developed by
Lambert19,21) and treating with 1 equiv of acetophenone.
In bromobenzene, the resulting species is soluble and
stable for several hours and can thus be examined
spectroscopically, although we were unable to procure its
isolation. 1H and 13C NMR spectroscopic studies are
completely consistent with clean formation of adduct 1
(eq 10). Most notably, the 13C resonance for the carbonyl
been shown that one-electron reduction of B(C6F5)3 is the
first step in its reaction with a zirconaazacyclobutane
derivative.25
In ter a ction of Su bstr a te w ith th e Bor a n e/Sila n e
Ad d u ct. As indicated in Scheme 1, once the borane/silane
complex is formed, we believe the carbonyl substrate
nucleophilically attacks the silicon center. This is indi-
cated by several competition experiments that show that
it is the most basic substrate that is reduced selectively,
albeit at a slower rate.26 For instance, while we have
established that benzaldehyde is hydrosilated at a slower
rate than acetophenone, when a 1:1 mixture of these two
substrates is hydrosilated with Ph3SiH, the benzaldehyde
is hydrosilated preferentially by a 6:1 margin (eq 9). The
observed rate constant for the consumption of silane in
this experiment (3.02(9) × 10-4 s-1) is similar to that
found for the hydrosilation of benzaldehyde itself (6.0(1)
carbon shifts downfield to 217.8 ppm from 196.8 ppm for
free acetophenone in C6D5Br, while the methyl protons
move upfield to 2.44 ppm as compared to 2.55 ppm in
unbound ketone.28 Clearly, when the counteranion is the
weakly coordinating [B(C6F5)4]- anion, ketone adducts of
trialkylsilylium ions are reasonable species and, like the
olefin adducts of R3Si+ reported by Lambert et al.,29 have
appreciable carbocationic character at the carbonyl car-
bon.
(25) Harlan, C. J .; Hascall, T.; Fujita, E.; Norton, J . R. J . Am. Chem.
Soc. 1999, 121, 7275.
(26) We assume here that the relative basicities of benzaldehyde,
acetophenone, and ethylbenzoate towards the borane/silane complex
mirror that found for B(C6F5)3 itself.14
(27) The smaller rate constant in the competition experiment is due
to the substrate inhibition phenonmenon.
(28) Hartman, J . S.; Stilbs, P.; Forsen, S. Tetrahedron Lett. 1975,
3497.
(29) Lambert, J . B.; Zhao, Y.; Wu, H. J . Org. Chem. 1999, 64, 2729.

B(C6F5)3-Catalyzed Hydrosilation
J . Org. Chem., Vol. 65, No. 10, 2000 3095
Mod e of Hyd r id e Deliver y. As mentioned above, two
routes for consummation of hydrosilation are plausible,
distinguished by whether hydride delivery to the coor-
dinated carbonyl function is performed by [HB(C6F5)3]-
(path a , Scheme 1) or by R3SiH (path b). Although we
are unable to conclusively prove which path is operative,
several lines of experimentation point strongly toward
the former path, i.e., borohydride attack of silylium
activated ketone.
That being said, reaction of acetophenone adduct 1
with a 1.5- to 2-fold excess of triethylsilane clearly
indicates that such an adduct is capable of abstracting
hydride from silane. However, as eq 11 shows, the
In these reactions, where Et3SiH is the only hydride
donor present, abstraction of hydride by the carbocation-
like ketone adducts of Et3Si+ regenerates the silylium ion
in the proximity of the newly formed silyl ether oxygen
(Scheme 2). Loss of Et3SiOSiEt3 from this species, which
is likely quite facile and irreversible,10 gives a carbocation
that rapidly abstracts hydride from silane to complete
the full deoxygenation of ketone. Because the product
distribution is so different in these reactions, the produc-
tion of silyl ether products exclusively in the B(C6F5)3-
catalyzed process suggests that the [HB(C6F5)3]- coun-
teranion is not simply a spectator but provides the
hydride necessary to complete hydrosilation of the car-
bonyl group.
Further evidence for this notion exists in the following
set of experiments. Borane B(C6F5)3 is unable to mediate
the hydrosilation of carbonyls (or the silation of alco-
hols12) when i-Pr3SiH is the silane employed, presum-
ably because the front strain associated with hydride
abstraction is too great to effect activation of this steri-
cally demanding silane. Nevertheless, when [i-Pr3Si]+-
[B(C6F5)4]-19 is used as a catalyst for acetophenone
reduction using i-Pr3SiH (2 equiv), rapid production of
ethylbenzene is observed (eq 13). Indeed, i-Pr3SiH is a
product distribution in this stoichiometric reaction is
quite different from that observed in the borane-catalyzed
process. The major acetophenone-derived product is
ethylbenzene, formed along with hexaethyldisiloxane and
arising from full deoxygenation of the ketone.10 Further-
more, these are the exclusive products of acetophenone
hydrosilation in reactions where silane is the only hy-
dride source in the system (eq 12). Whereas B(C6F5)3
superior hydride donor toward Ph3C+ in comparison to
Ph3SiH according to the scale developed by Mayr et al.24c
Thus, it is convincing evidence for hydridoborate delivery
of hydride that in the experiment shown in eq 14, PhCD-
gives the silyl ether reduction product exclusively, a 1:1
mixture of ethylbenzene/hexaethyldisiloxane and un-
reacted acetophenone are produced when [Et3Si]+[B-
(C6F5)4]-,19 [Ph3C]+[B(C6F5)4]-,30 or the diborane 1,2-
[B(C6F5)2]2C6F431 are employed as catalysts. In the latter
case, abstraction of H- from silane presumably leads to
silylium ion II, in which the hydride moiety in the
counteranion is chelated by the two borane centers32 and
is effectively unavailable for delivery to substrate; thus,
in this system “Et3Si+” is again the true catalyst.
(OSiPh3)CH3 is observed as the sole product.33
A related experiment (eq 15) is also strongly supportive
(30) (a) Bochmann, M.; Lancaster, S. J . J . Organomet. Chem. 1992,
434, C1. (b) Chien, J . C. W.; Tsai, W. M.; Rausch, M. D. J . Am. Chem.
Soc. 1991, 113, 8570.
(31) Williams, V. C.; Piers, W. E.; Clegg, W.; Collins, S.; Marder, T.
B. J . Am. Chem. Soc. 1999, 121, 3244.
(32) J ia, L.; Yang, X.; Stern, C.; Marks, T. J . Organometallics 1994,
13, 3755.
of borohydride delivery of hydride in the B(C6F5)3-
catalyzed reactions. A 1.5:0.5:0.5 mixture of acetophe-
none, (p-CH3C6H4)3SiH, and Ph3SiH was treated with a

3096 J . Org. Chem., Vol. 65, No. 10, 2000
Parks et al.
Sch em e 2
from the hydridoborate counteranion, rather than free
silane and believe that the cumulative evidence obtained
provides strong support for this proposal. This mecha-
nistic issue has implications for devising strategies aimed
at developing catalysts for performing diastereo and
enantioselective hydrosilations. For example, if path b
had been operative, a chiral silane (stoichiometric re-
agent) would presumably be necessary. However, the fact
that path a most likely predominates provides hope that
a chiral borane catalyst may be able to effect these
transformations asymmetrically; efforts are currently
being directed toward this goal.
Exp er im en ta l Section
Gen er a l. General procedures were as described previ-
ously.12,14 Substrates were purchased from Aldrich-Sigma and
dried and purified prior to use. Equilibrium constants for
adduct formation between B(C6F5)3 and the para-substituted
catalytic amount of B(C6F5)3, and the reaction was
monitored by H NMR spectroscopy. An excess of aceto-
1
phenone was employed to suppress B(C6F5)3 mediated
H/D scrambling between the silanes,33 which we have
shown is facile in the absence of Lewis bases (eq 4). Initial
spectra revealed peaks attributable to the products
1
acetophenone substrates were measured using H NMR spec-
troscopy as described previously.14 NMR data for the products
of hydrosilation of the substrates in entries 1-3, Table 1, were
reported in the Supporting Information deposited along with
the original communication.8 Data for the aldehyde products
of entries 7,39 8,40 and 941 were identical to reported data for
these compounds. LRMS was done on a Varian Star 3400 CX
GC with a Varian Saturn 2000 mass spec detector (electron
impact ionization); HRMS was performed on a Kratos MS-80
spectrometer.
34
PhCH(OSitol3)CH3 and PhCD(OSiPh3)CH3 in a ∼4:1
ratio.35 As the reaction progressed to completion, the two
silyl ethers were ultimately present in a 1:1 ratio. The
absence of the other possible isotopomeric products, i.e.,
PhCD(OSitol3)CH3 and PhCH(OSiPh3)CH3 strongly ar-
gues against hydride donation by silane in the B(C6F5)3-
catalyzed reaction, which would be expected to yield a
mixture of all four of these possible products.
Gen er a l P r oced u r e for Keton e a n d Ald eh yd e Hyd r o-
sila tion s. A stock Ph3SiH/2 mol % B(C6F5)3 solution was
prepared by dissolving 2.00 g of triphenylsilane and 0.079 g
of tris(pentafluorophenyl)borane in toluene in a 10.0 mL
volumetric flask. For each substrate, 0.5 mL of the stock
solution was placed into a 3 mL vial containing a magnetic
stir bar. For liquid substrates, 0.5 mL of toluene was added
to the vial prior to sealing with a rubber septum; reaction was
initiated by injection of 1 equiv of substrate. For solid
substrates, 1 equiv was dissolved in 0.5 mL of toluene, and
this solution was injected into the borane/silane mixture.
Progress of the reaction was monitored by GC analysis of the
reaction mixture. Upon completion of reaction the crude
reaction mixture was introduced directly onto a silica gel liquid
chromatography column and purified by column chromatog-
raphy using hexanes/ethyl acetate mixtures.
Gen er a l P r oced u r e for Ester Hyd r osila tion s. B(C6F5)3
(0.0023 mmol) was added to a 25 mL round-bottom flask
equipped with a magnetic stir bar. Benzene (3 mL) was added
to the flask followed by the ester (2.15 mmol), and then the
flask was sealed with a rubber septum. A solution of Ph3SiH
(2.4 mmol) in benzene (3 mL) was prepared and stored in a
syringe. The flask was clamped to a stir plate and the solution
was rapidly stirred. The silane solution was slowly added
dropwise over a period of 5 min to the ester solution. The
reaction was monitored by TLC and was quenched by the
addition of THF (9 mL) and H2O (1 mL) after all of the starting
material was consumed. The heterogeneous solution was
cooled to -8 °C, and then TBAF (5.4 mmol) in THF was added
to the stirred solution via syringe. The reaction mixture was
stirred at -8 °C for 30 min, then allowed to warm to room
temperature, and stirred for an additional 3 h. The reaction
mixture was poured into saturated NH4Cl solution (25 mL),
and the aqueous layer was extracted with Et2O (4 × 15 mL).
The combined organic extracts were dried over anhydrous
MgSO4, and then the solid was removed by suction filtration.
The solvent was removed in vacuo, giving the crude product.
Purification was performed by column chromatography using
Con clu sion s
Taken together, the above kinetic, competition, inter-
mediate modeling (i.e., the preparation of 1), isotopic
labeling, and crossover experiments provide solid support
for the silane activation mechanism by which the B(C6F5)3/
R3SiH hydrosilating system operates. Although this
Lewis acid binds to carbonyl functions in the usual
sense,14 the actual mechanism of addition of Si-H to the
CdO double bond follows the general path depicted in
Scheme 1. Contrary to expectations based on the general
perception that reactions such as this proceed through
the LA/carbonyl adduct, the key step in this reaction is
borane activation of the Si-H bond rather than CdO.
Although the results of this study caution against sweep-
ing generalizations, it appears that this type of mecha-
nism is operative in other B(C6F5)3-mediated reactions,
such as the silation12 and deoxygenation10 of alcohols, the
hydrosilation of imines,36 the hydrostannation of carbonyl
substrates,37 and allylsilation and stannation reac-
tions.38
We have taken pains to show that the final step in the
reduction involves addition of H- to the carbonyl carbon
(33) Although a trace of the proteo silyl ether is observed, this arises
from the trace amounts of Ph3SiH present in the d1-triphenylsilane.
It should also be noted that the rate of B(C6F5)3-catalyzed H/D
scrambling is very slow when one of the partners is i-Pr3SiH.
Furthermore, the rate of scrambling for silanes which are susceptible
to exchange is slowed considerably when a Lewis base is present.
(34) These products were easily distinguishable at 400 MHz and
were positively identified by separate synthesis and characterization.
(35) This observation is an indication that (p-CH3C6H4)3SiH is a
superior hydride donor in comparison to Ph3SiD, as expected on the
basis of both the isotope effect and the electronic properties of
p-CH3C6H4- vs C6H5-.
(36) Sonmor, E.; Blackwell, J . M.; Piers, W. E. unpublished results.
(37) (a) Ooi, T.; Uraguchi, D.; Maruoka, K. Tetrahedron Lett. 1998,
39, 8105. (b) Ooi, T.; Uraguchi, D.; Kagoshima, N.; Maruoka, K. J .
Am. Chem. Soc. 1998, 120, 5327. (c) Maruoka, K.; Ooi, T. Chem. Eur.
J . 1999, 5, 829.
(38) Lambert, J . B.; Zhao, Y.; Wu. H.; Tse, W. C.; Kuhlmann, B. J .
Am. Chem. Soc. 1999, 121, 5001.
(39) Ceruti, M.; Pegani, I.; Fochi, R. Synthesis 1987, 1, 79.
(40) Le Borgne, J .-F. J . Organomet. Chem. 1976, 122, 123.
(41) Koch, S. S. C.; Chamberlin, A. R. J . Org. Chem. 1993, 58, 2725.

B(C6F5)3-Catalyzed Hydrosilation
J . Org. Chem., Vol. 65, No. 10, 2000 3097
hexanes/ethyl acetate eluants, followed by distillation at
reduced pressure.
128.1 (m, 2F, Fortho); 143.6 (m, 1F, Fpara); 160.2 (m, 2F, Fpara).
19F NMR spectral data for 2 mg of B(C6F5)3 in 500 µL of Et3SiH
at -80 °C: 130.1 (m, 2F, Fortho); 150.3 (m, 1F, Fpara); 161.4 (m,
2F, Fmeta).
Com p u ta tion a l Stu d ies. All structures were built using
the SPARTAN molecular modeling program.42 All structures
were geometry-optimized with no constraints using semi-
empirical methods at the AM1 level of theory. Charges were
calculated at the AM1 level using Mulliken population analy-
sis.
Gen er a l P r oced u r e for Kin etic Stu d ies. Solutions con-
taining known concentrations of Ph3SiH and the standard
fluorene were prepared in toluene. Next, 1 µL of each solution
was injected into Et3N (100 µL), the resulting solution was
analyzed by GC using a standard temperature program, and
the ratio of areas for the two peaks were obtained. This
procedure was repeated three times for each concentration of
Ph3SiH, and the average of the area ratios was calculated. The
averages were used to obtain the response factor of Ph3SiH
with respect to fluorene by plotting the concentration of
Ph3SiH versus the ratio of areas multiplied by the concentra-
tion of fluorene. The slope of the line obtained corresponds to
the response factor of the system. Linear regression analysis
of the data gave R2 ) 0.997 and the response factor, RF )
0.68(2).
The kinetics of hydrosilation were followed quantitatively
by GC by monitoring the loss of Ph3SiH. A stock solution
consisting of Ph3SiH, 2 mol % B(C6F5)3, and fluorene was
prepared in the drybox by dissolving 2.00 g of Ph3SiH. 0.079
g of B(C6F5)3, and 0.5-1.0 g of fluorene in toluene in a 10.0
mL volumetric flask. Next, 0.5 mL of the stock solution (0.384
mmol Ph3SiH, 0.00772 mmol B(C6F5)3) was measured via
syringe and placed into a 1 dram vial containing a magnetic
stir bar, and the vial was sealed with a rubber septum. One
equivalent of substrate was dissolved in 0.5 mL of toluene,
and this solution was injected into the vial (total volume )
1.0 mL). After completion of addition of the substrate solution,
the time was recorded. Aliquots (1 µL) were extracted from
the reaction mixture with a microsyringe at various time
intervals and injected into a vial containing 100 µL of Et3N.
The moment at which the aliquot was injected into the Et3N
quenching solution was recorded. The quenched samples were
subsequently analyzed by gas chromatography. The ratios of
integrated area of Ph3SiH to fluorene (internal standard) were
calculated for the samples and then converted to concentra-
tions of Ph3SiH using the standardization curve obtained as
described above. The natural logarithm of the silane concen-
tration was plotted versus the time of quenching (in seconds)
giving a straight line over several half-lives.
Mea su r em en t of th e Deu ter iu m Kin etic Isotop e Ef-
fect. Ph3SiH (57.0 mg, 0.218 mmol), Ph3SiD (57.0 mg, 0.218
mmol), and B(C6F5)3 (2.0 mg, 0.0039 mmol) were dissolved in
C6D6 (0.6 mL) in a dry NMR tube. Acetophenone (25.4 µL,
0.218 mmol) was added to the solution via syringe. The
resulting solution was vigorously shaken to ensure complete
mixing,and then the 1H NMR spectrum was obtained. A
mixture of products PhCH(OSiPh3)CH3 and PhCD(OSiPh3)-
CH3 was observed, and the relative ratio of the two species
was found to be 1.4:1, respectively, by integration, correspond-
ing to a kinetic isotope effect found to be kH/kD ) 1.4(5).
In Situ Gen er a tion of [P h (CH3)CdO•SiEt3]+[B(C6F 5)4]-
(1). [Ph3C]+[B(C6F5)4]- (0.24 mg, 0.026 mmol) was suspended
in C6D6, and Et3SiH (4.2 µL, 0.026 mmol) was added via
syringe. Upon shaking, a clear liquid clathrate separated from
the benzene, and 1H NMR analysis revealed the presence of
Ph3CH. To this sample was added acetophenone (3.0 µL, 0.026
mmol) via syringe. Upon shaking, the liquid clathrate layer
turned light orange. The C6D6 was decanted from the oil, and
the oil was washed twice with 0.2 mL of C6D6. The oil was
subsequently dissolved in C6D5Br and analyzed by NMR
spectroscopy. 1H NMR: 7.68 (m, 2H); 7.44 (m, 1H); 7.09 (m,
2H); 2.44 (s, 3H); 0.7-1.3 (m, 15H). 13C{1H} NMR: 217.8,
144.7, 134.6, 131.9, 130.9, 149.1 (J C-F ) 242.6 Hz), 139.0 (J C-F
) 245.7 Hz), 137.1 (J C-F ) 243.6 Hz), 25.6, 6.1, 5.1. 19F NMR:
-131.9 (Fortho); -162.1 (Fpara); -166.1 (Fmeta). 11B{1H} NMR:
-16.8.
Deoxygen a tion of Acetop h en on e w ith Et3SiH w ith
Va r iou s Ca ta lysts (eq 12). Et3SiH (48 µL, 0.3 mmol) was
dissolved in C6D6, and catalyst (0.006 mmol) was added. Then
PhC(O)Me (35 µL, 0.3 mmol) in C6D6 was added via syringe,
and the reaction was followed by 1H NMR spectroscopy. In
each case, this analysis showed unreacted PhC(O)Me and
production of PhCH2CH3 with no PhCH(OSiEt3)Me observable.
Deoxygen a tion of Acetop h en on e w ith i-P r 3SiH Ca ta -
lyzed by [i-P r 3Si]+[B(C6F 5)4]-. To an orange-red, two-phase
solution of [Ph3C]+[B(C6F5)4]- (24 mg, 0.03 mmol) in C6D6 was
added i-Pr3SiH (60 µL, 0.30 mmol). The mixture remained
biphasic but became pale yellow in color. To this mixture was
added PhC(O)Me (15 mg, 0.13 mmol). 1H NMR analysis of the
top bright yellow layer within 5 min of mixing showed that
all PhC(O)Me was consumed and that PhCH2CH3 was formed
quantitatively.
Eth yl 3-Tr ip h en ylsiloxybu ta n oa te (Ta ble 1, en tr y 4).
The general procedure described above was used to prepare
this product as a white solid in 84% yield after column
1
chromatography. IR (KBr): 1735 (vs). H NMR (CDCl3): 7.61
(dd, J ) 1.9 Hz and J ) 7.5 Hz, 6H); 7.45-7.30 (m, 9H); 4.44
(ddq, J ) 5.7 Hz, J ) 6.1 Hz, and J ) 7.2 Hz, 1H,); 3.99 (ABXq,
J ) 7.1 Hz and J AB ) 10.8 Hz, 1H); 3.95 (ABXq, 1H); 2.60
(ABXq, J ) 7.2 Hz and J AB ) 14.7 Hz, 1H); 2.41 (ABXq, 1H);
1.21 (d, J ) 6.1 Hz, 3H); 1.14 (t, J ) 7.1 Hz, 3H). 13C NMR
(CDCl3): 171.2, 134.6, 135.5, 129.9, 127.8, 67.1, 60.2, 44.6, 23.7,
14.1. HRMS calcd for C24H26O3Si: 390.1651. Found: 390.1632.
Anal. Calcd for C24H26O3Si: C, 73.81; H, 6.71. Found: C, 73.91;
H, 6.54.
Hyd r osila tion of Acetop h en on e Usin g i-P r 3SiH/P h 3SiD
Mixtu r e. To PhC(O)Me (6 mg, 0.05 mmol), Ph3SiD (13 mg,
0.05 mmol), and i-Pr3SiH (10 µL, 0.05 mmol) dissolved in C6D6
was added B(C6F5)3 (5 mg, 0.01 mmol) as a solution in C6D6.
NMR analysis of the product mixture showed the formation
of PhCD(OSiPh3)Me and only trace quantities of PhCH-
(OSiPh3)3 interpreted to arise from a small amount of Ph3SiH
contaminant.
E t h yl 2-Tr ip h en ylsiloxy-1-cycloh exen eca r b oxyla t e
(Ta ble 1, en tr y 5). The general procedure described above
was used to prepare this product as a white solid in 78% yield
1
after column chromatography. IR (neat): 1713 (s). H NMR:
7.85 (m, 6H); 7.18 (m, 9H); 3.93 (q, J ) 7.0 Hz, 2H); 2.35 (m,
2H); 2.01 (m, 2H); 1.19 (m, 4H); 0.89 (t, 3H). 13C NMR 167.0,
157.9, 136.0, 130.3, 128.1, 134.8, 110.7, 59.5, 32.7, 25.9, 22.9,
22.4, 14.4. HRMS calcd for C27H28O3Si - C2H5O: 383.1467.
Found: 383.1477. Exact mass calcd for C27H28O3Si - C6H5:
351.1416. Found: 351.1420.
H yd r osila t ion of Acet op h en on e Usin g (t olyl)3SiH /
P h 3SiD Mixtu r e. Ph3SiD (13 mg, 0.05 mmol) and (p-CH3C6H4)3-
SiH (15 mg, 0.05 mmol) were dissolved in C6D6 in an NMR
tube. Acetophenone (18 µL, 0.15 mmol) was added followed
by a solution of B(C6F5)3 in C6D6 (5 mg, 0.01 mmol in 100 µL).
cis-2-Meth ylcycloh exa n oxytr ip h en ylsila n e (Ta ble 1,
en tr y 6). The general procedure described above was used to
prepare this product as a white solid in 75% yield after column
chromatography. 1H NMR: 7.80 (m, 6H); 7.20 (m, 9H); 4.03
(m, 1H); 1.90-1.10 (m, 9H); 0.96 (d, J ) 6.7 Hz, 3H). 13C
NMR: 135.7, 136.0, 130.1, 128.1, 73.5, 37.2, 33.0, 29.7, 24.9,
21.5, 17.8. HRMS calcd for C25H28OSi: 372.1909. Found:
372.1894. Anal. Calcd for C25H28OSi: C, 80.59; H, 7.58.
Found: C, 80.46; H, 7.29.
1
A H NMR spectrum was immediately obtained (1 min after
mixing), and spectra were collected periodically over a 75 min
time period. 1H NMR analysis revealed that only PhCH-
(OSitol3)Me and PhCD(OSiPh3)Me were formed. The identity
of the former compound was confirmed by independent syn-
thesis via B(C6F5)3-catalyzed silation of PhCH(OH)Me using
Evid en ce for Et3SiH/B(C6F 5)3 In ter a ction . 19F NMR
spectral data for 2 mg of B(C6F5)3 in 500 µL of Et3SiH at rt:
(42) Spartan Version 3.1; Wave function Inc.: Irvine, CA.

3098 J . Org. Chem., Vol. 65, No. 10, 2000
Parks et al.
(tol)3SiH. 1H NMR (C6D6): 7.75 (d, 6H, J ) 8.0 Hz); 7.39-
7.35 (m, 2H), 7.17-7.00 (m, 9H); 5.19 (q, 1H, J ) 6.3 Hz); 2.08
(s, 9H); 1.46 (d, 3H, J ) 6.3 Hz). 13C NMR (C6D6): 147.1, 140.2,
136.4, 132.5, 129.4, 127.5, 126.2, 72.6, 27.8, 22.0 (one aromatic
carbon missing).
W.E.P.. J .M.B. thanks the Izaak Walton Killam Foun-
dation and the University of Calgary for Fellowship
support. The authors thank Mr. Eric Sonmor for techni-
cal support. W.E.P. also thanks the Alfred P. Sloan
Foundation for a Research Fellowship (1996-00).
Ack n ow led gm en t. Financial support for this work
was provided by the Natural Sciences and Engineering
Council of Canada in the form of a Research Grant to
J O991828A