Hydrosilylation of alkenes and ketones catalyzed by nickel(II) indenyl complexes

Article abstract of DOI:10.1139/v03-133

Abstraction of Cl^-1 from the complexes (indenyl)Ni(PPh ₃)Cl generates cationic species that are effective precatalysts for the hydrosilylation of some olefins and ketones. For instance, the mixture of (1-Me-indenyl)Ni(PPh₃

Full text of DOI:10.1139/v03-133

1
299
Hydros ilylation of alke ne s and ke tone s catalyze d
by nicke l(II) inde nyl comple xe s 1
Fré dé ric -Ge orge s Fonta ine , Re né -Vie t Nguye n, a nd Da vit Za rga ria n
–
Abstract: Abstraction of Cl from the complexes (indenyl)Ni(PPh )Cl generates cationic species that are effective
3
precatalysts for the hydrosilylation of some olefins and ketones. For instance, the mixture of (1-Me-indenyl)Ni(PPh )Cl
3
and NaBPh (or methylaluminoxane) reacts at room temperature with ca. 100 equiv. each of PhSiH and styrene to
4
3
produce [1-phenyl-1-ethyl](phenyl)silane, PhCH(CH )(SiPhH ), in 50%–80% yield. The same system can also catalyze
3
2
the hydrosilylation of 1-hexene and norbornene, but the products arising from these substrates consist of mixtures of
regio- and stereoisomers. On the other hand, ketone hydrosilylation is regiospecific, giving the corresponding silyl
ethers in high yields. A number of experimental observations have indicated that the initially generated Ni-based cation
is not the catalytically active species. Indeed, the cationic initiators may be replaced by LiAlH or AlMe , which gen-
4
3
erate the corresponding Ni-H or Ni-Me derivatives, respectively. Moreover, the observed regioselectivity for the addi-
tion of PhSiH to styrene (i.e., predominant addition of the silyl fragment to the α-C) is opposite of what would be
3
expected if the reaction mechanism involved carbocationic intermediates. A new mechanism is proposed in which the
–
active species is a Ni-H species originating from the transfer of H from PhSiH to the initially generated Ni cation.
3
Key words: hydrosilylation, nickel indenyl complexes, cationic complexes, hydride intermediates.
Résumé : L’abstraction des ions chlorures des complexes (indényle)Ni(PPh )Cl génère des espèces cationiques qui sont
3
des précatalyseurs efficaces pour la réaction d’hydrosilylation de certaines oléfines et de certaines cétones. Par exemple,
le mélange de (1-Me-indényle)Ni(PPh )Cl et de NaBPh (ou de méthylaluminoxane) réagit à la température de la pièce
3
4
avec environ 100 équivalents de PhSiH et 100 équivalents de styrène pour donner le [1-phényl-1-éthyl](phényl)silane,
3
PhCH(CH )(SiPhH ) avec un rendement allant de 50 % à 80 %. Le même système peut également catalyser l’hydrosi-
3
2
lylation du 1-hexène et du norbonène, mais on obtient dans ces conditions des mélanges de régio- et de stéréo-isomères.
Par ailleurs, l’hydosilylation des cétones est régiospécifique, et conduit aux éthers silylés correspondants avec des ren-
dements élevés. Plusieurs observations expérimentales ont révélé que le cation à base de nickel généré initialement
n’est pas l’espèce active catalytiquement. Effectivement l’initiateur cationique peut-être remplacé par le LiAlH ou par
4
le AlMe qui génèrent respectivement les dérivés Ni-H ou Ni-Me correspondants. De plus, la régiosélectivité observée
3
lors de l’addition du PhSiH sur le styrène (i.e. l’addition prédominante du fragment silylé sur le carbone en α) est à
3
l’opposé de ce à quoi on pourrait s’attendre si le mécanisme réactionnel impliquait un carbocation intermédiaire. On
–
propose un nouveau mécanisme dans lequel l’espèce réactive est le Ni-H provenant du transfert de H à partir du
PhSiH vers le cation nickel généré initialement.
3
Mots clés : hydrosilylation, complexes indényle-nickel, complexes cationiques, hydrures intermédiaires.
[
Traduit par la Rédaction] Fontaine et al. 1306
Introduc tion
reductive elimination to give the final product. The main el-
ements of the Chalk–Harrod mechanism have served as use-
ful guiding principles for the development of numerous
hydrosilylation catalysts over the past four decades (2–4).
Based on their pioneering studies on the mechanism of the
olefin hydrosilylation reaction, Harrod and Chalk proposed,
in 1965, a general mechanistic scheme involving hydri-
do(silyl) intermediates of the type L M(H)(SiR ) (Scheme 1)
We became interested in the hydrosilylation reaction dur-
n
3
ing our studies on the oligomerization of PhSiH catalyzed
3
(
1). According to this scheme, insertion of the olefin into the
by Ni(II) indenyl precursors (5). The presumed involvement
M—H bond generates an alkyl intermediate that undergoes
of Ni-H or Ni-SiR intermediates in these reactions sug-
3
gested that the same Ni-indenyl systems might also promote
the hydrosilylation of olefins or ketones if these unsaturated
substrates were present in the reaction mixture. This asser-
tion was borne out by a series of tests, which indicated that
the presence of certain olefins effectively inhibited the
Si—Si bond formation step in favour of the hydrosilylation
reaction. These initial results prompted us to carry out a lit-
erature survey, which revealed that a number of related Ni
complexes are known to promote the hydrosilylation of ole-
Received 15 May 2003. Published on the NRC Research
Press Web site at http://canjchem.nrc.ca on 21 October 2003.
F.-G. Fontaine, R.-V. Nguyen, and D. Zargarian.2
Département de chimie, Université de Montréal, Montréal,
QC H3C 3J7 Canada.
1This article is part of a Special Issue dedicated to Professor
John Harrod.
2
Corresponding author (e-mail: zargarian.davit@umontreal.ca).
Can. J. Chem. 81: 1299–1306 (2003)
doi: 10.1139/V03-133
© 2003 NRC Canada

1
300
Can. J. Chem. Vol. 81, 2003
Scheme 1.
(PhCH CH SiPhH ) were barely perceptible at ca. 2.7 ppm.
2
2
2
These assignments are consistent with the reported data for
these products (9).
Subsequent experiments showed that the hydrosilylation
of styrene promoted by IndNi(PPh )Cl–MAO is fairly slug-
3
gish, requiring several hours of reaction time for ca. 70%
yield (see runs 1–4 of Table 1). Curiously, using more cata-
lyst or a higher temperature did not accelerate the catalysis
(
compare runs 5 to 2, and 6 to 4); on the other hand, com-
bining these two variations did increase the rate, but did not
affect the yield (compare run 7 to runs 6 and 5). It should be
noted here that the higher temperature reactions tend to give
higher proportions of the β-isomer (ca. 5%–10%); the signif-
icance of this observation for the probable mechanism of the
catalytic reaction will be discussed later. The reaction rate
also improved upon using the 1-i-Pr-Ind analogue of 1 as
precatalyst (compare run 8 to run 2), whereas using the
fins, as follows: the dimeric complex {CpNi(µ-CO)}2
catalyzes the addition of HSiCl to styrene to give PhCH-
3
(
SiCl )Me (6); CpNi(PPh )Ph reacts with butadiene and
3 3
Me SiH to give a mixture of products arising from 1,4-
3
hydrosilylation (MeCH=CHCH SiMe ), coupling–hydro-
silylation (2,6-octadienyltrimethylsilane), and dimerization–
PMe analogue of 1 did not offer any advantage (compare
run 9 to run 5).
Although the precise role of MAO in the present system is
not known yet, our previous studies (10) have shown that a
small excess of MAO (ca. 5 equiv.) serves primarily to
2
3
3
cyclization (1,5-cyclooctadiene) (7); CpNi(PR )R′ reportedly
3
catalyzes the hydrosilylation of methyl acrylate and allyl
formate with HSiCl (8). Although little is known about the
3
way these systems work, their mechanisms might involve in-
termediates similar to those operating in our systems.
Thus, we undertook to investigate the effectiveness of our
Ni(II) indenyl complexes in the hydrosilylation reaction and
examine the mechanism of these reactions. The present re-
port describes the hydrosilylation of alkenes and ketones cat-
methylate the Ni—Cl bond, while
a
large excess
(
>10 equiv.) leads to both methylation and ionization in
31
1
varying proportions. For example, P{ H} NMR analysis of
a 1:25 mixture of 1 and MAO showed the formation of an
approximately 50:50 mixture of the Ni-Me analogue of 1
(
(1-Me-Ind)Ni(PPh )Me) and the cationic complex [(1-Me-
3
alyzed by the precursors IndNi(PR )X (Ind = indenyl ligand
+
3
Ind)Ni(PPh ) ] . Previous studies have shown that the latter
3
2
and its substituted derivatives; R = Ph, Me; X= Cl, alkyl, or
positive charge).
complex forms when the highly electrophilic cation
+
[
IndNi(PPh )] is generated in the absence of suitable lig-
3
ands or nucleophilic substrates (11). We conclude, therefore,
that the MAO:Ni ratios used in our studies (ca. 10:1) convert
Re s ults a nd dis c us s ion
the Ni-Cl precursor to a mixture of the analogous Ni-Me de-
Our initial experiments focused on the reaction of PhSiH3
and styrene with a catalytic system based on the combina-
+
rivative and [IndNi(PPh )] . To determine which, if any, of
3
these in situ generated derivatives is crucial for the hydro-
silylation reaction, we tested the catalytic effectiveness of
each species separately, as follows.
tion of IndNi(PR )Cl and methylaluminoxane (MAO); the
3
selection of this Ni–MAO system was based on the knowl-
edge that it promotes the dehydrogenative polymerization of
Tests with the complex (1-Me-Ind)Ni(PPh )Me, either
PhSiH (5a). A typical experiment was carried out as fol-
3
3
preformed (12) or prepared in situ from 1 and AlMe3,
showed that it was less than half as effective as the combina-
tion of 1–MAO (compare runs 10 and 11 to run 4). The
weak activity of the Ni-Me derivative is perhaps not surpris-
ing since this compound is inert toward styrene insertion and
lows: to the toluene mixture of (1-Me-Ind)Ni(PPh )Cl (1),
3
styrene, and PhSiH (1:100:100 ratio) was added a toluene
3
solution of MAO (10 equiv. with respect to Ni); the wine-red
mixture darkened immediately, but no gas evolution was ob-
served, implying that the Si—Si bond formation reaction
reacts only very sluggishly with PhSiH (5b). On the other
(
and its concomitant formation of H gas) had been circum-
3
2
hand, the Ni-(i-Pr) derivative, which reacts faster with
vented by the presence of styrene.
PhSiH than the Ni-Me analogue (10b), showed a better
The reaction mixture was stirred for 8 h and worked-up;
analysis of the nonvolatile products by GC–MS confirmed
3
level of activity (run 12). Interestingly, the catalytic activity
of the Ni-Me complex improves somewhat when MAO is
added to the reaction medium (compare runs 10 and 11 to
run 13); this result implies that the main role of MAO in this
system is not simply methylating the Ni—Cl bond. We con-
cluded, therefore, that the reactivity of the Ni-Me complex
alone could not account for the activity levels shown by the
combination of 1 and MAO, and set out to assess the impor-
tance of ionization in the 1–MAO-catalyzed reactions.
3
that styrene had been hydrosilylated. Distillation gave a
colourless oil, which was shown to consist, almost exclu-
sively, of the α-isomer ((1-phenyl)ethylphenylsilane). This
1
assignment was based on the following features of the H
NMR spectrum of the final product (Fig. 1): the multiplet
due to the benzylic methyne proton at 2.55 ppm
(
PhCH(Me)SiH Ph), the signals due to the diastereotopic
2
2
9
SiH protons and their Si satellites at 4.35 ppm, and the
2
doublet at ca. 1.4 ppm assigned to the methyl protons
To examine the reactivity of the cationic species
3
+
(
JH-H = 7 Hz); the benzylic protons of the minor isomer
[IndNi(PPh )] in the absence of the Ni-Me derivative, we
3
3
Control experiments have shown that neither the Ni-Cl complexes nor the initiators used in this study (MAO, AlMe , NaBPh , LiAlH ) can
3
4
4
promote the hydrosilylation reaction when used alone.
©
2003 NRC Canada

Fontaine et al.
1301
1
Fig. 1. H NMR (C D ) of the product obtained from the Ni-catalyzed hydrosilylation of styrene with PhSiH . The peaks denoted by *
6
6
3
represent the satellites due to the J29Si-H coupling, while the peaks denoted by the # represent the internal standard (ca. 2.1 ppm), re-
sidual Et O (ca. 3.4 and 1.1 ppm), and SiMe .
2
4
–
generated this species in situ by the direct abstraction of Cl
not seem to improve the catalytic activity (run 5), but gave a
higher proportion of the β-isomer. Using the 1-i-Pr-Ind de-
rivative does not offer any advantages (runs 6 and 7), while
4
from 1 by NaBPh ; this approach gave results comparable
to those obtained with MAO (compare run 14 to run 4). To
our surprise, even the cation [(1-Me-Ind)Ni(PPh ) ] , which
4
+
using the PMe analogue of 1 led to considerably lower ac-
3
2
3
is normally quite inert in ligand substitution and other cata-
lytic reactions (10b), showed some reactivity in the absence
of MAO (run 15). These observations suggested that the
tivities (runs 8 and 9). The latter observation is interesting:
that the more active precatalyst is the one bearing the less
nucleophilic phosphine ligand might imply that the
electronically and coordinatively unsaturated species
hydrosilylation reaction involves a PR dissociation. Consis-
3
+
[
IndNi(PPh )] , which is generated in situ by the abstraction
tent with this possibility, the presence of added PPh3
3
–
of Cl by NaBPh or MAO, can initiate the hydrosilylation
([Ni]:PPh = 1:2) hindered the catalysis significantly (run
4
3
reaction. Thus, during the second round of the optimization
tests we focused our efforts on finding the best conditions
10). This point will be elaborated further during the discus-
sion of the mechanism of these reactions (vide infra).
for the catalytic hydrosilylation of styrene using NaBPh as
a cationic initiator, as described below.
4
The scope of the hydrosilylation reactions promoted by 1–
NaBPh has been explored briefly, as follows. Norbornene
4
The initial experiments showed that the ratio of 1:NaBPh4
used in the catalytic runs has a direct effect on the yield, ex-
and 1-hexene were hydrosilylated with PhSiH to ca. 1:1
3
mixtures of regio- (for 1-hexene) and stereoisomers (endo-
and exo-products from norbornene), while trans-β-Me-
styrene gave the α-isomer in ca. 10:1 ratio (ca. 50% overall
yield). On the other hand, cyclohexene, indene, and trans-
stilbene gave no hydrosilylation products; for the latter two
olefins, the reaction mixtures displayed broad signals in the
NMR spectra implying poly(olefin) formation, but no further
analyses were performed to confirm this possibility. Al-
cess of NaBPh giving higher yields (Table 2, runs 1–4).
4
Since only one equivalent of the initiator should be sufficient
for generating the cationic species, the need for a large ex-
cess (run 4) is presumably due to the limited solubility of
NaBPh . Even though a [Ni]:NaBPh ratio of 1:50 gave the
4
4
best yield, using such a large excess of the initiator is not
practical, and so we adopted a 1:10 ratio as a reasonable
compromise. As before, a higher reaction temperature does
though PhSiH has been used in almost all of our studies,
3
4
–
The abstraction of Cl from 1 by NaBPh is clean but generally sluggish, especially in nonpolar solvents such as toluene, in which NaBPh
4
4
–
has a very limited solubility. Although AgBF is much more efficient for Cl abstraction, its higher cost and the possibility of secondary re-
4
actions (e.g., electron transfer with the Ni complex and redistribution with PhSiH ) render it less practical than NaBPh for our purposes.
3
4
©
2003 NRC Canada

1
302
Can. J. Chem. Vol. 81, 2003
Table 1. Catalytic addition of PhSiH to styrene.*
3
$
Run
Catalyst –co-catalyst
[Ni] (%)
Time (h)
Temperature (°C)
Yield (%)
1
2
3
4
5
6
7
8
1–MAO (1:10)
1–MAO (1:10)
1–MAO (1:10)
1–MAO (1:10)
1–MAO (1:10)
1–MAO (1:10)
1–MAO (1:10)
1
1
1
1
2
1
2
1
3
5
7
16
5
16
5
25
25
25
25
25
65
65
25
28
39
42
69
38
61
66
58
(1-(i-Pr)-Ind)(PPh )Ni(Cl)–MAO (1:10)
5
3
9
(1-Me-Ind)(PMe )Ni(Cl)–MAO (1:10)
2
1
1
1
1
1
2
5
16
24
72
16
16
5
25
25
25
25
25
25
25
36
29
26
60
40
73
13
3
1
1
1
1
1
1
0
1
2
3
4
5
1–AlMe (1:1.1)
3
(1-Me-Ind)(PPh )Ni(Me)/—
3
(1-Me-Ind)(PPh )Ni(i-Pr)/—
3
(1-Me-Ind)(PPh )Ni(Me)–MAO (1:10)
3
1–NaBPh (1:10)
4
+
[(1-Me-Ind)Ni(PPh ) ] /—
3
2
*
The reactions were carried out under anaerobic conditions on NMR-scale (C D ), and the yields were determined relative to an
6 6
internal standard. The detailed procedure is described in the Experimental section.
$
1
= (1-Me-Ind)(PPh )Ni(Cl).
3
Table 2. Addition of PhSiH to styrene catalyzed by (1-R-Ind)Ni(PR ′)Cl–NaBPh .*
3
3
4
Run
Pre-catalyst
[Ni]:NaBPh :styrene:PhSiH
Yield (%)
4
3
1
2
3
4
5
6
1
1
1
1
1
1:1:50:50
1:2:50:50
1:10:50:50
1:50:50:50
27
50
69
86
64
50
$
1:10:50:50
(1-(i-Pr)-Ind)Ni(PPh )Cl
1:10:100:100
1:10:50:50
1:10:100:100
1:10:50:50
1:10:50:50
3
7
8
9
(1-(i-Pr)-Ind)Ni(PPh )Cl
61
10
36
19
3
(1-Me-Ind)Ni(PMe )Cl
3
(1-Me-Ind)Ni(PMe )Cl
3
1
0
1 + 2PPh3
*Unless otherwise indicated, all runs were carried out under anaerobic conditions in benzene at 25 °C for
5
h. Details of the procedure and yield determination are given in the Experimental section.
This run was carried out at 65 °C.
$
the reactivity of other silanes was also examined briefly.
ert towards the insertion of olefins and must, therefore, react
Thus, we found that tri-substituted silanes such as Et SiH
first with PhSiH . For simplicity, we begin the mechanistic
3
3
and (EtO) SiH are inactive in the hydrosilylation of styrene,
discussion with the latter systems.
3
but Ph SiH does add to styrene to give the products
2
2
The complexes (1-Me-Ind)(PR )Ni(Me) are known to con-
3
PhCH(CH )(Ph SiH) and PhCH CH (Ph SiH) in ca. 60:40
3
2
2
2
2
vert PhSiH (without a co-catalyst) to (PhSiH) (R = Me, Ph;
3
n
ratio (ca. 28%–48% overall yield). The inertness of tri-
substituted silanes is presumably due to steric hindrance.
Finally, acetophenone and 2-nonanone were hydrosilylated
n = 3–16) (5b). Whereas the reactions involving the PPh an-
3
alogue were quite sluggish, those of the PMe analogue pro-
3
ceeded at a conveniently rapid rate and were, therefore,
subjected to detailed kinetic and D-labeling studies. On the
basis of these studies, we have proposed that the oligo-
very efficiently to the corresponding PhH Si–ethers, giving
2
nearly quantitative conversions by NMR.
merization of PhSiH is initiated by a concerted, σ-bond me-
3
Mechanistic considerations
The results described above show that the hydrosilylation
reactions can be promoted by the in situ generated species
tathesis reaction as opposed to an oxidative addition –
reductive elimination sequence. Analysis of the side prod-
ucts of this reaction pointed to the formation of methane and
1-(SiPhH )-3-Me-Ind, but not PhMeSiH , suggesting that the
+
(
Ind)(PPh )Ni(Me) or [(Ind)(PPh )Ni] , the latter being more
3
3
2
2
effective in most cases. This section considers the various
ways in which these two species can initiate the catalysis. It
initial reaction leads to a Ni-silyl intermediate, as shown in
Scheme 2. The putative Ni-SiPhH intermediate would then
2
should be emphasized at the outset that whereas
react further with PhSiH to initiate the oligomerization pro-
cess, while its decomposition (by reductive elimination)
3
+
[
(Ind)(PPh )Ni] can, in principle, react with either the ole-
3
fin or the silane, the Ni-alkyl derivatives are known to be in-
could form 1-(SiPhH )-3-Me-Ind (Scheme 2) (5b).
2
©
2003 NRC Canada

Fontaine et al.
1303
Scheme 2.
Scheme 3.
To determine if the pathway depicted in Scheme 2 is also
followed by the PPh analogue, we did a similar analysis (by
3
1
TOCSY H NMR) of the products of the reaction of (1-Me-
Ind)(PPh )Ni(Me) with PhSiH . This study revealed the
3
3
presence of traces of PhMeSiH2 and 3-Me-Ind, but not
-(SiPhH )-3-Me-Ind. We infer from these observations
1
2
that, unlike its PMe3 analogue, the precursor (1-Me-
Ind)(PPh )Ni(Me) reacts with PhSiH3 to eliminate
3
PhSiMeH (instead of methane) and form the hydride deriv-
2
promote the observed hydrosilylation reactions. A recent ex-
ample of such Lewis-acid-catalyzed hydrosilylation has been
reported by Gevorgyan and co-workers (14) who have
shown that 5–10 mol% of B(C F ) can catalyze the hydro-
ative (eq. [1]; [Ni] = (1-Me-Ind)Ni(PPh )). The latter could
3
then react with styrene to initiate the hydrosilylation cataly-
sis, while its decomposition (by reductive elimination)
6
5 3
would produce 3-Me-Ind (instead of 1-(SiPhH )-3-Me-Ind)
2
silylation of a range of olefins with various aryl- or alkyl-
silanes. These authors have drawn on the findings of
Lambert et al. (15) and Piers and co-workers (16) to argue,
quite convincingly, that these B(C F ) -catalyzed reactions
(
eq. [2]). The precise reasons for this difference in reactivity
between the PMe and PPh derivatives are not known with
3
3
certainty. Modeling studies suggest that the greater steric
6
5 3
bulk of the PPh derivative might favour the transfer to the
+
–
3
are initiated by the formation of [R Si] [HB(C F ) ] . The
silylium cation R Si is believed to add to the C=C to gener-
3
6 5 3
Ni centre of the less-hindered side of PhH Si-H (i.e., the H,
+
2
3
see Scheme 3), but electronic factors cannot be ruled out
ate a carbocationic intermediate, which abstracts a hydride
(
the more electron-rich PMe analogue might stabilize the
–
3
from the anion [HB(C F ) ] to release the final product.
6
5 3
Ni—Si bond).
The regiochemistry of the hydrosilylation products obtained
by Gevorgyan and co-workers is consistent with a carbo-
cationic mechanism; for instance, styrene is converted to the
[
[
1]
PhSiH + [Ni]-Me → PhMeSiH + [Ni]-H
3 2
2]
3-Me-Ind + (PPh ) Ni(0) ← [Ni]-H
PhCH2CH2SiR3,
presumably via the intermediate
3
n
+
[
PhC HCH SiR ].
By analogy to Gevorgyan’s proposed abstraction of H by
2 3
→
hydrosilylation
–
To test the likelihood that a Ni-H derivative is involved in
the hydrosilylation catalysis promoted by the present system,
B(C F ) , we considered a scenario involving the abstraction
6
5 3
–
+
of H from PhSiH by [IndNi(PPh )] to give IndNi(PPh )H
3 3 3
we tested the effectiveness of LiAlH as co-catalyst (or initi-
and a silylium species; the main question is whether the ap-
4
+
–
ator, (instead of MAO or NaBPh ) in the hydrosilylation of
titude of the species [IndNi(PPh ] for abstracting H from
4
3
styrene. Thus, stirring a toluene solution of styrene–PhSiH3–
PhSiH3 can be assumed to be comparable to that of
B(C F ) . This question was investigated by the following
1
–LiAlH (100:100:1:2.5) at room temperature for 16 h re-
4
6 5 3
sulted in the usual product (PhCH(CH )(PhSiH )) in 41%
NMR experiments that allowed some measure of the relative
Lewis acidities of these electrophiles. First, monitoring mix-
tures of IndNi(PPh )Cl and B(C F ) (1:5) showed that the
3
2
yield. This result establishes that the hydrosilylation reaction
can be catalyzed by a Ni-H intermediate generated either
3
6 5 3
–
from 1–H or by the reaction of the Ni-Me precursor and
ionization of the Ni—Cl bond is very slow (<ca. 20%–30%
ionization in 30–60 min). Next, adding B(C F ) to solutions
PhSiH3.
6
5 3
Let us now turn to discussing the mechanism of the
of IndNi(PPh )Me (1:2) did not lead to ionization, forming
3
hydrosilylation reactions initiated by the in situ generated
instead what we believe is a species featuring a Ni•••Me•••B
+
5
+
cationic species [IndNi(PPh )] . Given the highly electro-
moiety. Hence, it appears that [IndNi(PPh )] and B(C F )
3
3 6 5 3
philic nature of this species and the fact that various Lewis
acids can catalyze hydrosilylation of olefins (13), we consid-
ered the likelihood of our Ni cations acting as Lewis acids to
have similar Lewis acidities and should have similar apti-
–
tudes for abstracting H from PhSiH . Unfortunately, it has
3
not been possible to find direct evidence for the formation of
5
1
NMR evidence supporting this assertion includes the observation of a broad signal in the upfield region of the H NMR spectrum, slightly
downfield of the original Ni-Me signal at ca. –0.7 ppm, and the absence of the characteristic AB signals for the formation of the species
+
[
IndNi(PPh ) ] . Significantly, none of the B- or Al-based Lewis acids tested in our study abstract the Ni-bound phosphine ligands, imply-
3
2
ing that the Ni centre is a strong Lewis acid.
©
2003 NRC Canada

1
304
Can. J. Chem. Vol. 81, 2003
Ni-H intermediates in the mixtures of PhSiH and in situ
system: (a) the Ni-H analogue of 1 is generated by the reac-
3
+
+
generated [IndNi(PPh )] , because the putative Ni-H species
tion of PhSiH with the in situ formed [IndNi(PPh )] or the
3
3
3
7
reacts with PhSiH at a faster rate than it is produced. How-
Ni-Me analogue; (b) insertion of the olefin (or ketone) into
the Ni—H bond of the intermediate generates a new Ni-
alkyl (or Ni-alkoxy) derivative, which reacts subsequently
with another molecule of silane to form the C—Si (or O—
3
ever, we have obtained indirect evidence for the generation
of such a Ni-H species, as described below.
Given that a number of late transition metal hydride bonds
react with C—Cl bonds, especially those of chloroform, we
reasoned that if the Ni-H species could be generated in a
chlorinated solvent, chlorination might give the Ni-Cl deriv-
atives, which could be easily detected. Thus, we monitored
8
Si) bond and regenerate the Ni-H intermediate. The regio-
chemistry of the hydrosilylation reaction is determined at the
insertion step; in the case of styrene, this gives the alkyl in-
termediates Ni-CH CH Ph and (or) Ni-CH(Me)Ph. Although
2
2
the NMR spectra of two CDCl samples, one containing
the latter intermediate should, in principle, be more suscepti-
3
the independently prepared cationic complex [(1-Me-
Ind)Ni(PPh ) ][BPh ] alone, the other containing the same
ble to β-H elimination, we propose that it is in fact more sta-
3
ble because of the possibility of reverting to an η -benzyl
3
2
4
complex in addition to ca. 10 equiv. of PhSiH . The NMR
derivative, as shown in Scheme 4. It is worth noting that
3
3
spectra showed that the cationic complex is stable in CDCl3
for at least 24 h in the absence of the silane (no new peaks
Brookhart and co-workers (18) have proposed such η -
benzyl intermediates for the addition of HSiR (R = Et, i-Pr)
3
3
1
1
in the P{ H} NMR spectrum), whereas the sample contain-
to styrene catalyzed by the cationic complexes [(1,10-
3
1
1
9
ing PhSiH displayed many new P{ H} NMR signals, in-
phenanthroline)Pd(Me)L][BAr ]. The likelihood of such in-
3
4
cluding that of the complex (1-Me-Ind)Ni(PPh )Cl (31.1 ppm).
termediates being involved in our system is supported by the
3
We believe that the latter compound likely arises from the
following observations: (a) Monitoring a catalytic run by
3
1
1
31
1
reaction of the Ni-H intermediate with CDCl . The P{ H}
P{ H} NMR spectroscopy showed that the initial signal of
3
NMR spectrum of the reaction mixture also showed signals
at 16–20 ppm, which is close to the spectral region associ-
ated with the signals for phosphonium salts such as
the precatalyst 1 (ca. 31 ppm) is replaced by the signal for
free PPh (ca. –4 ppm) and a number of new signals at ca.
3
3
1
1
43–45 ppm. The latter region is associated with the P{ H}
+
–
Ph MeP I (ca. 22 ppm); these signals might be due to
signals for complexes (1-Me-Ind)Ni(PPh )R, wherein R is a
3
3
phosphonium species such as [Ph P(SiR )][BPh ]. Finally,
secondary alkyl such as i-Pr or cyclohexyl (45 ppm), and
sec-Bu or neopentyl (42–43 pm) (5b). The presence of free
3
3
4
1
the H NMR spectrum of this sample showed traces of 3-Me-
Ind, which could originate from the reductive elimination of
a Ni-H intermediate.
PPh is also consistent with the earlier observation that the
3
PMe3 precursors are less effective for promoting the
hydrosilylation reaction; (b) Recall that using higher reac-
tion temperatures or Ph2SiH2 resulted in somewhat higher
proportions of the minor regioisomer, PhCH2CH2SiPhRH
(R = H, Ph). Higher temperatures should accelerate the β-H
elimination from Ni-CH(Me)Ph, thus favouring the Ni-
CH2CH2Ph intermediate; on the other hand, the more bulky
Ph2SiH2 might be expected to react more readily with the
less bulky Ni-CH2CH2Ph intermediate.
The above observations provide indirect support for the
proposal that the in situ generated Ni cations react first with
6
PhSiH (as opposed to styrene) to form Ni-H and silylium
3
species. The next question that arises is which one of these
intermediates would react with the olefin to initiate the
hydrosilylation reaction. On this question, the regioselect-
ivity of the Ni-catalyzed reactions can help rule out the pos-
sibility that these reactions proceed by the same reaction
pathway proposed for the B(C F ) system above, i.e., the
6
5 3
+
addition of R Si to the olefin. Thus, Ni-catalyzed hydro-
3
silylation of styrene gives α-silylation, which is the opposite
Conc lus ions
of that observed with the B(C F ) system discussed above;
6
5 3
moreover, the hydrosilylation of 1-hexene gives a 50:50
mixture of α- and β-silylation instead of the exclusive α-
silylation expected from a carbocationic path. On the other
hand, the regioselectivities observed in our system are con-
sistent with a Chalk–Harrod type mechanism involving M-H
intermediates. These considerations favour the insertion of
the olefinic substrate into the Ni—H bond over reaction with
The present study has shown that combining the com-
plexes IndNi(PPh )Cl with suitable cationic initiators gives
3
rise to an efficient catalytic system for the hydrosilylation of
olefins and ketones. This system is particularly attractive for
styrene because it gives, almost exclusively, one regioiso-
mer; this regioselectivity is comparable to that reported for
organolanthanide-catalyzed hydrosilylation reaction (9). On
the other hand, Pd(II){bis(imine)} complexes give the oppo-
site regioselectivity (18), while similar Ni precursors bearing
Cp ligands give complicated mixtures of products (6, 7). Fu-
ture studies will be aimed at expanding the scope of these
hydrosilylation reactions.
the silylium species. The latter is presumably stabilized by
–
[
BPh ] and (or) the solvent (17), but our results do not shed
4
any light on the fate of this species.
Therefore, we envisage the following sequence of steps
for the hydrosilylation reactions promoted by the present
6
The direct reaction of Ni-based cations with styrene would be expected to result in the oligo- or polymerization of styrene.
7
One of the reviewers of our manuscript proposed the following alternative mechanism: the redistribution reaction involving the intermediate
+
+
+
[
IndNi(PR )] might take place to produce [IndNi(PR ) ] and phosphine-free species such as “IndNiCl” or [IndNi] ; the latter might react
3
3
2
with the hydrosilane to generate a Ni-H species. This alternative scenario is consistent with our observations and merits consideration.
The latter step likely proceeds by a concerted, σ-bond metathesis pathway as opposed to an oxidative addition – reductive elimination route
involving Ni(IV) intermediates.
8
9
This Pd-based system leads to various mixtures of products arising from hydrosilylation and dehydrogenative silylation. The proposed
mechanism involves the insertion of styrene into the Pd—SiR bond as opposed to the Pd—H bond.
3
©
2003 NRC Canada

Fontaine et al.
1305
Scheme 4.
Aldrich and used as received. The NMR spectra were re-
corded using the following spectrometers: Bruker DMX600
1
29
1
13
1
(
1
2D H- Si), Bruker AMX400 ( H at 400 MHz, C{ H} at
3
1
1
00.56 MHz, and P{ H} at 161.92 MHz), and Bruker
1
AV300 ( H at 300 MHz for TOCSY). GC–MS analyses
were carried out on a Hewlett Packard 6890 series gas
chromatograph equipped with a split mode capillary injector
and a HP 5973 mass selective detector. The following oper-
ating parameters were used: the injector and detector tem-
peratures were 250 °C; the carrier gas was hydrogen
(
2 mL/min); the column used was HP-5MS, 5% phenyl
methyl siloxane; temperature program: 40 °C for 2 min,
0 °C/min up to 140 °C, 20 °C/min up to 280 °C.
1
Details of catalytic runs
Unless otherwise specified, all reactions were carried out
under a nitrogen atmosphere. The catalytic runs were con-
ducted as NMR-scale experiments or on a larger scale in
Schlenk vessels. For the NMR-scale reactions, the following
general procedure was followed. The Ni-Cl precursor com-
plex (ca. 8 mg, 0.016 mmol) and the internal standard
(
hexamethyl benzene, ca. 15 mg, 0.09 mmol) were dissolved
in C D (0.8 mL); styrene and PhSiH (100–200 µL, 50–100
6
6
3
equiv. of each) were then added to the resultant red solution.
1
Prior to initiating the catalysis, a H NMR spectrum was re-
corded to register the integral ratios for the signals of the in-
A number of observations, including the fact that LiAlH4
can be used instead of NaBPh and MAO to activate the Ni-
4
ternal standard and the reactants (Si-H of PhSiH and vinylic
3
Cl precursor, have pointed to the involvement of the corre-
sponding Ni-H derivative (X = H) as the active intermediate
in these reactions. In the case of the reactions involving
cationic initiators, we have proposed that this Ni-H interme-
protons of styrene). The NMR sample was then taken inside
the dry box to add the initiator. For the room temperature ex-
periments, the sample was kept in an ultrasonic bath
throughout the reaction time to ensure agitation and homo-
geneity; for the high-temperature experiments, the sample
was kept in a 65 °C water bath. If required, the progress of
–
diate is generated via the transfer of H from the silane to
the coordinatively unsaturated and highly electrophilic spe-
+
+
cies [IndNi(PPh )] . This Si → Ni hydride transfer process
3
1
the catalysis was monitored periodically by H NMR spec-
is very interesting because it provides a convenient route to
the reactive Ni-H intermediate.10 Studies are currently
troscopy. The final yields were determined by comparing the
intensities of the CH signals of the product and the internal
underway with the objective of elucidating the main parame-
ters controlling this reactivity and investigating the feasibil-
ity of catalytic hydroboration reactions with the same system
3
standard. The final values were checked against a calibration
curve prepared by plotting the integral ratios obtained for
various mixtures of the internal standard (C Me ) and the
(
i.e., transfer of hydride from R BH).
6
6
2
2
main product (PhCH(PhSiH )CH ) (slope = 1.069, R =
2
3
0
.9997).
Expe rim e nta l
The larger-scale catalytic runs were conducted as follows.
Inside the dry box, the Ni precursor (ca. 10 mg) and the ini-
tiator were weighed into a Schlenk vessel containing a stir-
ring bar; the capped vessel was then taken out of the dry
box. The reactants and the solvent (benzene, unless other-
wise specified) were then added under nitrogen in rapid suc-
cession (<1 min), and the mixture was stirred at room
temperature or 65 °C. The work-up consisted of removing
the volatiles on a rotoevaporator, extracting the residues with
General
Unless otherwise specified, all manipulations were per-
formed under an inert atmosphere of N or argon using stan-
dard Schlenk techniques and a dry box. Dry, oxygen-free
solvents were employed throughout. The complexes (1-Me-
Ind)Ni(PPh )Cl (1) (12), (1-Me-Ind)Ni(PMe )Cl (19), (1-i-
2
3
3
Pr-Ind)Ni(PPh )Cl (20), (1-Me-Ind)Ni(PPh )Me (12), and
3
3
[
(1-Me-Ind)Ni(PPh ) ][BPh ] (12) were prepared according
3 2 4
to previously published procedures. PhSiH was either pur-
water–Et O, and evaporating the combined organic layers.
3
2
chased from Aldrich and used as received, or prepared from
The resulting yellow oil was distilled under vacuum to give
the product(s) as a clear oil. When little product was ob-
tained, the yield was determined based on the H NMR
spectrum of a carefully prepared mixture of the product and
hexamethyl benzene as internal standard, according to the
protocol described above.
PhSiCl according to a previously published procedure (21).
3
1
Styrene was purchased from Aldrich and treated by (a) pass-
ing it through a plug of alumina, and (b) storing over 4 Å
molecular sieves prior to use. All other reagents, including
MAO, AlMe , LiAlH , and NaBPh , were purchased from
3
4
4
10
The same Ni-H intermediate might also be accessible from the reaction of PhSiH with Ind(PPh )Ni(Me), but this reaction is generally
3
3
sluggish.
©
2003 NRC Canada

1
306
Can. J. Chem. Vol. 81, 2003
Control experiments
Jacobsen, A. Pfaltz, and H. Yamamoto. Springer, Berlin. 1999.
Chap. 7; (b) H. Nishiyama. In Comprehensive asymmetric ca-
talysis. Edited by E.N. Jacobsen, A. Pfaltz, and H. Yamamoto.
Springer, Berlin. 1999. Chap. 6.
. (a) F.-G. Fontaine, T. Kadkhodazadeh, and D. Zargarian. J.
Chem. Soc., Chem. Commun. 1253 (1998); (b) F.-G. Fontaine
and D. Zargarian. Organometallics, 21, 401 2002.
. P. Svoboda, P. Sedlmeyer, and J. Hetflejs. Coll. Czech. Chem.
Commun. 38, 1783 (1973).
. S. Takahashi, H. Shibano, H. Kojima, and N. Hagihara.
Organometal. Chem. Syn. 1, 193 (1970).
. Z. Lou and L. Zhang. Fenzi Cuihua (Molecular catalysis,
China), 8, 468 (1994).
To ascertain that the initiators used in our studies (MAO,
NaBPh4, LiAlH4, and AlMe3) do not promote the
hydrosilylation reactions in the absence of Ni complexes, we
prepared mixtures of styrene, PhSiH , and the initiator (in a
molar ratio of 1:1:0.2) and analyzed them by H NMR spec-
troscopy. No hydrosilylation product was detected in any of
the experiments, and the ratio of styrene and PhSiH re-
mained unchanged. When AgOTf was tested with the sub-
strates, gas evolution was observed and analysis of the
sample showed that a small degree of silane redistribution
5
3
1
6
7
8
9
3
1
had taken place (e.g., PhH Si-SiPhH was detected by H
NMR).
2
2
. P.-F. Fu, L. Brard, Y. Li, and T.J. Marks. J. Am. Chem. Soc.
1
17, 7157 (1995).
Ac knowle dgm e nts
10. (a) L.F. Groux, D. Zargarian, L.C. Simon, and J.B.P. Soares. J.
Mol. Catal. A, 193, 51 (2003); (b) F.-G. Fontaine. Ph. D. the-
sis, Université de Montréal, Montréal, Qc. 2002.
The authors gratefully acknowledge the Natural Sciences
and Engineering Research Council of Canada (NSERC) and
Fonds pour la formation des chercheurs et l’aide à la recher-
che (Québec) (FCAR) for the financial support of this work,
and Professor H. Lebel and Valérie Paquet for the use of
their GC–MS instrument.
1
1. R. Vollmerhaus, F. Bélanger-Gariépy, and D. Zargarian.
Organometallics, 16, 4762 (1997).
1
2. For synthesis and complete characterization of this complex
see: T.A. Huber, M. Bayrakdarian, S. Dion, I. Dubuc, F.
Bélanger-Gariépy, and D. Zargarian. Organometallics, 16,
5
811 (1997).
3. (a) K. Oertle and H. Wetter. Tetrahedron Lett. 26, 5511 (1985);
b) K. Yamamoto and M. Takemae. Synlett, 259 (1990); (c) N.
Asao, T. Sudo, and Y. Yamamoto. J. Org. Chem. 61, 7654
1996); (d) T. Sudo, N. Asao, V. Gevorgyan, and Y.
1
Re fe re nc e s
(
1
. (a) A. Chalk and J.F. Harrod. J. Am. Chem. Soc. 87, 16
1965); (b) J.F. Harrod and A. Chalk. J. Am. Chem. Soc. 87,
133 (1965).
. For reviews on the hydrosilylation of alkenes and alkynes see:
a) I. Ojima. In The chemistry of organosilicon compounds.
Edited by S. Patai and Z. Rappoport. John Wiley, Chichester.
989. p. 1479; (b) B. Marciniec. Comprehensive handbook on
(
1
(
Yamamoto. J. Org. Chem. 64, 2494 (1999).
4. M. Rubin, T. Schweir, and V. Gevorgyan. J. Org. Chem. 67,
2
1
1
(
1
936 (2002).
5. (a) J.B. Lambert and Y. Zhao. J. Am. Chem. Soc. 118, 7867
1996); (b) J.B. Lambert, Y. Zhao, and H. Wu. J. Org. Chem.
4, 2729 (1999).
1
(
6
hydrosilylation. Vol. 8. Pergamon Press, Oxford. 1991. p. 763.
. For a sampling of hydrosilylation systems based on transition
metals and lanthanides see the following reports and the refer-
ences therein: (a) M.B. Carter, B. Schiøtt, A. Gultiérrez, and
S. Buchwald. J. Am. Chem. Soc. 116, 11 667 (1994); (b) X.
Verdaguer, U.E.W. Lango, and S. Buchwald. Angew. Chem.
Int. Ed. 37, 1103 (1998); (c) M.P. Doyle, K.G. High, C.L.
Nesloney, T.W. Clayton, Jr., and J. Lin. Organometallics, 10,
3
1
1
6. (a) J.M. Blackwell, K.L. Foster, V.H. Beck, and W.E. Piers. J.
Org. Chem. 64, 4887 (1999); (b) D.J. Parks, J.M. Blackwell,
and W.E. Piers. J. Org. Chem. 65, 3090 (2000).
7. Lambert et al. have demonstrated that aromatic solvents can
complex silylium cations when the Si substituents are not very
bulky: J.B. Lambert, S. Zhang, and S.M. Ciro. Organo-
metallics, 13, 2430 (1994).
8. A.M. Lapointe, F.C. Rix, and M. Brookhart. J. Am. Chem.
Soc. 119, 906 (1997).
9. F.-G. Fontaine, M.-A. Dubois, and D. Zargarian. Organo-
metallics, 20, 5145 (2001).
0. (a) M.-A. Dubois. M.Sc. Thesis, Université de Montréal,
Montréal, Qc. 2000; (b) R. Wang, L.F. Groux, and D.
Zargarian. Organometallics, 21, 5531 (2002).
1. R.A. Benkeser, H. Landesman, and D.J. Foster. J. Am. Chem.
Soc. 74, 648 (1952).
1
1
225 (1991); (d) R. Takeuchi and H. Yasue. Organometallics,
5, 2098 (1996); (e) M.L. Christ, S. Sabo-Etienne, and B.
1
1
2
Chaudret. Organometallics, 14, 1082 (1995); (f) J.F. Harrod
and S. Yun. Organometallics, 6, 1381 (1987); (g) R.S. Tannke
and R.H. Crabtree. Organometallics, 10, 415 (1991); (h) G.A.
Molander and C.P. Corrette. Organometallics, 17, 5504 (1998);
(
2
i) M.F. Lappert and R.K. Maskell. J. Organomet. Chem. 264,
17 (1984).
2
4
. For recent reviews of catalytic asymmetric hydrosilylation of
C=C, C=O, and C=N functionalities see: (a) T. Hayashi. In
Comprehensive asymmetric catalysis. Edited by E.N.
©
2003 NRC Canada