Catalytic dehydrogenative coupling of secondary silanes using Wilkinson's catalyst

Full text of DOI:10.1021/ja015697i

5120
J. Am. Chem. Soc. 2001, 123, 5120-5121
Catalytic Dehydrogenative Coupling of Secondary
Silanes Using Wilkinson’s Catalyst
Lisa Rosenberg,*,† Colin W. Davis, and Junzhi Yao
Department of Chemistry, UniVersity of Manitoba
Winnipeg, Manitoba, Canada R3T 2N2
ReceiVed February 20, 2001
ReVised Manuscript ReceiVed April 12, 2001
The pursuit of clean, straightforward alternatives to Wurtz
coupling reactions for the synthesis of polysilanes has focused
attention on dehydrocoupling reactions of silanes catalyzed by
transition metals.1 Among complexes investigated in their capacity
to produce these polymers2 two classes have emerged. One
includes group 4 metallocene derivatives, which are considered
the most active for catalytic dehydrocoupling, since they are
capable of producing oligomers and sometimes polymers from
primary silanes. Catalysts in the second class include many late
metal complexes, such as Wilkinson’s catalyst, (Ph3P)3RhCl, 1,
and other Rh(I) phosphine complexes;3 these are considered
relatively inactive for dehydrocoupling, since they do not produce
long chains, even for primary silane monomers.4 Complex 1 and
other late metal complexes catalyze the coupling of secondary
silanes to di- and trisilanes, but their application to the large-
scale synthesis of these useful, hydrido-substituted oligosilanes
has been discouraged by the known activity of such catalysts for
a competing reaction of silanes: redistribution of substituents at
silicon.5 In this communication we describe results indicating that
1 actually exhibits high activity and chemoselectivity for coupling
of secondary silanes, under appropriate conditions. We have
demonstrated that coupling in this system is extremely fast but is
not thermodynamically favored; chemoselectivity is therefore
exceedingly sensitive to the presence of hydrogen gas, also a
product of the coupling reaction.
Figure 1. Dependence of product distribution on catalyst concentration,
observed for small-scale reactions of Ph2SiH2 where (a) removal of
hydrogen was not rigorously controlled and (b) hydrogen was removed
efficiently (see text for details).
We therefore investigated the effects of varying both substrate
and catalyst concentration on the reactions of diphenylsilane in
the presence of 1. We obtained lower conversions of diphenyl-
silane, with higher relative amounts of redistribution, when we
used solvent in our initial experiments,7 so carried out subsequent
reactions in neat diphenylsilane. Preliminary, small-scale reac-
tions8 of the neat substrate (e.g. Figure 1a) consistently gave more
dehydrocoupling than redistribution9 but with much greater
selectivity for coupling at lower catalyst concentrations. At high
catalyst concentrations, an increase in the amount of redistribution
was apparently at the expense of coupled product.
However, in these small-scale reactions, the product distribu-
tions were surprisingly sensitive to the size of reaction vial relative
to the volume of substrate and to the rate and efficacy of stirring.
Chemoselectivity was enhanced, and its catalyst concentration
dependence became far less pronounced, when these conditions
were manipulated to allow more efficient removal of hydrogen
gas from the reaction mixtures (Figure 1b).10 We obtained
optimum conversion (75-80%, at which point the mixture
solidifies) and minimum redistribution using flat-bottomed vials,
wide enough to allow the substrate to be spread thinly, and high
stir rates, with stir bars of lengths approaching the inner diameter
of the vial. Thus, although redistribution of phenyl groups at
silicon can be a major competing reaction for diphenylsilane in
the presence of 1, it can be kept to a minimum with efficient
removal of H2(g).11 Based on the above results, which suggest
good catalyst activity and selectivity for coupling at low catalyst
concentrations, we prepared 2 on a 5-g scale from neat diphen-
ylsilane using 0.2 mol % 1 as catalyst.12 The reaction proceeds
readily at room temperature and is complete within 2 h, which
Our interest in preparing 1,2-bis(hydrido)-substituted disilanes
on a synthetically useful scale led us to reinvestigate the activity
of 1 in the dehydrocoupling of diphenylsilane (eq 1). Since
previous reports of this reaction invariably included the observa-
tion of considerable amounts of Ph3SiH, arising from redistribution
of the phenyl and hydride groups on silicon,6 we needed to identify
conditions that would favor dehydrocoupling over redistribution.
(7) Product distributions, showing 26% redistribution, were similar to those
reported in ref 3d for this reaction.
† New address as of July 1, 2001: Department of Chemistry, University
of Victoria, P.O. Box 3065, Victoria, B.C., Canada, V8W 3V6.
(1) (a) Tilley, T. D. Comments Inorg. Chem. 1990, 10, 37. (b) Corey, J. Y.
AdV. Silicon Chem. 1991, 1, 327. (c) Gauvin, F.; Harrod, J. F.; Woo, H. G.
AdV. Organomet. Chem. 1998, 42, 363. (d) Reichl, J. A.; Berry, D. H. AdV.
Organomet. Chem. 1999, 43, 197.
(2) Polysilanes are of interest for their unusual electronic and optical
properties, arising from extensive electron delocalization along the all-silicon
backbone: (a) West, R. J. Organomet. Chem. 1986, 300, 327. (b) Miller, R.
D.; Michl, J. Chem. ReV. 1989, 89, 1359. (c) Hamada, T. J. Chem. Soc.,
Faraday Trans. 1998, 94, 509.
(3) Ojima, I.; Inaba, S.; Kogure, T. J. Organomet. Chem. 1973, 55, C7.
(b) Lappert, M. F.; Maskell, R. K. J. Organomet. Chem. 1984, 264, 217. (c)
Brown-Wensley, K. A. Organometallics 1987, 6, 1590. (d) Chang, L. S.;
Corey, J. Y. Organometallics 1989, 8, 1885. (e) Rosenberg, L.; Fryzuk, M.
D.; Rettig, S. J. Organometallics 1999, 18, 958.
(4) For exceptions to this classification of late metal catalysts, see: (a)
Fontaine, F.-G.; Kadkhodazadeh, T.; Zargarian, D. Chem. Commun. 1998,
1253 (b) Chauhan, B. P. S.; Shimizu, T.; Tanaka, M. Chem. Lett. 1997, 785.
(5) Curtis, M. D.; Epstein, P. S. AdV. Organomet. Chem. 1981, 19, 213.
(6) Phenylsilane is also produced by this redistribution reaction, but this
more volatile product is less often detected.
(8) Reactions were carried out at room temperature in small, open vials in
the glovebox. Fixed amounts of substrate (50-200 mg) and reaction times
(1-4 h) were used for each series of experiments, with reactions being
quenched by hexanes elution through Florisil.
(9) Product distributions were determined from the amounts of 2 and Ph3-
1
SiH observed by H NMR spectroscopy.
(10) The apparent dependence on catalyst concentration of chemoselectivity
initially observed (Figure 1a) suggests that very different mechanistic pathways
exist for these rhodium-mediated dimerization and redistribution reactions. A
system exhibiting such concentration dependence presents an attractive
opportunity to identify differences in key intermediates in these competing
reactions. However, almost complete disappearance of the concentration
dependence of this chemoselectivity with efficient removal of hydrogen (Figure
1b) suggests that rhodium intermediates in the two mechanisms may not
actually be so different. Neither mechanism has been firmly established for
the late transition metals: proposed mechanisms rely on either the intermediacy
of transition metal silylene fragments or a series of oxidative addition/reductive
elimination steps (refs 1 and 5), and the involvement of dinuclear intermediates
has not been ruled out (ref 3e, for example).
(11) The maximum amount of Ph2SiH2 undergoing redistribution in these
experiments was 9%.
10.1021/ja015697i CCC: $20.00 © 2001 American Chemical Society
Published on Web 05/05/2001

Communications to the Editor
J. Am. Chem. Soc., Vol. 123, No. 21, 2001 5121
Table 1. Dehydrocoupling of Other Secondary Silanes in the
Presence of 1a
Scheme 1
reaction
time (h)
total monomer
coupled (TOFc)
silane
product (yieldb)
MePhSiH2
0.58
disilane (11%), trisilane
(59%), Me2PhSiH (<1%),
MePh2SiH (<1%)
70% (600)
represents an average turnover frequency (TOF) of 240 (moles
of monomer consumed/moles of catalyst/hour). This is in marked
contrast to a TOF of ∼3 observed for the same reaction mediated
at 90 °C by a titanocene-based catalyst system.13 This activity of
1 for dehydrocoupling of diphenylsilane is also several orders of
magnitude higher than is observed for the coupling of MePhSiH2
to di- and trisilanes,14 and for dehydropolymerization of primary
silanes,15 by group 4 metallocenes.
Et2SiH2
(n-Hex)2SiH2
1.0
1.5
disilane (44%), trisilane (9%)
disilane (60%), trisilane (13%)
57% (290)
73% (250)
a Reactions were run at room temperature, without solvent, using
0.2 mol % 1. b Based on integration of H NMR. c Average value, in
1
mol monomer/mol catalyst/h.
hydrogen gives a room-temperature equilibrium constant for the
coupling (forward) reaction of Kc ) 2 × 10-4, indicating that
the equilibrium lies significantly toward monosilane under these
conditions.17
The differences between a and b in Figure 1 can be explained
by the effects of an equilibrium between diphenylsilane and the
products of coupling, 2 and H2(g), which lies heavily toward the
substrate monomer (Scheme 1). Local concentrations of hydrogen
formed in the reaction mixture, if not dispersed by rapid stirring
or by other means, will allow this monomer-dimer equilibrium
to be established, thus limiting the amount of 2 produced,
regardless of the rate of formation (i.e. the forward rate of the
coupling reaction). At higher catalyst concentrations, the rates of
both coupling and redistribution of diphenylsilane will be greater
than at lower catalyst concentrations. Higher local concentrations
of hydrogen gas, resulting from fast coupling, will maintain
increased monomer concentrations in these runs, which should
give higher relative rates of redistribution (i.e. poorer chemo-
selectivity). At lower catalyst concentrations the hydrogen is
produced more slowly and is therefore readily dissipated with
“normal” stirring, allowing isolation of 2, the kinetic product.
This situation explains the observed erratic and lower yields of 2
and the steadily increasing yields of Ph3SiH with increasing
catalyst concentration, when the removal of hydrogen (i.e. stir
rate and sample depth) is not stringently controlled (Figure 1a).
When hydrogen is more efficiently removed (Figure 1b), the
relative amounts of 2 and Ph3SiH observed within a fixed reaction
time more accurately reflect chemoselectivity, since the real rate
of dimerization is in effect in this controlled situation. Clearly
redistribution of phenyl and hydrido groups in Ph2SiH2 proceeds
very slowly, relative to dimerization, at all catalyst concentrations.
Evidence for the participation and importance of the above
equilibrium in these rhodium-catalyzed reactions is provided by
experiments in which H2(g)-saturated, d6-benzene or d8-toluene
solutions containing 2 and catalytic amounts of 1 are sealed in
an NMR tube under an atmosphere of hydrogen. Within minutes
a significant amount of Ph2SiH2 is observed by 1H NMR
spectroscopy, with equilibrium being attained after 1 week.16
Integration of the relevant Si-H signals and that due to dissolved
Our technique for promoting chemoselectivity for the dehy-
drocoupling of silanes by 1 is general. Preliminary, small-scale
reactions indicate that under the conditions we developed for the
reactions of diphenylsilane, similar or improved chemoselectivity
is possible for a range of secondary silanes in the presence of 1
(Table 1). Dialkylsilanes react less vigorously than the aryl-
substituted silanes, yet for all three substrates we see significant
amounts of trisilane, as well as disilane. Group 4 metallocene
catalysts are not active for the dehydrocoupling of dialkylsilanes.18
We consistently observe more phenyl/hydride redistribution than
methyl/hydride redistribution for MePhSiH2 in the presence of
1. Two experiments, one in a sealed reaction vessel, the other
open to a nitrogen manifold and being flushed with N2(g), confirm
that this is not simply due to reaction conditions that facilitate
the departure of MeSiH3(g) preferentially to PhSiH3(l). The
preferred redistribution of aryl over alkyl groups is further
illustrated by the fact that for the dialkyl substrates Et2SiH2 and
(n-Hex)2SiH2 we see no products arising from redistribution in
the presence of 1.
Our results indicate rates of silane dehydrocoupling catalyzed
by a late metal complex that are much higher than previously
thought, and much higher than those reported for early metal
dehydrocoupling catalysts. The barrier to coupling in this system
is low enough, relative to that for redistribution, to allow almost
complete selectivity for coupling, under appropriate conditions.
This work suggests that other late metal complexes that are known
to catalyze silane redistribution may actually be extremely active
for dehydrocoupling, under conditions that shift the monosilane/
disilane equilibrium via highly efficient removal of hydrogen.
Work is underway to optimize large-scale syntheses of disilanes
and trisilanes (both useful building blocks in organosilicon
chemistry) for a range of substrates using 1, and to probe
mechanistic issues through further study of the reactions with
silanes of 1 and related Rh systems.
(12) To remove hydrogen sufficiently quickly in this larger volume of Ph2-
SiH2, N2(g) was passed through a gas dispersion tube immersed in the reaction
mixture. Twice, small amounts of toluene were added as the mixture solidified,
to allow it to be stirred. This procedure gave >98% (by 1H NMR spectroscopy)
conversion to disilane, which was isolated by recrystallization from hexanes
in 78% yield, after removal of the catalyst on a Florisil column.
(13) Corey, J. Y.; Kraichely, D. M.; Huhmann, J. L.; Braddock-Wilking,
J.; Lindeberg, A. Organometallics 1995, 14, 2704. Calculated TOF is based
on the reported isolated yield.
Acknowledgment. This work was supported by grants from the
Natural Science and Engineering Research Council of Canada and the
University of Manitoba.
Supporting Information Available: Full experimental details (PDF).
This material is available free of charge via the Internet at http://pubs.acs.org.
(14) (a) Corey, J. Y.; Zhu, X.-H.; Bedard, T. C.; Lange, L. D. Organo-
metallics 1991, 10, 924. (b) Imori, T.; Woo, H.-G.; Walzer, J. F.; Tilley, T.
D. Chem. Mater. 1993, 5, 1487.
JA015697I
(15) These estimated TOFs are based on total conversion of monomer to
oligomer or polymer in the presence of group 4 catalysts, and thus are not
strictly analogous to TOFs for dimerization. Longer chain lengths probably
arise from the coupling of additional monosilane with oligomers, but these
condensations are likely to be increasingly slower than dimerization (ref b
below). Still, given the much faster coupling normally observed for primary
versus secondary silanes for any one catalyst, the differences in these TOFs
are striking. See, for example: (a) Aitken, C.; Barry, J.-P.; Gauvin, F.; Harrod,
J. F.; Malek, A.; Rousseau, D. Organometallics 1989, 8, 1732. (b) Woo, H.-
G.; Walzer, J. F.; Tilley, T. D. J. Am. Chem. Soc. 1992, 114, 7047. (c)
Dioumaev, V. K.; Harrod, J. F. Organometallics 1994, 13, 1548. (d)
Grimmond, B. J.; Corey, J. Y. Organometallics 2000, 19, 3776.
(16) The absence of any signal due to dissolved H2(g) until this time
suggests that hydrogenolysis of 2 occurs as rapidly as the gas dissolves in
deuterated solvent. Since diffusion of gases in NMR samples is notoriously
slow, the real rate of approach to equilibrium in these reactions is probably
much higher.
(17) The disilane/monosilane equilibrium ratio is ∼0.03:1.
(18) The activity of a titanocene-based catalyst for the coupling of (n-
Pr)2SiH2 has been reported, but requires the presence of cyclic olefins, which
are hydrogenated in tandem with the coupling process. The resulting product
mixtures are complex, including hydrosilation products. Corey, J. Y.; Zhu,
X.-H. Organometallics 1992, 11, 672.