Hydrosilylation of alkenes with early main-group metal catalysts

Article abstract of DOI:10.1002/anie.200504164

(Chemical Equation Presented) Clean conversion of conjugated alkenes with exclusive formation of one regioisomer is achieved by using a new class of hydrosilylation catalysts based on early main-group metals (Ca, Sr, and K). The regioselectivity can be sw

Full text of DOI:10.1002/anie.200504164

Angewandte
Chemie
highly active SiH4 and high pressure, which result in multiple
hydrosilylation.[7] The only representative catalysts from the
main-group elements are the Lewis acids AlCl3 and
B
(C6F5)3.[8]
G
As organometallic complexes of the heavier alkaline-
earth metals (Ca, Sr, and Ba)show similarities with organo-
lanthanides[9] and are increasingly used in catalysis,[10–12] they
could be potential candidates for early main-group metal
hydrosilylation catalysts. A possible catalytic cycle could be
analogous to that established for the hydrosilylation with
lanthanide catalysts (Scheme 1). This mechanism poses two
Synthetic Methods
DOI: 10.1002/anie.200504164
Hydrosilylation of Alkenes with
Early Main-Group Metal Catalysts**
Scheme 1. Established catalytic cycle for hydrosilylation with lanthanide
catalysts.
tentative problems: the proposed catalytically active species,
a heteroleptic metal hydride complex, is currently unknown
for the heavier alkaline-earth metals. Our previous attempts
to isolate well-defined soluble heteroleptic Ca hydrides
resulted in the precipitation of insoluble CaH2.[13] A highly
reactive, short-lived Ca hydride complex as an intermediate in
a catalytic cycle, however, remains feasible. The second
problem is that early main-group metal alkyl complexes, and
possibly also hydrides, only add to conjugated alkenes (e.g.,
styrene and butadiene), which also readily polymerize under
these conditions.
Frank Buch, Julie Brettar, and Sjoerd Harder*
The catalytical hydrosilylation of alkenes is of great impor-
tance in the production of silicon compounds.[1] There are
numerous methods to initiate this highly atom-efficient key
transformation. Classical thermal and radical-initiated reac-
tions often lead to oligomers, especially when readily
polymerizable alkenes (e.g., styrenes)are used. [2] The discov-
ery of the Speier hydrosilylation catalyst [H2PtCl6]·6H2O/
iPrOH represented a major breakthrough in silicon chemis-
try.[3] This catalyst was later replaced by the more active and
selective Karstedt catalyst.[4] Nevertheless, in some cases poor
regiocontrol and side reactions, such as alkene isomerization
and hydrogenation, are still an issue.[5] The last decade has
seen the development of lanthanide-based catalysts.[6]
Although these catalysts show poor regioselectivity in some
cases, they certainly feature several advantages, such as
1)tunability of the regiochemistry by metal size and ligand
choice and 2)enantioselective hydrosilylation. Hydrosilyla-
tion catalyzed by LiAlH4 has been described but requires
Calcium-mediated hydrosilylation was first attempted
with 1,1-diphenylethylene (DPE), a substrate which does
not polymerize. Addition of 5 mol% of the heteroleptic
catalyst 1 to a mixture of DPE and PhSiH3 at 508C resulted in
[*]F. Buch, Dr. J. Brettar, Prof. Dr. S. Harder
Anorganische Chemie
Universität Duisburg-Essen
Universitätsstrasse 5, 45117 Essen (Germany)
Fax: (+49)201-183-2621
E-mail: sjoerd.harder@uni-due.de
[**]We kindly acknowledge H. Bandmann, Dr. T. Schaller, and W. Karow
for measurement of the NMR (500 MHz) spectra, GCMS analyses,
and many fruitful discussions.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2006, 45, 2741 –2745
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2741

Communications
Scheme 2. Initiation and possible catalytic cycles for hydrosilylation with early main-group metal catalysts.
complete conversion of the a-Me3Si-2-(Me2N)-benzyl
(DMAT)ligand into 5 (Scheme 2). The intense red color of
the reaction mixture suggests addition of the resulting hydride
to DPE and formation of MePh2Cꢀ, an intermediate in
catalytic cycle A. However, an overnight reaction gave only
10% conversion into hydrosilylation product 6 (R1 = R2 = Ph;
Table 1, entry 1).
conversion within 2 h. The more reactive Sr analogue 3 gave
similar results with a catalyst loading of 2.5 mol% (Table 1,
entry 4).
The cleanliness and speed of the reactions indicate that
soluble heteroleptic R-M-H species play a role in the catalytic
cycle. The attempted isolation of a heteroleptic [(DMAT)-
CaH] complex was not successful, and there is evidence that
hydride-rich clusters of [{(DMAT)<1CaH>1}n] exist in solu-
tion.[14] It is, however, not excluded that small amounts of
CaH2 precipitate in a very fine highly reactive form and, after
a heterogeneous reaction, are reintroduced in the catalytic
cycle (in some cases a fine precipitate was observed). Freshly
ground, commercially available CaH2 was found not to be
active as a hydrosilylation catalyst (Table 1, entry 5).
The scope of hydrosilylation with polar early main-group
catalysts was further investigated by probing different sub-
strates. The hydrosilylation of a-Me-styrene with PhSiH3
proceeded slowly with Ca catalyst 2, but use of the Sr
analogue 3 gave fast conversion (Table 1, entries 6 and 7).
Runs with a bulkier secondary silane (Table 1, entry 8)also
gave good conversion, albeit slower. As expected, no
conversion was observed in the hydrosilylation of unactivated
alkenes, such as allylbenzene and norbornene.
As homoleptic 2 generally shows higher reactivity than
heteroleptic 1, this compound was also tested as a potential
In contrast to our expectations, the hydrosilylation of
styrene with PhSiH3 did not give polystyrene by-products, but
clean and very fast regioselective hydrosilylation was
observed already at room temperature (Table 1, entries 9
and 10). A decrease in the catalyst loading to 0.5 mol%
(Table 1, entry 11)and reaction with a secondary silane
(Table 1, entry 12)both gave a good conversion; furthermore,
the heteroleptic calcium catalyst 1 was found to be active
(Table 1, entry 13). The hydrosilylation of styrene is appa-
rently much faster than its polymerization.
catalyst (Table 1, entries 2 and 3). Despite the possible
formation of insoluble CaH2 in the initiation step, a clean
and complete conversion of DPE and PhSiH3 into the
expected silane 6 was observed overnight (catalyst loading:
2.5 mol%; no traces of the other regioisomer could be found).
An increase in the catalyst loading to 10 mol% gave full
The alkaline-earth-metal catalysts 2 and 3 also showed
very high activity in the hydrosilylation of the conjugated
double bond in hexadiene, thus exclusively yielding the
2742
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 2741 –2745

Angewandte
Chemie
Table 1: Summary of results for the hydrosilylation of alkenes with various main-group metal catalysts.
monosilylated product with a Si
atom in the allylic position
(Table 1, entries 14 and 15).
The “in situ” generation of
highly reactive alkaline-earth-
metal hydrides drew our attention
to the use of organopotassium
compounds as potential catalysts
in alkene hydrosilylation. The reac-
Entry Substrates
Product
Catalyst
G
U
T
t
Conver.
[%]
[8C] [h]
1
2
3
4
DPE
1
5
50
50
50
50
16
10
>98
>98
>98
PhSiH3
DPE
PhSiH3
2
2
3
2.5
16
2
DPE
PhSiH3
10
tion of 5 mol% of K(DMAT) 4
A
DPE
2.5
2
with DPE and PhSiH3 immediately
produced the initiation product 5,
and complete conversion of the
substrates was reached within 2 h
(Table 1, entry 16). Analysis, how-
ever, showed exclusive formation
of the other regioisomer 7. Inter-
estingly, similar results were
obtained with commercially avail-
able KH; however, a 90-min induc-
tion period was observed (Table 1,
entry 17). Hydrosilylation of other
alkenes with the potassium catalyst
4 either resulted in no reaction
(allylbenzene, norbornene), a slug-
gish reaction (cyclohexadiene), or
partial alkene polymerization (sty-
rene and a-Me-styrene).
PhSiH3
[a]
5
6
DPE
PhSiH3
–
CaH2
25
50
50
50
50
20
20
50
20
48
24
0
=
Ph(Me)C CH2
PhSiH3
2
3
3
2
3
2
2
2.5
20
=
7
Ph(Me)C CH2
PhSiH3
2.5
2.5
2.5
2.5
0.5
2.5
2.5 >98
=
8
Ph(Me)C CH2
Ph(Me)SiH2
24
>98
9
styrene
PhSiH3
<0.1 >98
<0.1 >98
1.5 >98
10
11
12
styrene
PhSiH3
styrene
PhSiH3
It is unlikely that regioisomer 7
is formed by the hydride cycle
(cycle A)through a 1,2-insertion
which produces the resonance-
unstabilized Ph2CHCHꢀ2 ion. The
fast and exclusive formation of
styrene
Ph(Me)SiH2
<0.1 >98
diastereomers[b]
13
14
15
16
styrene
PhSiH3
1
2
3
4
5
50
20
20
50
20
>98
cyclohexadiene
PhSiH3
2.5
2.5
5
<0.1 >98
<0.1 >98
hydrosilylation product
7
(no
traces of 6 could be detected)
indicates that the potassium-medi-
ated hydrosilylation might operate
through a different mechanism. A
possible route could be the silanide
pathway (cycle B), a mechanism
that has been rejected for lantha-
nide-catalyzed hydrosilylation[6a]
but has been proposed in transi-
tion-metal-catalyzed hydrosilyla-
tion.[15] The first step is well estab-
lished: KH reacts with PhSiH3 to
cyclohexadiene
PhSiH3
DPE
2
>98
PhSiH3
17
18
DPE
KH[c]
25
50
50
4[d] >98
PhSiH3
4
5
16
>98
PhSiH3
form
hypervalent
PhSiHꢀ4 K+,
which yields PhSiHꢀ2 K+ after the
loss of H2.[16] Addition of the sila-
nide to DPE (cycle B)gives the
resonance-stabilized intermediate
PhSiH2CH2CPhꢀ2 K+, which gives 7
after s-bond metathesis with
PhSiH3.
19[e]
20[e]
21[e]
DPE
2
3
4
2.5
2.5
5
50
50
20
3
2
>98
>98
PhSiH3
DPE
PhSiH3
DPE
<0.1 >98
PhSiH3
A
recently reported crystal
structure of Ph3SiHꢀ2 K+[17] raises
the possibility that potassium-
mediated hydrosilylation could
[a]Commercially obtained CaH 2 was freshly ground under nitrogen. [b]Both diastereomers were
obtained in an approximate ratio of 1:1. [c]KH was isolated from a commercially available suspension in
paraffin oil. [d]A 90-min induction period was observed. [e]Reaction in THF.
Angew. Chem. Int. Ed. 2006, 45, 2741 –2745
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
2743

Communications
proceed through a concerted insertion reaction of the alkene
products. A complete change in regioselectivity was also
observed for the Sr catalyst 3 in THF (Table 1, entry 20). The
reaction in Et2O only gave regioisomer 7, which is in
with the pentavalent PhSiHꢀ4 intermediate (Scheme 2,
cycle C). Steric interactions between the Ph rings in DPE
and the silanide anion would result in the exclusive formation
of regioisomer 7. This mechanism is analogous to the well-
established mechanism for the KF-catalyzed hydrosilylation
of ketones (Scheme 3).[18]
ꢀ
agreement with the larger ionicity of the Sr C bond. The
addition of THF to the DPE/PhSiH3/4 system gave essentially
the same product as found in the THF-free system; however,
the conversion was much faster (Table 1, entry 21). Phenyl-
cyclohexane could not be hydrosilylated using catalysts 2 and
3 in THF (also not without solvent). This behavior indicates
that the metal plays an important role in the catalytic species
and rules out the ion-pair mechanism (cycle C).
It is hitherto unclear how bond polarity redirects the
catalytic reaction from the hydride cycle to the silanide cycle.
The reaction is initiated in all cases by formation of 5 and a
metal hydride, which either reacts with the alkene or the
silane. A reasonable explanation could be that large cations
and polar solvents favor the formation of the intermediate ion
ꢀ
pair [M+]
A
In summary, the first hydrosilylation catalysts based on
early main-group metals (Ca, Sr, and K)are very effective for
the conversion of conjugated double bonds. The catalytic
reactions are initiated in all cases by the formation of a highly
reactive metal hydride, which either adds to an alkene or to
the silane. The regiochemistry for the hydrosilylation of DPE
can be completely controlled by either the polarity of the
solvent or metal choice. Further research will be directed to
the isolation of well-defined heteroleptic alkaline-earth-metal
hydrides and other possible intermediates; therefore, better
insight into the mechanisms involved may be gained. In
addition, possible enantioselective hydrosilylation with these
catalysts will be explored.
Scheme 3. Mechanism for the KF-catalyzed hydrosilylation of ketones.
Hydrosilylation of a cyclic alkene gave strong support for
a silanide mechanism: overnight reaction of 1-Ph-cyclohex-
ene with PhSiH3 using potassium catalyst 4 (Table 1, entry 18)
resulted in the formation of 1-Ph-2-PhSiH2-cyclohexane
exclusively as a cis diastereomer, from which a trans addition
of the silane can be concluded. This result essentially rules out
a concerted addition as in cycle C, but is in agreement with the
two-step silanide mechanism in cycle B. Exclusive formation
of the cis isomer can be explained by s-bond metathesis in
which PhSiH3 approaches from the least-hindered side. This
preferred stereochemistry is opposite to that in the hydro-
boration and therefore allows access to the complementary
alcohols after a Tamao–Fleming oxidation. The additional
formation of approximately 10% of cis-1-Ph-2-SiH3-cyclo-
hexane and even larger amounts of Ph2SiH2 also points to a
concomitant disproportionation of the silane substrate during
the overnight reaction (2PhSiH3!Ph2SiH2 + SiH4, etc.)and
formation of KSiH3. This base-catalyzed disproportiona-
tion,[19] which is proposed to proceed through a hypervalent
intermediate,[16] could be reproduced by reaction of 4 with
excess PhSiH3 in the absence of an alkene.[20] The formation
of KSiH3 is in agreement with the recent finding that
prolonged reaction of KH with PhSiH3 yields KSiH3.[21]
The complete change in regioselectivity observed in
switching from Ca (or Sr)catalysts to the corresponding K
catalyst indicates that metal–carbon bond polarity could be a
decisive factor. For this reason, we also studied the effect of
the solvent on the regioselectivity. Whereas hydrosilylation of
DPE with Ca catalyst 2 in benzene gave essentially similar
results as the solvent-free runs (that is, sole formation of
product 6), experiments in THF gave fast and exclusive
formation of regioisomer 7 (Table 1, entry 19). The reaction
in Et2O produced a mixture of isomers 6 (75%)and 7 (25%),
which shows that regioselectivity can be tuned by solvent
polarity. The use of other styrenic substrates led to oligomeric
Experimental Section
All experiments were carried out using standard Schlenk techniques
and freshly dried solvents. Alkene substrates were dried over CaH2
and freshly distilled prior to use. The silanes PhSiH3 (Fluka)and
Ph(Me)SiH2 (Aldrich)were used as received. The catalysts were
prepared according to our previously reported methods: 1,[10b] 2,[10a]
3,[23] and 4.[23]
A typical hydrosilylation experiment: a dry Schlenk tube was
charged with the alkene substrate (2.0 mmol)and the silane
(2.0 mmol). After addition of the catalyst (generally 2.5 mol%) and
heating of the solution to the required temperature (generally 508C),
a color change to red was usually observed. In some cases, a fine
material, presumably a metal hydride, was deposited. To follow the
conversion, samples were taken at regular time intervals and analyzed
by 1H NMR spectroscopy and GC MS. In all cases (except entry 18 in
Table 1), the crude reaction products consisted of essentially pure
hydrosilylation product with traces of the product of initiation 5 and
in some cases small amounts of one of the substrates. All hydro-
silylation products and initiation products were isolated as pure
1
compounds and completely characterized by H, 13C, and 2D NMR
spectroscopy and by GC MS (see the Supporting Information). In a
typical larger scale experiment (20 mmol of styrene and PhSiH3; Ca
catalyst 2), product 6 (R1 = Ph, R2 = H)was obtained after isolation
by column chromatography in 92% yield.
Received: November 22, 2005
Published online: March 20, 2006
2744 www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 2741 –2745

Angewandte
Chemie
359, C33; d)B. Becker, R. Corriu, C. GuØrin, B. J. L. Henner, J.
Organomet.Chem. 1989, 369, 147.
[17] M. J. Bearpark, G. S. McGrady, P. D. Prince, J. W. Steed, J.Am.
Chem.Soc. 2001, 123, 7737.
[18] a)J. Boyer, R. Corriu, R. Perz, C. Reye, Tetrahedron 1981, 37,
2165; b)J. Boyer, R. Corriu, R. Perz, M. Poirier, C. Reye,
Synthesis 1981, 558; c)R. Schiffers, H. B. Kagan, Synlett 1997, 10,
1175; d)C. Chuit, R. Corriu, C. Reye, C. Young, Chem.Rev.
1993, 93, 1371.
[19] a)J. A. Morrison, M. A. Ring, Inorg.Chem. 1967, 6, 100; b)M.
Itoh, K. Inoue, J. Ishikawa, K. Iwata, J.Organomet.Chem. 2001,
629, 1.
Keywords: alkaline-earth metals · homogeneous catalysis ·
hydrides · hydrosilylation · silanides
.
[1] I. Ojima, Z. Li, J. Zhu in The Chemistry of Organosilicon
Compounds, Vol.2 (Eds.: Z. Rappoport, Y. Apeloig), Wiley,
New York, 1998, p. 1687.
[2] D. A. Armitage in Comprehensive Organometallic Chemistry,
Vol.2 (Eds.: F. G. A. Stone, E. W. Abel), Pergamon, Oxford,
1982, p. 115.
[3] J. L. Speier, Adv.Organomet.Chem. 1979, 17, 407.
[4] a)B. D. Karstedt, (General Electric), US Patent 3 715 334, 1973;
b)P. B. Hitchcock, M. F. Lappert, N. J. W. Warhurst, Angew.
Chem. 1991, 103, 439; Angew.Chem.Int.Ed.Engl. 1991, 30, 438.
[5] It has been suggested that this is due to the formation of colloidal
Pt species: a)J. Stein, L. N. Lewis, Y. Gao, R. A. Scott, J.Am.
Chem.Soc. 1999, 121, 3693; b)I. E. Markó, S. StØrin, O. Buisine,
G. Berthon, G. Michaud, B. Tinant, J.-P. Declercq, Adv.Synth.
Catal. 2004, 346, 1429.
[20] Reaction of 4 with neat PhSiH3 (20-fold excess; 508C)gave the
initiation product 5 and resulted in gas development. Analysis of
the product showed Ph2SiH2 and minor amounts of Ph3SiH. The
evolved gas, which ignited spontaneously in air, was quenched in
ethanol (with catalytical amounts of LiOEt)and gave Si (OEt)4.
R
No sign of dehydropolymerization reactions (2PhSiH3!
PhSiH2SiH2Ph + H2)were observed.
[21] Z. Hou, Y. Zhang, M. Nishiura, Y. Wakatsuki, Organometallics
2003, 22, 129.
[22] The stability of Ca2+·6THF ions is well known: a)W. L.
Driessen, M. Den Heijer, Inorg.Chim.Acta 1979, 33, 261; b)S.
[6] a)P.-F. Lu, L. Brard, Y. Li, T. J. Marks, J.Am.Chem.Soc. 1995,
117, 7157; b)H. Schumann, M. R. Keitsch, J. Demtschuk, G. A.
Molander, J.Organomet.Chem.
1999, 582, 70; c)T. I.
Harder, F. Feil, T. Repo, Chem.Eur.J.
Perruchas, F. Simon, S. Uriel, N. Avarvari, K. Boubekeur, P.
Batail, J.Organomet.Chem. 2002, 643–644, 301; d)I. G.
2002, 8, 1991; c)S.
Gountchev, T. D. Tilley, Organometallics 1999, 18, 5661;
d)A. A. Trifonov, T. P. Spaniol, J. Okuda, Organometallics
2001, 20, 4869; e)O. Tardif, M. Nishiura, Z. Hou, Tetrahedron
2003, 59, 10525; f)Y. Horino, T. Livinghouse, Organometallics
2004, 23, 12.
Fedushkin, A. N. Lukoyanov, S. Dechert, H. Schumann, Eur.J.
Inorg.Chem. 2004, 12, 2421; e)S. Harder, S. Mꢀller, E. Hꢀbner,
Organometallics 2004, 23, 178.
[7] M. Kobayashi, M. Itoh, Chem.Lett. 1996, 1013.
[23] F. Feil, S. Harder, Organometallics 2001, 20, 4616.
[8] a)K. Oertle, H. Wetter, Tetrahedron Lett. 1985, 26, 5511; b)M.
Rubin, T. Schwier, V. Gevorgyan, J.Org.Chem. 2002, 67, 1936.
[9] a)R. A. Williams, T. P. Hanusa, J. C. Huffman, Organometallics
1990, 9, 1128; b)M. Rieckhoff, U. Pieper, D. Stalke, F. T.
Edelmann, Angew.Chem. 1993, 105, 1102; Angew.Chem.Int.
Ed.Engl. 1993, 32, 1079; c)C. Eaborn, P. B. Hitchcock, K. Izod,
Z.-R. Lu, J. D. Smith, Organometallics 1996, 15, 4783; d)F.
Weber, H. Sitzmann, M. Schultz, C. D. Sofield, R. A. Andersen,
Organometallics 2002, 21, 3139; e)S. Harder, Angew.Chem.
2004, 116, 2768; Angew.Chem.Int.Ed. 2004, 43, 2714.
[10] a)S. Harder, F. Feil, A. Weeber, Organometallics 2001, 20, 1044;
b)F. Feil, S. Harder, Angew.Chem. 2001, 113, 4391; Angew.
Chem.Int.Ed. 2001, 40, 4261.
[11] Z. Zhong, P. J. Dijkstra, C. Birg, M. Westerhausen, J. Feijen,
Macromolecules 2001, 34, 3863; M. H. Chisholm, J. C. Gallucci,
K. Phomphrai, Inorg.Chem. 2004, 43, 6717.
[12] M. R. Crimmin, M. S. Hill, J.Am.Chem.Soc. 2005, 127, 2043.
[13] J. Brettar, F. Buch, S. Harder, unpublished results.
[14] Reaction of [(DMAT)2Ca]·2THF (2)with one equivalent of
PhSiH3 cleanly yields PhH2Si(DMAT)and a clear yellow
G
solution presumably of [(DMAT)CaH]. Attempted crystalliza-
tion of this product resulted in large yellow crystalline blocks
which were determined to be 2. The mother liquor was still
largely clear at this stage and no fine CaH2 precipitate could be
detected. The mother liquor gave more well-defined crystals of 2
after repeated concentration and must have consisted of a
soluble hydride-rich Ca cluster [{(DMAT)<1CaH>1}n] at this
stage. At later stages, a fine white powder precipitated, which
reacts vigorously with ethanol under gas evolution. Soluble
hydride-rich clusters have been recently observed in lanthanide
chemistry; for a review, see: Z. Hou, Bull.Chem.Soc.Jpn. 2003,
76, 2253.
[15] A. M. LaPointe, F. C. Rix, M. Brookhart, J.Am.Chem.Soc.
1997, 119, 906, and references therein.
[16] a)D. J. Hajdasz, R. R. Squires, J.Am.Chem.Soc.
1986, 108,
3139; b)J. L. Brefort, R. Corriu, C. GuØrin, B. Henner, J.
Organomet.Chem. 1989, 370, 9; c)B. Becker, R. Corriu, C.
GuØrin, B. J. L. Henner, Q. Wang, J.Organomet.Chem. 1989,
Angew. Chem. Int. Ed. 2006, 45, 2741 –2745
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
2745