The first alkene-platinum-silyl complexes: Lifting the hydrosilation mechanism shroud with long-lived precatalytic intermediates and true Pt catalysts

Article abstract of DOI:10.1021/ja0127335

The synthesis, characterization, and exploratory chemistry of two classes of alkene-platinumsilyl complexes, which have been postulated as hydrosilation intermediates, are described in this report. The unique dimeric complexes 1, [R₃Si(u-Cl)(η²-COD)Pt]₂ {R₃Si = Et₃Si, MeCl₂Si, Me₂ClSi, "(EtO)₃Si", PhMe₂Si, and (Me₃SiO)Me₂Si: COD = cycloocta-1,5-diene}, and the bis-silyl complexes 2, (η⁴-COD)Pt-(SiR₃)₂ (R₃Si = Cl₃Si, MeCl₂Si, Me₂ClSi, and PhMe₂Si), are formed from the sequential reaction of 2 and 4 equiv of the corresponding hydrosilanes, respectively, with Pt(COD)Cl₂ in the presence of a small excess of COD. Complexes 1 are stable for many days in solution at room temperature but decompose via slow elimination of chlorosilane. Some of the bis-silyl compounds 2 are stable for extended periods under inert atmosphere and especially below 0 °C, either in the solid state or in solution (in the presence of a small excess of free COD). Complexes 2 display catalytic activity as discrete, molecular, and mononuclear species for hydrosilation and isomerization reactions. Compound 2c (R₃Si = MeCl₂Si) MeCl2Si) was fully characterized via multinuclear NMR spectroscopy and x-ray crystal structure analysis. The facile H-transfer rather than Si-transfer to bound COD provides experimental support for the sequence of insertive steps in the Chalk-Harrod catalytic cycle, at least for Pt-catalyzed hydrosilation.

Full text of DOI:10.1021/ja0127335

Published on Web 07/16/2002
The First Alkene-Platinum-Silyl Complexes: Lifting the
Hydrosilation Mechanism Shroud with Long-Lived Precatalytic
Intermediates and True Pt Catalysts
Aroop K. Roy* and Richard B. Taylor
Contribution from Central Research and DeVelopment, Dow Corning Corporation,
Midland, Michigan 48686-0994
Received December 20, 2001
Abstract: The synthesis, characterization, and exploratory chemistry of two classes of alkene-platinum-
silyl complexes, which have been postulated as hydrosilation intermediates, are described in this report.
The unique dimeric complexes 1, [R3Si(µ-Cl)(η2-COD)Pt]2 {R3Si ) Et3Si, MeCl2Si, Me2ClSi, “(EtO)3Si”,
PhMe2Si, and (Me3SiO)Me2Si; COD ) cycloocta-1,5-diene}, and the bis-silyl complexes 2, (η4-COD)Pt-
(SiR3)2 (R3Si ) Cl3Si, MeCl2Si, Me2ClSi, and PhMe2Si), are formed from the sequential reaction of 2 and
4 equiv of the corresponding hydrosilanes, respectively, with Pt(COD)Cl2 in the presence of a small excess
of COD. Complexes 1 are stable for many days in solution at room temperature but decompose via slow
elimination of chlorosilane. Some of the bis-silyl compounds 2 are stable for extended periods under inert
atmosphere and especially below 0 °C, either in the solid state or in solution (in the presence of a small
excess of free COD). Complexes 2 display catalytic activity as discrete, molecular, and mononuclear species
for hydrosilation and isomerization reactions. Compound 2c (R3Si ) MeCl2Si) was fully characterized via
multinuclear NMR spectroscopy and X-ray crystal structure analysis. The facile H-transfer rather than Si-
transfer to bound COD provides experimental support for the sequence of insertive steps in the Chalk-
Harrod catalytic cycle, at least for Pt-catalyzed hydrosilation.
Scheme 1. Chalk-Harrod Mechanism
Introduction
Platinum-catalyzed hydrosilation is a versatile reaction in
organosilicon chemistry and is second only to the Rochow
process for its importance to silicon-carbon bond formation
in the silicone industry.1 Since Speier’s extensive work with
hexachloroplatinic acid as a catalyst for the addition of hy-
drosilanes to alkenes,2,3 numerous salts, complexes, and sup-
ported forms of platinum have been used to add hydrosilanes
to a variety of unsaturated compounds, yet little is known about
the nature of the reactive intermediates involved in the catalytic
cycle. Some years following Speier’s initial reports, Chalk and
Harrod proposed a mechanism for the catalytic cycle based on
the similarities between transition metal-catalyzed hydrogenation
and hydrosilation (Scheme 1).4
migratory insertion between a metal-element bond, and reduc-
tive elimination. The nondescript species A is visualized to be
the active, catalytic metal complex in its resting state. Of the
above four species in this catalytic cycle with platinum, to our
knowledge B is known only with phosphines as the ancillary
ligands on platinum,5 with the exception of one unique series
of compounds containing the aromatic η5-cyclopentadienyl
ligand, and a few more recent examples with nitrogen-based
donors.6 Only a few compounds of the type D (again, with
phosphines and select alkyl/aryl groups) have been reported in
The elementary steps of the Chalk-Harrod catalytic cycle
consist of the formation of reactive intermediates B, C, and D
via oxidative addition, alkene coordination to the metal center,
* Address correspondence to this author. E-mail: (personal) arooproy@
aol.com. Current address: 505 Stillmeadow Ln., Midland, MI 48642.
(1) (a) Marciniec, B., Ed. ComprehensiVe Handbook on Hydrosilylation;
Pergamon Press: Oxford, England, 1992, and the review references therein.
(b) Ojima, I. In The Chemistry of Organic Silicon Compounds; Patai, S.,
Rappoport, Z., Eds.; Wiley: New York, 1989; p 1479. (c) Harrod, J. F.;
Chalk, A. J. In Organic Synthesis Via Metal Carbonyls; Wender, I., Pino,
P., Eds.; Wiley: New York, 1977; p 673.
(2) Speier, J. L.; Webster, J. A.; Barnes, G. H. J. Am. Chem. Soc. 1957, 79,
974.
(3) Speier, J. L. In AdVances in Organometallic Chemistry; Stone, F. G. A.,
West, R., Eds.; Academic Press: New York, 1979; Vol. 17, pp 407-447.
(4) Chalk, A. J.; Harrod, J. F. J. Am. Chem. Soc. 1965, 87, 16.
(5) Corey, J. Y.; Braddock-Wilking, J. Chem. ReV. 1999, 99, 175.
(6) (a) Boardman, L. D. Organometallics 1992, 11, 4194. (b) Reinartz, S.;
White, P. S.; Brookhart, M.; Templeton, J. L. Organometallics 2000, 19,
3748. (c) Fang, X.; Scott, B. L.; Watkin, J. G.; Kubas, G. J. Organometallics
2000, 19, 4193.
9
9510
J. AM. CHEM. SOC. 2002, 124, 9510-9524
10.1021/ja0127335 CCC: $22.00 © 2002 American Chemical Society

Alkene−Platinum−Silyl Complexes
A R T I C L E S
Scheme 2. Modified Chalk-Harrod Mechanism
Because dark-colored solutions are often observed in Pt-
catalyzed hydrosilation and precipitation of black, metallic Pt
at the end of the reaction or even during catalysis is a common
occurrence, the role of homogeneous catalysis has been unclear.
In 1986, Lewis and Lewis proposed18 that colloidal Pt particles
(formed via reduction of Pt complexes with the reactant silane)
were a prerequisite for catalytic activity for most platinum
complexes, except for clearly homogeneous ones such as
(Ph3P)4Pt. Subsequent studies by Lewis’s group19,20 using
Karstedt’s catalyst, Pt(1,3-divinyl-1,1,3,3-tetramethyldisilox-
ane)3, as well as H2PtCl6 led to the following conclusions: (1)
The vinylsiloxane ligands surrounding platinum in the Karstedt
catalyst are hydrosilated away prior to reaction of the substrate
alkene. This constitutes the induction period. (2) During the
most active stages of hydrosilation, only species containing
Pt-C and Pt-Si bonds (but no Pt-Pt bonds) were observed,
irrespective of the nature or stoichiometry of the reagents: thus,
only mononuclear Pt species were postulated to be catalytic.
(3) The platinum end products depended strongly on reagent
stoichiometry and alkene coordination strength. With excess
strongly coordinating alkene, only species with Pt-C bonds
were present. But, for weakly coordinating alkenes (even in
excess over SiH) or with excess hydrosilane, only species with
Pt-Si and Pt-Pt bonds were observed at the end. These results
were also identical when chloroplatinic acid was used as the
precatalyst. (4) On the basis of these observations and others
reported in the paper, the researchers conclude that mononuclear,
homogeneous, and likely structurally identical Pt species are
responsible for hydrosilation catalysis using ordinary metal
complexes. Further, oxygen acts to break up the Pt-Pt bonded
species formed during hydrosilation of weakly coordinating
alkenes, such that a constant supply of the catalytic mononuclear
compounds can be maintained, and finally, Pt-Pt multinuclear
species are likely responsible for olefin isomerization.
the context of hydrosilation.7,8 Species A and C have never been
observed for Pt, although a few examples of C are known for
the much less active metals Ir and Rh.5,9,10
Although the basic aspects of the Chalk-Harrod mechanism
still enjoy wide acceptance, a number of phenomena, including
an induction period for many precatalysts and the formation of
vinylsilanes, are not explained by this mechanism.11 Scheme 2
describes a catalytic cycle based on silyl transfer to alkene
occurring first (E to F), followed by oxidative addition of silane
and reductive elimination of the silylated alkane from G. This
order of migratory insertion and elimination is the opposite of
that proposed in the Chalk-Harrod mechanism and has been
suggested by several researchers on the basis of identification
of intermediates in hydrosilations catalyzed by metals such as
Fe, Co, Rh, and Pd.1c,12-17 This representation of the modified
Chalk-Harrod mechanism implies that a metal-silyl species,
A′, is the active catalyst. However, the scheme can just as easily
be redrawn to accommodate species A, B, and C from the
Chalk-Harrod cycle. The main difference between the two
mechanisms is in the intermediates D and F (formed via hydride
transfer to alkene or silyl transfer to alkene, respectively). From
Scheme 2, the formation of vinylsilane products can be
explained via â-hydride transfer to the metal in F, followed by
vinylsilane decomplexation. It is noteworthy that intermediates
of the type E, F, or G have not been reported in the literature
for platinum, and vinylsilane formation to any significant extent
in Pt-catalyzed alkene hydrosilation is rare. On the contrary,
alkene isomerization, a common phenomenon in Pt-catalyzed
hydrosilation, can be readily explained from species D via
â-hydride transfer to the metal followed by isomerized alkene
decomplexation. Thus, the modified Chalk-Harrod mechanism
is a plausible pathway in hydrosilation catalyzed by metals of
lower activity in this reaction, but its applicability to Pt-catalyzed
hydrosilation as the primary mechanistic pathway is doubtful,
particularly in light of the findings reported here.
Just prior to the publication of the work of Lewis et al. in
1999, a report appeared on the efficient, homogeneous hydrosi-
lation of olefins using Karstedt’s catalyst modified with electron-
deficient ligands such as naphthoquinones, tetracyanoethylene,
fumaronitrile, etc.21 Not only were all the naphthoquinone-
modified catalysts more active than Karstedt’s catalyst, but the
reactions remained completely homogeneous, whereas darkening
of solution color and precipitation of Pt was observed during
the course of hydrosilation with Karstedt’s catalyst alone. In
addition, a significantly higher total turnover was achieved with
excess quinone-based ligand present in the system. An even
more active catalyst was one where the chelating siloxane ligand
was replaced with two norbornene ligands, the third alkene on
Pt being a methylnaphthoquinone. Homogeneity and heteroge-
neity test results showed that reactions with the modified
catalysts were clearly homogeneous but hydrosilations with
Karstedt’s catalyst were apparently homogeneous at the begin-
ning and likely a combination of homogeneous and heteroge-
neous as the reaction progressed. The report concludes that the
electron-deficient ligands help stabilize and solubilize the active
(7) Ozawa, F.; Hikida, T.; Hayashi, T. J. Am. Chem. Soc. 1994, 116, 2844.
(8) Ozawa, F. J. Organomet. Chem. 2000, 611, 332.
(9) Carlton, L.; Molapisi, J. J. J. Organomet. Chem. 2000, 609, 60.
(10) Duczmal, W.; Maciejewska, B.; Sliwinska, E.; Marciniec, B. Transition
Met. Chem. 1995, 20, 162.
(11) For a concise but excellent review of mechanistic studies, see: Chan, D.
Ph.D. Thesis, University of York, York, UK, 1999.
(12) Seitz, F.; Wrighton, M. S. Angew. Chem., Int. Ed. Engl. 1988, 27, 289.
(13) Tanke, R. S.; Crabtree, R. H. Organometallics 1991, 10, 415.
(14) Bergens, S. H.; Whelan, P. N. J. Am. Chem. Soc. 1992, 114, 2128.
(15) Duckett, S. B.; Perutz, R. N. Organometallics 1992, 11, 90.
(16) Yasue, T. R. Organometallics 1996, 15, 2098.
(18) Lewis, L. N.; Lewis, N. J. Am. Chem. Soc. 1986, 108, 7228.
(19) Lewis, L. N.; Stein, J.; Gao, Y.; Colborn, R. E.; Hutchins, G. Platinum
Metals ReV. 1997, 41, 66.
(20) Stein, J.; Lewis, L. N.; Gao, Y.; Scott, R. A. J. Am. Chem. Soc. 1999, 121,
3693.
(17) LaPointe, A. M.; Rix, F. C.; Brookhart, M. J. Am. Chem. Soc. 1997, 119,
906.
(21) Steffanut, P.; Osborn, J. A.; DeCian, A.; Fisher, J. Chem. Eur. J. 1998, 4,
2008.
9
J. AM. CHEM. SOC. VOL. 124, NO. 32, 2002 9511

A R T I C L E S
Roy and Taylor
Figure 1. Expanded 13C NMR spectrum of 1a (reaction mixture) showing Pt-coupled carbons and signals due to other products.
catalytic species, and that monomeric and dimeric (hydrido-
bridged) species are likely involved in the catalytic cycle.
Preamble to the Present Work. The observation that the
cyclic olefins cyclooctene (COE) and cycloocta-1,5-diene (COD)
are highly resistant to hydrosilation even at low concentration
levels, but in their presence hydrosilations using platinum
precatalysts such as H2PtCl6 appeared to remain completely
homogeneous on repeated use of the charge of platinum, led us
to postulate that these olefins remained coordinated to platinum
throughout the hydrosilation cycle. Thus, they could potentially
serve as chemical tracers for studying and “trapping” reactive
intermediates in Pt-catalyzed hydrosilation. Accordingly, this
initial study was designed to investigate stoichiometric reactions
of a series of hydrosilanes with simple platinum complexes in
the presence of COD in particular. The objective was to attempt
to identify and characterize intermediates and reaction products
from these interactions via multinuclear NMR spectroscopy.
PtCl2 and Pt(COD)Cl2 were used as starting platinum com-
pounds in the presence of COD. A deliberate effort was made
to use a relatively high platinum concentration in solution
(0.05-0.3 M) for faster analysis and easier observation with
NMR spectroscopy.
allylic region (due in part to the presence of COD, hydrogenated
COD (COE), and isomerized COD), only two showed clear
coupling to Pt. Farther downfield, a Pt-coupled signal was
present at 74.8 ppm, with JPtC ) 231 Hz. Since complexation
with Pt typically produces significant upfield shifts for olefin
carbon signals, the peak at 74.8 ppm was suspected to be due
to the binding of COD to Pt. However, examination of the
integrals showed that only one CdC of COD was represented
by the signal at 74.8 ppm.
Over 1-5 days, the complex gradually decomposed to
produce more Et3SiCl (the sole species in the 29Si NMR
spectrum at full decomposition) and a small quantity of
Pt(COD)Cl2, together with black metallic Pt. Upon careful,
partial removal of volatiles from a fresh reaction mixture, a
second Pt-coupled olefin signal could be seen at 127.5 ppm
(only about 1 ppm upfield of pure COD) with weak coupling
to Pt (JPtC ) 15.3 Hz), and with a peak integral identical to
that for the Pt-olefin peak at 74.8 ppm (Figure 1). Integral
comparisons now showed the composition of the Pt complex
to consist of one Pt, one COD, and one Et3Si group per “unit”.
Taking into account that the complex on decomposition
produced Et3SiCl, it appeared certain that it also contained one
Cl per Pt as well. Additionally, the 195Pt NMR spectrum showed
only one Pt signal (at -4216 ppm), with the expected 29Si
satellites indicating that only one Pt complex was present. This
required that both pairs of Pt-coupled alkene carbon signals
originate from the same complex. Integration showed that the
ratios of the two Pt-coupled alkene groups and the Et3Si group
was 1:1:1, meaning that there was one silyl group and one COD
per Pt. Additionally, the COD was unsymmetrically bound to
platinum! This is a most unusual bonding mode for COD. In
all metal complexes of cycloocta-1,5-diene that we were aware
of, the two olefin groups were known to be symmetrically
bonded to the metal in an η4 fashion.22 However, in this
complex, with a nearly imperceptible upfield shift of one alkene
signal, it appeared that this alkene group was not directly bonded
to Pt and that the coordination mode for COD was η2 rather
than η4. The most probable structure of this complex, 1a, as a
chloride-bridged dimer, is shown below. It represents the first
Results
Reaction of Et3SiH with PtCl2 (or with PtCODCl2) in the
Presence of COD. Since the formation of highly reactive
species was anticipated, CDCl3 was used as the solvent for
reaction to facilitate NMR analysis of reaction mixtures. A
reaction between PtCl2, Et3SiH, and COD in a 1:1:1 ratio yielded
no observable platinum complex. With 2 equiv of silane and
2-6 equiv of COD per platinum, analysis of the reaction mixture
within 1-2 h of reaction by 1H, 13C, and 29Si NMR spectroscopy
showed the presence of a Pt-alkene complex, COE, COD,
isomerized COD (both 1,4- and 1,3-cyclooctadiene), and Et3-
SiCl. Closer analysis of the NMR spectra showed the following
interesting and unusual features. In the 29Si NMR spectrum,
only two signals were observedsone was determined to be due
to Et3SiCl (against authentic sample), and the other, at 22.9 ppm,
showed distinct satellites due to Pt coupling with 1JPtSi ) 1014
Hz. In the 13C NMR spectrum, two sets of Et3Si signals were
presentsone due to Et3SiCl and a second with Pt coupling to
both the CH3 and CH2 carbons. Among several signals in the
(22) A very recent publication reported the first examples of complexes of COD
with η2 bonding, to our knowledge. Oberbeckmann, N.; Merz, K.; Fischer,
R. A. Organometallics 2001, 20, 3265.
9
9512 J. AM. CHEM. SOC. VOL. 124, NO. 32, 2002

Alkene−Platinum−Silyl Complexes
A R T I C L E S
of Cl, Cl2, and COD from the Pt dimer were observed to be
among the strongest ions in the mass spectrum. The addition
of three protons to a parent ion is unusual, but since NMR had
unambiguously established the species around Pt, the APCI MS
data clearly showed the Pt complex to be the four-coordinate
Cl-bridged Pt dimer.
Reaction of (EtO)3SiH with Pt(COD)Cl2 occurred at 58-60
°C to produce a light yellow solution which turned brown in a
few hours. Although a single Pt-coupled signal among the three
observed in the 29Si NMR initially indicated a clean reaction,
the presence of Si(OEt)4 as a significant silane redistribution
product and several Pt-olefin signals in the 13C NMR spectrum
were clear signs of a complex mixture of products (one of which
could possibly be the expected dimer). Reactions with PhMe2-
SiH and Me3SiOSiMe2H, however, were very clean. Both the
arylsilane and the siloxysilane produced the dimeric platinum
compounds with the same NMR spectral patterns as for the two
chlorosilanes and triethylsilane. For PhMe2SiH, in addition to
the dimer signals, a very small set of secondary Pt-coupled
signals as for the chlorosilanes was evident in the 13C NMR
spectrum. Equation 1 summarizes the reactions of common
classes of silanes with Pt(COD)Cl2, in the presence of an
additional equivalent of COD, to yield the dimers 1. 13C and
29Si NMR chemical shifts for 1 are compiled in Table 1. It is
particularly noteworthy that, for all silanes (except triethoxysi-
lane), the Pt complexes 1 formed in an essentially quantitative
manner as determined by NMR.
Pt complex ever seen containing a simple alkene-Pt-Si moiety.
Though PtCl2 was initially used as the starting platinum
compound, Pt(COD)Cl2 was found to be a much better substrate.
Not only is this white, crystalline COD complex more soluble
in CDCl3, but its reactions were also easier to follow visually
and were cleaner, leading to simpler NMR spectra.
Reaction of Other Representative Organosilanes with Pt-
(COD)Cl2 in the Presence of COD. In accordance with the
original plan of examining a full series of tertiary silanes in
reactions with Pt compounds, representative members of hy-
drochlorosilanes, hydroalkoxysilanes, hydroarylsilanes, and hy-
drosiloxanes were examined. Trichlorosilane failed to react with
Pt(COD)Cl2, even under refluxing conditions or at room
temperature over several days. Very weak signals for possible
Pt-alkene interactions were observable in the 13C NMR
spectrum, but not much else could be deduced. However, both
Me2SiHCl and MeSiHCl2 reacted very cleanly in the presence
of only 1 mol of excess COD per mole of Pt(COD)Cl2. While
dimethylchlorosilane reacted at 55-58 °C, methyldichlorosilane
required refluxing at 65-70 °C for 45 min to 1 h for complete
reaction. In both cases, the reaction mixture was a white
suspension in the beginning but became totally clear and
colorless over a 2-5 min period at the reaction temperature.
For dimethylchlorosilane, the clear solution then quickly turned
yellow to light brown, while for methyldichlorosilane only a
permanent light yellow color developed. The solutions were
cooled quickly after reaction to room temperature or below using
an ice bath. In view of the clean reactions with these two silanes,
it is likely that trichlorosilane failed to react with Pt(COD)Cl2,
in part because of the silane’s low boiling point, but its reaction
under pressure was not examined in this study.
1H, 13C, 29Si, (and 195Pt) NMR spectra of the reaction mixtures
from the two methylhydrochlorosilanes showed signal patterns
identical to those seen for 1a. Inequivalent olefin signals and
allylic carbon signals appeared in the same regions as for 1a,
again clearly pointing to the preference for η2-bonding of COD
to Pt. Additionally, another set of small olefin and MeSi signals
showing coupling to Pt appeared downfield of the corresponding
signals for the main dimer compounds in the 13C NMR spectra
for both chlorosilane reactions. Single signals (with respective
satellites) corresponding to these smaller 13C peaks were also
seen in the 29Si and 195Pt NMR spectra (see below).
The clean formation of structurally identical platinum dimers
from a representative range of common silanes strongly suggests
that this is a common path taken by platinum-chloride-based
precatalysts in their initial reaction with silanes in the presence
of alkene. Various aspects of this striking observation leading
to the first series of platinum complexes containing simulta-
neously bound alkene and silyl groups (alkene-Pt-Si) are
addressed in the Discussion section.
As NMR analysis showed in each of the above three cases,
the initial Pt complex formed contained one monodentate COD,
one silyl group, and one Cl per Pt. The question remaining was
whether the complex is monomeric or a Cl-bridged dimer. APCI
MS was performed on the solution resulting from the reaction
of MeSiHCl2 with Pt(COD)Cl2 and 1 equiv of excess COD.
The most abundant ion observed in the high mass envelope was
909 Da, which corresponds to the Pt dimer plus three protons.
The isotope pattern observed was similar to the calculated
pattern. No ion was observed in the mass range of a monomeric
Pt complex, but several fragment patterns corresponding to loss
Brief attempts were made to prepare monoolefin (cyclooctene,
COE) as well as norbornadiene (NBD) equivalents of 1.
However, these reactions led to mixtures containing only
chlorosilane (indicating unstable intermediates) or produced the
COD-based dimer only. A few reactions of Pt(COD)Ph2 (readily
soluble in CDCl3) with Et3SiH were also carried out to quickly
examine the effect of using a non-chloride-based starting Pt
compound. Reaction was extremely slow under conditions
similar to those used for eq 1. Forcing conditions produced
mostly decomposition products, although some Pt-olefin signals
could be clearly detected.
9
J. AM. CHEM. SOC. VOL. 124, NO. 32, 2002 9513

A R T I C L E S
Roy and Taylor
Table 1. 13C and 29Si NMR Shifts of the Indicated Signal (in ppm) and Pt Coupling Constants (in Hz) for the Pt Dimer Series 1
13C
29Si
Pt dimer
CdC(1), JPtC
CdC(2), JPtC
allyl(1), JPtC
allyl(2), JPtC
Me, JPtC
SiCH , JPtC
silyl, JPtSi
2
1a
1c
1d
“1e”
1f
127.5, 15.3
126.7, <1
126.4, none
83.6, 260a
127.0, 12.8
127.0, 15
74.8, 230.7
86.2, 200.2
82.7, 218.1
80.7, 253a
77.3, 230.2
77.8, 235.6
32.3, 31.7
32.0, 24.9
31.9, 27.8
34.6, 35
32.0, 30.4
31.3, 31.8
26.1, 31.1
26.8, 21.2
26.6, 27.1
24.2, 52
26.2, 29.3
26.4, 31.8
8.8, 18.3
10.6, 88.9
7.0, 62.2
5.4, 36.0
22.9, 1014
18.2, 1791
31.2, 1329
-23.5, 2659
1.4, 1053
0.1, 37.7
5.2, 66.2
1g
1.56, <1b
8.2, 1151
a Several Pt-alkene resonances were observed in the 13C spectrum. Peak assignments for “1e” are tentative. b SiCH3.
Reactivity of Bridged Platinum Dimers toward Additional
Silane. Observation of the formation of structurally identical
Pt dimers from a range of silanes immediately raised the
question of whether these compounds were just initial inter-
mediates or actual catalytic species. To address this, interaction
of these dimers with additional silane (and alkene) needed to
be examined. Early probing of interaction of 1a with excess
Et3SiH had indicated complete decomposition of the dimer (the
likely reason that Lewis and co-workers and other researchers
before them did not observe compounds of type 1 in their studies
with Pt(COD)Cl2 and silanes). However, the second set of
apparently paired Pt-alkene and Pt-SiMe signals observed in
the 13C NMR spectra during dimer formation with MeSiHCl2,
Me2SiHCl, and PhMe2SiH raised the question of whether these
signals were due to precursors of 1c, 1d, and 1f, respectively,
or whether they were due to subsequent reaction products.
To simplify NMR spectral analysis of products, dimer 1f was
chosen first for interaction with additional silane to minimize
reactive Si-Cl bonds in the system. A measured aliquot of 1f
in CDCl3 was transferred to a 16 mm Si-free NMR tube at -70
°C. One equivalent of cold PhMe2SiH (per Pt) was added, and
the tube was shaken on a vortex mixer and quickly inserted
into the NMR probe, which was precooled to -65 °C. 1H, 13C,
and 29Si NMR spectra were acquired. These spectra all showed
the presence of PhMe2SiH and 1f in 1:1 ratio. Upon raising the
temperature to -50 °C, clear evidence of reaction was observed.
Comparisons with the 13C and 29Si NMR spectra from the
preparation of 1f showed that about half of 1f had been
converted to the new Pt complex 2f with concomitant loss of 1
equiv of PhMe2SiCl. Although the 13C NMR spectrum was
rather crowded due to the presence of many signals from
remaining 1f and byproducts from its synthesis, new Pt-coupled
olefin (δ 110.4 ppm) and SiMe2 signals due to 2f were now
prominent. The new signals exactly corresponded to those seen
during the preparation of 1f. The ratio between these signals
was roughly 1:1, suggesting a normally (η4) bound COD and
two PhMe2Si groups on Pt. The 29Si NMR spectrum showed
only three signalssdue to PhMe2SiCl, 1f, and a new signal with
much larger platinum coupling to silicon (JPtSi ) 1532 Hz) than
in 1f itself (JPtSi ) 1053 Hz). Further structural characterization
via NMR of 2f was difficult due to the complexity of the 13C
NMR spectrum and onset of decomposition above ca. -30 °C.
Much clearer structural insight on this new type of platinum
complex was obtained when the reaction was repeated with
MeSiHCl2. One equivalent of MeSiHCl2 (per Pt) was added to
a CDCl3 solution of 1c (prepared as in eq 1) in an NMR tube
in the spectrometer probe at -60 °C. The temperature was then
raised in 20-30 °C steps. No reaction was apparent until 40-
45 °C, when, as with PhMe2SiH and 1f, a growing set of new
Pt-olefin and Pt-SiMeCl2 signals was observed in the 13C
NMR spectrum. These new signals appeared at chemical shifts
identical to those seen in the preparation of 1c above. The
formation of compound 2c, with 13C NMR spectral signal and
integration patterns identical to those of 2f, was further
confirmed from the 29Si NMR spectrum, where a new signal
with larger Pt-Si coupling (JPtSi ) 2110 Hz) than in 1c (JPtSi
) 1791 Hz) was prominent. Since only about half of 1c had
reacted with 1 equiv of silane, accompanied by the formation
of an equivalent of MeSiCl3, it was apparent that an additional
equivalent of MeSiHCl2 would likely convert 1c completely to
2c. This was confirmed experimentally, with a second equivalent
of silane reacting to transform the remaining 1c to 2c. Further,
a separate experiment showed that when 2 equiv of MeSiHCl2
per Pt were added at one time, 1c was completely converted to
2c. Even more significantly, Pt(COD)Cl2 could be converted
completely, in a near quantitative fashion, to 2c via reaction
directly with 4 equiv of MeSiHCl2, in the presence of 2-3 equiv
of excess COD. The excess COD is needed because COD is
lost due to both hydrogenation and isomerization reactions. The
significant implications of these last two processes with COD
are discussed later. The synthesis of the bis-silylplatinum (bs-
Pt) species 2c via various but equivalent paths is summarized
in Scheme 3.
The clean synthesis of 2c and the ability to remove most
byproducts and solvent yielded excellent NMR spectra. From
the 13C NMR spectrum of 2c, it was now clear that this new Pt
complex contained only symmetrically (η4) bound COD (as
evidenced by a single olefin carbon signal with weak Pt-alkene
coupling at 117.4 ppm and a single allylic carbon signal) and
two SiMeCl2 groups per Pt. Further, since only Si coupling and
no hydride coupling to the single 195Pt signal (-4577 ppm, JPtSi
) 2110 Hz) was observed and no Pt-hydride signal was
1
discernible in the H NMR spectrum, any oxidatively added
Si-Pt-H-type complex could be ruled out. Figure 2 shows the
29Si and 13C NMR spectra for 2c. Chemical shifts and coupling
constants are tabulated in Table 2. While molecular ions for 2c
could not be directly observed via mass spectrometry despite
several attempts using TOFSIMS as well as APCI MS tech-
niques, single-crystal X-ray structure analysis showed that the
structure of this compound is, indeed, that of the expected
“square-planar” Pt(II) complex (see Discussion section).
Although a direct attempt was not made in this study to
specifically prepare the Me2ClSi- analogue 2d, its formation
during the preparation of the corresponding dimer 1d is readily
apparent from the small but identical signal patterns in the 13C,
29Si, and 195Pt spectra for 1d, as mentioned in the previous
section (see Tables 2 and 3 for spectral data on 2d). Platinum
complex 2b, based on -SiCl3 groups on platinum, and which
was not readily accessible via the above route, was prepared
via a different procedure, as discussed below. It should be
9
9514 J. AM. CHEM. SOC. VOL. 124, NO. 32, 2002

Alkene−Platinum−Silyl Complexes
A R T I C L E S
Scheme 3. Formation of 2c via Equivalent Routes
Table 3. 195Pt NMR Shifts for Pt Dimer Series 1 and the bs-Pt
Series 2
Pt dimer
ppm
bs-Pt
ppm
1a
1c
1d
“1e”
1f
-4212
-4042
-4134
-4126a
-4193
-4168
2c
2d
-4573
-4558
1g
a Identification of “1e” is tentative.
Figure 2. (a) 29Si NMR spectrum of 2c (crude). (b) 13C NMR spectrum of
2c (recrystallized).
Table 2. 13C and 29Si NMR Shifts of the Indicated Signal (in ppm)
and Pt Coupling Constants (in Hz) for Bis-silylplatinum (bs-Pt)
Complex Series 2
13C
29Si
bs-Pt
vinyl, JPtC
allyl, JPtC
SiMe, JPtC
silyl, JPtSi
2c
2d
2f
117.4, 15.7
114.7, <1
110.4, <1
29.1, <1
buried
28.9, <1
15.0, 180
10.2, 137.2
3.8, 95.1
35.0, 2110
39.9, 1694
-2.9, 1532
pointed out at this juncture that the compound identified as “1e”
(via 29Si NMR) in eq 1 could possibly be the bis-silyl-Pt-COD
complex “2e” instead, since the coupling constant for the Si-
Pt interaction is fairly large (JPt-Si ) 2659 Hz for “1e”)
compared to the values for the other dimers 1.
The scarcity of 195Pt NMR data (and even 29Si NMR data)
on Pt-Si complexes has been a handicap in understanding
structure-property relationships in this area.5 Thus, in this study,
195Pt NMR spectra were acquired for both the Pt dimer and
bs-Pt series where possible, and stacked plots of expanded
spectra are presented in Figure 3. For the dimers, the spectra
are arranged to reflect an apparent effect of silyl substituents
Figure 3. Stacked plots of 195Pt NMR spectra of the Pt dimer series 1 and
the bis(silyl)Pt series 2.
on Pt chemical shift (Table 3). The fortuitous appearance of
signals for both 1 and 2 in the same chemical shift window
greatly facilitated their identification. A normal electronic effect
of silicon substituents on the trend in 195Pt chemical shifts for
the dimers 1 is apparent, but what seems to be the opposite
effect in 2 cannot be confirmed without data on more members
of the bis-silyl complexes.
Are Complexes 2 Catalytic? Reactions with Silane and
Alkene. As with the dimers 1, the question of whether
9
J. AM. CHEM. SOC. VOL. 124, NO. 32, 2002 9515

A R T I C L E S
Roy and Taylor
complexes 2 were catalytic in nature was addressed in light of
this study’s focus to identify and characterize reactive inter-
mediates in Pt-catalyzed hydrosilation. Since a deliberate
strategy was to examine reactions at stoichiometric levels of Pt
using concentrations readily detectable via NMR, a reaction
between 1-hexene and MeSiHCl2 was examined in the presence
of 0.5 equiv of 2c as prepared in solution from Pt(COD)Cl2
and MeSiHCl2. The alkene was added at -30 °C to an NMR
tube containing freshly prepared 2c. This precipitated some of
the 2c. No interaction of 1-hexene alone with the Pt complex
was observed up to 25 °C. The NMR tube was cooled to -65
°C, and 1 equiv of MeSiHCl2 was added. Upon raising the
temperature in ca. 30 °C steps, a slow reaction began at 35 °C
and was complete in about 3 h at 50 °C. At this point, the NMR
tube contained only n-hexylmethyldichlorosilane and 2c, in
addition to the byproducts already present from the synthesis
of 2c (eq 2). It is likely that the excess olefin byproducts present
fast rate of reaction that was actually observed. Thus, the bis-
(silyl)platinum complex, 2c, is a true, active hydrosilation
catalyst (in its resting state).
The catalytic nature of 2c was further corroborated by the
observation that with excess COD (but no hydrosilatable alkene)
in the system, addition of 1 equiv of MeSiHCl2 and storage of
the mixture at room temperature for 5 days caused extensive
isomerization of COD to essentially all 1,3-cycloctadiene with
some hydrogenation to cyclooctene. Only a very small loss of
2c and MeSiHCl2 was observed. Upon heating this mixture at
50 °C for 15 h, the silane was consumed via what appears to
be hydrosilation of 1,3-cyclooctadiene, but again little loss of
2c was noted.
Silyl-Exchange Reactions with 1c and 1d. Observations on
the relative stabilities of the complexes within the series 1 and
2 were indicative of higher strength of the Pt-Si bond in those
complexes with greater electronegativity of the substituents (and
the number of such substituents) on silicon. As expected, when
dimers 1c and 1d were allowed to react with HSiCl3 and
MeSiHCl2 (2 equiv per Pt), respectively (eqs 4 and 5), the
predominant (>90%) bis-silyl-Pt complex formed was the one
containing the more electronegative silyl group. Although 2b
2 MeSiHCl2
2c + 2 hexene-1
8 2 n-hexylMeSiCl2 + 2c (2)
50 °C
slowed the hydrosilation reaction by effectively competing with
1-hexene for Pt complexation. No unreacted 1-hexene or any
of its isomerization products was observed, nor was any
decomposition of 2c detected. This provided the first clue that
2c was likely a true catalyst by definition, i.e., a species
facilitating transformation of the reactants to product but not
exhibiting structural change itself at the end. Given the observed
susceptibility of compounds 2 to undergo further interaction with
silane (that leads to rapid decomposition at elevated tempera-
tures) in the absence of hydrosilatable alkenes, it is unlikely
that all of the 2c initially present would have survived in the
presence of silane if 2c were not playing a true catalytic role in
the conversion of 1-hexene to n-hexylsilane.
To dispel the possibility that a catalytic amount of another
Pt complex, rather than 2c, was responsible for the observed
catalysis, an aliquot of the above reaction mixture containing
50 µmol of 2c per mole of alkene was used to catalyze the
hydrosilation of 1-hexene (0.1 mol) with MeSiHCl2 in a flask
under nitrogen at ca. 60 °C. The reaction was complete at the
end of silane addition (no reflux of unreacted silane). NMR
analysis of the light yellow solution showed complete conversion
of 1-hexene to n-hexylmethyldichlorosilane with no trace of
1-hexene or any isomerized product (eq 3). The solution
is highly insoluble in common, nonreactive solvents (which
thwarted purification and elemental microanalysis), its weak
NMR signals (in CDCl3 or C2D2Cl4) followed the patterns for
2c, 2d, and 2f, so that combined with NMR signal of byproducts
(>90% MeSiCl3, <10% SiCl4, all HSiCl3 and 1c consumed),
its identity was not in question. Despite its initial insolubility
in the reaction medium, 2b readily catalyzes hydrosilation
reactions, just like 2c. For example, at 60 µmol of Pt, 2b acts
as a fast catalyst for the hydrosilation of styrene with MeSiHCl2.
Interestingly, the ratio of internal to terminal hydrosilation
products is much higher (58:42) than seen with the usual
homogeneous Pt precatalysts, and this ratio is unchanged on
using the original hydrosilation product mixture as heel for a
second run.
50 µmol of 2c
(aliquot from eq 2
product mixture)
60 °C
hexene-1 + MeSiHCl2
8 n-hexylMeSiCl2 (3)
remained yellow, with no Pt precipitation observable for several
weeks, unlike the brown solutions and platinum precipitation
within a day that is typically observed with common catalysts
such as H2PtCl6.
Since a homogeneous Pt complex can be readily detected
even at close to 1% level by NMR (depending on ligand NMR
characteristics), it is clear that the above aliquot containing 50
µmol of 2c would contain less than about 0.5 µmol of another
homogeneous Pt complex (no other homogeneous Pt species
was evident by NMR in the stoichiometric reaction). At this
low level, catalysis of the 1-hexene hydrosilation by such a
second species would be expected to be quite slow, unlike the
Discussion
I. Alkene-Pt-Si Dimeric Complexes 1. The series of
dimeric Pt complexes, 1, has provided a first clear glimpse of
a class of platinum compounds that contain a π-bonded simple
alkene and silyl groups simultaneously bound to platinum.
Hence, several aspects of their synthesis, characterization, and
structural characteristics are discussed below.
9
9516 J. AM. CHEM. SOC. VOL. 124, NO. 32, 2002

Alkene−Platinum−Silyl Complexes
A R T I C L E S
A. Formation of 1. The likely reason complexes of the type
1 were not detected in earlier studies4,18,23,24 with Pt-COD
complexes is a combination of two reaction condition factors:
the lack of sequential and controlled addition of equivalents of
silane and the presence of excess COD. Excess silane (often
used in catalyst preparation) beyond the precise equivalents per
Pt necessary to form 1 and 2 quickly leads to the decomposition
of 2, especially at elevated temperatures.
Several observations in connection with the formation of
complexes 1 call for comment. Although essentially all the
reactions with PtCl2 and Pt(COD)Cl2 were carried out in CDCl3,
a few reactions with Et3SiH were performed in refluxing toluene.
These clearly showed the presence of 1a in the product mixture,
albeit in very small amount due to the rapid decomposition of
1a at the elevated temperature used. Thus, the formation of 1
as a common reactive intermediate is likely not dependent on
a strong solvent effect, provided the solvent does not possess
coordinating characteristics.
As was commonly observed in this study, hydrogenation of
COD to COE, and even to some cyclooctane, has been reported
by Lewis and Lewis for silane reactions with Pt(COD)Cl2.18
Hydrogenation of styrene to ethylbenzene in (styrene)2PtCl2 by
2 equiv of Et3SiH/Pt, producing 2 equiv of Et3SiCl and the
complex (styrene)3Pt in the presence of excess styrene, has been
reported by Caseri and Pregosin.23 Thus, it is apparent that
complexes of Pt(2+) containing chloride and alkene ligands or
added alkene undergo this general reaction of alkene hydrogena-
tion and reductive elimination of chlorosilane to produce Pt(0)
species stabilized by complexation with suitable alkene. In this
study, however, COD presents a special situation where the
hydrogenated COD (COE) can be very rapidly displaced by
the much stronger chelating COD present in solution. Without
excess COD present, only some dimer is observed, with
significant precipitation of metallic Pt. However, the hydrogena-
tion of COD during the formation of 1 is not efficient (often
30% or less, the rest of the hydrogen escaping as gaseous H2).
Though the exact sequence of steps and intermediates is not
yet clear, it is likely that via some sort of steric-electronic
stabilization, COD provides an opportunity for the dimeric Pt-
(2+) complexes 1 to form and remain relatively stable to
chlorosilane elimination. It might be noted in this connection
that we are aware of only a small number of other examples of
stable platinum complexes containing a silyl group and chloride
simultaneously bound to platinum (Si-Pt-Cl), and these contain
phosphines as auxiliary ligands but no alkene.5,25,26 Another
intriguing aspect of the reaction between Pt(COD)Cl2 and the
silanes that were examined is the long induction period
(particularly for the chlorohydridosilanes) before any visible
reaction. Yet once the reaction started, essentially all of the Pt-
(COD)Cl2 reacted within 1-3 min. It is possible that, at least
for chloride-based platinum precatalysts, this represents part of
the induction period before the formation of the active catalyst.
B. Structural Aspects of 1. The unique structural feature of
compounds 1, with η2-bonded COD, is rather intriguing,
although their similarity to Zeise’s dimer-type complexes is
apparent. First, all of the clearly identified members of the series
appear to be rather stable to monomer formation. 13C NMR
signals, particularly for the Pt-alkene and allylic carbons,
remain sharp over a wide temperature range (from -65 to 55
°C for 1c), maintaining constant JPtC coupling. This is indicative
of the absence of both rapid olefin exchange and monomer-
dimer equilibrium. The rather small or nonexistent coupling of
the nonbonded olefin of COD to Pt and the insignificant upfield
shift of these carbons (relative to free COD) are suggestive of
interaction through σ bonds with platinum rather than a weak
π-interaction, although the latter cannot be completely ruled
out from the available data. Since several bis(alkene)PtCl2
complexes are known in their monomeric form, the question
arises as to why complexes 1 would be dimeric. It is highly
probable that the answer lies in the well-known, relatively strong
trans-influence of silicon. The CdC of COD that would be trans
to the silyl group in a monomeric complex would experience a
significant bond-weakening effect in its interaction with Pt. As
a result, the complex might be expected to prefer bridging
chloride σ-coordination to satisfy preferred four-coordinate
geometry. The dimer-monomer equilibrium that has been
reported for (styrene)2PtCl2 by Pregosin et al.23,27 is disfavored
for 1 due to the trans-influence of silicon. Again, because of
this trans-influence, it might be expected that dimeric complexes
1 would prefer a trans geometry for the silyl and η2-COD ligands
across the two Pt centers. This would ensure that the two
bridging chloride ligands are equivalently bonded to the two Pt
nuclei. However, this aspect of the structure is less certain
without support from direct crystal structure evidence. Because
the salient NMR chemical shifts for all the complexes 1 follow
an identical pattern, the structural aspects of these complexes
are also expected to be identical. Additionally, perhaps the
strongest indirect support for the structural features of 1 is
evident in a report by Pregosin and co-workers, where the
preparation (via different chemistry) of a series of anionic Pt
complexes containing various mono-olefin, chloride, and stannyl
groups on Pt is described.28 These anionic stannyl-Pt-alkene
complexes are direct equivalents of Zeise’s salt and can also
be visualized as anionic monomeric analogues of 1. Indeed, one
might anticipate that such anionic and monomeric complexes
can be made from 1 via appropriate transformations or directly
from Na2PtCl4.
C. Decomposition Pathway. Although byproducts from the
synthesis of several of the complexes can be partially removed,
and the solids left behind are stable for timely NMR analysis,
the slow but continuous elimination of chlorosilane from 1
hampered our ability to purify and maintain these complexes
for crystal structure or elemental microanalysis. The most visibly
noticeable change as 1 decomposes is the steady formation of
black, “metallic platinum”. The major pathway for decomposi-
tion is unquestionably via loss of chlorosilane, as evidenced
from NMR. However, for all members of the series, a small
amount of the starting preparative complex, Pt(COD)Cl2, is also
produced. Monitoring the preparation and decomposition of 1
via 13C NMR showed the total disappearance of the Pt-vinyl
signal due to Pt(COD)Cl2 upon completion of the preparative
reaction, followed by gradual reappearance and growth of this
signal over time (2-15 days) as the signals due to 1 diminished.
The formation of Pt(COD)Cl2 as a decomposition product of 1
(23) Caseri, W.; Pregosin, P. S. Organometallics 1988, 7, 1373.
(24) Caseri, W.; Pregosin, P. S. J. Organomet. Chem. 1988, 356, 259.
(25) Brost, R. D.; Bruce, G. C.; Joslin, F. L.; Stobart, S. R. Organometallics
1997, 16, 5669.
(27) Albinati, A.; Caseri, W.; Pregosin, P. S. Organometallics 1987, 6, 788.
(28) Albinati, A.; Pregosin, P. S.; Ru¨egger, H. Angew. Chem., Int. Ed. Engl.
1984, 23, 78.
(26) Yamashita, H.; Tanaka, M.; Goto, M. Organometallics 1997, 16, 4696.
9
J. AM. CHEM. SOC. VOL. 124, NO. 32, 2002 9517

A R T I C L E S
Roy and Taylor
requires that an accounting for the silyl groups be made. Disilane
formation, an obvious pathway to account for the silicon, was
not detected in any of the decomposition reactions, nor was
any other soluble silicon compound identifiable via NMR.
Therefore, the only logical explanation is the formation of
insoluble Pt-Si species together with the metallic platinum. It
is tempting to speculate whether such Pt-Si species might not
retain the ability to be resurrected to active catalytic species in
the presence of silane and alkene. This mode of silyl loss is
consistent with Lewis’s observation via EXAFS of Pt-Si bonds
(together with Pt-Pt bonds) at the end of hydrosilation with
excess silane20 and might explain why hydrosilation mixtures
that show visible precipitation of “Pt” can still be used to
catalyze further hydrosilation at appreciable rates.
D. Mechanistic Implications. Perhaps the most scientifically
useful aspect of the ability to synthesize relatively stable
complexes of the type 1 lies in unraveling the structural,
compositional, orientatational, and other critical characteristics
of the hydrosilation catalytic pathway. This is an important
primary step toward building a foundation for structure-activity
relationships in hydrosilation catalysis. The observation during
the synthesis of 1 that COD is readily hydrogenated and
isomerized by hydrosilanes but that the silyl group on Pt does
not transfer preferentially or even competitively to the alkene
bound to Pt suggests that hydride transfer to alkene is facile
and is kinetically favored in a competitive mode, at least for
Pt-catalyzed hydrosilation. This directly reinforces the Chalk-
Harrod mechanism. In this context, it should be mentioned that
the strong resistance of COD and COE to hydrosilation can now
clearly be seen to be due to the inability of the silyl group to
transfer to either the alkene or the alkyl group formed via
hydride transfer. Early reports in the literature of COD hydrosi-
lation using Pt as catalyst under ordinary conditions are clearly
questionable. In fact, it is likely (from NMR observations) that
not COD but its isomerization product 1,3-cyclooctadiene
undergoes slow hydrosilation (likely in the 1,4 fashion of
conjugated dienes), as has been suggested by Yamamoto and
Kumada.29 The origin of the resistance of COD and COE to
hydrosilation is most probably steric in nature, as is believed
to be the case for most internal olefins.
II. True Hydrosilation Catalytic Species, (COD)Pt(SiR3)2,
2. Although the true catalysts 2 are remarkably simple in
compositional and structural characteristics, because of their
intrinsic importance to hydrosilation, aspects related to their
formation, structural features, stability, catalytic behavior, and
implications to catalysis necessitate further analysis and discus-
sion.
A. Formation of 2 from 1. Two observations on the synthesis
of 1 and 2 reported in the Results section are very insightful.
First, during the synthesis of 1c, 1d, and 1f, formation of 5-10%
2c, 2d, and 2f, respectively, was typically detected by NMR.
This suggests that the species 2 are preferential products in the
sequential and consecutive reaction of chloride-based platinum
precatalysts with silane, even when there is a stoichiometric
deficiency of silane in the medium. This was further cor-
roborated by a second observation (made during reaction of
silane with 1) that adding only 1 equiv of silane per Pt consumes
one-half of the dimer 1 to yield a 1:1 mixture of 1 and 2. Thus,
the intermediate formed from reaction of silane with 1 must be
much more reactive toward silane than 1 itself. Otherwise, the
intermediate species, or its decomposition products, would likely
have been detected. It is very pertinent to note here that, in a
recent study with cationic phenanthroline-Pd complexes,
Brookhart et al. reported on just such behavior upon reaction
of the Pd precatalysts with 1 or 2 equiv of silane per Pd.17
Indeed, these researchers reported the detection and character-
ization of not only Pd-silyl species that are true hydrosilation
catalysts, but also bis-silyl-Pd complexes containing a hydride
moiety. In the present case with platinum, no hydride resonance
was observed even at temperatures as low as -65 °C. This is
almost certainly due to the fact that the hydride was quickly
consumed via hydrogenation of COD, unlike in the Pd
complexes containing phenanthroline as the ancillary ligand.
The lack of an available group for hydride transfer trapped the
hydride from the second equivalent of silane in the Pd
compounds. This suggests that Pt-H-containing species also
bearing alkene and silyl ligands might be accessible via use of
alkenes that are sterically resistant to hydrogenation, provided
oxidative addition of the silane to Pt is possible. Just such an
opportunity may possibly be presented by the sterically demand-
ing derivative of COD, dibenzo[a,e]cyclooctatetraene (DCT).30
B. Bonding and Structure. The bonding in 2 is expected to
be much like bonding in complexes of the general formula cis-
(alkene)2PtX2, where X is a halide or other anionic ligand. Thus,
structural features of 2 and Pt(COD)Cl2 would be similar, except
that a weaker coordination of COD to Pt in 2 would be
anticipated on the basis of the strong trans-influence of the silyl
groups in a square-planar geometry. This weaker binding of
COD to Pt in 2 is, indeed, evidenced by a significant downfield
shift of the Pt-olefin signal in the 13C NMR spectrum relative
to Pt(COD)Cl2 (110-118 ppm for 2 compared to 100 ppm for
the chloride complex). In a formal sense, the oxidation state of
Pt in 2 is (+2), since the compound can be viewed as formed
from the oxidative addition of a disilane to a coordinatively
unsaturated (COD)Pt(0). Figure 4a and b represent two different
ORTEP views of the crystal structure for 2c. Some pertinent
crystal data are summarized in Table 4.
From the crystal structure for 2c, a nearly square-planar
geometry around platinum is obvious, if one visualizes the
centroids of the two olefinic bonds of COD to be part of the
square plane. Of particular note is the longer (by about 0.17 Å)
Pt-C bonds in 2c compared with those in Pt(COD)Cl2.31 This
lengthening of the Pt-C π-bonds (which are nearly sym-
metrical) may be attributed to the trans-influence of silicon and
correlates well with the Pt-olefin chemical shifts and Pt-C
coupling in the 13C NMR spectra for these two complexes.
Another point to note is the existence of packing disorder in
the crystal with respect to one Cl and the methyl group at each
silicon. The Cl and Me groups above and below the Si-Pt-Si
plane alternate in up and down positions with 50% probability
at each silicon, but only chloride groups [Cl(3) and Cl(3)′] are
observed in the “in-plane” positions at each silicon. The effect
of the differences in bonding mode of COD (in 2) and of 1,1,3,3-
tetramethyl-divinyldisiloxane (in Karstedt’s catalyst) with plati-
num is clearly borne out by the inability of the chelating
disiloxane to displace COD from 2c, even at a 2:1 concentration
advantage in favor of the disiloxane. Yet, it was found that the
(30) Anton, D. R.; Crabtree, R. H. Organometallics 1983, 2, 855.
(31) Ashfaquzzaman, S.; Stevens, E. D.; Cruz, S. G. Inorg. Chem. 1984, 23,
3673.
(29) Yamamoto, K.; Kumada, M. J. Organomet. Chem. 1968, 13, 131.
9
9518 J. AM. CHEM. SOC. VOL. 124, NO. 32, 2002

Alkene−Platinum−Silyl Complexes
A R T I C L E S
hydrosilation catalytic intermediates generated from these
precatalysts. The trans-influence of silyl groups may be of
paramount importance in these areas.
C. Stability of 2: Effect of Silicon Substituents. Complex
2c is relatively quite stable in solution over periods of months
(at -20 °C) in the presence of some excess free COD, which
imparted the stability needed during shipping so that its crystal
structure could be determined. The trichlorosilyl-based complex
2b appears to have much higher stability to spontaneous
decomposition in an inert atmosphere, as no degradation of the
white solid was observed on storage in a flask at room
temperature for a year. Part of this stability may be attributed
to the high insolubility of 2b but perhaps partly also to a smaller
overall trans-influence of silyl groups with more electronegative
substituents. Conversely, the weakening of Pt-COD π-bonding
may be expected with increasing electron-donating substituents
on silicon. This effect is probably reflected in the lower
stabilities for 2 with more methyl (or ethyl) substituents on
silicon. Thus, the triethylsilyl-based 2a was not observed in this
work, at least at room temperature, although this may also in
part be due to steric effects of the Et3Si group. Higher bond
strength for the Pt-Si bond and weaker trans-influence of the
corresponding silyl group with electronegative substituents have
been predicted by Sakaki et al., and most of our observations
related to stability and reactivity of the alkene-Pt-silyl
complexes are in line with these predictions.32 Although a
thorough understanding of the effects of silicon substituents on
catalytic aspects of transition metal behavior is lacking, it
appears that studies are beginning to address the subject of
structure-property relationship in reactions of catalytically
active transition metal complexes with silanes.33
D. Catalytic Behavior: Proposed Activation and Catalytic
Cycle. The singular characteristic of 2 that is of the highest
interest for understanding structural aspects of reactive inter-
mediates in Pt-catalyzed hydrosilation is the true catalytic nature
of this series of silyl-Pt-COD complexes. In addition to
preliminary demonstration of the catalytic nature of 2c, it was
also shown that hydrosilations, such as that of styrene by
MeSiHCl2, are catalyzed by 2b which, though initially insoluble
even at the catalytic level, dissolves in the reaction medium
either upon raising the temperature or upon reaction with the
silane. The reaction mixture remains homogeneous and light
yellow essentially throughout the course of the hydrosilation.
Interestingly, for 2b in particular, precipitation of the off-white
catalyst after consumption of all alkene was observed. This is
in line with the high insolubility of this complex. An important
question that arises with respect to the ability of compounds 2
to behave as true hydrosilation catalysts is whether these
complexes with two silyl groups on Pt are capable of accom-
modating a third silyl group on Pt via oxidative addition of SiH
to 2. Indirect but very strong evidence for this ability is provided
by the known anionic complexes of Pt containing η4-bound
COD and three trichlorostannyl groups, such as [Ph4As]+[(COD)-
Pt(SnCl3)3]-.28 If platinum can accommodate three stannyl
groups and an η4-bound COD, then the simultaneous bonding
of three silyl groups (smaller than equivalent stannyl groups)
and an η4-coordinated COD on platinum can be readily
visualized. Thus, a much clearer catalytic cycle for hydrosilation
Figure 4. (a) “Bird’s-eye” ORTEP view of the X-ray crystal structure of
2c showing that Cl(3)/Cl(3)′ locations are fixed, while the positions of Cl-
(1)/Cl(1)′ and Cl(2)/Cl(2)′ with respect to the Si-Pt-Si plane are up and
down with 50% probability at the two silicon atoms. Hydrogen atoms have
been shown at idealized positions to help visualize the potential for steric
interactions between Si-Me groups and bound COD. (b) ORTEP side-
view of the X-ray crystal structure of 2c showing an essentially square-
planar geometry for ligand bonding around platinum, with the two CdC
centroids as part of the square plane.
Table 4. Basic Crystal Data for Complex 2c
empirical formula
formula weight
crystal color, habit
crystal dimensions
crystal system
C10H18Cl4Si2Pt
531.33
colorless, cut block
0.56 × 0.31 × 0.15 mm3
orthorhombic
primitive
lattice type
lattice parameters
a ) 8.260(2) Å
b ) 16.961(3) Å
c ) 11.406(2) Å
V ) 1597.9(4) Å3
Pbcn (No. 60)
4
space group
Z value
Dcalc
2.208 g/cm3
1008.00
F000
GOF indicator
µ(Mo KR)
2.40
95.41 cm-1
divinyldisiloxane was readily hydrosilated in the presence of
2c without any decomposition or loss of 2c. This, of course,
has important implications in understanding the platinum orbitals
that are used in precatalysts of various oxidation states and in
(32) Sakaki, S.; Mizoe, N.; Sugimoto, M. Organometallics 1998, 17, 2510.
(33) Lemke, F. R.; Galat, K. J.; Youngs, W. J. Organometallics 1999, 18, 1419.
9
J. AM. CHEM. SOC. VOL. 124, NO. 32, 2002 9519

A R T I C L E S
Roy and Taylor
Scheme 4. Formation of Catalytic Species 2 from Starting
Precatalyst: The “Induction Period”
ate with bridging ligands is formed may depend on the nature
of the other ligands at the coordination sites occupied by COD.
Scheme 4 depicts possible pathways for the formation of 1
and 2 from Pt(COD)Cl2 and 4 equiv of silane in the presence
of excess COD. For the formation of 1, both pathways A and
B seem probable, although B may be favored since inefficient
hydrogenation of COD to COE and the visible loss of a good
amount of hydrogen gas is observed. For the formation of 2
from 1, however, a pathway similar to A′ may be preferred,
since during this part of the activation process hydrogenation
of COD to COE is very efficient and suggests that dihydrido-
Pt species are likely not present. The pentacoordinate Pt(4+)
reactive intermediates are visualized on the basis that Jennings
and co-workers have reported a similar Pt(4+) metallacycle
compound where a bicyclic cyclooctene-derived moiety is both
σ- and π-bonded to platinum.34,35 However, in this work, with
additional alkene/diene present in the system, it is possible that
the pentacoordinate intermediates depicted are actually hexa-
coordinate Pt(IV) species.
Comparison of Chalk and Harrod’s original proposed mech-
anism (Scheme 1) with the structurally much-clarified version
that we propose (Scheme 5) by simply substituting 2 for [M]
shows that ours is also fully compatible with the tenets of
fundamental transitional metal chemistry, namely, oxidative
addition, dynamic metal-ligand equilibrium, insertion into
metal-ligand bond, and reductive elimination. The visualization
of the η2-COD complex T η4-COD complex rapid equilibrium
(especially with trans-silyl groups) is supported by Ozawa and
co-workers’ findings of the inhibitory effect of added phosphine
in hydrosilations catalyzed by bis(phosphine)platinum com-
plexes, where the prior dissociation of a phosphine ligand from
Pt is a prerequisite for migratory insertion.7,8,36 In fact, such an
inhibitory effect is often observed when COD is in excess for
hydrosilations catalyzed by Pt-COD complexes.
The delineation of the Chalk-Harrod mechanism that the use
of COD has permitted now allows consideration of the pos-
sibility that the hydrosilation of many alkenes using initial Pt-
(II) and Pt(IV) complexes may actually proceed via catalytic
cycles containing bis-silyl-Pt catalytic species. However, it also
appears from the work of Pregosin et al. with (styrene)3Pt, Lewis
and co-workers’ recent report using Karstedt’s catalyst, and the
results of Osborn’s group with electron-deficient-alkene-modi-
fied Karstedt’s catalyst that tris(alkene)Pt(0) complexes [and
likely other Pt(0) complexes] do not require a complex series
of activation steps (since a reductive elimination to remove
species such as chlorosilanes is not necessary). Thus, catalytic
cycles based on these types of complexes may involve inter-
mediates with only one or two silyl groups on platinum at any
point in the catalytic cycle. Nevertheless, it may be reasonably
inferred at this point in the evolution of Pt-catalyzed hydrosi-
lation understanding that Pt(0) T Pt(II) or Pt(II) T Pt(IV)
transitions could equally efficiently effect hydrosilation catalytic
cycles. The nature of the auxiliary ligands present in the system
might determine the (resting) oxidation state of the active
catalyst species. This flexibility in oxidation states for catalytic
behavior likely partially rationalizes why a myriad of platinum
compounds are such versatile hydrosilation precatalysts.
Scheme 5. Chalk-Harrod Mechanism with Bis(silyl)Pt(COD), 2,
as a True Catalyst
with chloride-based Pt(2+) precatalysts containing COD or
another non-hydrosilatable ligand may now be proposed (Schemes
4 and 5). These proposed catalytic schemes not only are
consistent with the results of this work but also take into account
various observations reported by the groups of Lewis,20
Brookhart,17 Osborn,21 and Pregosin.23,24 The same catalytic
cycle with only minor modifications may well be generally
applicable to common Pt(2+) precatalysts containing other
halide or various oxo species as anionic ligands that are also
capable of bridging coordination. Whether a dimeric intermedi-
(34) Parsons, E. J.; Larsen, R. D.; Jennings, P. W. J. Am. Chem. Soc. 1985,
107, 1793.
(35) Jennings, P. W.; Johnson, L. L. Chem. ReV. 1994, 94, 2241.
(36) Ozawa, F.; Kamite, J. Organometallics 1998, 17, 5630.
9
9520 J. AM. CHEM. SOC. VOL. 124, NO. 32, 2002

Alkene−Platinum−Silyl Complexes
A R T I C L E S
A last but very important point about the hydrosilation
catalytic cycle with discrete molecular precatalysts is the
question of whether the catalysis is homogeneous or heteroge-
neous. The reports of the groups of Pregosin, Lewis, and Osborn,
when taken together, point toward a fully homogeneous catalytic
cycle. Now, the work documented in this report, describing the
identification and characterization of soluble reactive intermedi-
ates and true catalysts, clearly provides direct evidence that the
catalytic process is, indeed, homogeneous with molecular Pt
compounds, as long as Pt-stabilizing and -solubilizing ligands
are present in the reaction medium. Employing a fairly reliable
method reported by Crabtree and Anton,37 where the bulky
dibenzo[a,e]cyclooctatetraene (DBCOT or dct) is used to bind
with and sterically block various metal complexes, Osborn
showed that their Karstedt-based catalyst systems are signifi-
cantly retarded by DBCOT but not by mercury, which is a
known poison for heterogeneous catalysts.21,38 Such tests on
Karstedt’s catalyst seem to suggest both homogeneous and
heterogeneous reactions, and we are currently investigating
hydrosilations with this precatalyst in the presence of COD.
Some interesting and surprising results are emerging, which will
be reported in future publications. Nevertheless, in line with
the suggestions of the groups of Osborn and Lewis in their recent
publications, we conclude that, with molecular Pt compounds,
the catalytic process is completely homogeneous as long as the
Pt can be kept soluble using stabilizing ligands. On isomerization
of alkenes, it is clear from this study that mononuclear
complexes can isomerize alkenes: multinuclear species do not
need to be invoked for isomerization, contrary to Lewis and
co-workers’ suggestion.20 However, the findings of the present
study concur with Lewis’s group’s conclusion that one cause
of deactivation of homogeneous Pt hydrosilation catalytic
species is very likely the formation of metal-metal bonded
aggregates in the absence of stabilizing alkenes or other ligands,
or in the presence of excess hydrosilane.20
of the catalytic cycle that visualize simultaneous bonding of
alkene and silyl groups to the Pt center. Further, the reactivities
shown by 1 and 2, including the lack of any facile interaction
between 2 and 1-hexene, styrene, or vinylsiloxane in the absence
of silane, together with the reported difficulty of silyl group
transfer to ordinary alkenes in other silylplatinum complexes,26
support the sequence of hydride and silyl transfer visualized in
the Chalk-Harrod mechanism. It also appears that the trans-
influence of silicon in metal-silyl complexes may play an
important role in transition-metal-catalyzed hydrosilation.
The identification of species 1 and 2 and implications of their
further utility raise the prospects of additional chemical
investigations with these and similar Pt complexes and other
starting Pt compounds. It is hoped that this modest beginning
in molecular catalytic species identification and structure-
property relationship development will invite further studies that
lead to better understanding of catalyst behavior and improved
reaction control in hydrosilation reactions.
Experimental Section
General. Starting Materials. The following chemicals and solvents
were obtained from internal or commercial sources and used without
further purification unless otherwise noted. Sigma-Aldrich Chemical:
triethylsilane, phenyldimethylsilane, triethoxysilane, COD, COE, 1,3-
cyclooctadiene, chlorotriethylsilane, 1-hexene, styrene, n-hexane, an-
hydrous chloroform (amylene stabilized), anhydrous toluene. Dow
Corning Corp.: trichlorosilane, methyldichlorosilane, dimethylchlo-
rosilane. Gelest, Inc.: pentamethyldisiloxane. Johnson Matthey Co.:
Pt(COD)Cl2, Pt(COD)Ph2. PGP: PtCl2. Isotec, Inc.: CDCl3 (99.8%
deuterated), stored over 4 Å molecular sieves from Fisher Scientific.
No special effort was made to purify or dry reagents, except to use
CDCl3 that was stored over molecular sieves and dry glassware to avoid
reaction with surface moisture.
General NMR Spectroscopic Analysis Conditions. 13C and 29Si
NMR spectra were generally run on a MercuryVx 400 MHz spectrom-
eter equipped with a 16 mm Si-free Nalorac PulseTune probe and
Varian SMS autosample changer. Samples were run in homemade Si-
free NMR tubes previously described.39 1H and 195Pt spectra were
acquired on a MercuryVx 300 MHz spectrometer equipped with a
Varian 5 mm switchable 1H/BB PFG probe. 13C, 29Si, and 195Pt spectra
were acquired with 90° pulses and Waltz decoupling during the
acquisition period only. Relaxation delays of 30, 60, and 4 s were
Concluding Remarks
This study was undertaken with the objective of elucidating
structural and chemical aspects of the hydrosilation catalytic
cycle with homogeneous platinum complexes. Employing a
strategy of using “non-hydrosilatable” COD as an ancillary but
tracer ligand on Pt, two related series of the first complexes
containing alkene and silyl groups simultaneously bound to
platinum were identified and unambiguously characterized. The
dimeric compounds 1 are semistable, precatalytic intermediates
formed initially from the reaction of hydridosilanes with PtCl2
or Pt(COD)Cl2 in the presence of additional COD. The nature
of COD bonding to Pt in these complexes is unique, but the
complexes themselves are analogous to well-known compounds
such as Zeise’s dimer. The second series of compounds, 2,
formed from 1 via further reaction with silane, were shown to
be the first truly catalytic species observed in Pt-catalyzed
hydrosilation studies. The relatively high stability exhibited by
1 and 2 is truly remarkable, considering that such complexes
have never been observed, even as transitory species. Identifica-
tion and characterization of complexes 1 and 2 provide direct
experimental evidence for the various mechanistic postulations
1
utilized for quantitation, respectively. For H spectra, 90° pulses and
15 s relaxation delays were used. All samples were prepared and run
in CDCl3, which was used as an internal reference for 1H and 13C spectra
at 7.25 and 77.0 ppm, respectively. 29Si spectra were externally
referenced to tetramethylsilane at 0 ppm. 195Pt spectra were externally
referenced to H2PtCl6 at 0 ppm. Spectra were run at room temperature
unless otherwise indicated.
General MS Acquisition Conditions. Atmospheric pressure chemi-
cal ionization (APCI) MS studies were carried out on a PE Sciex triple-
quadrupole API 350 mass spectrometer operating at unit mass resolving
power across the entire mass range (30-4500 Da). The APCI source
housing was operated at ambient pressure and temperature of 23 °C.
The APCI sprayer tube temperature was critically optimized such that
solvent evaporation of the sprayed liquid solution could take place
efficiently with minimal thermal decomposition of the protonated
analytes. Sprayer tube temperatures over the range of 75-200 °C were
evaluated. The optimum sprayer tube temperature was found to be 100-
110 °C. A 150 ppm (wt/vol) solution of the Pt sample was prepared in
molecular-sieve-dried CHCl3 (Fisher HPLC grade). The CHCl3 solutions
containing the analytes were directly infused into the APCI ion source
(37) Crabtree, R. H.; Anton, D. R. Organometallics 1983, 2, 855.
(38) Whitesides, G. M.; Hackett, M.; Brainard, R.; Lavalleye, J.-P. P.; Sowinski,
A. F.; Izumi, A. N.; Moore, S. S.; Brown, D. W.; Staudt, E. M.
Organometallics 1985, 4, 1819.
(39) Taylor, R. B.; Parbhoo, B.; Fillmore, D. M. In The Analytical Chemistry
of Silicones; Smith, A. L., Ed.; John Wiley & Sons: New York, 1991.
9
J. AM. CHEM. SOC. VOL. 124, NO. 32, 2002 9521

A R T I C L E S
Roy and Taylor
at a constant flow rate of 30 µL/min using a Harvard Apparatus (model
22) syringe pump, whereby solvent vaporization preceded ionization
of the analytes. APCI MS experiments were performed in the positive-
ion mode. The nebulizing current applied to the nebulizing needle within
the ion source was set to 1.5 µA, and the chemical ionization gas used
was 99.9997% pure CH4 gas mixed with ambient air (present within
the ion source). The observed protonated molecular ions within the
APCI mass spectra are formed by low-translational-energy ion-
molecule reactions that occur in the chemical ionization plasma region
of the ion source. The protonated molecular ions formed in these studies
were subsequently mass analyzed by the PE Sciex API 350 mass
spectrometer and detected utilizing a channeltron detector operating at
a voltage of 1800 V (gain ≈ 105), whereby 10 individual mass spectra
were signal-averaged to provide a single summed mass spectrum.
General Reaction Conditions for the Synthesis of 1 and 2. [R3Si-
(µ-Cl)(η2-COD)Pt]2, (1). Typically, Pt(COD)Cl2 (0.005 mol, 1.87 g)
was weighed out into an oven-dried, three-necked, 100 mL round-
bottom flask, equipped with an egg-shaped Teflon-coated stirring bar.
The flask was fitted with a dropping funnel, a water condenser topped
with a gas inlet, and a 10 mL separatory funnel stoppered with a rubber
septum (for cannulating the reaction mixture directly into NMR tubes).
All joints were connected using Teflon sleeves. The flask containing
Pt(COD)Cl2 was evacuated for 10-15 min at room temperature and
1-2 mmHg. The flask was then backfilled with dry nitrogen. Cycloocta-
1,5-diene (cod or COD, 0.005-0.010 mol) was added to the flask
through the septum on the dropping funnel, using a syringe. Deute-
riochloroform (10 mL) was then added through the dropping funnel to
wash down the COD. The respective silane (0.0102 mol, 2% excess)
was transferred via syringe to the dropping funnel and diluted with
CDCl3 (5 mL). The flask was heated using a water or silicone oil bath
to the temperature range (45-75 °C) used for the various silanes, and
the silane was added dropwise, usually over 10-20 min to the stirring
slurry of Pt(COD)Cl2, and the mixture heated until all Pt(COD)Cl2
reacted (typically, in a rapid exotherm), to yield a yellow-to-brown
solution. (Fairly steady bubbling of hydrogen at the oil bubbler
connected to the flask via the gas inlet was always observed throughout
the silane addition process.) A slight positive pressure of nitrogen was
maintained in the flask. The transparent reaction mixture was maintained
at the reaction temperature for 2-5 min following disappearance of
all suspended Pt(COD)Cl2 and then quickly cooled to room temperature
or below, using an ice bath under the flask. For NMR analysis, an
18-gauge stainless steel cannula was used to transfer the solution in
the flask to oven-dried 16 mm (Si-free) and 5 mm glass NMR tubes
stoppered with rubber septa. Usually, NMR analysis showed nearly
quantitative conversion of Pt(COD)Cl2 to 1. As needed for further low-
temperature reactions or analysis, the NMR tubes were kept in dry ice/
IPA baths, and the reaction mixture (cooled to -78 °C) was transferred
cold to the NMR tubes.
and the mixture became clear and light yellow over 3-5 min following
reflux for 40-50 min. After reflux was maintained for an additional 5
min, the flask was quickly cooled to room temperature with an ice
bath. (Prolonged heating beyond 5-10 min at temperatures above 55
°C leads to slow decomposition of 1c, rendering a gray, metallic color
to the reaction mixture.) 1H NMR (CDCl3, 40C): δ 6.02 (m, 2H, JPt-H
) 28 Hz, olefin), 4.86 (m, 2H, JPt-H ) 71 Hz, Pt-olefin).
[ClMe2Si(µ-Cl)(η2-COD)Pt]2 (1d). The silane was added over 10-
20 min at room temperature. The bath was heated to 55-58 °C. The
mixture became transparent deep yellow at this temperature and was
cooled to room temperature after being maintained at 58 °C for 5-8
1
min following reaction of all Pt(COD)Cl2. H NMR (CDCl3): δ 5.91
(m, 2H, JPt-H ) 26 Hz, olefin), 4.68 (m, 2H, JPt-H ) 75 Hz, Pt-olefin).
“[(EtO)3Si(µ-Cl)(η2-COD)Pt]2” (“1e”). The silane was added over
10 min at room temperature. The bath temperature was raised slowly,
and at 62 °C the mixture became transparent light yellow. The solution
was cooled to room temperature and became light brown in 2-3 h
(NMR analysis was performed on the yellow solution). 1H NMR
(CDCl3): δ 6.12 (m, 2H, JPt-H ) 32 Hz, olefin), 4.84 (m, 2H, JPt-H
86 Hz, Pt-olefin), 4.74 (m, 2H, JPt-H ) 100 Hz, Pt-olefin).
)
[PhMe2Si(µ-Cl)(η2-COD)Pt]2 (1f). The silane was added over 10
min at room temperature (or at 45 °C). Upon heating, the mixture
became transparent yellow and then quickly turned brown at 50-52
°C. After 1-2 min at 50-52 °C, the mixture was quickly cooled to
1
room temperature. H NMR (CDCl3): δ 7.7-7.3 (m, 5H, Ph), 5.95
(m, 2H, JPt-H ) 25 Hz, olefin), 4.07 (m, 2H, JPt-H ) 75 Hz, Pt-olefin),
0.60 (s, 3H, JPt-H ) 17.4 Hz, CH3).
[Me3SiOMe2Si(µ-Cl)(η2-COD)Pt]2 (1g). The siloxyhydride was
added over 10 min at 45-50 °C (the mixture started to clear rapidly
beyond 60-70% silane addition). The reaction mixture was cooled to
room temperature quickly, immediately following completion of silane
addition. NMR analysis showed that even though all Pt(COD)Cl2 was
consumed, nearly 60% of 1g had already decomposed. 1H NMR
(CDCl3): δ 5.86 (m, 2H, JPt-H ) 24 Hz, olefin), 4.39 (m, 2H, JPt-H
80 Hz, Pt-olefin).
)
Synthesis of 2. As reported in the Results section, only 2c (the most
stable of the soluble complexes 2 that have been observed in this study)
was prepared in separate experiments for the purpose of isolation,
purification, and full characterization. The others, 2d and 2f, were
generated and identified in situ. The insoluble complex 2b was prepared
via a different route, as described below.
Preparation, Isolation, and Purification of (η4-COD)Pt(SiMeCl2)2.
A 100 mL, three-neck 14/20 round-bottom flask, equipped with Teflon
sleeves, water condenser, dropping funnel, and Teflon stirring bar, was
charged with Pt(COD)Cl2 (0.015 mol, 5.61 g) and evacuated for 10
min at 1-2 mmHg. After backfill with dry nitrogen, COD (5.5 mL,
0.045 mol) and sieves-dried CDCl3 (10 mL) were added to the flask
via syringe. MeSiHCl2 (6.6 mL, 0.0612 mol) was transferred to the
dropping funnel and diluted with 5 mL of CDCl3. The flask was heated
with water or silicone fluid bath to 60-65 °C, and the silane was added
over 10 min to the stirring slurry in the flask. The funnel was rinsed
with 2 mL of CDCl3. The mixture was heated at 72-75 °C until it
became completely transparent and clear (30-45 min). The bath
temperature was then quickly dropped to 48-50 °C and the mixture
heated at this temperature for 2-3 h, midway through which a white
precipitate of the product formed (depending on concentration in
CDCl3). The mixture was stirred overnight at 30-32 °C. Complete
conversion to product was then achieved (as evidenced from NMR
analysis). 1H NMR (CDCl3): δ 6.16 (m, 4H, JPt-H ) 35 Hz, Pt-olefin),
2.75 (m, 4H, CH2), 2.4 (m, 4H, CH2), 1.11 (s, JPt-H ) 20.8 Hz, CH3).
Purification of 2c was accomplished by adding anhydrous hexanes
(30 mL) to the flask and stirring the mixture at -78 °C for 1 h to
facilitate complete precipitation of the product, followed by cold
(medium porosity glass frit) filtration under nitrogen and washing with
three 12-15 mL portions of dry hexanes. The microcrystalline solid
on the filter was dried directly on the frit under vacuum and transferred
For the attempts to isolate several 1, the flask was connected to a
dry ice/IPA trap, and volatiles were slowly removed under reduced
pressure at temperatures no higher than about 40-45 °C. The solids
obtained with 1a, 1c, 1d, and 1f were always contaminated with some
black metallic platinum and residual high-boiling COD, COE, byproduct
chlorosilane, or small quantities of 2 (2c and 2d, respectively, in the
case of MeSiHCl2 and Me2SiHCl).
Reaction Conditions for Individual 1. [Et3Si(µ-Cl)(η2-COD)Pt]2
(1a). The silane was added over 20 min at 45-47 °C. Following about
90% silane addition, a fast exotherm was noted, and the reaction mixture
quickly became transparent yellow over 1-2 min, and then turned light
brown. 1H NMR (CDCl3): δ 5.89 (m, 2H, JPt-H ) 26 Hz, olefin), 4.68
(m, 2H, JPt-H ) 77 Hz, Pt-olefin), 1.0 (m, 9H, CH3), 0.77 (m, 6H,
CH2).
[Cl2MeSi(µ-Cl)(η2-COD)Pt]2 (1c). The suspension of Pt(COD)Cl2
was heated to 60-65 °C, and the silane was added over 10 min
(alternatively, the silane can be added all at once at room temperature,
followed by heating). The bath temperature was raised to 70-75 °C,
9
9522 J. AM. CHEM. SOC. VOL. 124, NO. 32, 2002

Alkene−Platinum−Silyl Complexes
A R T I C L E S
back to the original flask. Yield of crude product: 94-95%. If desired,
the crude solid may be further purified via dissolution in CDCl3 (30
mL, heating to 30-35 °C) containing added COD at 1-1.5 mol of
COD/mol of Pt, a second filtration, and washing with 5-10 mL of
CDCl3, followed by recrystallization via storage in a freezer at -20
°C to yield colorless crystals of pure Pt(COD)(SiMeCl2)2. Several 1-2
mm size crystals, together with mother liquor containing excess COD,
were transferred to a clean, oven-dried 8 mm diameter glass ampule
inside a glovebag under nitrogen atmosphere. The ampule was attached
by using Tygon tubing to a small separatory funnel with a Teflon
stopcock and stoppered with a rubber septum, removed from the
glovebag, attached to a vacuum line, and flame-sealed after a single
freeze-thaw cycle using liquid nitrogen. It was then shipped for single-
crystal X-ray structure analysis of 2c.
tion in 12 months. Attempts were made to obtain solid-state NMR data
on a sample of 2b, but it failed to tune well, an explanation for which
is not apparent.
2c from 1d. 1d (0.005 mol Pt) was prepared in solution as described
above (a 4:1 ratio of COD:Pt was used, and 30-45 min heating at 60
°C was required for complete reaction of Pt(COD)Cl2). To the stirred
solution of 1d at 50 °C was added MeSiHCl2 (0.0102 mol, 1.1 mL)
from the dropping funnel over 10 min. The mixture was heated at 45-
48 °C for 2 h and at 40 °C for 1 h. The solution was light yellow.
NMR analysis showed the predominant formation of 2c, with the ratio
of Me2SiCl2:MeSiCl3 being >10:1. Even the 5-10% 2d originally
present with 1d had reacted with MeSiHCl2 to yield 2c.
Hydrosilation of Styrene with MeSiHCl2 Using 2b as Catalyst.
Using glassware as with hydrosilation of 1-hexene, platinum complex
2b (0.007 g, 60 µmol of white powder) was transferred to the flask
under nitrogen, followed by styrene (0.2 mol, 22.9 mL). MeSiHCl2
(0.204 mol, 21.9 mL) was transferred to the addition funnel via syringe.
The flask was warmed to 50 °C (the catalyst appeared to dissolve at
ca. 30 °C in styrene), and the silane was added over about 25 min.
Reaction was almost instantaneous, producing a mild exotherm to 54
°C, and was over by the completion of silane addition. A sample for
NMR analysis was withdrawn and showed (by 1H, 13C, and 29Si NMR)
clean hydrosilation of styrene to produce internal and terminal adduct
in the ratio ca. 58:42. No residual styrene was detected, but the excess
silane used was clearly identified in the product mixture. To the heel
from this first reaction was added a second charge of styrene (0.1 mol,
11 mL), and this was again hydrosilated with MeSiHCl2 (0.102 mol,
11 mL). This second hydrosilation was over in 10 min (addition time
for silane), again producing a mild exotherm. NMR analysis of this
second product mixture showed complete hydrosilation with ap-
proximately the same ratio (58:42) of internal to terminal adducts.
Note: Chloroform containing ethanol or even amylenes as stabilizer
will cause decomposition of 2c. However, amylene-stabilized CHCl3
may be used for recrystallization, provided at least one COD per Pt is
added to the chloroform. It may also be possible to use amylene-
stabilized chloroform for the synthesis, but excess COD imparts
significant stability and “shelf life” to solutions of 2c.
Hydrosilation of 1-Hexene with MeSiHCl2 in the Presence of 0.5
Equiv of 2c. A solution of 2c containing 0.005 mol of complex in 18
mL of CDCl3 was prepared, and an aliquot containing 0.0014 mol of
complex (and byproducts cyclooctene, MeSiCl3 and excess COD) was
cannulated to a Si-free Teflon NMR tube. 1-Hexene (350 µL, 0.0028
mol) was added, followed by MeSiHCl2 (302 µL, 0.0029 mol), and
the mixture was heated in the NMR probe at 50 °C. NMR analysis
after about 4 h indicated complete hydrosilylation of the 1-hexene to
n-hexylmethyldichlorosilane, with no isomerization of 1-hexene and,
more importantly, no measurable loss of 2c to any reaction with the
substrates or the product.
Hydrosilation of 1-Hexene with MeSiHCl2 in the Presence of a
Catalytic Level of 2c. A 100 mL, three-neck flask equipped with a
Teflon stirring bar, addition funnel, and water condenser was purged
with nitrogen and charged with 1-hexene (12.5 mL, 0.10 mol).
MeSiHCl2 (ca. 11.0 mL, 0.102 mol) was transferred to the additional
funnel. An aliquot of the above “stoichiometric” reaction mixture
containing 50 µmol of Pt with respect to 1-hexene was added to the
1-hexene in the flask using a microliter syringe. The flask was heated
to 60 °C and the silane added over 20 min. The hydrosilylation reaction
began within a few minutes and was essentially over at the end of the
silane addition, as evidenced by the lack of any refluxing of unreacted
silane (bp 41 °C) in the flask. The light yellow solution was heated for
another 30 min, following which NMR analysis showed complete
hydrosilylation, with no trace of residual 1-hexene or 2-hexene
(isomerization).
X-ray Crystal Structure Analysis for 2c. A colorless cut from a
block crystal of C10H18Cl4Si2Pt (see purification of 2c, above) having
approximate dimensions of 0.56 × 0.31 × 0.15 mm3 was mounted
using oil (Infineum V8512) on a glass fiber. All measurements were
made on a SMART-1000 CCD area detector with graphite-monochro-
mated Mo KR radiation. The data were collected at a temperature of
-120 ( 1 °C to a maximum 2θ value of 56.5°. Data were collected in
0.30° oscillations with 10.0 s exposures. The crystal-to-detector distance
was 50.00 mm. The detector swing angle was 28.00°.
The structure was solved by direct methods and expanded using
Fourier techniques. The disordered chlorine and carbon atoms were
refined isotropically, while the remaining non-hydrogen atoms were
refined anisotropically. The hydrogen atoms on the disordered methyl
group were included in fixed positions from the difference Fourier map,
while the remaining hydrogen atoms were refined isotropically. The
final cycle of full-matrix least-squares refinement on F2 was based on
1637 observed reflections and 95 variable parameters and converged
(largest parameter shift was 0.00 times its esd) with unweighted and
weighted agreement factors of
Silyl-Exchange Reactions To Prepare 2b and 2c from 1c and
1d, Respectively. 2b from 1c. 1c (0.005 mol Pt) was prepared in
solution as described above (except that the water condenser was topped
with a dry ice condenser for the setup, and a 4:1 molar ratio of COD:
Pt was used). To the stirred solution of 1c maintained at 45 °C was
added HSiCl3 (0.0102 mol, 1.0 mL) from the dropping funnel over 2
min. Within 2 min after addition of HSiCl3, a white solid fell out of
solution. The mixture was stirred overnight at 40 °C. NMR analysis of
the supernatant liquid showed very weak 13C and 29Si signals attributable
to 2b, but MeSiCl3 and SiCl4 were formed in the ratio of 10-12:1,
with no HSiCl3 remaining. The essentially complete consumption of
HSiCl3 and 1c and the highly selective formation of MeSiCl3 provided
strong evidence for the preferential formation of 2b. Thus, 2b, which
cannot be made via direct reaction of HSiCl3 with Pt(COD)Cl2, due to
the extremely slow first step to form 1b, can be made in very high
yield via the indirect silyl-exchange reaction. Most of the solvent and
byproducts were cannulated out, and the rest was removed under
vacuum at 45 °C. The crude off-white 2b was stored in a flask
(stoppered using Teflon sleeves and then wrapped with Parafilm at the
necks) at room temperature and showed no visible sign of decomposi-
R1 )
||Fo| - |Fc||/ |Fo| ) 0.024
∑
∑
wR2 ) [ (w(Fo2 - Fc ) )/ w(Fo2)2]1/2 ) 0.049
2 2
∑
∑
Acknowledgment. This work is dedicated to John L. Speier,
Alan J. Chalk, and John F. Harrod for their pioneering
contributions to the science and technology of Pt-catalyzed
hydrosilation. The authors acknowledge help from Ron Teck-
lenburg of Dow Corning for carrying out mass spectral analysis
on the platinum dimer 1c. Assistance in setting up several NMR
runs, “macro” creation, and low-temperature NMR work
preparations from Don Eldred of the Dow Corning NMR team
is gratefully acknowledged. We are also very grateful to
9
J. AM. CHEM. SOC. VOL. 124, NO. 32, 2002 9523

A R T I C L E S
Roy and Taylor
Professor Tobin Marks and Ms. Charlotte Stern of Northwestern
University, Evanston, IL, for carrying out the vital crystal
structure analysis and structure interpretation of 2c.
the preparation of 1c, containing 1c as the predominant platinum
complex; and all information for X-ray crystallographic analysis
of 2c, including tables of bond distances and bond angles (PDF
and CIF). This material is available free of charge via the
Internet at http://pubs.acs.org.
Supporting Information Available: 1H NMR spectra for
1a,c,f,g and 2c; stacked plots of 13C and 29Si NMR spectra for
all 1 identified in the study; APCI MS of reaction mixture from
JA0127335
9
9524 J. AM. CHEM. SOC. VOL. 124, NO. 32, 2002