Mechanistic insights into the hydrosilylation of allyl compounds - Evidence for different coexisting reaction pathways

Article abstract of DOI:10.1016/j.jcat.2012.06.006

The hydrosilylation of allyl compounds is often accompanied by the formation of high amounts of byproducts. The formation processes have not been fully understood so far. In this work, the allyl hydrosilylation mechanism is investigated in detail and experimental and theoretical evidence for multiple, coexisting reaction pathways is provided. Based on earlier reports and the observations during an extensive catalytic study, different pathways, leading to the observed byproducts, were identified and proven by labeling experiments and DFT calculations. Oxidative addition of the silane and the insertion of the allyl compound into the Pt-H bond turned out to be the crucial, selectivity-determining steps within the catalytic cycle. Based on these findings, it should be possible to systematically influence these steps and pave the way to a rational and straightforward design of more selective catalysts.

Full text of DOI:10.1016/j.jcat.2012.06.006

Journal of Catalysis 295 (2012) 1–14
Contents lists available at SciVerse ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Mechanistic insights into the hydrosilylation of allyl compounds – Evidence
for different coexisting reaction pathways
Peter Gigler a,b, Markus Drees b, Korbinian Riener b, Bettina Bechlars b, Wolfgang A. Herrmann b,
,
⇑
Fritz E. Kühn b,c,
⇑
a WACKER-Institut für Siliciumchemie, Department of Chemistry, Technische Universität München, Ernst-Otto-Fischer-Straße 1, 85747 Garching b. München, Germany
b Inorganic Chemistry, Catalysis Research Center, Department of Chemistry, Technische Universität München, Ernst-Otto-Fischer-Straße 1, 85747 Garching b. München, Germany
c Molecular Catalysis, Catalysis Research Center, Department of Chemistry, Technische Universität München, Ernst-Otto-Fischer-Straße 1, 85747 Garching b. München, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 16 March 2012
Revised 1 June 2012
Accepted 7 June 2012
Available online 25 September 2012
The hydrosilylation of allyl compounds is often accompanied by the formation of high amounts of
byproducts. The formation processes have not been fully understood so far. In this work, the allyl hydro-
silylation mechanism is investigated in detail and experimental and theoretical evidence for multiple,
coexisting reaction pathways is provided. Based on earlier reports and the observations during an exten-
sive catalytic study, different pathways, leading to the observed byproducts, were identified and proven
by labeling experiments and DFT calculations. Oxidative addition of the silane and the insertion of the
allyl compound into the Pt–H bond turned out to be the crucial, selectivity-determining steps within
the catalytic cycle. Based on these findings, it should be possible to systematically influence these steps
and pave the way to a rational and straightforward design of more selective catalysts.
Keywords:
Hydrosilylation
Platinum catalyst
Allyl compounds
Chlorosilanes
Ó 2012 Published by Elsevier Inc.
Mechanism
Density functional theory
1. Introduction
to d and e exists so far. The Chalk–Harrod mechanism (CH), which
is widely accepted for the Pt-catalyzed hydrosilylation of olefins
The hydrosilylation of olefins is among the most widespread
reactions in silicon chemistry [1–10]. In contrast to simple hydro-
genations, the addition of a Si–H bond to an olefin leads to func-
tionalized compounds, prone to further conversions. For example,
hydrosilylations are used for the synthesis, modification and
cross-linking of polysiloxanes [1]. The catalytic hydrosilylation of
[17] and the modified variant thereof (mCH) [18], respectively,
can solely explain the formation of the desired product
(Scheme 2).
c
The byproduct formation, however, has only been investigated
in a handful of scientific publications [8,13–15,19]. It was first de-
scribed by Wagner in 1953 [19], who proposed two possible reac-
tion sequences, depicted in Scheme 3, to explain the observed
product pattern.
The consecutive sequence, shown in Scheme 3a, is starting with
the formation of the desired product c, followed by a metathesis
with b yielding the byproducts d and f. Although the formation
of propene has not been discussed by Wagner, it is imaginable that
e is formed in a similar way by the reaction of a and b.
A few years later, Speier disproved this reaction sequence,
showing that the metathesis step does not occur under the applied
reaction conditions. No reaction was observed upon treatment of
benzyl chloride with dichloromethylsilane in the presence of chlo-
roplatinic acid (Scheme 4a) [14]. Accordingly, the simultaneous
formation of the desired product and the byproducts d and e, fol-
lowed by a subsequent hydrosilylation of e to f (Scheme 3b), was
the only plausible reaction sequence, later confirmed by Marciniec
et al. while investigating the kinetics of the hydrosilylation of allyl
chloride with trichlorosilane on Pt/C particles [15].
allyl compounds is of particular interest [11]. The resulting c-func-
tionalized organosilanes can be used as binders or as reactants for
further conversions, for example, hydrolyses and nucleophilic sub-
stitution reactions [2,12]. Unfortunately, the Pt-catalyzed reaction
of an allyl compound (a) with tertiary silanes (b), leading to the de-
sired hydrosilylation product c, is accompanied by varying
amounts of unwanted byproducts, namely R3SiX (d), propene (e),
and propylsilane (f) (Scheme 1) [13–16].
Although the reaction is frequently used on large industrial
scale, no deep mechanistic understanding of the processes leading
⇑
Corresponding authors. Address: Inorganic Chemistry, Catalysis Research
Center, Department of Chemistry, Technische Universität München, Ernst-Otto-
Fischer-Straße 1, 85747 Garching b. München, Germany (F.E. Kühn). Fax: +49 89
289 13473.
E-mail addresses: wolfgangherrmann@ch.tum.de (W.A. Herrmann), fritz.-
kuehn@ch.tum.de (F.E. Kühn).
0021-9517/$ - see front matter Ó 2012 Published by Elsevier Inc.
http://dx.doi.org/10.1016/j.jcat.2012.06.006

2
P. Gigler et al. / Journal of Catalysis 295 (2012) 1–14
(a)
(b)
(c)
(d)
(e)
(f)
Scheme 1. Pt-catalyzed hydrosilylation of allyl compounds.
(A)
(B)
Fig. 1. Different transition states proposed by Speier [14].
(a)
(c)
Depending on the polarization of the double bond, the reaction
proceeds via transition state A to the desired hydrosilylation prod-
uct – as it exclusively occurs in case of 2-methyl allyl chloride – or,
according to an allylic rearrangement reaction via transition state
B, to the observed byproducts.
A different explanation was given by Marciniec et al. [15]. The
byproduct formation was attributed to a 2,1-insertion step of allyl
chloride into the Pt–H bond in contrast to the 1,2-insertion occur-
ring during the classical Chalk–Harrod mechanism. The resulting
Markovnikov product C was assumed to undergo an elimination
reaction to form the observed byproducts d and e (Scheme 5).
Most recently, the original mechanistic proposal of Speier was
taken up by Roy [8]. He could show that during the hydrosilylation
of crotyl chloride (g) with deuterodichloromethylsilane, the bideu-
terated butylsilane h was formed exclusively (Scheme 6).
Most surprisingly, the reaction was observed to be as fast as
hydrosilylation reactions of terminal olefins, which is quite uncom-
mon for the hydrosilylation of internal olefins [2]. Hence, a plati-
num-assisted allylic rearrangement, as depicted in Scheme 7,
seemed to be the appropriate explanation for this reaction
behavior.
Nevertheless, a deep and comprehensive understanding of the
mechanistic processes leading to the byproduct formation during
the hydrosilylation of allyl compounds is still missing. Based on
the aforementioned studies and various catalytic and mechanistic
experiments, supported by DFT calculations, a comprehensive
mechanistic picture on the underlying processes is presented in
this work. It enables the identification of the crucial elementary
(b)
Scheme 2. (Modified) Chalk–Harrod mechanism [17,18] applied to the hydrosily-
lation of allyl compounds.
(a)
(a)
(b)
(b)
(c)
(c)
(d)
(f)
(b)
(c)
(a)
(b)
(f)
(e)
(d)
Scheme 3. Consecutive (a) and parallel (b) reaction sequences to explain the
observed byproducts.
(d)
(e)
(c)
(a)
(b)
(c)
(d)
(e)
(f)
(c)
Scheme 5. Proposed reaction mechanism for the hydrosilylation of allyl chloride
with trichlorosilane on Pt/C particles [15].
Scheme 4. Key observations of Speier investigating the hydrosilylation of allyl
compounds [14].
Another contradictory observation was made by Speier et al.:
while the hydrosilylation of allyl chloride led to the aforemen-
tioned byproducts d–f, the reaction of 2-methyl allyl chloride
exclusively gave the desired product (Scheme 4b and 4c). The
formation of isobutene and isobutylsilane, respectively, was not
observed. This behavior was ascribed to the two different transi-
tion states A and B shown in Fig. 1.
(g)
(h)
Scheme 6. The hydrosilylation of crotyl chloride leading to the exclusive formation
of butylsilane.

P. Gigler et al. / Journal of Catalysis 295 (2012) 1–14
3
2.4. NMR-scale reactions
(d)
(e)
(a)
2.4.1. Scrambling reaction
A NMR tube was equipped with one drop (ꢀ1–3
l
mol) of Kar-
stedt catalyst or 1 mg (1.45 lmol) of Markó catalyst, 20 mg of i
(0.25 mmol) and 0.5 mL of C6D6. The mixture was heated to 60 °C
in an oil bath, and the reaction progress was followed by 1H
NMR and 2H NMR spectroscopy.
2.4.2. Catalytic hydrosilylation reactions
(b)
In a typical experiment, a NMR tube was equipped with one
drop (ꢀ1–3
lmol) of Karstedt catalyst or 1 mg (1.45 lmol) of
Markó catalyst, the allyl compound (0.25 mmol), silane
(0.25 mmol), and 0.5 mL of C6D6. The sample was heated to 60 °C
for 16 h, and the products were identified by 1H, 2H, 13C and 29Si
NMR spectroscopy.
Scheme 7. Proposed allylic rearrangement causing the byproducts [8].
steps within the catalytic cycle, being responsible for the formation
of the byproducts and thus should open the way to a straightfor-
ward design of more selective catalysts for the hydrosilylation of
allyl compounds.
2.5. Computational methods
All calculations were performed with GAUSSIAN-03 [25] using
the density functional/Hartree–Fock hybrid model Becke3LYP
[26–29] and the split valence double-f (DZ) basis set 6-31G⁄⁄
[30] except for Pt where the Stuttgart-Dresden-ECP [31,32] has
been used. No symmetry or internal coordinate constraints were
applied during optimizations. All reported intermediates were ver-
ified as being true minima by the absence of negative eigenvalues
in the vibrational frequency analysis. Transition-state structures
(indicated by TS) were located using the Berny algorithm [33] until
the Hessian matrix had only one imaginary eigenvalue. The identi-
ties of all transition states were confirmed by IRC calculations and
by animating the negative eigenvector coordinate with MOLDEN
[34] and GaussView [35].
2. Experimental methods
2.1. General remarks
All manipulations were carried out under an atmosphere of dry
argon using standard Schlenk techniques. Solvents were dried by
standard methods and distilled under argon [20]. 1H, 2H, 13C, 29Si
NMR spectra were recorded on a Bruker AMX 400 MHz and a Bru-
ker Avance III 400 MHz spectrometer at room temperature, and
referenced to the residual solvent signal [21] or to TMS (29Si). Spec-
tral assignments were achieved by combination of 13C DEPT,
1H–13C HMQC and 1H–13C HMBC experiments. The Karstedt cata-
lyst (20 wt.% Pt) was provided by the Wacker Chemie AG and both
the Markó catalyst and allyl alcohol were synthesized according to
literature procedures [22–24].
Approximate free energies (DG) and enthalpies (DH) were ob-
tained through thermochemical analysis of frequency calculations,
using the thermal correction to Gibbs free energy as reported by
GAUSSIAN-03. This takes into account zero-point effects, thermal
enthalpy corrections, and entropy. All energies reported in this pa-
per, unless noted otherwise, are free energies or enthalpies at
298 K in gas-phase, using unscaled frequencies. All transition
states are maxima on the electronic potential energy surface
(PES), which may not correspond to maxima on the free energy
surface. The coordinates and energies calculated for each interme-
diate or transition state can be requested by the authors.
2.2. Catalytic hydrosilylation
The catalyst (10 lmol) and the allyl compound (31 mmol) were
dissolved in 10 mL of dry toluene, and the reaction mixture was
warmed to 40 °C. At this temperature, the silane (62 mmol) was
added and the mixture was stirred for 16 h at 60 °C, before it
was allowed to cool to room temperature. The product selectivity
was determined by 29Si NMR spectroscopy, integrating the signal
of the desired product c against the byproduct d formed in parallel.
The assignment of the signals in the 29Si NMR spectra, summarized
in Table S1 (Supplementary material), was achieved by separate
measurements of the single compounds.
3. Results and discussion
The present study is divided into three main parts. First, the cat-
alytic performance of different allyl compounds and silanes was
examined. Second, based on these results, specific experiments
and calculations were performed to elucidate and verify possible
reaction pathways. In a third step, the overall mechanism is dis-
cussed and crucial elementary steps are pointed out.
2.3. 3,3-dideutero allyl chloride (i)
1.20 mL (1.88 g, 13.7 mmol) of PCl3 was added dropwise to a
precooled (0 °C) mixture of 1.99 g (34.3 mmol) 3,3-dideutero allyl
alcohol and 0.67 ml (0.66 g, 8.35 mmol) of pyridine. The product
3.1. Catalytic results
was subsequently distilled (b.p. 34 °C) to afford 1.68 g (64%) of i
3.1.1. Hydrosilylation of various allyl compounds
3
as a colorless liquid. 1H NMR (400 MHz, CDCl3): d = 5.22 (d, JHH
=
In order to produce reliable catalytic results that elucidate the
influence of each reaction component, different silanes (trichloros-
ilane, dichloromethylsilane, chlorodimethylsilane, and triethoxysi-
lane) were employed in the hydrosilylation of various allyl
compounds (allyl choride, allyl ethyl ether, allyl acetate, and N,N-
dimethyl allyl amine) under identical reaction conditions. The
reaction was catalyzed by the electron-deficient Karstedt catalyst
[36] or the electron-rich Markó catalyst [22,37] (Fig. 2).
3
10.1 Hz, 1H, CHCH2/E), 5.35 (d, JHH = 16.9 Hz, 1H, CHCH2/E), 5.97
(dd, JHH = 10.1 Hz, JHH = 16.9 Hz, 1H, CHCH2/F). 1H NMR
3
3
3
(400 MHz, C6D6): d = 4.80 (d, JHH = 10.1 Hz, 1H, CHCH2/E), 4.93
3
3
3
(d, JHH = 16.9 Hz, 1H, CHCH2/E), 5.57 (dd, JHH = 16.9 Hz, JHH
=
10.1 Hz, 1H, CHCH2/F). 2H NMR (CDCl3, 61 MHz): d 4.04 (s, CD2).
13C {1H} NMR (100 MHz, CDCl3): d = 44.9 (qu, 1JCD = 23.0 Hz, CD2Cl),
118.8 (s, CHCH2), 133.9 (s, CHCH2).

4
P. Gigler et al. / Journal of Catalysis 295 (2012) 1–14
catalyzed hydrosilylation with HSiCl3 depends on the silane-to-
substrate ratio, whereas none of the other silanes and none of
the Markó catalyzed hydrosilylations show this trend.
Although the same reaction products are observed for all inves-
tigated systems, a different dependence on the silane concentra-
tion can be seen for the Karstedt catalyzed hydrosilylation of
allyl chloride with trichlorosilane. Therefore, at least two different
reaction pathways have to be operative, leading to the observed
byproducts. Otherwise, all systems would be expected to show
the same influence of the silane concentration.
Fig. 2. Platinum catalysts applied in the hydrosilylation of allyl compounds.
The product selectivity (with respect to the employed olefin)
was determined by integrating the 29Si NMR signal of the desired
product c in comparison with the signal of simultaneously formed
quaternary silane d. An overview of the selectivities achieved is
given in Table 1.
3.2. Mechanistic aspects
Based on the aforementioned catalytic results, specific experi-
ments and DFT calculations were performed for allyl chloride
(X = Cl) to obtain a profound mechanistic understanding and to
identify crucial elementary steps during the catalytic cycle. For
the DFT calculations, the active species of the Karstedt catalyst
was modeled as a Pt(0) complex with the corresponding olefin
(e.g. allyl chloride) as the sole ligand, and the Markó catalyst bear-
ing one methyl-substituted N-heterocyclic carbene (NHC) ligand,
respectively. Besides the catalyst, the silane species was varied
from HSiCl3 via MeSiHCl2 to Me2SiHCl.
Based on these results, the influence of each reaction compo-
nent (allyl compound, silane, and catalyst) on the product selectiv-
ity was evaluated and further studied. As can be seen in Table 1, a
significant influence of all three components is evident. Regarding
different allyl compounds, allyl chloride is by far the least selective
substrate, indicating a crucial role of the allyl substituent within
the catalytic cycle. In contrast, the product selectivities using
different silanes seem to depend on the nature of the employed
catalyst and vice versa (Table 1, entries 1 and 2). For the hydrosi-
lylation of allyl chloride with different chlorosilanes, an opposite
trend is observed for the two catalysts, suggesting an electronic
interplay of catalyst and silane.
3.2.1. Product formation
As mentioned in Section 1, the product formation can be ex-
plained with the classical Chalk–Harrod mechanism for olefins
[17] and its modified variant [18], respectively (Scheme 2). In order
to get a deeper insight into the product formation and to distin-
guish both variants, DFT calculations considering the underlying
catalytic cycles were performed. The Chalk–Harrod mechanism,
applied to the hydrosilylation of allyl chloride, is shown in
Scheme 8.
3.1.2. Influence of the silane concentration
The hydrosilylation of allyl chloride represents the most unse-
lective reaction. Therefore, this particular substrate was investi-
gated in more detail by varying the stoichiometry of the reaction
partners. As shown in Table 2, only the selectivity of the Karstedt
As the mono-coordinated Pt(0) species was claimed to be the
starting point of the catalytic cycle, only a small or no energy bar-
rier is expected for the coordination of the allyl chloride to the plat-
Table 1
Product selectivities (% c) in the hydrosilylation of different allyl compounds with
various silanes using the Karstedt and Markó catalyst.
inum center (ICH G = 0.9 kcal molꢁ1 for Karstedt, no barrier for
, D
Markó). The subsequent oxidative addition (IICH) of the silane to
form a four-coordinated Pt intermediate shows barriers of 9–
12 kcal molꢁ1 using the Markó catalyst. In contrast to this, the
insertion into the Pt–H bond (IIICH) only shows significant barriers
of 4–6 kcal molꢁ1. For the Karstedt catalyst, the transition states
IICH and IIICH have energies that are lower than the preceding
3
3
3
3
Entry
X
Catalyst
HSiCl3
MeSiHCl2
Me2SiHCl
HSi(OEt)3
1
2
3
4
5
6
7
8
Cl
Cl
Karstedt
Markó
Karstedt
Markó
Karstedt
Markó
Karstedt
Markó
20
70
55
66
75
97
92
100
66
46
92
100
97
27a
26b
90
(b)
OEt
OEt
OAc
OAc
NMe2
NMe2
100
100
100
100
88
100
100
90d
100e
100
c
c
c
–
–
–
c
c
c
–
–
–
a
b
c
42%, considering a subsequent Cl/OEt exchange of the product.
41%, considering a subsequent Cl/OEt exchange of the product.
Formation of silyl ammonium chloride.
d
e
48%, considering the observed
58%, considering the observed
a
a
-product as a byproduct (see Section 3.2.2.2).
-product as a byproduct (see Section 3.2.2.2).
(a)
Table 2
(c)
Karstedt catalyzed hydrosilylation of allyl chloride. Product selectivities (% c) using
varying amounts of silane.
Equiv. silane
HSiCl3
MeSiHCl2
Me2SiHCl
HSi(OEt)3
0.5
1.0
1.5
2.0
67
63
34
20
51
55
66
66
32
27
Scheme 8. Computational investigation of the Chalk–Harrod mechanism.

P. Gigler et al. / Journal of Catalysis 295 (2012) 1–14
5
Fig. 3. Free energy profiles along the Chalk–Harrod mechanism for (a) the Karstedt catalyst and (b) the Markó catalyst. Legend: black line for HSiCl3, dashed line for HSiMeCl2,
and gray line for HSiMe2Cl as substrate.
ground state and hence are regarded as barrierless.1 Finally, the
(b)
reductive elimination (IVCH
) shows barriers of around 16–
23 kcal molꢁ1 for the Karstedt catalyst and therefore represents the
rate-determining step for this system. If the Markó catalyst is used,
the barriers for the product release are lower than for the oxidative
addition, which is why the latter is the rate-determining step accord-
ing to calculations. The free energy profiles along the Chalk–Harrod
mechanism are shown in Fig. 3.
In contrast to the classical Chalk–Harrod mechanism, the mod-
ified variant comprises an insertion into the Pt–Si bond (IIImCH) in-
stead of the Pt–H bond (Scheme 9).
(a)
In order to prove the relevance of the alternative insertion reac-
tion, the hydrosilylation of allyl chloride with trichlorosilane was
calculated for the modified Chalk–Harrod mechanism. The com-
parison between both mechanistic pathways is given in Fig. 4.
Regarding the overall catalytic cycles, the computational results
show that both mechanisms are feasible, but displaying individual
rate-determining steps. In case of the classic Chalk–Harrod mech-
anism, this r.d.s. is depending on the catalyst used (Karstedt cata-
lyst ) reductive elimination IVCH / Markó catalyst ) oxidative
addition IICH. The insertion into the Pt–Si bond (IIImCH)) has the
highest barrier for all steps of the modified Chalk–Harrod mecha-
nism for both catalyst systems. However, in contrast to the modi-
fied variant, the barriers of the insertion into the Pt–H bond are
significantly lower for the Chalk–Harrod mechanism, predicting
an exclusive insertion into the Pt–H bond. Hence, based on the cal-
culations and in agreement with previous theoretical results [38–
(c)
Scheme 9. Computational investigation of the modified Chalk–Harrod mechanism.
40] and the following experimental findings, the modified Chalk–
Harrod mechanism does not play a significant role in the hydrosi-
lylation of allyl compounds.
3.2.2. Byproduct formation
3.2.2.1. Allyl mechanism. The hydrosilylation of allyl chloride is the
least selective reaction compared to the other investigated allyl
compounds (Table 1). Therefore, the nature of the allyl substituent
obviously has a significant influence on the byproduct formation. A
reasonable explanation for this influence is the oxidative addition
of the allyl compound to the platinum center instead of the silane
(Scheme 10).
1
Alternatively, a
r-bond metathesis instead of the oxidative addition of the silane
is possible. However, the unsaturated active species forces the oxidative addition,
which is confirmed by the low or absent barriers calculated for this reaction step. A
r
-
bond metathesis has a significantly higher barrier according to our calculations,
which is why this mechanistic alternative has not been further considered.

6
P. Gigler et al. / Journal of Catalysis 295 (2012) 1–14
Fig. 4. Hydrosilylation of allyl chloride with trichlorosilane. Free energy profiles along the Chalk–Harrod mechanism (solid line) and the modified Chalk–Harrod mechanism
(dashed line) for (a) the Karstedt catalyst and (b) the Markó catalyst.
bling. Instead, the allyl chloride is shown to coordinate to the plat-
(b)
inum center by substituting divinyltetramethyldisiloxane, as
observed in the 1H NMR spectrum.
Hence, from an experimental point of view, an activation of the
C–Cl bond as the initial elementary step of a possible side-reaction
pathway has to be considered for the Karstedt catalyst. For the
Markó catalyst, this reaction pathway seems to be not as likely,
although it cannot be excluded in the presence of substrate under
catalytic conditions.
(a)
An appropriate mechanistic pathway (allyl mechanism, AM)
based on the oxidative addition of the allyl chloride is proposed
in Scheme 12.
After the coordination of the olefin (IAM), the allyl compound is
Scheme 10. Competing Si–H vs. C–Cl activation as
byproduct formation.
a
potential source of the
oxidatively added to the metal center (IIAM). In a third step, d can
be formed via
r
-bond metathesis (IIIAM),2 followed by reductive
elimination of propene and regeneration of the catalyst (IVAM). Pro-
pene can be subsequently hydrosilylated according to the classical
Chalk–Harrod mechanism.
The DFT calculations support the experimental results dis-
cussed before. The observed scrambling in case of the Karstedt cat-
alyst (Fig. 5) can be explained by comparing the calculated barriers
of the oxidative addition of the allyl chloride to the platinum cen-
ter and its reverse reaction. For both catalysts, the barrier for the
C–Cl activation is similar (14.5 kcal molꢁ1 and 16.2 kcal molꢁ1),
whereas the barrier for the reverse reaction (Cl transfers back to
the olefin) is considerably higher for the Markó catalyst
(38.8 kcal molꢁ1 vs. 27.7 kcal molꢁ1), thus supporting the experi-
mental findings. Considering the energy barriers calculated for
(i)
(j)
Scheme 11. Expected H/D-scrambling as a result of a reversible C–Cl activation.
Comparing the investigated allyl compounds, the chloro substi-
tuent represents the best leaving group and therefore an activation
of the C–Cl bond is very likely. Transition metal complexes in gen-
eral and in particular those of palladium and platinum are well
known for this reactivity, which is used as a standard synthetic
method to synthesize transition metal allyl complexes [41]. The
most famous application of this reaction is the Pd-catalyzed Tsu-
ji–Trost reaction [42].
each step of the discussed allyl mechanism, the
r-bond metathesis
turns out to be the rate-determining step of the catalytic cycle for
both catalysts (Fig. 6).
In order to verify the proposed C–Cl activation, catalytic
amounts of the Karstedt and the Markó catalyst, respectively, were
reacted with 3,3-dideutero allyl chloride, (i) and the reaction pro-
gress was monitored by 1H and 2H NMR spectroscopy. As shown in
Scheme 11, the reversible activation of the C–Cl bond is expected
to end up in a 1:1 distribution of 3,3-dideutero allyl chloride (i)
and 1,1-dideutero allyl chloride (j).
After 24 h, almost complete deuterium scrambling has occurred
in the presence of the Karstedt catalyst (i:j = 57/43). Accordingly,
the intensity of the signal for the methylene protons of j at
3.53 ppm in the 1H NMR spectrum increases while the intensity
of the signals at 4.79 and 4.92 ppm for the vinyl protons of i de-
creases (Fig. 5).
3.2.2.2. a-Mechanism. As denoted in the footnotes of Table 1, a pri-
mary product other than d or e was observed in the hydrosilylation
of N,N-dimethyl allyl amine with triethoxysilane (Table 1, entries 7
and 8). This product was unambiguously identified as the so-called
a-product (or Markovnikov product) k (Scheme 13a). The forma-
tion of the -product has already been reported for the PtO2-cata-
a
lyzed hydrosilylation of methyl allyl amine and allyl n-butyl amine
with methyl diethoxysilane [43].
In contrast to the Chalk–Harrod mechanism, with the b-addi-
tion product c being formed via a 1,2-insertion step, the a-product
2
Alternatively, the reaction can proceed via a Pt(IV) species. However, according to
our calculations such a species is thermodynamically unfavored. From the kinetic
point of view the oxidative addition to Pt(IV) turned out to be possible, but the barrier
of the subsequent silane elimination is significantly higher than for the back reaction
and the metathesis step, respectively. Therefore, an equilibrium between the Pt(II)
and the Pt(IV) species can exist, however, the latter does not have an impact on the
catalytic performance due to its high barriers.
The broadening of the signal for the CH-group at 5.56 ppm can
be attributed to overlapping signals for the methine protons of the
two coexisting deuterated species i and j. The reverse change in
intensities was observed in the 2H NMR spectra. In contrast to
the Karstedt catalyst, the Markó catalyst causes no H/D-scram-

P. Gigler et al. / Journal of Catalysis 295 (2012) 1–14
7
Fig. 5. 1H NMR spectrum of 3,3-dideutero allyl chloride in C6D6 (a) and time-resolved reaction progress using catalytic amounts of Karstedt catalyst (b and c).
(b)
(d)
(a)
(e)
Scheme 12. Proposed mechanistic pathway leading to the observed byproducts via oxidative addition of the allyl compound (allyl mechanism).
is generated via a 2,1-insertion reaction of the olefin into the Pt–H
bond (Scheme 13b).
confirming the instability of any intermediate species along the
2,1-insertion pathway for X = Cl.
Although the double bond polarities of the employed allyl com-
pounds are supposed to be quite similar and hence the a-product
should be formed in all cases, it is exclusively observed when using
the amine. Accordingly, a decomposition reaction along the 2,1-
insertion pathway for the allyl compounds with X – NMe2 has to
be taken into account. Its existence was proven by the reaction
of 2-chloro propene (l) and trichlorosilane (Scheme 14).
Mechanistically, 5 is generated via a 2,1-insertion of the olefin
into the Pt–H bond (Scheme 13b). Thereafter, two different elimi-
nation pathways are feasible, which could explain the observed
instability. Simple reductive elimination would yield the a-product
k, which then could undergo an outer-sphere b-elimination to d
and e (Scheme 15a). Alternatively, a b-Cl-elimination could pro-
ceed at the metal center followed by an elimination of d and e
(Scheme 15b).
The constitution of this particular substrate and the resulting
stronger polarization of its double bond inhibit both the formation
of a platinum allyl complex (Scheme 14a) and the 2,1-insertion
product m (Scheme 14b). Consequently, 2-chloro-1-silylpropane
The outer-sphere b-elimination, as shown in Scheme 15a, is not
very likely due to the fact that k, even for X = Cl, exhibit sufficient
stability. Accordingly, very high barriers of about 50 kcal molꢁ1 for
this outer-sphere elimination step were found by DFT calculations.
Interestingly, none of the published syntheses of k [44] is per-
formed via the common platinum-catalyzed hydrosilylation reac-
tion, indicating the instability of the intermediate platinum silyl
complex 5K/5M. Considering further that vinyl chloride, exhibiting
exactly the same b-Cl silyl structure as the 2,1-insertion product 5,
undergoes a platinum-mediated b-elimination delivering ethylene,
(k0), an isomer of the
a-product, should be formed exclusively. If
an intermediate, for example, 5K0, or the product itself is unstable
under the applied reaction conditions, R3SiX (d) and propene (e)
are expected to be formed. The latter should be further converted
to the corresponding propylsilane f. In fact, the hydrosilylation of
2-chloro propene with trichlorosilane, proceeding already at room
temperature, solely leads to the formation of propylsilane, thus

8
P. Gigler et al. / Journal of Catalysis 295 (2012) 1–14
Fig. 6. Free energy profiles along the allyl mechanism for (a) the Karstedt catalyst and (b) the Markó catalyst. Legend: black line for HSiCl3, dashed line for HSiMeCl2, and gray
line for HSiMe2Cl as substrate.
(c)
(k)
(a)
(b)
(a)
(k)
(d)
(e)
(d)
(e)
(b)
Scheme 15. Possible mechanistic elimination pathways of the 2,1-insertion
product.
Scheme 13. (a) Formation of
a-product k during the hydrosilylation of N,N-
dimethylallylamine with triethoxysilane and (b) mechanistic pathways leading
either to the a- or b-addition product.
the reaction pathway depicted in Scheme 15b represents a valid
alternative to the allyl mechanism for the formation of the ob-
served byproducts.
The experimental findings are also confirmed by DFT calcula-
tions according to Fig. 7, with the transfer of the allyl substituents
X = NMe2 and X = Cl to the metal center (Scheme 15b) being com-
(a)
(b)
pared to the competing reductive elimination of the
(Scheme 15a).
a-product
(k’)
For X = Cl, low barriers of 1–7 kcal molꢁ1 and an exergonic reac-
tion with
D
G of around ꢁ13 to ꢁ27 kcal molꢁ1 are calculated for
a
IV M. The competing reductive elimination of the
a-product
a
(VII M), however, shows high barriers of 26–29 kcal molꢁ1 for the
Karstedt and 20–36 kcal molꢁ1 for the Markó catalyst. In contrast,
(e)
(f)
(d)
the transfer of the amine to the metal center (IV M) leads to con-
a
siderably higher barriers of 34–42 kcal molꢁ1 and to an endergonic
reaction, whereas the reductive elimination of the
a
-product
(VII M) shows lower barriers of 25–33 kcal molꢁ1. Accordingly,
the 2,1-insertion product 5 for X = Cl is unstable and the -product
a
a
cannot be formed due to the high barriers for the reductive elimi-
Scheme 14. Hydrosilylation of 2-chloro propene in order to verify the stability of
the intermediate species along the 2,1-pathway for allyl compounds with X = Cl.
nation. The b-elimination of the amine, however, is prohibited by

P. Gigler et al. / Journal of Catalysis 295 (2012) 1–14
9
Fig. 7. Comparison of the two elimination pathways of 5 for X = Cl (solid line) and X = NMe2 (dashed line) for (a) the Karstedt catalyst and (b) the Markó catalyst. The free
energy profiles are exemplarily shown for HSiCl3.
Fig. 8. Free energy profiles along the a-mechanism for (a) the Karstedt catalyst and (b) the Markó catalyst. Legend: black line for HSiCl3, dashed line for HSiMeCl2, and gray
line for HSiMe2Cl as substrate.
high barriers and unfavorable reaction thermodynamics, leading to
the reductive elimination of the observed -product.
quence: the first one via a stepwise and the second via a concerted
pathway.
a
Hence, a 2,1-insertion of the allyl compound into the Pt–H bond
has to be considered for all investigated systems. Depending on the
As mentioned in the introduction, the main reason for proposing
a concerted reaction pathway was the reaction rate Roy found for
the hydrosilylation of crotyl chloride [8]. In contrast to the signifi-
cantly lower reaction rates, which are usually observed for internal
olefins, the reaction of crotyl chloride and dichloromethylsilane was
as fast as in case of terminal olefins, exclusively forming the byprod-
ucts. In order to find out whether this observation is in agreement
nature of the allyl substituent, either the
a-product is obtained
(X = NMe2) or the byproducts are formed (e.g. X = Cl). Scheme 16
summarizes the mechanistic pathways discussed above.
The computational results for the
a-mechanism (aM) are
shown in Fig. 8. Independently of the employed catalyst, the
rate-determining step is represented by the reductive elimination
with the a-mechanism as well, DFT calculations for this particular
a
of R3SiX (VI M) in the last step of the catalytic cycle.
substrate employing the Karstedt catalyst were performed. In addi-
tion, the barriers and reaction energies for propene and 2-butene
were calculated to evaluate the role of the chloro substituent.
3.2.2.3.
Scheme 17, the
proposed by Speier and Roy describe both a similar reaction se-
a-Mechanism vs. allylic rearrangement. As pointed out in
With regard to the product formation, either b- (c) or
a-prod-
a-mechanism and the allylic rearrangement (AR)
uct (k), the reductive elimination of the hydrosilylation product

10
P. Gigler et al. / Journal of Catalysis 295 (2012) 1–14
(b)
(a)
(k)
(d)
(e)
Scheme 16. Mechanistic pathways of the
Scheme 18. Different reaction pathways based on 5K. The b-elimination is only
a-mechanism, proceeding via 2,1-
possible for X = Cl.
insertion of the allyl compound into the Pt–H bond.
elimination of the hydrosilylation products (c or k), a b-chloro elim-
a
inationofthe2,1-insertionproduct5(IV )andsubsequentreductive
a
a
elimination of the byproducts (V and VI ) can occur (Scheme 18).
(a)
This pathway is favored over reductive elimination
(30.1 kcal molꢁ1) of k (VII M) by its significantly lower barriers of
a
9.5 kcal molꢁ1 (IV M), 12.9 kcal molꢁ1 (V M), and 20.1 kcal molꢁ1
a
a
(VI M), respectively. The rate-determining step is given by the re-
a
lease of d and is therefore independent of the nature of the em-
ployed olefin. This is clearly reflected when comparing the
barriers of crotyl chloride as an internal olefin and allyl chloride
as a terminal olefin (Fig. 9).
(b)
Consequently, no significant difference with respect to the reac-
tion rate is expected for crotyl chloride forming the observed
byproducts. A preceding isomerization process of crotyl chloride
to the corresponding terminal olefin can be excluded, due to the
high barriers of this process calculated for 2-butene.
Scheme 17. (a) Stepwise and (b) concerted reaction pathway to 6.
represents the rate-determining step for all substrates (IVCH and
VII M, Table S7). The preceding elementary steps, however, have
a
Hence, besides the allylic rearrangement, the observations of
only small barriers or are even barrierless. Therefore, the different
reaction rates, observed for unsubstituted, internal and terminal
olefins, for example, propene and 2-butene, are determined by
the differing barriers of the corresponding rate-determining steps.
Comparing crotyl chloride with its unsubstituted analog 2-bu-
tene, both substrates can react via an almost barrierless pathway
to the 1,2- (4K) and 2,1-insertion product (5K), respectively. How-
ever, incontrastto2-butene, theallylicsubstituentopensanalterna-
tive reaction pathway for crotyl chloride. Besides the reductive
Roy can also be explained by the stepwise
a-mechanism. Never-
theless, for the comprehensive discussion intended, both pathways
have to be considered.
3.2.2.4.
a-Mechanism vs. allyl mechanism. In order to differentiate
between
a-mechanism and allyl mechanism, further labeling
experiments were performed. The product distribution resulting
from the hydrosilylation of 3,3-dideutero allyl chloride (i) with
Fig. 9. Comparison of the free energy profiles of crotyl choride (solid line) and allyl chloride (dashed line) along the a-mechanism for the Karstedt catalyzed hydrosilylation
with chlorodimethylsilane.

P. Gigler et al. / Journal of Catalysis 295 (2012) 1–14
11
(n)
(p)
(A)
(B)
(o)
(q)
(q)
(o)
Scheme 19. Expected deuterated byproducts using i as substrate.
Table 3
various chlorosilanes was determined. As shown in Scheme 19, the
product distribution expected for the different reaction pathways
differ.
Selectivities (%) and thereof derived ratio between both side-reaction pathways.
a
Catalyst
HSiCl3
MeSiHCl2
M/AR p:q
Me2SiHCl
M/AR p:q AM:
13:87 15:85
The oxidative addition of 3,3-dideutero allyl chloride is ex-
pected to result in a loss of the chemical information about the ori-
ginal position of the chloro substituent, leading to a 1:1 mixture of
3,3-dideutero propene (n) and 1,1-dideutero propene (o)
(Scheme 19a). In contrast, the 2,1-insertion/b-Cl-elimination and
the allylic rearrangement, respectively, would solely lead to 1,1-
dideutero propene (o, Scheme 19b). The analysis of the product
distribution of the subsequently formed propylsilanes p and q via
2H NMR spectroscopy allows to distinguish between the main side-
reaction pathways.
p:q
AM:
a
AM:
a
aM/AR
Karstedt 49:51 96:04
Markó 15:85 18:82
39:61 36:74
20:80 25:75
b
b
–
–
a
Integration of n and o due to overlapping signals of p and q.
Determination not possible, due to overlapping signals.
b
In summary, the experiments confirm the coexistence of the
proposed side-reaction pathways: the allyl mechanism and the
-mechanism.
a
The ratio of the deuterated byproducts p and q was determined
by integrating the signals of the CD2-groups in a- and c-position.
The 2H NMR spectrum of the Karstedt catalyzed hydrosilylation
of i with chlorodimethylsilane is exemplarily shown in Fig. 10.
Based on the expected 1:1 ratio for the allyl mechanism, the
contribution of each side-reaction pathway can be determined.
The obtained results are summarized in Table 3.
3.3. Overall reaction mechanism
From the experimental and theoretical results discussed above,
in accordance with the mechanistic proposals mentioned in the
introduction, Scheme 20 shows the overall reaction network.
Generally, the described side-reaction pathways coexist, with
the a-mechanism being the dominating process. Only the byprod-
After coordination of the allyl compound (I), the oxidative addi-
a
tion of either the silane to 3 (IICH/ M) or the allyl compound (IIAM
)
ucts of the Karstedt catalyzed hydrosilylation of allyl chloride with
trichlorosilane are (almost) exclusively formed via the allyl mech-
anism. Accordingly, this reaction shows exclusively the aforemen-
tioned influence of the silane concentration on the product
selectivity (Section 3.1.2).
to 8 represents the first selectivity-determining step within the
mechanistic cycle. If the allyl compound is added, the byproducts
are formed via the allyl mechanism, proceeding via a
r-bond
metathesis of 8 and the silane (IIIAM) as the rate-determining step.
If the oxidative addition of the silane takes place, three reaction
Fig. 10. 2H NMR spectrum of the Karstedt catalyzed hydrosilylation of i with chlorodimethylsilane. The peak at 3.09 ppm is attributed to the CD2Cl group of the desired
product.

12
P. Gigler et al. / Journal of Catalysis 295 (2012) 1–14
Scheme 20. Overall reaction mechanism for the hydrosilylation of allyl compounds.
pathways of 3 are possible, thus exhibiting the second selectivity-
determining step. According to the Chalk–Harrod mechanism, the
desired product c can be formed via a 1,2-insertion reaction of
the olefin into the Pt–H bond (4) followed by a reductive elimina-
tion. However, a 2,1-insertion leads to the formation of 5, which
reaction pathways to the overall side-product formation by per-
forming the reaction with 3,3-dideutero allyl chloride (Sec-
tion 3.2.2.4), a similar trend is observed. As listed in Table 3, it
becomes obvious that the contribution of the allyl mechanism de-
creases with increasing electron density of the employed silane
when the Karstedt catalyst is used, while the opposite trend is ob-
served for the Markó catalyst.
Hence, the favored reaction pathway is determined by the
electronic interplay between the catalyst and the silane (for a
given allyl compound), due to the underlying competition of
the oxidative addition of the silane vs. the allyl compound
(Scheme 10).
For example, the oxidative addition of the electron-poor trichlo-
rosilane to the platinum center of the Karstedt catalyst would re-
sult in an extremely electron-deficient platinum (II) species,
which therefore should be thermodynamically unfavored. The oxi-
dative addition of the allyl compound, however, would lead to an
electronic stabilization of the formed allyl complex. Thus, the oxi-
dative addition of the silane should be electronically favored, when
an electron-deficient silane is employed in the presence of an elec-
tron-rich catalyst and vice versa.
In that case, a corresponding theoretical interpretation turned
out to be not very reasonable, considering the fact that the calcu-
lations were performed under slightly different reaction conditions
(e.g. room temperature) and therefore, the values of the energy
barriers can only give an order of magnitude. Nevertheless, regard-
ing all the reaction free enthalpies as well as the energy barriers,
both side-reaction pathways (oxidative addition of the silane and
the allyl compound) are generally feasible from the computational
point of view.
undergoes a reductive elimination for X = NR2 yielding the a-prod-
a
uct k (VII M). For X – NR2, the byproducts d and e are generated
a
after a b-elimination at the platinum center (IV M) and two subse-
a
M
a
quent elimination steps (V
and VI M), whereas the rate-deter-
mining step is attributed to the reductive elimination of d (VI M).
a
a
M
Instead of the 2,1-insertion/b-elimination process (VII
and
IV M), the formation of the byproducts can alternatively proceed
via a concerted allylic rearrangement step (IIIAR), representing
the rate-determining step within this reaction sequence.
a
Therefore, besides the Chalk–Harrod mechanism leading to the
desired product, three different pathways yielding the observed
byproducts are apparent: the ‘‘allyl mechanism,’’ the ‘‘
nism,’’ and the ‘‘allylic rearrangement’’.
a-mecha-
As mentioned before, two crucial elementary steps during the
catalytic cycle are responsible for the observed byproduct forma-
tion: first, the oxidative addition of the silane vs. the allyl com-
pound and second, the insertion of the olefin into the Pt–H bond.
In order to find ways to improve the catalytic performance, that
is, to achieve higher product selectivities, these catalytic steps
are focused on and potential solutions discussed.
3.3.1. Oxidative addition of the silane vs. the allyl compound
Regarding the catalytic results discussed in Section 3.1.1, an
opposite trend with respect to the product selectivity was ob-
served for the catalysts using different silanes in the hydrosilyla-
tion of allyl chloride: While the product selectivity increases
with increasing electron density of the silane in case of the Kar-
stedt catalyst, the opposite behavior is observed for the Markó cat-
alyst. Considering further the contribution of the different side-
The electronic interplay of the silane and the catalyst clearly re-
veals the possibility to influence the catalytic performance. The
contribution of the allyl mechanism could be reduced either by
the choice of the allyl compound itself (see Table 1) or for a given

P. Gigler et al. / Journal of Catalysis 295 (2012) 1–14
13
allyl compound (e.g. allyl chloride) by choosing the right catalyst
for a specific application. For example, the reduced formation of
the byproducts in case of the hydrosilylation of allyl chloride with
trichlorosilane would require an electron-rich catalyst, while for
the same reaction with chlorodimethylsilane, the Karstedt catalyst
is expected to give the better results. Alternatively, by choosing the
right stoichiometry, that is, avoiding an excess of silane, a higher
selectivity could also be achieved, as shown for the Karstedt cata-
lyzed hydrosilylation of allyl chloride with trichlorosilane in
Section 3.1.2.
and Dr. Dennis Troegel (Wacker Chemie AG) for helpful discus-
sions. The Leibniz Rechenzentrum (LRZ) of the Bavarian Academy
of Sciences is acknowledged for providing the necessary computing
time on their Linux cluster farms.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2012.06.006.
3.3.2. Insertion reaction
References
The second selectivity-determining step within the catalytic cy-
cle is represented by the insertion of the olefin into the metal-hy-
dride bond. While a 1,2-addition leads to the desired product, the
2,1-addition product is unstable and decomposes to the observed
byproducts. The different regioselectivity of this reaction can be
attributed to the similar polarities of the olefinic carbon atoms,
as the DFT calculations confirm.
Therefore, the computational data of the H-transfer from the
platinum center to either the terminal or the internal position were
compared. From a kinetic point of view, there is no significant dif-
ference comparing the energy barriers for the two possible inser-
tion steps. Thermodynamically, the 2,1-insertion product is
preferred, because of an additional Pt–Cl contact, initiating the
subsequent Cl transfer discussed above.
[1] B. Marciniec, Hydrosilylation: A Comprehensive Review on Recent Advances,
Springer, London, 2009.
[2] B. Marciniec, Comprehensive Handbook on Hydrosilylation, Pergamon Press,
Oxford, 1992.
[3] B. Marciniec, in: B. Cornils, W.A. Herrmann (Eds.), Applied Homogeneous
Catalysis with Organometallic Compounds, Wiley-VCH, Weinheim, 2002.
[4] B. Marciniec, Silicon Chem. 1 (2002) 155–174.
[5] B. Marciniec, in: B. Cornils, W.A. Herrmann (Eds.), Applied Homogeneous
Catalysis with Organometallic Compounds, Wiley-VCH, Weinheim, 2008, pp.
487–506.
[6] I. Ojima, in: S. Patai, Z. Rappoport (Eds.), Organic Silicon Compounds, John
Wiley & Sons, Chichester, 1989.
[7] I. Ojima, Z. Li, J. Zhu, in: Y. Apeloig, Z. Rappoport (Eds.), The Chemistry of
Organic Silicon Compounds, John Wiley & Sons, Chichester, 1998.
[8] A.K. Roy, Adv. Organomet. Chem. 55 (2007) 1–59.
[9] M.A. Brook, Silicon in Organic Organometallic and Polymer Chemistry, Wiley,
New York, 2000.
[10] J.A. Reichl, D.H. Berry, Adv. Organomet. Chem. 43 (1999) 197–265.
[11] D. Troegel, J. Stohrer, Coord. Chem. Rev. 255 (2011) 1440–1459.
[12] U. Deschler, P. Kleinschmit, P. Panster, Angew. Chem. Int. Ed. 25 (1986) 236–
252.
[13] J.L. Speier, J.A. Webster, G.H. Barnes, J. Am. Chem. Soc. 79 (1957) 974–979.
[14] J.W. Ryan, G.K. Menzie, J.L. Speier, J. Am. Chem. Soc. 82 (1960) 3601–3604.
[15] B. Marciniec, H. Maciejewski, W. Duczmal, R. Fiedorow, D. Kityn´ ski, Appl.
Organomet. Chem. 17 (2003) 127–134.
[16] M. Jankowiak, H. Maciejewski, J. Gulinski, J. Organomet. Chem. 690 (2005)
4478–4487.
[17] A.J. Chalk, J.F. Harrod, J. Am. Chem. Soc. 87 (1965) 16–21.
[18] M.A. Schroeder, M.S. Wrighton, J. Organomet. Chem. 128 (1977) 345–358.
[19] G.H. Wagner, US Patent 2637738, 1953 (to Union Carbide and Carbon
Corporation).
In contrast to the previously discussed oxidative addition step,
the similar polarities of the olefinic carbon atoms leading to the
undefined regioselectivity of the insertion reaction are supposed
to be an intrinsic property of allyl compounds, having the for-
mula CH2CHCH2X. Considering further that this reaction step is
mainly responsible for the formation of the byproducts, the de-
mand for appropriate catalysts to overcome this problem is obvi-
ous. The goal should be to develop catalysts, which either
electronically or sterically force the insertion reaction into the
right direction.
[20] A.B. Pangborn, M.A. Giardello, R.H. Grubbs, R.K. Rosen, F.J. Timmers,
Organometallics 15 (1996) 1518–1520.
4. Conclusion
[21] G.R. Fulmer, A.J.M. Miller, N.H. Sherden, H.E. Gottlieb, A. Nudelman, B.M. Stoltz,
J.E. Bercaw, K.I. Goldberg, Organometallics 29 (2010) 2176–2179.
[22] G. Berthon-Gelloz, O. Buisine, J.-F. Brière, G. Michaud, S. Stérin, G. Mignani, B.
Tinant, J.-P. Declercq, D. Chapon, I.E. Markó, J. Organomet. Chem. 690 (2005)
6156–6168.
[23] D.V. McGrath, R.H. Grubbs, Organometallics 13 (1994) 224–235.
[24] R.C. van der Drift, M. Vailati, E. Bouwman, E. Drent, J. Mol. Catal. A: Chem. 159
(2000) 163–177.
[25] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman,
J. Montgomery, J. A., T. Vreven, K.N. Kudin, J.C. Burant, J.M. Millam, S.S. Iyengar,
J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G.A.
Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa,
M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J.E. Knox,
H.P. Hratchian, J.B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R.E.
Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W. Ochterski, P.Y.
Ayala, K. Morokuma, G.A. Voth, P. Salvador, J.J. Dannenberg, V.G. Zakrzewski, S.
Dapprich, A.D. Daniels, M.C. Strain, O. Farkas, D.K. Malick, A.D. Rabuck, K.
Raghavachari, J.B. Foresman, J.V. Ortiz, Q. Cui, A.G. Baboul, S. Clifford, J.
Cioslowski, B.B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R.L.
Martin, D.J. Fox, T. Keith, M.A. Al-Laham, C.Y. Peng, A. Nanayakkara, M.
Challacombe, P.M.W. Gill, B. Johnson, W. Chen, M.W. Wong, C. Gonzalez, J.A.
Pople, 2004. Gaussian 03, Revision C.02, Gaussian, Inc., Wallingford, CT.
[26] A.D. Becke, J. Chem. Phys. 98 (1993) 5648–5652.
In summary, a comprehensive mechanistic picture of the hydro-
silylation of allyl compounds is given. Besides the product forma-
tion, which proceeds via the classical Chalk–Harrod mechanism,
three different reaction pathways leading to the observed byprod-
ucts have to be considered. It turned out that the contribution of
each path strongly depends on the employed reactants as well as
the catalyst. Nevertheless, two crucial elementary steps were iden-
tified, determining the overall product selectivity. The first one, oxi-
dative addition of the allyl compound vs. silane addition can be
controlled to a certain extent either by the choice of the allyl com-
pound itself, by choosing the right combination of silane and cata-
lyst for a given allyl compound or by varying the stoichiometry of
the reaction partners. Controlling the regioselectivity of the inser-
tion reaction, exhibiting the second selectivity-determining step,
however, turns out to be the real challenge of this catalytic reaction.
For this purpose, the mechanistic work presented herein, and in
particular, the identification of the selectivity-determining steps
opens up the way to a straightforward design of well-defined cata-
lysts to overcome the problem of byproduct formation during the
hydrosilylation of allyl compounds.
[27] C. Lee, W. Yang, R.G. Parr, Phys. Rev. B.: Condens. Matter 37 (1988)
785–789.
[28] P.J. Stephens, F.J. Devlin, C.F. Chabalowski, M.J. Frisch, J. Phys. Chem. 98 (1994)
11623–11627.
[29] S.H. Vosko, L. Wilk, M. Nusair, Can. J. Phys. 58 (1980) 1200–1211.
[30] W.J. Hehre, R. Ditchfield, J.A. Pople, J. Chem. Phys. 56 (1972) 2257–2261.
[31] M. Dolg, H. Stoll, H. Preuss, R.M. Pitzer, J. Phys. Chem. 97 (1993) 5852–5859.
[32] M. Dolg, U. Wedig, H. Stoll, H. Preuss, J. Chem. Phys. 86 (1987) 866–872.
[33] H.B. Schlegel, J. Comput. Chem. 3 (1982) 214–218.
Acknowledgments
[34] G. Schaftenaar, J.H. Noordik, J. Comput. Aided Mol. Des. 14 (2000) 123–134.
[35] R. Dennington II, T. Keith, J. Millam, K. Eppinnett, W.L. Hovell, R. Gilliland,
2003. GaussView, Version 3.09. Shawnee Mission, KS.
We are grateful for the financial support of the Wacker Chemie
AG and thank Prof. Peter Härter (TUM), Dr. Aroop K. Roy (Wacker
Chemical Cooperation), Dr. Jürgen Stohrer, Dr. Florian Hoffmann,
[36] B.D. Karstedt, US Patent 3 775 452, 1973 [to G.E. Company].

14
P. Gigler et al. / Journal of Catalysis 295 (2012) 1–14
[37] I.E. Markó, S. Stérin, O. Buisine, G. Mignani, P. Branlard, B. Tinant, J.-P. Declercq,
Science 298 (2002) 204–206.
[41] R.H. Crabtree, The Organometallic Chemistry of the Transition Metals, John
Wiley & Sons, Hoboken, 2009.
[38] S. Sakaki, N. Mizoe, M. Sugimoto, Organometallics 17 (1998) 2510–2523.
[39] S. Sakaki, N. Mizoe, M. Sugimoto, Y. Musashi, Coord. Chem. Rev. 190–192
(1999) 933–960.
[40] S. Sakaki, M. Ogawa, Y. Musashi, T. Arai, J. Am. Chem. Soc. 116 (1994) 7258–
7265.
[42] B.M. Trost, D.L. Van Vranken, Chem. Rev. 96 (1996) 395–422.
[43] N. Sabourault, G. Mignani, A. Wagner, C. Mioskowski, Org. Lett. 4 (2002) 2117–
2119.
[44] ReaxysÒ reaction search.