Mechanism of olefin hydrosilylation catalyzed by [RuCl(NCCH3)5]+: A DFT study

Article abstract of DOI:10.1016/j.jorganchem.2007.01.060

The hydrosilylation reaction between methyldimethoxysilane and methylvinyldimethoxysilane, catalyzed by the cationic species chloropenta(acetonitrile)ruthenium(II)⁺ (C1), was investigated with density functional theory (DFT). The Chalk-Harrod,

Full text of DOI:10.1016/j.jorganchem.2007.01.060

Journal of Organometallic Chemistry 692 (2007) 2282–2290
www.elsevier.com/locate/jorganchem
Mechanism of olefin hydrosilylation catalyzed by [RuCl(NCCH3)5]+:
A DFT study
a,
a
a,
b
b
Tell Tuttle *, Dongqi Wang , Walter Thiel *, Jutta Ko¨hler , Marco Hofmann ,
b
Johann Weis
a
Max-Planck-Institut fu¨r Kohlenforschung, D-45470 Mu¨lheim an der Ruhr, Germany
Consortium fu¨r Elektrochemische Industrie GmbH, Zielstattstrasse 20, D-81379 Mu¨nchen, Germany
b
Received 14 November 2006; accepted 30 January 2007
Available online 11 February 2007
Abstract
The hydrosilylation reaction between methyldimethoxysilane and methylvinyldimethoxysilane, catalyzed by the cationic species chlo-
ropenta(acetonitrile)ruthenium(II)+ (C1), was investigated with density functional theory (DFT). The Chalk–Harrod, Glaser–Tilley and
r-bond metathesis mechanisms were considered as mechanistic possibilities for the reaction and enthalpy profiles of each pathway were
computed for the active form of C1. In contrast to the commonly accepted Chalk–Harrod mechanism of hydrosilylation, the computa-
tional results indicate that a r-bond metathesis mechanism, involving the formation of a hydride analogue of C1, is most favored. The
B3LYP calculated activation enthalpy for this pathway (DHact = 13.1 kcal/mol) is consistent with the experimental observation that C1
is a reasonable catalyst for this reaction under the applied experimental conditions.
Ó 2007 Elsevier B.V. All rights reserved.
Keywords: Density functional theory; B3LYP; Ruthenium; Hydrosilylation; Chalk–Harrod mechanism; Glaser–Tilley mechanism; r-Bond metathesis
1. Introduction
activity and the ability to control their reactivity by altering
the experimental conditions [1–4,8]. One avenue for con-
trolling the activity of platinum based compounds is
through the manipulation of the active species generated
in the induction phase [8]. This type of selectivity was suc-
cessfully applied for a particular class of platinum based
catalysts, bis(alkynyl)(1,5-cyclooctadiene)platinum com-
plexes (COD)Pt(CCR)2 [9].
Other transition metal compounds can catalyze the
hydrosilylation reaction as well, e.g. those based on ruthe-
nium [10–19]. However, such catalysts generally suffer from
a loss of activity and/or selectivity relative to platinum
catalysts, and they may induce side reactions of the sub-
strates including olefin isomerization, dehydrogenative sily-
lation and hydrogenation. Nonetheless, some very selective
ruthenium catalysts have been found for the regioselective
hydrosilylation of alkynes, especially the cationic ruthe-
nium species [Cp*Ru(NCMe)3]PF6 which has been
extensively used by Trost and Ball [13,14]. Another ionic
Hydrosilylation, the addition of hydrosilanes across
unsaturated bonds [1–4], is an important reaction in the
process of creating organosilicon building blocks. These
building blocks are subsequently used in the construction
of commercially available silicon based products, such as:
silicone rubber, liquid injection molding compounds, paper
release coatings, pressure sensitive adhesives, binders and
coupling agents [2,5–7].
Hydrosilylation reactions are normally performed under
mild conditions in the presence of a catalyst [1–4]. The
most common catalysts for these reactions have tradition-
ally been platinum based compounds owing to their high
*
Corresponding authors. Tel.: +49 208 306 2163; fax: +49 208 306 2996
(T. Tuttle).
E-mail addresses: tell@mpi-muelheim.mpg.de (T. Tuttle), thiel@mpi-
muelheim.mpg.de (W. Thiel).
0022-328X/$ - see front matter Ó 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2007.01.060

T. Tuttle et al. / Journal of Organometallic Chemistry 692 (2007) 2282–2290
2283
ruthenium species, [RuCl(NCCH3)5][RuCl4(NCCH3)2] has
been shown to be a very effective catalyst in the hydrosily-
lation of CO2 with tri- and diorganosilanes [20,21]. These
studies prompted us to investigate the activity of the latter
catalyst in the hydrosilylation of C@C-functional sub-
strates. A preliminary test reaction of diethoxymethylsilane
with diethoxymethylvinylsilane indicated that the activity
of the catalyst was indeed reasonable (see Section 2).
To optimize ionic catalysts such as [RuCl(NC-
CH3)5][RuCl4(NCCH3)2] it is essential to understand the
catalytic mechanism. Thus, we present in this report the
results of a computational investigation of the mechanism
by which the Ru(II) cation [RuCl(NCCH3)5]+ (C1) cata-
lyzes the reaction between dimethoxymethylsilane (R1)
and dimethoxymethylvinylsilane (R2) to form 1,2-
bis(dimethoxymethylsilyl)ethane (P1), Scheme 1, as the
first step towards the fine-tuning of the catalyst for optimal
activity. We assume in our model calculations that the
Ru(II) cation is the active species, rather than the Ru(III)
anion [RuCl4(NCCH3)2]ꢀ, since there are no Ru(III) com-
plexes among all known Ru hydrosilylation catalysts
(excluding RuCl3, which is not the active species itself).
The Chalk–Harrod [22] (CH) and modified Chalk–Har-
rod [16,23–27] (mCH) catalytic cycles are the most com-
mon mechanisms in the literature for explaining the
catalysis of the hydrosilylation reaction with late transition
metal based compounds [8,13,14,17,28–33]. Both the CH
and mCH mechanisms involve an initial oxidative addition
of an SiH-functional silane to the transition metal, Ru (step
I, Scheme 2). In the CH mechanism, this is followed by
subsequent insertion of a C@C functional substrate into
the Ru–H bond (step IIa), whereas in the mCH mechanism
the C@C bond is inserted into the Ru–Si bond (step IIb).
The final step (step III) for both mechanisms is the reduc-
tive elimination of the hydrosilylation product.
Scheme 2. Chalk–Harrod and modified Chalk–Harrod catalytic cycles.
Scheme 3. Glaser–Tilley catalytic cycle.
and Ru). Given the cationic nature of our catalyst, this
mechanism was also tested.
2. Experimental
2.1. General considerations
The silanes diethoxymethylvinylsilane and diethoxy-
methylsilane were purchased from GELEST and used as
received. The ruthenium catalyst [RuCl(NCCH3)5][RuCl4-
(NCCH3)2] was prepared by the method of Pitter et al.
[20]. 1H and 29Si NMR spectra were recorded at room tem-
perature on a Bruker Avance 300 spectrometer in CDCl3.
All chemical shifts are in ppm referenced to the residual
proton solvent resonance at d 7.24 ppm (1H), or to the
external standard TMS (29Si). Routine GC analysis was
performed with an AGILENT 6890 N gas chromatograph
equipped with an RtX-200 column (Restek GmbH). Cali-
bration was performed with pure samples of the silanes
diethoxymethylvinylsilane, diethoxymethylsilane, and tri-
ethoxymethylsilane (purchased from GELEST) as well as
bis(diethoxymethylsilyl)ethane, which was independently
prepared by the literature method [52]. GC–MS spectra
were measured on an AGILENT 6890/MSD 5973 GC/
mass spectrometer at 70 eV.
An alternative to the Chalk–Harrod scheme is the r-
bond metathesis (SBM) mechanism [34,35], which is well
established for early transition metal catalysts [36–40].
More recently, it has also been shown to be active in a
number of late transition metal catalysts as well [19,41–
51]. Finally, in 2003, Glaser and Tilley proposed another
mechanism for a cationic silylene Ru(II) catalyst in the
hydrosilylation of alkenes [11], and subsequent computa-
tional studies supported this proposal [10]. The Glaser–Til-
ley (GT) mechanism (Scheme 3) involves the insertion of
the alkene directly into an Si–H bond remote from the
metal center (i.e. no direct interaction between the alkene
2.2. Catalytic run
A 25 mL two-necked round-bottomed flask equipped
with a reflux condenser, argon inlet and a magnetic stir-
Scheme 1. Chemical structure and notation of the reactants, products and
catalyst.

2284
T. Tuttle et al. / Journal of Organometallic Chemistry 692 (2007) 2282–2290
ring bar was charged under argon with 17.1 mg (25.6 lmol)
of [RuCl(NCCH3)5][RuCl4(NCCH3)2]. 1.75 g (10.9
the-identity (RI)-DFT approach [64], as implemented in
TURBOMOLE [65–68]. All structures were subsequently
refined in GAUSSIAN-03 [69], by carrying out geometry
optimizations with the B3LYP hybrid functional
[61,63,70–73], which has been shown to produce reliable
thermochemical data for ruthenium-based compounds
[74–76].
GAUSSIAN-03 was employed in the refinement process as
the search for TSs often made use of the synchronous tran-
sit-guided quasi-Newton method [77,78], available in this
package. Frequency calculations were performed on all
optimized structures, using the B3LYP functional, to char-
acterize the stationary points as minima or TSs, as well as
for the calculation of zero-point energies (ZPE), enthalpies
(H), entropies (S), and Gibbs free enthalpies (G) at 298 K.
A
mmol) amount of Me(Vi)Si(OEt)2 and 1.46 g (10.9 mmol)
of Me(H)Si(OEt)2 were added by syringe and the mixture
was stirred and heated at 100 °C. The progress of the reac-
tion was monitored by GC. After 2 h the starting SiH-func-
tional silane was completely consumed. After cooling to
room temperature, the reaction mixture was analyzed by
GC and GC/MS. The crude mixture was purified by distil-
lation leaving pure (EtO)2MeSi–CH2CH2–SiMe(OEt)2 in a
1
yield of 1.77 g (55%). The H NMR data of (EtO)2MeSi–
CH2CH2–SiMe(OEt)2 compare well with the literature
values [53].
2.3. (EtO)2MeSi–CH2CH2–SiMe(OEt)2
1H NMR (300.1 MHz, CDCl3)
d
3.72 (q, 8H,
4. Results and discussion
J = 7.0 Hz, H2CO), 1.17 (t, 12H, J = 7.0 Hz, H3CCH2),
0.54 (s, 4H, CH2CH2), 0.07 (s, 6H, H3CSi). 29Si NMR
(59.6 MHz, CDCl3) d ꢀ4.42 (s). GC/MS: m/z (relative
abundance): 279 (4%, MꢀCH3).
The addition of R1 to R2 in an uncatalyzed hydrosilyla-
tion reaction occurs in a concerted mechanism, with the
H-abstraction from R1 and the Si–C bond formation
occurring concomitantly. In our previous study [19], this
reaction was found to have a prohibitively high barrier
(DHact = 53.6 kcal/mol) and a catalyst is therefore needed
to lower the activation enthalpy for the hydrosilylation to
occur under normal operating conditions.
2.4. Reaction products
The hydrosilylation of diethoxymethylvinylsilane with
diethoxymethylsilane was performed at 100 °C in the pres-
ence of [RuCl(NCCH3)5][RuCl4(NCCH3)2] (0.0024 equiv.)
without solvent monitoring the progress of the reaction via
GC. After 2 h, no hydrosilane was detectable by GC. How-
ever, the formation of the hydrosilylation product (EtO)2-
MeSi–CH2CH2–SiMe(OEt)2 (60% yield by GC) is
accompanied by several by-products, especially MeSi-
(OEt)3 (12%), presumably as a result of a H/OEt exchange
reaction of diethoxymethylsilane. This would also explain
that diethoxymethylvinylsilane is not fully consumed (7%
left in the reaction mixture). The product distribution also
contains 21% of unidentified compounds. However, it is
noteworthy that the product of a dehydrogenative silyla-
tion, i.e. (EtO)2MeSi–CH@CH–SiMe(OEt)2 [54] is not
formed, which implies that a b-hydrogen elimination reac-
tion as side reaction path consecutively to step IIb in
Scheme 2 (according to the modified Chalk–Harrod mech-
anism) does not occur. This may also imply that the mCH
mechanism is not valid for this catalyst (see below).
4.1. The catalyst
The cation C1 is produced as an ion pair [20,21] – trans-
[RuIICl(MeCN)5][RuIIICl4(MeCN)2] – with a Ru(II) cation
and a Ru(III) anion, from the reaction of RuCl3 with di-
methylphenylsilane in acetonitrile. As mentioned before,
we only consider the Ru(II) cation C1 as catalyst for the
hydrosilylation reaction.
The optimized structure of C1 (Fig. 1) is in good agree-
ment with the available X-ray structure. The average Ru–N
˚
bond length, cis to the Ru–Cl bond, is 2.03 A, in excellent
˚
agreement with the reported X-ray structure value (2.03 A)
3. Computational methods
Density functional theory (DFT) [55,56] was employed
for the calculation of all reactants, transition states (TSs),
intermediates and products. In all calculations, the ruthe-
nium atom was described by a small-core, quasi-relativistic,
effective core potential with the associated (7s6p5d)/
[5s3p3d] valence basis set [57], while the 6-31G(d,p) basis
set [58–60] was used for all other atoms. Initial geometry
optimizations of the minima along the reaction path were
performed with the gradient-corrected BP86 functional
[61–63], in order to take advantage of the resolution-of-
˚
Fig. 1. Optimized structure of C1. Bond length in A and angles in °.

T. Tuttle et al. / Journal of Organometallic Chemistry 692 (2007) 2282–2290
2285
[20]. The agreement between the computed and X-ray
structure is equally good for the Ru–Cl bond (2.40 A).
generated by a subsequent direct r-bond metathesis reac-
tion between the dissociation product RuCl2(CO)2(PPh3)
(A7) and R1 which required only very little activation
(DHact = 1.9 kcal/mol) followed by a rearrangement of
the resulting penta-coordinated complex RuHCl(CO)2-
(PPh3) (A8). The induction mechanisms for the previously
and presently studied catalysts thus show some similarities:
the active species C7 and A16 are analogous penta-coordi-
nated complexes that contain a Ru–H bond which is intro-
duced by a chloride/hydride exchange through a r-bond
metathesis reaction with R1. However, there are also differ-
ences: in the present case, additional hexa-coordinated
complexes (C4 and C5) are encountered on the induction
pathway, and in the initial induction step (C1 ! C4), a
direct substitution is enthalpically favored because of the
higher barrier to ligand dissociation. The induction is
enthalpically less facile in the present case, since the largest
single-step activation enthalpies are 15.4 and 23.5 kcal/mol
for A1 and C1, respectively, while the highest points on the
corresponding enthalpy profiles are 17.3 and 25.6 kcal/mol,
respectively. Under the normally applied experimental con-
ditions, such barriers are surmountable in the induction
phase.
˚
The computed structure has a slightly elongated Ru–N
˚
bond in trans position to the Cl ligand (2.06 A), consistent
with the strong r-donor capabilities of the chloride ligand.
˚
In the X-ray structure (Ru–N = 2.02 A) such an elongation
was not observed. Nonetheless, the overall agreement
between the computed and X-ray structure is excellent
and indicates that the current level of theory is capable of
adequately describing the cationic Ru(II) catalyst.
4.2. The induction period
Generating the active form of C1 requires the dissocia-
tion of at least one of the six ligands in the octahedral com-
plex. In the classic version of the CH or mCH mechanisms,
two free coordination sites are required to bind the reaction
partners. The dissociation enthalphy (DHd (298 K)) for
removing two acetonitrile ligands is 70.5 kcal/mol, hence
this process is prohibitive enthalpically (Scheme 4).
An alternative induction mechanism begins with the
replacement of one of the acetonitrile ligands by R1, which
yields the hexa-coordinated complex C4 (Fig. 2). The direct
substitution via the transition state TS(C1–C4) requires an
activation of 23.5 kcal/mol and is thus enthalpically more
favorable than the corresponding dissociation/association
mechanism, which has an enthalpic cost of 31.2 kcal/mol
for the initial dissociation C1 ! C2 while the subsequent
addition C2 + R1 ! C4 is barrierless and exothermic.
The complex C4 is destabilized by 13.4 kcal/mol relative
to C1. It can be transformed into a more active form
(C5) by a slightly exothermic r-bond metathesis reaction
(DH = ꢀ2.3 kcal/mol), in which the Si–H bond of R1 is
replaced by the Si–Cl bond, and the Ru–Cl bond by the
Ru–H bond (via TS(C4–C5), DHact = 12.2 kcal/mol). The
hexa-coordinated complex C5 can lose the chlorosilane
ligand rather easily (DH = 8.9 kcal/mol) to form the
penta-coordinated complex C7 (see below). Both C5 and
C7 are thus enthalpically accessible and can serve as start-
ing points for the catalytic cycle.
In a recent study [19] of the same hydrosilylation reac-
tion with the RuCl2(CO)2(PPh3)2 catalyst (A1), the initial
dissociation of a triphenylphosphine ligand was found to
be much more facile, with an enthalpic cost of only
15.4 kcal/mol, i.e. about half of that required for removing
an acetonitrile ligand from the current catalyst C1. This
difference is due to the electron-deficient Ru(II) environ-
ment created by the acetonitrile ligands in C1. In the previ-
ous study [19], the active form of the catalyst (A16) was
In the following, we address three possible catalytic
mechanisms for our current catalyst: the CH and GT
mechanisms starting from C5, and an SBM mechanism
starting from C7. The role of entropic effects will be dis-
cussed thereafter.
4.3. Catalytic mechanisms
The CH mechanism involves the insertion of an alkene
into the metal-hydride bond followed by the product form-
ing step where the metal–silyl bond is broken (Scheme 2).
In C5 the alkene (R2) is able to insert into the Ru–H bond
(Fig. 3) to form C6. However, a Ru–Si bond does not exist
in C5, which excludes the mCH mechanism (Scheme 2) for
this active species. The optimized mCH intermediate, con-
structed from the four-coordinate Ru(II) complex, is
23.8 kcal/mol less stable than the analogous CH intermedi-
ate. Thus, the relative stability of the two intermediates
also disfavors the possibility of an mCH mechanism for
this catalyst.
The insertion of R2 into C5 to form the CH intermediate
is an exothermic reaction (DH(C6) = ꢀ6.8 kcal/mol) but
requires the surmounting of an extremely large barrier
(DHact(TS(C5–C6)) = 56.7 kcal/mol). The high activation
enthalpy for the olefin insertion process arises from the ste-
ric crowding around the metal center. The approaching
olefin is not able to form a pre-complex with the metal
before abstracting the hydride due to the saturation of
the coordination sites in C5 (Fig. 4). Consequently, the
hydride abstraction precedes the formation of the Ru–C
bond in C6, which is a highly unfavorable process given
the electron deficient acetonitrile environment. This is in
contrast to our findings on the RuCl2(CO)2(PPh3)2 catalyst
where the metal center had been oxidized to a Ru(IV) state
Scheme 4. Dissociation of acetonitrile ligands (L = NCMe), dissociation
enthalpy DHd (298 K) given in kcal/mol.

2286
T. Tuttle et al. / Journal of Organometallic Chemistry 692 (2007) 2282–2290
Fig. 2. Induction mechanism for C1: Replacement of an acetonitrile ligand by R1. L: NCMe, [Si]: Si(OMe)2Me.
Fig. 4. TS structure for the olefin insertion in the CH mechanism. Bond
˚
lengths in A.
did not invest time into obtaining the TS of the product
forming step, although the reaction is again exothermic
(DH(C2) = ꢀ3.6 kcal/mol).
Fig. 3. Enthalpy profile for the CH mechanism. L: NCMe, [Si]:
Si(OMe)2Me.
The GT mechanism also requires a formal metal–silyl
bond in its original form [11]. Nonetheless, in C5 the
hydride and silyl groups are well positioned to allow them
to interact directly with R2, even though there is no binding
between R2 and the metal center in the GT-type TS
(Fig. 5). The simultaneous abstraction of both the silyl
and hydride ligands leads directly to P1 and the active form
of the catalyst, C2. The addition of R1 to C2 occurs in a
barrierless reaction (DH(C4) = ꢀ17.8 kcal/mol) to generate
C4 (see Fig. 2). However, the activation enthalpy for this
upon addition of the silane in the CH mechanism [19]. The
abstraction of the hydride to form the CH intermediate
therefore occurred with only a small barrier (6.7 kcal/
mol), although, the final reductive elimination step
excluded the CH mechanism for this catalyst as well [19].
In the current catalyst it is the barrier to forming the CH
intermediate that excludes this mechanism as a viable
pathway for the hydrosilylation reaction. Therefore, we

T. Tuttle et al. / Journal of Organometallic Chemistry 692 (2007) 2282–2290
2287
Scheme 5. SBM catalytic mechanism. L: NCMe, [Si]: Si(OMe)2Me.
coordination site in C9 is amenable to the attack of R1,
which stabilizes the complex by 2.4 kcal/mol via a dative
bond involving its hydrogen atom (C10, Fig. 6) in an
unhindered process. The last step in the catalytic cycle is
the rate-determining step. The creation of the product
occurs through a final SBM reaction where the H–Si bond
in C10 is replaced by the C–Si bond in P1 and the Ru–C
bond is replaced by a Ru–H bond as the active species,
C7, is reformed.
This catalytic mechanism is analogous to that in the pre-
viously studied catalyst [19]. However, contrary to the
induction period, the present catalyst is predicted to be
more efficient in the catalytic cycle, particularly with regard
to the rate-determining product formation step. The over-
all barrier in the catalytic cycle for the current catalyst
(DHact = 16.0 kcal/mol, relative to C8), is significantly
lower than that obtained for the RuCl2(CO)2(PPh3)2 cata-
lyst (DHact = 21.8 kcal/mol) [19], which reflects the activity
ordering obtained experimentally for the two systems. The
differences in the barrier heights during induction thus do
not play a significant role in determining the catalytic abil-
ity of the two compounds – it is the product formation step
in the catalytic cycle that is decisive.
˚
Fig. 5. TS structure for the GT-type mechanism. Bond lengths in A.
reaction is 47.4 kcal/mol, which is clearly too high to
explain the observed catalytic activity of this catalyst.
Generally, silyl ligands have strong r-donor tenden-
cies; in the GT mechanism (Scheme 3) the electron dona-
tion to the metal centre is enhanced by forming a double
bond to the metal center. This electron donation weakens
the Si–H bond, which allows the alkyl group to abstract
the H-atom from Si and form a Si–C bond. That is, by
forming a double bond to the metal center the Si-atom
is activated and consequently the barrier is lowered for
the hydrosilylation reaction. In C3 the Si-atom has not
been activated by Ru, rather the interaction between
the Ru- and Si-atoms is very weak and the chlorosilane
ligand is energetically stable. The barrier to the hydrosily-
lation reaction in this mechanism is therefore only slightly
lower than in the uncatalyzed reaction (DHact = 53.6 kcal/
mol) [19].
The CH and GT-type pathways discussed above provide
helpful insights to the requirements that the actual mecha-
nism must fulfill. To wit, an additional free coordination
site is required to allow enough space around the metal
center for the reactants to interact with Ru; and, at least
R1 or R2 must directly bind to the metal center in order
to lower the activation barrier. We have incorporated these
features into a catalytic mechanism that is based on a series
of SBM reactions (Scheme 5).
The active species in this mechanism is C7 (Fig. 6) which
is generated from C5 by the liberation of chlorosilane in an
endothermic process (DH(C7) = 8.9 kcal/mol). The olefin
is able to bind to C7 in a barrierless, exothermic reaction
to form C8 (DH(C8) = ꢀ29.5 kcal/mol). The intramolecu-
lar H-transfer reaction to form C9 occurs with a low bar-
rier (DHact(TS(C8–C9)) = 6.9 kcal/mol), however, the
intermediate alkyl-ruthenium complex is produced in an
endothermic reaction (DH(C9) = 5.3 kcal/mol). The free
4.4. Entropic effects
Equilibria and reaction rates depend on the Gibbs free
enthalpies, which differ from the reported relative enthal-
pies (DH) through the inclusion of the entropic contribu-
tions (DG = DH ꢀ TDS). These contributions have been
calculated by applying the harmonic oscillator/rigid rotor
approximation (see Supplementary material); however,
the results should be viewed with some caution, for two
reasons. First, the computed entropic contributions refer
to the gas phase and thus neglect solvation and desolvation
effects in solution, which may be substantial. Second, the
harmonic oscillator/rigid rotor approximation is known
to be problematic in the case of weakly bound complexes
(such as C4 and C10) due to the large number of low-
energy vibrational modes, which in turn have large contri-
butions to the entropy. Therefore, the following discussion

2288
T. Tuttle et al. / Journal of Organometallic Chemistry 692 (2007) 2282–2290
Fig. 6. Enthalpy profile for the SBM mechanism. L: NCMe, [Si]: Si(OMe)2Me.
focuses on how the computed enthalpy profile (Fig. 7) is
qualitatively affected by the entropic contributions. The
main such effect is that association reactions suffer from
an entropic penalty because of the loss of translational
and rotational degrees of freedom (typically around
10 kcal/mol at 298 K in the gas phase), while dissociation
reactions are entropically favored in an analogous manner.
In solution, these entropic effects will be less pronounced
than in the gas phase due to solvation and desolvation,
but they will be present to some extent.
In the induction phase (Fig. 7, black part) the direct sub-
stitution of acetonitrile (C1 ! C4) becomes less facile on
the DG scale because of the entropic contributions associ-
ated with the loss of rotational and translational freedom
when coordinating R1 (i.e. TS(C1–C4) is a seven coordi-
nate complex). However, on the DG scale the dissociation
Fig. 7. Relative enthalpy profile of the predicted mechanism for the hydrosilylation of R2 by R1 with the C1 catalyst. Induction period in black, catalytic
cycle in red. The numerical values given are barriers in the case of the TSs and reaction enthalpies relative to the preceding minimum otherwise (kcal/mol).
(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

T. Tuttle et al. / Journal of Organometallic Chemistry 692 (2007) 2282–2290
2289
of one of the acetonitrile ligands to form C2 becomes more
favorable (DG(C2) = 21.2 kcal/mol). The subsequent asso-
ciation of R1 remains as an unhindered process, although it
is not as stabilizing on the DG scale due to the destabilizing
entropic contribution (C2 ! C4: DH = ꢀ17.8 kcal/mol,
DG = ꢀ5.5 kcal/mol). Thus, the inclusion of entropic
effects should lead to a change in the preferred mechanism
for the formation of C4 from a direct mechanism
(C1 ! C4), which is favored enthalpically, to a dissocia-
tion/association mechanism (C1 ! C2 ! C4), which is
more facile when free enthalpies are considered.
dissociation of an acetonitrile followed by the association
of R1 to form C4 is more favorable from a free enthalpy
perspective (DHact = 31.2 kcal/mol; DGact = 21.1 kcal/
mol). Thus, either of these induction mechanisms are
clearly favored over the spontaneous dissociation of two
acetonitrile ligands due to the electron deficient environ-
ment of Ru(II).
The subsequent dissociation of the chlorosilane to form
C7 represents the starting point of the SBM catalytic mech-
anism. A stabilizing g2-coordination of R2 (ꢀ29.5 kcal/
mol) in a barrierless reaction initiates the catalytic cycle,
where the initial transformations (C7 ! C8 ! C9 ! C10)
are facile. The final SBM step that generates the product
(C10 ! C7) is rate-determining and the corresponding bar-
rier of DHact = 13.1 kcal/mol (DGact = 22.5 kcal/mol) is
consistent with the experimentally observed activity.
The SBM reaction in the induction step (C4 ! C5) is
not significantly affected by the inclusion of entropic effects
(DHact = 12.2 kcal/mol, DGact = 14.8 kcal/mol) due to the
intramolecular nature of the reaction. The dissociation of
chlorosilane in the formation of C7 is favored from an
entropic perspective (C5 ! C7: DH = 8.9 kcal/mol,
DG = ꢀ3.0 kcal/mol). On the DG scale, formation of C4
thus remains as the rate-determining step of the induction
phase, and the five-coordinate intermediate (C7), formed
in the latter part of the reaction, is entropically favored.
Considering the SBM catalytic cycle (Fig. 7, red part1)
we first note that the overall reaction (C7 + R1 +
R2 ! C7 + P1) combines two reactant molecules into
one product molecule so that the entropic contributions
must be positive, which is indeed found (DH = ꢀ29.0
kcal/mol, TDS = 12.9 kcal/mol, DG = ꢀ16.1 kcal/mol).
Similarly, the intermediates in the SBM catalytic cycle
(C8 + R1, C9 + R1, C10) all have fewer molecules than
the entry point (C7 + R1 + R2). Therefore, in the DG pro-
file of the catalytic cycle, all intermediates and the product
are shifted upwards compared with the DH profile, but the
other features remain qualitatively unchanged. In particu-
lar, the barrier to form the alkyl intermediate (C9) remains
similar on the DH and DG scales, and product formation
remains as the rate-determining step. The qualitative con-
clusions drawn from the DH profile (Fig. 7) thus remain
valid also after considering entropic effects in the catalytic
cycle.
Acknowledgement
We thank Dr. Hans-Jurgen Eberle for helpful
¨
discussions.
Appendix A. Supplementary material
Energies, thermochemical data and Cartesian coordi-
nates of all optimized structures and the complete Ref.
[69] are provided as supplementary material. Supplemen-
tary data associated with this article can be found, in the
online version, at doi:10.1016/j.jorganchem.2007.01.060.
References
[1] D. Armitrage, in: G. Wilkinson, F.G.A. Stone, W.E. Abel (Eds.),
Comprehensive Organometallic Chemistry, Pergamon Press, Oxford,
1982, p. 117.
[2] B. Marciniec, J. Gulinski, W. Urbaniak, Z.W. Kornetka, Compre-
hensive Handbook on Hydrosilylation;, Pergamon Press, Oxford,
1992.
[3] I. Ojima, in: S. Patai, Z. Rappoport (Eds.), The Chemistry of Organic
Silicon Compounds, Wiley, New York, 1989, p. 1479.
[4] J.L. Speier, in: F.G.A. Stone, W.E. Abel (Eds.), Advances in
Organometallic Chemistry, Academic Press, New York, 1979, p. 407.
[5] L.N. Lewis, J. Stein, Y. Gao, R.E. Colborn, G. Hutchins, Platinum
Met. Rev. 41 (1997) 66.
5. Conclusions
[6] W. Noll, Chemistry and Technology of Silicones, Academic Press,
New York, 1968.
[7] F.O. Stark, J.R. Falender, A.P. Wright, in: G. Wilkinson, F.G.A.
Stone, W.E. Abel (Eds.), Comprehensive Organometallic Chemistry,
Pergamon Press, Oxford, 1982, p. 306.
[8] M.N. Jagadeesh, W. Thiel, J. Ko¨hler, A. Fehn, Organometallics 21
(2002) 2076.
[9] A. Fehn, F. Achenbach, E.P. Appl, Curable Silicone Rubber
Compositions Containing Organoplatinum Catalysts, Wacker-Che-
mie GmbH, Germany, 2000, EP 994159.
The comparison of the different reaction mechanisms
involving C1 shows that the SBM mechanism provides
the most favorable route to the hydrosilylation of R2 by
R1. The induction period of the catalyst involves the
replacement of an acetonitrile ligand (either by a substitu-
tion reaction or a dissociation/association process) by R1,
followed by a SBM reaction that causes the exchange of
the H-atom of R1 for the Cl-atom and introduces a
hydride ligand on the metal center (C5). In the induction
phase, the direct substitution of an acetonitrile ligand by
R1 has an activation enthalpy of 23.5 kcal/mol. The initial
[10] C. Beddie, M.B. Hall, J. Am. Chem. Soc. 126 (2004) 13564.
[11] P.B. Glaser, T.D. Tilley, J. Am. Chem. Soc. 125 (2003) 13640.
[12] L.W. Chung, Y.D. Wu, B.M. Trost, Z.T. Ball, J. Am. Chem. Soc.
125 (2003) 11578.
[13] B.M. Trost, Z.T. Ball, J. Am. Chem. Soc. 125 (2003) 30.
[14] B.M. Trost, Z.T. Ball, J. Am. Chem. Soc. 123 (2001) 12726.
[15] Y. Maruyama, K. Yamamura, T. Sagawa, H. Katayama, F. Ozawa,
Organometallics 19 (2000) 1308.
1
For interpretation of the references to colour in Fig. 7, the reader is
referred to the web version of this article.

2290
T. Tuttle et al. / Journal of Organometallic Chemistry 692 (2007) 2282–2290
[16] I. Ojima, T. Fuchikami, M. Yatabe, J. Organomet. Chem. 260 (1984)
335.
[47] F. Maseras, M. Duran, A. Lledos, J. Bertran, J. Am. Chem. Soc. 114
(1992) 2922.
[17] Y. Maruyama, K. Yamamura, I. Nakayama, K. Yoshiuchi, F.
Ozawa, J. Am. Chem. Soc. 120 (1998) 1421.
[48] A. Milet, A. Dedieu, G. Kapteijn, G. van Koten, Inorg. Chem. 36
(1997) 3223.
[18] B. Marciniec, C. Pietraszuk, Top. Organomet. Chem. 11 (2004) 197.
[19] T. Tuttle, D. Wang, W. Thiel, J. Ko¨hler, M. Hofmann, J. Weis,
Organometallics 25 (2006) 4504.
[20] A. Jansen, H. Gorls, S. Pitter, Organometallics 19 (2000) 135.
[21] A. Jansen, S. Pitter, J. Mol. Catal. A 217 (2004) 41.
[22] A.J. Chalk, J.F. Harrod, J. Am. Chem. Soc. 87 (1965) 16.
[23] M.J. Fernandez, M.A. Esteruelas, M.S. Jimenez, L.A. Oro, Organo-
metallics 5 (1986) 1519.
[49] D.G. Musaev, A.M. Mebel, K. Morokuma, J. Am. Chem. Soc. 116
(1994) 10693.
[50] P.E.M. Siegbahn, R.H. Crabtree, J. Am. Chem. Soc. 118 (1996) 4442.
[51] M.D. Su, S.Y. Chu, J. Am. Chem. Soc. 119 (1997) 5373.
[52] V.G. Lakhtin, V.M. Nosova, A.V. Kisin, P.V. Ivanov, E.A. Cherny-
shev, Russ. J. Gen. Chem. 71 (2001) 1252.
[53] H. Watanabe, M. Asami, Y. Nagai, J. Organomet. Chem. 195 (1980)
363.
[24] A. Onopchenko, E.T. Sabourin, D.L. Beach, J. Org. Chem. 48 (1983)
5101.
[54] B. Marciniec, H. Maciejewski, J. Gulinski, L. Rzejak, J. Organomet.
Chem. 362 (1989) 273.
[25] C.L. Randolph, M.S. Wrighton, J. Am. Chem. Soc. 108 (1986) 3366.
[26] M.A. Schro¨der, M.S. Wrighton, J. Organomet. Chem. 128 (1977)
345.
[27] F. Seitz, M.S. Wrighton, Angew. Chem., Int. Ed. Engl. 27 (1988) 289.
[28] S. Sakaki, N. Mizoe, Y. Musashi, M. Sugimoto, J. Mol. Struct.
(THEOCHEM) 462 (1999) 533.
[29] S. Sakaki, N. Mizoe, M. Sugimoto, Organometallics 17 (1998) 2510.
[30] S. Sakaki, N. Mizoe, M. Sugimoto, Y. Musashi, Coord. Chem. Rev.
192 (1999) 933.
[31] S. Sakaki, M. Ogawa, Y. Musashi, T. Arai, J. Am. Chem. Soc. 116
(1994) 7258.
[55] W. Kohn, L.J. Sham, Phys. Rev. 140 (1965) 1133.
[56] R.G. Parr, W.T. Yang, Density Functional Theory of Atoms and
Molecules, Oxford University Press, New York, 1989.
[57] D. Andrae, U. Ha¨ussermann, M. Dolg, H. Stoll, H. Preuss, Theor.
Chim. Acta 77 (1990) 123.
[58] P.C. Hariharan, J.A. Pople, Theor. Chim. Acta 28 (1973) 213.
[59] W.J. Hehre, R. Ditchfield, J.A. Pople, J. Chem. Phys. 56 (1972)
2257.
[60] R. Ditchfield, W.J. Hehre, J.A. Pople, J. Chem. Phys. 54 (1971)
724.
[61] A.D. Becke, Phys. Rev. A 38 (1988) 3098.
[62] J.P. Perdew, Phys. Rev. B 33 (1986) 8822.
[63] S.H. Vosko, L. Wilk, M. Nusair, Can. J. Phys. 58 (1980) 1200.
[64] K. Eichkorn, O. Treutler, H. Ohm, M. Ha¨ser, R. Ahlrichs, Chem.
Phys. Lett. 240 (1995) 283.
[32] S. Sakaki, M. Sumimoto, M. Fukuhara, M. Sugimoto, H. Fujimoto,
S. Matsuzaki, Organometallics 21 (2002) 3788.
[33] S.M. Maddock, C.E.F. Rickard, W.R. Roper, L.J. Wright, Organo-
metallics 15 (1996) 1793.
[34] R.H. Crabtree, Chem. Rev. 95 (1995) 987.
[35] B.A. Arndtsen, R.G. Bergman, T.A. Mobley, T.H. Peterson, Acc.
Chem. Res. 28 (1995) 154.
[36] P.F. Fu, L. Brard, Y.W. Li, T.J. Marks, J. Am. Chem. Soc. 117
(1995) 7157.
[37] M.R. Kesti, R.M. Waymouth, Organometallics 11 (1992) 1095.
[38] N.S. Radu, T.D. Tilley, J. Am. Chem. Soc. 117 (1995) 5863.
[39] A.Z. Voskoboynikov, I.N. Parshina, A.K. Shestakova, K.P. Butin,
I.P. Beletskaya, L.G. Kuzmina, J.A.K. Howard, Organometallics 16
(1997) 4041.
[40] H.G. Woo, J.F. Walzer, T.D. Tilley, J. Am. Chem. Soc. 114 (1992)
7047.
[41] J.F. Hartwig, S. Bhandari, P.R. Rablen, J. Am. Chem. Soc. 116
(1994) 1839.
[65] TURBOMOLE, V. 5.7.1, COSMOlogic GmbH & Co. KG, Leverkusen,
Germany, 2004.
[66] R. Ahlrichs, M. Ba¨r, M. Ha¨ser, H. Horn, C. Ko¨lmel, Chem. Phys.
Lett. 162 (1989) 165.
[67] K. Eichkorn, O. Treutler, H. Ohm, M. Ha¨ser, R. Ahlrichs, Chem.
Phys. Lett. 242 (1995) 652.
[68] O. Treutler, R. Ahlrichs, J. Chem. Phys. 102 (1995) 346.
[69] M.J. Frisch et al., GAUSSIAN-03, V.C.01, Gaussian, Inc., Wallingford,
CT, 2004.
[70] A.D. Becke, J. Chem. Phys. 98 (1993) 5648.
[71] C.T. Lee, W.T. Yang, R.G. Parr, Phys. Rev. B 37 (1988) 785.
[72] P.J. Stephens, F.J. Devlin, C.F. Chabalowski, M.J. Frisch, J. Phys.
Chem. 98 (1994) 11623.
[73] R.H. Hertwig, W. Koch, Chem. Phys. Lett. 268 (1997) 345.
[74] J. Baker, M. Muir, J. Andzelm, A. Scheiner, in: Chemical Applica-
tions Of Density Functional Theory, American Chemical Society
Symposium, Washington, DC, 1996, p. 342.
[42] F. Hutschka, A. Dedieu, M. Eichberger, R. Fornika, W. Leitner, J.
Am. Chem. Soc. 119 (1997) 4432.
[43] C.N. Iverson, M.R. Smith, J. Am. Chem. Soc. 117 (1995) 4403.
[44] J.C. Lee, E. Peris, A.L. Rheingold, R.H. Crabtree, J. Am. Chem.
Soc. 116 (1994) 11014.
[75] S.Q. Niu, M.B. Hall, Chem. Rev. 100 (2000) 353.
[76] S.E. Vyboishchikov, M. Buhl, W. Thiel, Chem. Eur. J. 8 (2002) 3962.
¨
[45] X.L. Luo, R.H. Crabtree, J. Am. Chem. Soc. 111 (1989) 2527.
[46] T.B. Marder, N.C. Norman, C.R. Rice, E.G. Robins, J. Chem. Soc.,
Chem. Commun. (1997) 53.
[77] C.Y. Peng, P.Y. Ayala, H.B. Schlegel, M.J. Frisch, J. Comput.
Chem. 17 (1996) 49.
[78] C.Y. Peng, H.B. Schlegel, Isr. J. Chem. 33 (1993) 449.