Mechanism of ketone hydrosilylation by Cu(I) catalysts: A theoretical study

Article abstract of DOI:10.1016/j.jorganchem.2009.08.017

The plausibility of the catalytic cycle suggested for the hydrosilylation of ketones by Cu(I) hydrides has been investigated by a theoretical DFT study. A model system made up of a CuH(PH₃)₂ catalyst, acetone and SiH₄ give

Full text of DOI:10.1016/j.jorganchem.2009.08.017

Journal of Organometallic Chemistry 694 (2009) 3943–3950
Contents lists available at ScienceDirect
Journal of Organometallic Chemistry
journal homepage: www.elsevier.com/locate/jorganchem
Mechanism of ketone hydrosilylation by Cu(I) catalysts: A theoretical study
*
Thomas Gathy, Daniel Peeters, Tom Leyssens
Laboratoire de Chimie Quantique, Université Catholique de Louvain, Place Louis Pasteur 1, B-1348 Louvain-la-Neuve, Belgium
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 1 July 2009
Received in revised form 3 August 2009
Accepted 7 August 2009
Available online 18 August 2009
The plausibility of the catalytic cycle suggested for the hydrosilylation of ketones by Cu(I) hydrides has
been investigated by a theoretical DFT study. A model system made up of a CuH(PH3)2 catalyst, acetone
and SiH4 gives us the necessary insight into the intrinsic reactivity of the system. This reactivity is con-
firmed, by introducing the more rigid CuH[C4H4(PH2)2] catalyst, as well as tetra-coordinated, CuH(PH3)3
and CuH(PH3)[C4H4(PH2)2] compounds. Computations show the activation of the copper fluoride pre-cat-
alyst, as well as both steps of the catalytic cycle to involve a 4 center metathesis transition state as sug-
gested in literature. These results show the reaction to be favored by the formation of a Van der Waals
complex resembling the transition states. The formation of these latter is induced by stabilizing electro-
static interactions between those atoms involved in the bond breaking and bond forming. Both steps of
the actual catalytic cycle show a free energy barrier of about 10 kcal/mol with respect to the isolated
reactants, hereby confirming the plausibility of the suggested cycle. We have found a substantial overall
exothermicity of the catalytic cycle of about 35 kcal/mol.
Keywords:
CuH(phosphine)2/3 complexes
Cu(I) catalysis
Hydrosilylation
Catalytic cycle
Theoretical study
Ó 2009 Elsevier B.V. All rights reserved.
1. Introduction
this system is air-sensitive, in situ preparations of CuH systems re-
mains attractive. As the goal of the current paper is to present a
Selective reduction of unsaturated double bonds through
homogenous catalysis, is of great interest in synthetic organic
chemistry. It was not until the ‘70s, that the first Rh(I) based cata-
lyst, used for reductive hydrosilylation of ketones and imines,
yielding, respectively, alcohols and amines were developed [1–3].
The softer reaction conditions of hydrosilylation, turned out to be
a major advantage over other reduction reactions, such as hydroge-
nation. The cost of Rh catalysts, however, formed a major stum-
bling-block for their large scale industrialization. Over the last
two decades alternative efficient asymmetric hydrosilylation reac-
tions using titanium [4–7], zinc [8,9], tin [10], titanocene [5,11],
manganese [12], and copper [11,13–17] were developed.
first computational mechanistic study of these systems, a system
was chosen for which the CuH species is created in situ, as hereby
one not only investigates the actual catalytic cycle, but also the
activation mechanism of the pre-catalyst. CuF2 systems were re-
ported by Carreira and co-worker [27], as well as Riant and co-
workers [28] as interesting precursors to copper hydride. These lat-
ter showed that a CuF2/Binap recipe used for the hydrosilylation of
ketones, leads to a CuH system, that is not only stable towards oxy-
gen, but shows accelerated reduction upon injection of oxygen into
the system [29]. These features make this type of system an inter-
esting candidate for a mechanistic study.
A scheme has been proposed to explain the observed reactivity.
This scheme involves the formation of a pre-catalyst through a two
step process, starting of with the transformation of CuF2 to a
FCu(PPh3)3 complex, which through ligand exchange will lead to
the final diphosphine ligand as presented in Scheme 1.
In 1988, Stryker was the first to introduce the use of copper hy-
dride complexes for the reduction of
a,b-unsaturated compounds
[18–20] using [CuHPPh3]6, as a stoichiometric reducing reagent.
This so called ‘‘Stryker’s reagent” calls for CuCl plus an equivalent
of NaOt-Bu to generate CuOt-Bu. To avoid this extremely air-sensi-
tive system advances have been made to form the Cu–H species
in situ [21,16,17,22]. By using silanes, as a stoichiometric hydride,
only catalytic quantities of the copper species are required (for a
review of CuH-catalyzed reactions see Refs. [14,15,18,31] and ref-
erences herein). Recently Lipshutz and Frieman [26] described a
chiral CuH species, which is exceptionally stable and can be stored
in solution if kept under argon. This so called ‘‘CuH in a bottle” can
be used directly without in situ preparation of a CuH species. As
A
r-bond metathesis between the copper fluoride and the
hydrosilylation reagent then expectedly generates the copper hy-
dride species (Scheme 2) [30].
This catalyst then enters the suggested catalytic cycle [23–25]
shown in Scheme 3. In this cycle, the copper hydride reacts with
the ketone to form a copper alkoxyde. Next, the copper alkoxyde
undergoes a
r-bond metathesis with the hydrosilylation agent to
regenerate the Cu–H complex and the silyl ether.
To verify the plausibility of the above mentioned cycle, we use
computational chemistry in this paper to examine energetic, elec-
tronic, and structural properties of the suggested cycle. Our calcu-
lations use two different models. First we consider, the CuH(PH3)2
* Corresponding author.
E-mail address: tom.leyssens@uclouvain.be (T. Leyssens).
0022-328X/$ - see front matter Ó 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2009.08.017

3944
T. Gathy et al. / Journal of Organometallic Chemistry 694 (2009) 3943–3950
icon atoms.2 This effect is no longer observed for the actual catalytic
MeOH reflux
CuF2 + 3.5 PPh3 + 0,5 H2O
FCu(PPh3)3 + 0.5 OPPh3 + HF
cycle (Tables 4 and 6). As the main objective of this paper is to study
the plausibility of the actual catalytic cycle, with the possibility of
transposing the basis set used to larger systems in order to study
the enantioselectivity, the smaller basis set was chosen throughout
this work.
P
P
P
Cu
F
FCu(PPh3)3
+
+ 3-n PPh3 n = 0,1
P
(PPh3)n
diphosphine
ligand
3. Results and discussion
3.1. Formation of the pre-catalyst
Scheme 1. Formation of the pre-catalyst.
The pre-catalyst is formed by placing a CuF2/PPh3 mixture at re-
flux in methanol. The structure of the formed CuF(PPh3)3 com-
pound has been determined experimentally by X-ray diffraction
of both the FCu(PPh3)3Á2MeoH [36] and FCu(PPh3)3Á2EtoH [37]
solvates.3
P
P
Cu
F
Cu
H
+ R3Si-H
+ R3Si-F
P
P
(PPh3)n
(PPh3)n
Table 1 compares some theoretical and experimental bond
lengths for these complexes.
Scheme 2. Catalyst formation.
Table 1 shows a shorter Cu–F bond compared to Cu–P bonds.
The discrepancies between theoretical and experimental bond
lengths, can be due to a multitude of effects, such as crystalline
packing effects, or the fact that calculations have been performed
in a gas phase model using the DFT level of theory.
Although Kagan and co-worker [38] showed that asymmetric
hydrosilylation of ketones is possible using a Cu hydrate complex
coordinated by monophosphine ligands, the low enantiomeric ex-
cess obtained, favored the use of diphosphine ligands. These latter
were furthermore shown to have a positive influence on the reac-
tion kinetics.
catalyst, which is convenient for a mechanistical study as it in-
volves fewer atoms than the large catalytic systems used experi-
mentally, but will give us the necessary insight into the intrinsic
reactivity of the system. Using smaller systems facilitates the iden-
tification of stationary points. This would become cumbersome
when using a true experimental system, without prior approxi-
mate knowledge of these points. Secondly, by introducing
CuH[C4H4(PH2)2] catalysts we strive towards a conformationally
more rigid system. Doing so, one can determine whether or not
the mechanism of the cycle studied is strongly affected by the nat-
ure of the diphosphine used. This information can guide further
computational studies on large experimental systems. As concern
might rise, whether or not the actual catalyst contains a supple-
mentary PH3 ligand, hence being tetra-coordinated, CuH(PH3)3
and CuH(PH3)[(C4H4(PH2)2] compounds are also considered.
The diphosphine complexes are usually obtained by ligand ex-
change as presented in scheme 1. As mentioned above, in this pa-
per we study the CuH(PH3)[(C4H4(PH2)2] complex to mimic these
substances. The free energy difference (
shows the overall process to be spontaneous.
D
G°) of À4.4 kcal/mol
3.2. Activation of the pre-catalyst
2. Computational details
The active species for the selective reduction of ketones to alco-
hols is a copper hydride. This later is formed through a bond
metathesis between the hydrosilane compound and the copper
fluoride pre-catalyst. As shown on Fig. 1, the formation of a bond
r
All structures were fully optimized using Becke’s three parame-
ter hybrid functional (B3LYP), as implemented in the GAUSSIAN 03
[32] series of programs. Optimized geometries are provided in
the Supplementary material. Force constants were determined
for the stationary points to characterize these latter, as well as to
determine their entropies and free energies, based on a statistical
thermodynamical treatment. Solvent effects are not yet included,
as the main purpose of this paper is to verify the plausibility of
the suggested mechanism. The Cu atom was described using an
effective core potential to represent all but the valence nd and
(n + 1)s and outer core ns and np electrons [33]. The latter were de-
scribed with a triple zeta contraction of the original double zeta ba-
sis set, this combination is referred to as the LANL2DZ basis set. All
non-metal atoms were described using the standard 6-31G(d,p)
basis set (with only the five spherical harmonic d functions). Atom-
ic charges were calculated using the NBO method [34]. As concern
may rise concerning the adequacy of the chosen basis set, calcula-
tions using a larger, all electron basis set were performed for the
CuX(PH3)2 model system.1 Results (Table 2) show a consistent shift
of about 5 kcal/mol for species involved in the transformation of the
pre-catalyst to the catalyst, with respect to isolated reactants, and is
due to a basis set superposition error involving the fluorine and sil-
r
between the copper and hydrogen atom, occurs through trans-
metalation [18,17,30] passing by a 4 center transition state.
For the tetra-coordinated complexes, two energetically and
structurally similar transition states are observed, which are con-
nected through a Van der Waals complex being a mere 0.5 kcal/
mol more stable than both transition states. This intermediate
should therefore be considered as a very labile complex, that will
not be observed experimentally and which origin can be related
to the almost planar energy surface around the transition states.
For simplicity the highest energy transition state will be discussed
hereafter.
To arrive at the 4 center transition state, the pre-catalyst forms
an energetically favored Van der Waals complex with the SiH4
hydrogenating agent. Table 2 shows the relative energy, enthalpy
2
The BSSE for the CuF(PH3)2/SiH4 and CuH(PH3)2/CH2O is a mere 1.5 kcal/mol
using the TZVP/6-31++G(d,p) basis set, but increases to 3.8 kcal/mol for the
CuH(PH3)2/CH2O and as high as 9.3 kcal/mol for the CuF(PH3)2/SiH4 complex using
the LANL2DZ/6-31G(d,p) basis set. This important BSSE for the CuF species can be
reduced when introducing diffuse basis functions on the F atom. The BSSE contributes
to an overestimation of the stabilization of these complexes with respect to isolated
reactants. Excluding this contribution, the Van der Waal complexes remain stabilized
with respect to isolated reactants, and the overall discussion therefore remains
unaltered.
1
For this system a TZVP [35] basis set was used for the Cu atom, and a 6-
31++G(d,p) basis set for all other atoms. Structures were optimized using this basis
set, and transition states checked for their validity.
3
Theoretical structures are given in Supplementary material.

T. Gathy et al. / Journal of Organometallic Chemistry 694 (2009) 3943–3950
3945
P
R3Si
O
Cu
H
O
P
(PPh3)n
*
R
R1
R
R1
R
*
R1
P
P
Cu
O
R3Si-H
(PPh3)n
Scheme 3. Catalytic cycle.
Table 1
transferred to the Cu atom, no significant increase in electron den-
sity is observed along the Cu–H axis at the transition state. The for-
mation of the copper hydride bond and breaking of the Si–H bond
Experimental X-ray distances (Å) for FCu(PPh3)3Á2MeoH and FCu(PPh3)3.2EtoH and
theoretical distances obtained for the FCu(PH3)3 complex.
Distance (Å)
FCu(PPh3)3Á2MeoH
FCu(PPh3)3Á2EtoH
FCu(PH3)3
is thus expected to be less advanced. The 4
r bond metathesis can
therefore be characterized as an asynchronous concerted reaction,
with the breaking of the Cu–F bond in favor of the formation of the
Si–F bond occurring slightly earlier than the transfer of the silicon
linked hydrogen to the cupper atom.
Cu–F
Cu–P
2.11
2.32
2.06
2.32
1.87
2.39
For the formation to take place a free energy barrier of about
6 kcal/mol and 9 kcal/mol4 has to be crossed for, respectively, the
tri- and tetra-coordinated copper reactants.5 The more important
barrier for the tetra-coordinated copper pre-catalysts, can be ex-
plained by an increased steric effect and slight differences in bonding
H
SiR3
LCuF + HSiR3
LCuH + FSiR3
LCu
F
to the Cu atom explained by an additional
effect of the phosphine ligand.
r donating and p acceptor
Fig. 1. Activation of the pre-catalyst.
The activation mechanism of the pre-catalyst is exothermic by
about 8 kcal/mol (
H° = À8.10 kcal/mol for the CuF(PH3)2) pre-cat-
D
alyst) and spontaneous as shown by the difference in free energy of
about À10 kcal/mol and À8 kcal/mol for, respectively, the tri- and
tetra-coordinated copper reactants (Table 2). The fluorine atom
seems essential to the activation of the pre-catalyst, as other cop-
per halogen complexes show little to no reduction of ketones to
alcohols. This observation could be explained by the less important
electronegativity of the other halogen atoms, leading to less stabi-
lized initial Van der Waals complexes due to decreased electro-
static interactions, but other factors such as bond strengths,
steric effects, etc. could also be of importance. Further work is
ongoing to understand these observations.
and free energy of the Van der Waals complexes and transition
states.
Table 2 shows, the formation of the initial Van der Waals com-
plex between the copper fluoride pre-catalyst and the SiH4 reac-
tant to be energetically favored, although not spontaneous. The
electrostatic interaction between the negatively charged fluorine
atom and the positively charged copper atom, as well as the nega-
tively charged hydrogen atom and the positively charged silicon
atom as shown in Table 3, explains the driving force for this ener-
getically favorable formation of the Van der Waals complex. The
alignment of both dipoles orientates the formation of the 4 center
transition state as shown on Fig. 2 for the CuF[C4H4(PH2)2] – SiH4
and CuF(PH3)2 – SiH4 complexes.
3.3. Catalytic cycle
As shown by Fig. 2 and Table 2 the transition states lie energet-
ically and structurally close to the Van der Waals complexes. As ex-
Once activated, the copper hydride enters the catalytic cycle.
The proposed mechanism for the hydrogenation of ketones by
these hydride catalysts [14,15,31] occurs through a two step cycle
as presented in Fig. 4. A first step concerns the formation of a cop-
pected for
a r bond metathesis, the transition states are
characterized by increased Cu–F and Si–H bond lengths, while
Cu–H and Si–F bonds become shorter.
Chemical reactions occur through reorganization of electron
density between different nuclei. Investigation of how this density
changes upon formation of the transition state therefore helps clar-
ify the reaction mechanism. Fig. 3 shows the differential density
between, respectively, the transition state, and the sum of both
reactants at transition state geometry. Red and blue zones indicate,
respectively, an increase and decrease in electron density. As
shown on this figure, the transition state is characterized by a
strong decrease of electron density around the copper atom, as
well as a strong increase in the F–Si area, which corresponds to
the breaking of the Cu fluoride bond, in favor of the formation of
the Si–F bond. Although one observes a polarization of the electron
density of the Si atom towards the hydrogen atom that will be
per alkoxyde, through a
r metathesis transition state similar to
that observed for the activation of the pre-catalyst. In a second
step, this alkoxyde reacts with a hydrosilane to yield a silylated
ether, and the reactivated catalyst. Once more, the suggested
mechanism passes through a four center transition state. The final
alcohol is generated by hydrolysis of the silylated ether. Due to the
fact that the alkoxyde complex is not observed experimentally the
reduction of the ketone is suggested to be the rate limiting step
[14].
4
Free energy barriers are reported with respect to isolated reactants.
Calculations were performed in the gas phase. As a consequence experimental
5
D
S° and DG° values are expected to be smaller.

3946
T. Gathy et al. / Journal of Organometallic Chemistry 694 (2009) 3943–3950
Table 2
Relative energy (DE°), enthalpy (DH°) and free energy (DG°) with respect to pre-catalyst and SiH4 reactants during the activation of the pre-catalyst. DE° (DH°, DG°) in kcal/mol.
(Values in italic are those for the larger basis set (TZVP on Cu and 6-31++G(d,p) on all other atoms.)
Pre-catalyst
CuF(PH3)2
Van der Waals complex
TS
Van der Waals complex
Products
À6.5 (À4.2; 2.7)
À2.8 (À1.4; 8.5)
À4.6 (À3.3; 3.64)
À6.3 (À4.4; 4.8)
À6.8 (À5.5; 4.2)
À4.6 (À3.6; 4.9)
0.1 (0.5; 11.4)
À3.6 (À3.3; 6.6)
À4.1 (À3.2; 8.9)
À3.4 (À3.3; 8.5)
À10.8 (À9.2; À5.1)
À5.2 (À4.7; 2.5)
À7.9 (À8.1; À10.7)
À2.3 (À3.0; À2.6)
À7.3 (À7.9; À10.0)
À7.7 (À8.4; À7.8)
À6.5 (À7.1; À7.4)
CuF[C4H4(PH2)2]
CuF(PH3)3
CuF(PH3)[C4H4(PH2)2]
À10.0 (À10.1; À2.0)
À9.4 (À8.2; À0.4)
À9.2 (À8.7; À1.0)
Table 3
Atomic charges on F, Cu, Si, and H atoms for the initial Van der Waals complexes.
Pre-catalyst
F
Cu
Si
H
CuF(PH3)2
CuF[C4H4(PH2)2]
CuF(PH3)3
À0.71
À0.72
À0.72
À0.72
0.72
0.72
0.72
0.74
0.85
0.67
0.75
0.77
À0.34
À0.19
À0.23
À0.23
CuF(PH3)[C4H4(PH2)2]
As for the activation of the pre-catalyst, a stabilizing Van der
Waals complex is observed between the catalyst and the reacting
ketone during the first step of the cycle. In this complex, the orien-
tation of the molecules favors the formation of the 4 center transi-
tion state. Table 4 shows the relative energies, enthalpies, as well
as free energies of the species involved in the first step of the cat-
alytic cycle.
Fig. 3. Density difference between transition state and sum of reactants at
transition state geometry. Red and blue zones indicate, respectively, an increase
and decrease in electron density. Isodensity surfaces given at 0.01 and 0.008 eÀ/V.
Plots were made using the GAUSSVIEW program [39]. (For interpretation of the
references to colour in this figure legend, the reader is referred to the web version of
this article.)
The results show the formation of the van de Waals complex to
be energetically favored by about 5 kcal/mol. The slightly less
important stabilization of the tetra-coordinated complexes could
once more be due to an increased steric hindrance, as well as to
changes in electronic density around the copper atom coming from
interaction with the added phosphine ligand.
As shown on Fig. 6, the Van der Waals complexes are structur-
ally close to the transition state activated complexes, hence
explaining the small energetic differences between both. At the
transition state the complexes are characterized by elongated, C–
O and Cu–H bonds, as well as reduced Cu–O and C–H bonds.
As for the activation of the pre-catalyst, the density differences
at the transition state (Fig. 5), point towards an asynchronous con-
certed reaction mechanism, with the Cu–H–C rearrangement being
more advanced than the formation of the Cu–O bond. Fig. 5 shows
a decrease in electron density along the Cu–H axis, as well as an in-
crease along the H–C axis. The electron density of the C@O double
bond is furthermore displaced towards a lone pair on the oxygen
atom. In a second instance, this electronic density on the oxygen
which is more important than the barrier observed for the activa-
tion of the pre-catalyst. It can therefore be assumed that the total
of the Cu–F species will have been transformed into Cu–H catalyt-
ical species at the onset of the ketone reduction. As for the activa-
tion of the pre-catalyst, the tetra-coordinated species are
characterized by a slightly more important free energy barrier. A
replacement of two PH3 ligands by a diphosphine ligand, slightly
lowers the free energy barrier. As shown on Fig. 6, an increase in
P–Cu–P bond angle, could release the tension of the cisoid diphos-
phine ligand, hereby stabilizing the transition state with respect to
reactants, and thus lower the free energy barrier. Such an effect of
changes in bite angle on rate and selectivity is common to diphos-
phine ligands [40].
The formation of the copper alkoxyde is furthermore strongly
exothermic (
(see Table 5).
D
H° = À26.3 kcal/mol for the CuH(PH3)2 molecule)
The second step of the catalytic cycle concerns the regeneration
of the catalyst, through the reaction of the Cu-alkoxyde species
with a silane molecule, yielding a silylated ether product. As for
will be transferred to the Cu atom to complete the
metathesis.
r bond
the previous steps, a
r metathesis reaction is suggested with a 4
center transition state. Computations indeed show a stabilizing
energetic interaction of about 5 kcal/mol (Table 6) between the al-
For the reaction to take place, a free energy barrier of about
10 kcal/mol with respect to the isolated reactants has to be crossed
H
1.49 (1.69)
H
H
H
Si
PH2
2 . 5 4 ( 1 . 7 7 )
H
1.57 (1.68)
H
3.07 (1.70)
Si
H
1 . 8 9 ( 1 . 7 7 )
Cu
F
H
1.82 (2.13)
H3P
1.87 (1.72)
F
Cu
2.01 (2.15)
PH2
H3P
Fig. 2. Alignment of both dipoles for the CuF[C4H4(PH2)2] – SiH4 and CuF(PH3)2 – SiH4. Bond lengths are given in Å for the Van der Waals complexes, as well as the transition
states (between brackets).

T. Gathy et al. / Journal of Organometallic Chemistry 694 (2009) 3943–3950
3947
O
PH2
H3C
SiH3
O
Cu
H
H
H
PH3
PH2
H
O
H
SiH3
O
H2P
H2P
H2P
H2P
H
PH3
+
PH3 +
Cu
Cu
H
CH3
H2P
H2P
CH3
Cu
O
SiH4
PH3
Fig. 4. The proposed mechanism for the hydrogenation of ketones by copper hydride catalysts.
Table 4
Relative energy (
and ketone reactants during the first step of the catalytic cycle.
D
E°), enthalpy (
D
H°) and free energy (
D
G°) with respect to catalyst
E° ( H°, G°) in
D
D
D
kcal/mol. (Values in italic are those for the larger basis set (TZVP on Cu and 6-
31++G(d,p) on all other atoms.)
Catalyst
Van der Waals
complex
TS
Products
CuH(PH3)2
À6.0
À3.1
À33.1
(À4.9; 4.6)
À6.2
(À1.8; 9.9)
À4.5
(À27.9; À18.4)
À32.1
(À4.7; 6.0)
À5.3
(À3.4; 8.2)
À1.8
(À27.2; À17.4)
À32.8
CuH[C4H4(PH2)2]
CuH(PH3)3
(À4.4; 5.1)
À4.8
(À0.3; 8.5)
À1.9
(À27.7; À15.9)
À31.2
Fig. 5. Density difference between transition state and sum of reactants at
transition state geometry. Red and blue zones indicate, respectively, an increase
and decrease in electron density. Isodensity surfaces given at 0.004 and 0.002 e-/V.
Plots were made using the GAUSSVIEW program [39]. (For interpretation of the
references to colour in this figure legend, the reader is referred to the web version of
this article.)
(À3.8; 5.1)
À4.8
(À0.8; 11.9)
À2.8
(À26.3; À16.2)
À32.1
CuH(PH3)[C4H4(PH2)2]
(À4.2; 5.7)
(À1.7; 10.2)
(À27.1; À16.8)
koxyde catalyst and the silane molecule, which is comparable to
the energetic stabilization observed for Van de Waals complexes
obtained during the activation of the pre-catalytic CuF species, as
well as that obtained during the first step of the catalytic cycle.
This interaction orientates the compounds once more to the 4 cen-
ter transition state, characterized by shortened Cu–H and Si–O
bonds, as well as elongated Si–H and Cu–O bonds, as shown on
Fig. 7.
Density differences (Fig. 8), once more indicate an asynchro-
nous transition state with the Cu–O bond breaking/O–Si bonding
being more advanced compared to the hydrogen transfer towards
the Cu atom. Although a polarization of electron density from the
silicon atom towards the transferable hydrogen is observed, no
density increase is yet observed along the H–Cu axis.
do not allow identifying the rate limiting step. Steric effects most
likely have an increasing importance, for more complex diphos-
phine ligands (e.g. the binap ligand), as well as with increasing ke-
tone size. The identification of the rate limiting step, will be
discussed in future contributions. Energetic barriers, are neverthe-
less substantially lower than the free energy barrier of 50.16 kcal/
mol calculated for the non catalyzed hydrosilylation of acetone.
The low energetic barriers calculated in this work show the plausi-
bility of the suggested catalytic cycle, and furthermore explain the
high turnover number observed for these types of catalyst.
The tetra-coordinated species are once more characterized by a
slightly more important free energy barrier. The second step of the
For the reaction to take place, a free energy barrier of about
9 kcal/mol and 12 kcal/mol with respect to isolated reactants for,
respectively, the tri- and tetra-coordinated complexes has to be
crossed, comparable to the barriers observed during the activation
of the pre-catalyst, as well as the first step of the catalytic cycle.
The comparable energy barriers for the two steps of the cata-
lytic cycle are shown in Fig. 9 for the CuH(PH3)2 complex. They
catalytic cycle shows an exothermicity of (
D
H°) À8.4 kcal/mol for
the CuH(PH3)2 molecule, comparable to the exothermicity ob-
served during the preparation of the catalysis from the CuF species.
The overall catalytic cycle shows an exothermicity of about
35 kcal/mol (
which is driven by the transformation of a
Si–H bond, into more stable C–H and O–Si bonds.
D
H° = À34.7 kcal/mol for the CuH(PH3)2 molecule),
p
C@O bond and a
r
r
r

3948
T. Gathy et al. / Journal of Organometallic Chemistry 694 (2009) 3943–3950
1.22 (1.27)
H
H
C
O
PH2
H
2 . 7 3 ( 1 . 7 8 )
1.23 (1.27)
C
O
H
2.77 (2.18)
Cu
H
H3P
84° (94°)
1.55 (1.64)
2.54 (2.29)
112° (124°)
H
Cu
1.56 (1.63)
H3P
PH2
Fig. 6. Alignment of both dipoles for the CuH[C4H4(PH2)2] – H2CO and CuH(PH3)2 – H2CO complexes. Bond lengths are given in Å for the Van der Waals complexes, as well as
the transition states (between brackets).
Table 5
Atomic charges on H, Cu, C, and O atoms for the Van der Waals complexes at the first
step of the catalytic cycle.
Catalyst
H
Cu
C
O
CuH(PH3)2
CuH[C4H4(PH2)2]
CuH(PH3)3
À0.49
À0.47
À0.55
À0.53
0.55
0.47
0.60
0.57
0.12
0.15
0.20
0.21
À0.59
À0.57
À0.57
À0.57
CuH(PH3)[C4H4(PH2)2]
Fig. 8. Density difference between transition state and sum of reactants at
transition state geometry. Red and blue zones indicate, respectively, an increase
and decrease in electron density. Isodensity surfaces given at 0.008 and 0.004 eÀ/V.
Plots were made using the GAUSSVIEW program [39]. (For interpretation of the
references to colour in this figure legend, the reader is referred to the web version of
this article.)
Table 6
Relative energy (
catalyst and silane reactants during the second step of the catalytic cycle.
G°) in kcal/mol. (Values in italic are those for the larger basis set (TZVP on Cu and 6-
31++G(d,p) on all other atoms.)
D
E°), enthalpy (
D
H°) and free energy (
D
G°) with respect to alkoxyde
E° ( H°,
D
D
D
Catalyst
Van der Waals
TS
Products
complex
wards these transition states, through the formation of energeti-
cally favored Van der Waals complexes, which energetically and
structurally resemble the transition states. The driving force be-
hind the formation of the Van der Waals complexes is a stabilizing
electrostatic interaction. The transition states are characterized by
an increase in bond length of those bonds that are breaking, and a
decrease in bond length of the newly formed bonds. Density differ-
CuH(PH3)2
À4.6 (À4.5; 7.4)
À4.1
À7.0
(À3.4; 8.7)
À3.2
(À7.4; À6.7)
À6.2
À3.8 (À2.5; 9.0)
À4.8 (À4.6; 6.2)
À5.8 (À5.0; 6.0)
À6.0 (À5.3; 5.8)
(À2.4; 10.1)
À4.1
(À6.5; À6.1)
À7.3
CuH[C4H4(PH2)2]
CuH(PH3)3
(À3.4; 8.2)
À2.8
(À7.5; À9.2)
À8.9
(À1.6; 12.0)
À4.3
(À9.0; À8.9)
À8.0
ences furthermore show that the
r metathesis reactions studied
CuH(PH3)[C4H4(PH2)2]
(À3.3; 7.0)
(À8.1; À8.3)
here should be described as asynchronous concerted mechanisms,
with the transfer from Cu linked atoms being more advanced than
the transfer towards this metal center.
For the reductive hydrosilylation of a acetone by SiH4 to take
place, a free energy barrier of about 10 kcal/mol with respect to
the isolated reactants has to be crossed. Comparable energy barri-
ers are obtained for the two successive steps of the actual catalytic
cycle. To identify the rate limiting step, larger ligands will have to
be introduced in order to account for the steric effects. Introduc-
tion of larger ligands, might also shown more important differ-
4. Conclusions
In this paper, we have shown the theoretical validity of the sug-
gested catalytic cycle for the hydrosilylation of ketones using Cu(I)
hydride catalysts. The activation of the catalyst from a copper-fluo-
ride complex, as well as both steps of the catalytic cycle involve a
metathesis 4 center transition state. The reactants are guided to-
r
H
H
1.57 (1.64)
H
H
Si
H
PH2
1.58 (1.64)
H
1 . 9 4 ( 1 . 8 6 )
Si
1 . 9 3 ( 1 . 8 6 )
1.90 (1.78)
H
H
H3P
Cu
O
1.91 (1.79)
1.98 (2.04)
O
Cu
2.01 (2.06)
H3P
CH3
PH2
CH3
Fig. 7. Alignment of both dipoles for the H3COCu[C4H4(PH2)2] – SiH4 and H3COCu(PH3)2 – SiH4. Bond lengths are given in Å for the Van der Waals complexes, as well as the
transition states (between brackets).

T. Gathy et al. / Journal of Organometallic Chemistry 694 (2009) 3943–3950
3949
Fig. 9. Calculated B3LYP/6-31GÃÃ potential energy surface (kcal/mol) for the catalytic cycle of hydrosilylation of formaldehyde by the HCu(PH3)2 model catalyst.
[8] O. Riant, N. Mostefa, J. Courmarcel, Synthesis 18 (2004) 2943.
[9] H. Mimoun, J.Y. de Saint Laumer, L. Giannini, R. Scopelliti, C. Floriani, J. Am.
Chem. Soc. 121 (1999) 6158.
[10] N. Lawrence, S.M. Bushell, Tetrahedron Lett. 41 (2000) 4507.
[11] J. Yun, S.L. Buchwald, J. Am. Chem. Soc. 121 (1999) 5640.
[12] S.U. Son, S.-J. Paik, I.S. Lee, Y.-A. Lee, Y.K. Chung, Organometallics 18 (1999)
4114.
[13] J. Wu, J.-X. Ji, A.S.C. Chan, Proc. Nat. Acad. Sci. 102 (2005) 3570.
[14] B.H. Lipshutz, W. Chrisman, K. Noson, J. Organomet. Chem. 624 (2001) 367.
[15] D.-W. Lee, J. Yun, Tetrahedron Lett. 45 (2004) 5415.
[16] B.H. Lipshutz, A. Lower, K. Noson, Org. Lett. 4 (2002) 4045.
[17] B.H. Lipshutz, C.C. Caires, P. Kuipers, W. Chrisman, Org. Lett. 5 (2003) 3085.
[18] W.S. Mahonney, D.M. Bretensky, J.M. Stryker, J. Organomet. Chem. 110 (1988)
291.
[19] J.-X. Chen, D.M. Daeuble, J.M. Stryker, Tetrahedron 56 (2000) 2789.
[20] J.-X. Chen, D.M. Daeuble, J.M. Stryker, Tetrahedron 56 (2000) 2153.
[21] B.H. Lipshutz, K. Noson, W. Chrisman, J. Am. Chem. Soc. 123 (2001) 12917.
[22] B.H. Lipshutz, K. Noson, W. Chrisman, A. Lower, J. Am. Chem. Soc. 125 (2001)
8779.
[23] B.H. Lipshutz, in: N. Krause (Ed.), Modern Organocopper Chemistry, Springer,
Weinheim, 2002, pp. 167–187. and references cited therein.
[24] S. Rendler, M. Oestreich, Angew. Chem., Int. Ed. 46 (2007) 498.
[25] C. Deutsch, N. Krause, B.H. Lipshutz, Chem. Rev. 108 (2008) 2916.
[26] B.H. Lipshutz, B.A. Frieman, Angew. Chem. 117 (2005) 6503;
B.H. Lipshutz, B.A. Frieman, Angew. Chem., Int. Ed. 44 (2005) 6345.
[27] (a) C. Czekelius, E.M. Carreira, Org. Process Res. Dev. 11 (2007) 633;
(b) C. Czekelius, E.M. Carreira, Org. Lett. 6 (2004) 4575;
ences between tri- and tetra-coordinated copper complexes, which
show little to no difference for the small systems studied here. As
mentioned previously, reduction by this system is accelerated by
oxygen, invoking a possible change in nature of the catalyst. These
points will be discussed in future work.
Finally, we have found a substantial overall exothermicity of the
catalytic cycle of about 35 kcal/mol, driven by the transformation
of a
p C@O bond and a r Si–H bond, into more stable r C–H and
r
O–Si bonds.
Acknowledgements
The authors are indebted to the Université Catholique de Lou-
vain (Thomas Gathy), as well as the Belgian National Fund for Sci-
entific Research (F.N.R.S.) for its financial support to this research
(Tom Leyssens is a Postdoctoral Researcher). They would also like
to thank the F.N.R.S for its support to access computational facili-
ties project (Project No. 2.4502.05).
Appendix A. Supplementary material
(c) C. Czekelius, E.M. Carreira, Angew. Chem. 115 (2003) 4941;
(d) C. Czekelius, E.M. Carreira, Angew. Chem., Int. Ed. 42 (2003) 4793.
[28] (a) S. Sirol, J. Courmarcel, N. Mostefa, O. Riant, Org. Lett. 3 (2001) 4111;
(b) J. Courmarcel, N. Mostefa, S. Sirol, S. Choppin, O. Riant, Israel J. Chem. 41
(2001) 231.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jorganchem.2009.08.017.
Reference
[29] N. Mostefa, S. Sirol, J. Courmarcel, O. Riant, Israel Synth. 349 (2007) 1797.
[30] H. Ito, T. Ishizuka, T. Okumura, H. Yamanaka, J. -I Tateiwa, M. Sonoda, A.
Hosomi, J. Organomet. Chem. 574 (1999) 102.
[31] H. Kaur, F. Zinn, E.D. Stevens, P.S. Nolan, Organometallics 23 (2004) 1157.
[32] GAUSSIAN 03, Revision C.02, M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria,
M.A. Robb, J.R. Cheeseman, J.A. Montgomery Jr., 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.
[1] I. Ojima, M. Nihonyanagi, Y. Nagai, J. Chem. Soc., Chem. Commun. (1972) 938.
[2] N. Langlois, T.P. Dang, H.B. Kagan, Tetrahedron Lett. 49 (1973) 4865.
[3] W. Dumont, J.-C. Poulin, H.B. Kagan, J. Am. Chem. Soc. 95 (1973) 8295.
[4] R. Halterman, T.M. Ramsey, Z. Chen, J. Org. Chem. 59 (1994) 2642.
[5] M.B. Carter, B. Schiott, A. Gutiérrez, S.L. Buchwald, J. Am. Chem. Soc. 116 (1994)
11667.
[6] J.F. Harrod, S. Xin, Can. J. Chem. 73 (1995) 999.
[7] H. Imma, M. Mori, T. Nakai, Synlett (1996) 1229.

3950
T. Gathy et al. / Journal of Organometallic Chemistry 694 (2009) 3943–3950
Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C.
(c) A.E. Reed, R.B. Weinstock, F. Weinhold, J. Chem. Phys. 83 (1985) 735;
(d) A.E. Reed, F. Weinhold, J. Chem. Phys. 83 (1985) 1736.
[35] (a) A. Schaefer, H. Horn, R. Ahlrichs, J. Chem. Phys. 97 (1992) 2571;
(b) A. Schaefer, C. Huber, R. Ahlrichs, J. Chem. Phys. 100 (1994) 5829.
[36] P.C. Healy, J.V. Hanna, J.D. Kildea, B.W. Skelton, A.H. White, Aust. J. Chem. 44
(1991) 427.
[37] D.J. Gulliver, W. Levason, M. Webster, Inorg. Chim. Acta 52 (1981) 153.
[38] F. Lagasse, H.B. Kagan, Chem. Pharm. Bull. 48 (2000) 315.
[39] T. Keith, J. Millam, GAUSSVIEW, Version 4.1.2, R. Dennington, II, Semichem, Inc.,
Shawnee, 2006.
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, GAUSSIAN, Inc., Wallingford CT, 2004.
[33] P.J. Hay, W.R. Wadt, J. Chem. Phys. 82 (1985) 720.
[34] (a) J.P. Foster, F. Weinhold, J. Am. Chem. Soc. 102 (1980) 7211;
(b) A.E. Reed, F. Weinhold, J. Chem. Phys. 78 (1983) 4066;
[40] P.W.N.M. Van Leeuwen, P.C.J. Kamer, J.N.H. Reek, P. Dierkes, Chem. Rev. 100
(2000) 2741.