Mechanistic insights into the iridium-catalyzed hydrosilylation of allyl compounds

Article abstract of DOI:10.1016/j.jcat.2015.09.003

The hydrosilylation of allyl compounds is very important for the industrial production of γ-substituted propylsilanes; however, it is also a process known to suffer from either substantial selectivity issues or short catalyst lifetimes. While there are reports on the platinum-catalyzed variant, this study is the first comprehensive work on the more recently employed iridium-catalyzed process. A combination of stoichiometric and catalytic experiments as well as reactions with isotope-labeled compounds is used to elucidate the critical parameters influencing the catalytic performance and to identify the main deactivation pathways. This report is intended to pave the way toward the optimization of current iridium-catalyzed processes and the design and synthesis of improved catalyst structures.

Full text of DOI:10.1016/j.jcat.2015.09.003

Journal of Catalysis 331 (2015) 203–209
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Mechanistic insights into the iridium-catalyzed hydrosilylation of allyl
compounds
a
a
b
a
a,
⇑
Korbinian Riener , Teresa K. Meister , Peter Gigler , Wolfgang A. Herrmann , Fritz E. Kühn
a
Chair of Inorganic Chemistry/Molecular Catalysis, Catalysis Research Center and Department of Chemistry, Technische Universität München, Lichtenbergstr. 4, 85747 Garching
bei München, Germany
Wacker Chemie AG, Consortium für elektrochemische Industrie, Zielstattstraße 20, 81379 München, Germany
b
a r t i c l e i n f o
a b s t r a c t
Article history:
The hydrosilylation of allyl compounds is very important for the industrial production of c-substituted
Received 3 August 2015
Revised 6 September 2015
Accepted 8 September 2015
propylsilanes; however, it is also a process known to suffer from either substantial selectivity issues or
short catalyst lifetimes. While there are reports on the platinum-catalyzed variant, this study is the first
comprehensive work on the more recently employed iridium-catalyzed process. A combination of
stoichiometric and catalytic experiments as well as reactions with isotope-labeled compounds is used
to elucidate the critical parameters influencing the catalytic performance and to identify the main
deactivation pathways. This report is intended to pave the way toward the optimization of current
iridium-catalyzed processes and the design and synthesis of improved catalyst structures.
Ó 2015 Elsevier Inc. All rights reserved.
Keywords:
Allyl compounds
Deactivation
Hydrosilylation
Iridium catalysis
Mechanism
Silanes
1
. Introduction
The hydrosilylation of C–C multiple bonds is the key catalytic
exhibit short catalyst lifetimes, resulting in low yields and the
necessity for high metal loadings, which is the main drawback
concerning their economic use [1,2]. The catalyst loadings can be
reduced by addition of co-catalysts such as cyclooctadiene (COD),
but a detailed mechanistic picture to understand the catalytic sys-
tem on a molecular level is still missing. To date, patent literature
is almost the sole source of information [11–23]. In this work, the
first comprehensive mechanistic study on the iridium-catalyzed
hydrosilylation of allyl compounds is presented, also revealing
the main deactivation pathways. The results from both stoichio-
metric and catalytic experiments will be helpful to pave the way
toward the design and implementation of more efficient catalyst
systems.
reaction to access organosilicon compounds that are of high indus-
trial importance for the production of organofunctionalized silanes
and silicones and are an integral part of polymers, crosslinkers and
adhesives [1–3]. Within this field, the hydrosilylation of allyl com-
pounds to yield c-substituted propylsilanes is of particular interest
due to the multiple functionalities of the obtained products. Their
efficient synthesis remains one of the current challenges in
hydrosilylation catalysis (Scheme 1) [1,2,4].
The common platinum-based hydrosilylation catalysts were
employed in these transformations and mechanistic studies have
revealed several reaction pathways, in some cases leading to by-
product formation and thus low selectivities [4–10]. This holds
true especially for the hydrosilylation of allyl chloride, which is
the most widely used allyl compound in industry [1,2]. Since these
selectivity issues are an intrinsic problem of platinum catalysis, it
appears worthwhile to focus on other systems. Iridium catalysis
was reported as an efficient tool to overcome the problems con-
cerning selectivity, with the most prominent catalyst precursor
2. Experimental methods
2.1. General remarks
All reactions were performed in an argon atmosphere using
standard Schlenk and glove box techniques. All solvents and
reagents were purified, dried and degassed in line with standard
2
being [{IrCl(COD)} ] [11–23]. However, these catalyst systems
purification techniques [24]. Allyl chloride-d
according to a literature procedure [4]. [{IrCl(COD)}
2
was synthesized
] was pur-
2
⇑
Corresponding author. Fax: +49 89 289 13473.
E-mail address: fritz.kuehn@ch.tum.de (F.E. Kühn).
chased from Sigma Aldrich and used as received. NMR spectra were
acquired on a Bruker Avance Ultrashield 400 MHz spectrometer. All
http://dx.doi.org/10.1016/j.jcat.2015.09.003
021-9517/Ó 2015 Elsevier Inc. All rights reserved.
0

2
04
K. Riener et al. / Journal of Catalysis 331 (2015) 203–209
2
.3.2. Reaction of [{IrCl(COD)}
In an NMR tube equipped with a J. Young valve, a mixture of
2.1 L Me SiHCl (27.6 mg, 292 mol, 10 equiv.) and 30.5
(33.6 mg, 292 mol, 10 equiv.) in 1.0 mL CD Cl is added
to a solution of 19.6 mg [{IrCl(COD)} ] (29.2 mol, 1.0 equiv.) in
is used as internal standard for NMR analysis.
2 2 2
] and a mixture of Me SiHCl/MeSiHCl
3
l
2
l
lL
MeSiHCl
2
l
2
2
2
l
1
2 2 3
.6 mL CD Cl . CHCl
Scheme 1. Ir- and Pt-catalyzed hydrosilylation of allyl compounds [1,2,4].
2
.3.3. Synthesis of [Ir(allyl)(Cl)
.11 g [{IrCl(COD)} ] (1.65 mmol, 1.0 equiv.) is dissolved in
20 mL CH Cl . 1.35 mL allyl chloride (1.26 g, 16.5 mmol, 10 equiv.)
is added, the reaction mixture is stirred for 15 min at room temper-
ature and all volatiles are removed in vacuum. [Ir(allyl)(Cl) (COD)]
is obtained in quantitative yield. H NMR (major isomer; 400 MHz,
CD Cl , 213 K): d = 1.66 (m , 1 H, CH -COD), 1.87 (m , 1 H, CH -COD),
.37–2.51 (m, 3 H, CH -COD), 2.62–2.77 (m, 1 H, CH -COD), 3.00
H,H = 8.2 Hz, 1 H, CH -COD), 3.22 (m , 1 H,
H,H = 8.2 Hz, 1 H, CH-COD), 3.49 (d,
2
(COD)]
1
2
1H and 13C chemical shifts are reported in parts per million (ppm)
2
2
relative to TMS, with the residual solvent peak serving as internal
reference [25]. Elemental analyses were performed in the microan-
alytical laboratory of TUM-CRC and FAB mass spectrometry was
carried out using a Finnigan MAT 90. A Jasco VT-550 photometer
was used to conduct the UV–Vis measurements in dichloro-
methane. TEM images were recorded at a nominal magnification
of 250,000 on a JEOL JEM 2011 electron microscope operated at
20 kV. The samples were prepared by adding a drop of the iridium
nanoparticles suspended in ethanol on copper grids covered with a
Quantifoil Multi A holey carbon film and a 2 nm carbon film on top.
2
1
2
2
c
2
c
2
2
2
2
2
3
(
dd,
J
H,H = 15.4 Hz,
J
2
c
3
CH
2
-COD), 3.42 (pseudo-q, J
3
3
J
H,H = 13.9 Hz, 1 H, anti CH
2
-allyl), 3.76 (d, JH,H = 10.1 Hz, 1 H, anti
H,H = 7.2 Hz, H, syn CH -allyl), 4.62
H,H = 8.2 Hz, H, CH-COD), 4.92 (pseudo-t,
H,H = 7.3 Hz, 1 H, CH-COD), 4.97 (d,
1
3
CH
2
-allyl), 4.32 (d,
J
1
2
3
(
pseudo-q,
J
J
1
3
3
J
H,H = 8.6 Hz, 1 H, syn
H,H = 7.3 Hz, 1 H, CH-COD), 5.67–5.77
m, 1 H, CH-allyl). C { H} NMR (major isomer; 101 MHz, CD Cl
-COD),
-allyl), 85.9 (CH-COD),
8.5 (CH-COD), 92.1 (CH-COD), 93.8 (CH-COD), 118.4 (CH-allyl).
3
2
CH -allyl), 5.07 (pseudo-t, J
(
2
3
1
3
1
2
2
,
2
2
.2. Catalytic reactions
13 K): d [ppm] = 25.5 (CH
6.8 (CH -COD), 43.6 (CH
2
-COD), 28.4 (CH
-allyl), 84.2 (CH
2 2
-COD), 35.3 (CH
2
2
2
.2.1. General protocol for hydrosilylation experiments
A mixture of [{IrCl(COD)} ], cyclooctadiene (COD) and the allyl
2
8
1
H NMR (minor isomer; 400 MHz, CD
2
Cl
2
, 213 K): d [ppm] = 1.26
compound (55.0 mmol, 1.0 equiv.) is stirred (500 rpm) at 40 °C
for 10 min in a reaction flask equipped with a reflux condenser.
At this temperature, the silane (55.0 mmol, 1.0 equiv.) is added
using a Landgraf LA-30 syringe pump, followed by the addition
of 1.28 mL mesitylene (1.10 g, 9.15 mmol, 6.0 equiv.) as internal
standard. The reaction mixture is further stirred for 10 min at
⁄
2
(
m
J
c
, 1 H, CH
2
-COD), 1.66 (1 H, CH
2
-COD), 2.18 (dd, JH,H = 14.7 Hz,
3
H,H = 8.9 Hz, 1 H, CH
2
-COD), 2.25–2.33 (m, 1 H, CH
2
-COD), 2.37–
⁄
⁄
2
.51 (1 H, CH
2
-COD), 2.62–2.77 (1 H, CH
H,H = 8.0 Hz, 1 H, CH
2
-COD), 2.88 (dd,
2
3
⁄
J
H,H = 15.6 Hz,
J
2
-COD), 2.97–3.03 (1 H,
3
⁄
CH
2
-COD), 4.24 (pseudo-q, JH,H = 8.0 Hz, 1 H, CH-COD), 4.31–4.34
3
(1 H, anti CH
2
-allyl), 4.49 (d,
.59–4.65 (1 H, CH-COD), 4.72 (m
JH,H = 4.5 Hz, 1 H, syn CH
2
-allyl),
4
0 °C and an aliquot is taken for NMR analysis.
⁄
⁄
4
c
, 1 H, CH-COD), 5.05–5.09
3
(2 H, CH-allyl, anti CH
2
-allyl), 5.17 (d,
JH,H = 4.5 Hz, 1 H, syn
3
3
13
CH
{
2
-allyl), 5.54 (dd,
J
H,H = 6.0 Hz,
J
H,H = 8.0 Hz, 1 H, CH-COD).
Cl , 213 K): d [ppm] = 25.9
-COD), 37.3 (CH -COD), 45.3
-allyl), 86.2 (CH-COD), 87.1 (CH-COD), 88.1 (CH-COD), 93.3
CH-COD), 94.8 (CH -allyl), 110.4 (CH-allyl). MS (FAB): m/z (%)
713 (51) [2 M – allyl – 2 Cl] , 377 (51) [M – Cl] , 335 (100)
C
2
.2.2. Hydrosilylation of allyl chloride-d
In an NMR tube equipped with a J. Young valve, 4.10 mg [{IrCl
COD)} ] (6.10 mol, 14.6 mol% Ir) is dissolved in 0.7 mL CD Cl
L allyl chloride-d (6.58 mg, 83.8 mol, 1.0 equiv.) is added
2 2
and Me SiHCl
1
H} NMR (minor isomer; 101 MHz, CD
-COD), 28.4 (CH -COD), 34.0 (CH
2
2
(
(
(
CH
CH
2
2
2
2
(
7
2
l
2
2
.
2
.00
l
2
l
2
to the reaction mixture and the NMR tube is shaken for 15 min, fol-
lowed by addition of 9.20 L Me SiHCl (7.91 mg, 83.6 mol,
.0 equiv.). Mesitylene is used as internal standard for NMR
analysis.
+
+
=
l
2
l
+
[M – allyl – Cl – H] . Elemental analysis (%): calc.: C, 32.04; H, 4.16.
1
⁄
found: C, 31.82; H, 4.25. Overlapping signals with major isomer.
2
.3.4. Reaction of [Ir(allyl)(Cl)
In an NMR tube equipped with a J. Young valve, 22.2 mg
Ir(allyl)(Cl) (COD)] (53.8 mol, 1.0 equiv.) is dissolved in 1.6 mL
and 47.4 L Me SiHCl (40.8 mg, 431 mol, 8.0 equiv.) in
Cl is added to the reaction mixture. CHCl is used as
2 2
(COD)] and Me SiHCl
2
2
.3. Reactivity studies
[
CD
1
2
l
2 2
.3.1. Reaction of [{IrCl(COD)} ] and Me SiHCl
2
Cl
2
l
2
l
NMR-scale: In an NMR tube equipped with a J. Young valve,
.10 mg [{IrCl(COD)} (12.1 mol, 1.0 equiv.) is dissolved in
.5 mL CD Cl and 21.4 Me SiHCl (18.3 mg, 193 mol,
6 equiv.) is added to the reaction mixture. CHCl is used as inter-
.0 mL CD
2
2
3
8
0
1
2
]
l
L
internal standard for NMR analysis.
2
2
l
2
l
3
nal standard for NMR analysis.
TEM: 153 mg [{IrCl(COD)} ] (228
in 10 mL CH Cl and 401 Me
6 equiv.) are added to the reaction mixture. After stirring for
0 min at room temperature, the volatiles are removed at reduced
3. Results and discussion
2
l
mol, 1.0 equiv.) is dissolved
2
2
l
L
2
SiHCl (345 mg, 3.65 mmol,
3.1. Influence of catalyst/co-catalyst concentration and the rate of
silane addition on catalytic performance
1
3
pressure and the solid is dried at 200 °C for 5 h. For TEM analysis,
the particles are suspended in ethanol.
The commonly used model system in iridium-catalyzed
hydrosilylation of allyl compounds is the reaction of allyl chloride
UV–Vis: 2.45 mg [{IrCl(COD)}
solved in 23 mL CH Cl and 1.5 mL of the solution is transferred
to a 1 cm UV–Vis cuvette for UV–Vis analysis. The solutions are
combined and 6.42 L Me SiHCl (5.52 mg, 58.3 mol, 16 equiv.)
is added. Again, 1.5 mL of the solution is transferred to a 1 cm
UV–Vis cuvette for UV–Vis analysis.
2
] (3.65
l
mol, 1.0 equiv.) is dis-
2 2
and Me SiHCl, with [{IrCl(COD)} ] and cyclooctadiene (COD) as
catalyst/co-catalyst combination, to yield chloro(3-chloropropyl)
dimethylsilane (Scheme 2) [11–22].
Typically, the silane (1.0 equiv.) is slowly added to a mixture of
the allyl compound (1.0 equiv.), the catalyst and the co-catalyst at
40 °C under neat conditions. Interestingly, three parameters differ
2
2
l
2
l

K. Riener et al. / Journal of Catalysis 331 (2015) 203–209
205
loadings and the quite substantial yield increase starting around
0 ppm. The results do not depend on the reaction time, i.e. if the
4
reaction mixture is further stirred at 40 °C after silane addition
and aliquots for analysis are taken every hour, the yields remain
unaltered.
2
Scheme 2. Hydrosilylation of allyl chloride and Me SiHCl to yield chloro(3-
chloropropyl)dimethylsilane [11–22].
The rate of silane addition was the influence factor studied next,
with an iridium concentration of 20 ppm and 10,000 ppm of COD
(Fig. 2).
significantly in the experimental sections of the patent reports
11–22]:
[
With an increasing rate of silane addition, the yields of the
desired hydrosilylation product drop drastically, in an inverse
exponential fashion. This strengthens the assumed relationship
between silane concentration and deactivation, with the latter
being caused by a local ‘‘overconcentration” of silane in solution.
The average values for the yields range from over 60% yield at
1
. Catalyst concentration: 40–360 ppm.
2
3
. Co-catalyst concentration: 80–2300 ppm.
. Rate of silane addition: 0.0028–0.011 equiv. min
À1
.
À1
À1
In order to better understand this model system, it appeared to
0.0038 equiv. min
to under 2% at 0.12 equiv. min . It seems
be necessary to reproduce the results from patent literature and
then elucidate the influence of the abovementioned parameters
on the catalytic performance. In a typical catalytic experiment,
worth noting that the standard deviations at very slow addition
rates are quite substantial. One reason might be the fact that these
data points are within the steepest regime of the curve, possibly
leading to relatively large differences in yield caused by even small
perturbations in the reaction setup. Within all experiments carried
out in this study, these are by far the most significant fluctuations.
When the average yield of chloro(3-chloropropyl)dimethylsilane is
plotted against the iridium concentration over the rate of silane
addition, the graph implies a linear relation (see Supplementary
Material). This again shows the importance of the iridium to silane
ratio.
[
4
{IrCl(COD)}
0 °C, before Me
2
], COD and allyl chloride were stirred for 10 min at
SiHCl was added over a defined period of time.
2
When the addition was completed, mesitylene was added as inter-
nal standard and after further 10 min, an aliquot was taken for
NMR analysis.
The first parameter of interest was the catalyst concentration,
2
i.e. the employed amount of [{IrCl(COD)} ]. With differing iridium
concentrations at a constant COD concentration (10,000 ppm) as
À1
well as a fixed rate of Me
00 mmol/h), the yield of the obtained
was determined and the results are shown in Fig. 1.
2
SiHCl addition (0.061 equiv. min
,
Finally, the influence of the amount of COD was studied. The
2
c
-substituted propylsilane
iridium concentration was again kept at 20 ppm, with a silane
À1
addition rate of 0.015 equiv. min
(50 mmol/h). Since the
While low amounts of iridium up to 20 ppm do not result in
significant product formation (3% yield), a considerable increase
can be observed at 40 ppm (20% yield) with an overall sigmoidal
progression up to 80 ppm with quantitative product yield. This
indicates substantial catalyst deactivation at low iridium concen-
trations in the reaction mixture, which might point toward a
dependence on the employed iridium to silane ratio. Another rea-
son might be impurities in the mixture, which deactivate a certain
amount of the catalyst leading to low yields at small iridium
reported COD concentrations in patent literature vary strongly,
the employed co-catalyst concentrations range over six orders of
magnitude and are shown in Fig. 3 with a logarithmic x-axis.
The overall curve shows a sigmoidal progression with the first
significant product yield increase at 10,000 ppm of COD (17% yield),
while up to 1,000 ppm, the yields are comparably low at 3–5%. A
further increase to 25% yield can be achieved at [COD] =
100,000 ppm; however, this seems unreasonable from an
Fig. 1. Average yield of chloro(3-chloropropyl)dimethylsilane as a function of the
Fig. 2. Average yield of chloro(3-chloropropyl)dimethylsilane as a function of the
rate of silane addition: iridium concentration of 20 ppm and COD concentration of
10,000 ppm; at least three catalytic runs per data point with the standard deviation
as ±y error bars (R2 = 0.997).
iridium concentration: COD concentration of 10,000 ppm and rate of Me
2
SiHCl
À1
addition at 0.061 equiv. min ; at least three catalytic runs per data point with the
2
standard deviation as ±y error bars (R = 0.999).

2
06
K. Riener et al. / Journal of Catalysis 331 (2015) 203–209
respectively. This experiment illustrates that the increased polarity
via chlorination of the silane significantly increases its ability to
reduce ligand and metal, leading to decomposition of the catalyst
as soon as the iridium to silane ratio is too small, i.e. either small
amounts of catalyst are present or the silane addition rate is high.
A combination of TEM and UV–Vis measurements was used to
verify the formation of iridium nanoparticles (Figs. 4 and 5). These
are common techniques in the characterization of iridium parti-
cles, which are usually obtained from reactions of various iridium
precursors and dihydrogen [29–32].
The iridium nanoparticles for TEM analysis were obtained from
the reaction solution by evaporation of all volatile compounds at
high temperature. For UV–Vis measurements, a spectrum of
[
{IrCl(COD)}
2
] was recorded and further monitored over time after
SiHCl. The absorption bands between 300 and
00 nm instantaneously disappear and only one strong band at
40 nm remains, which is characteristic for iridium nanoparticles
addition of Me
5
2
2
[
31,32].
The formation of particles queried whether the process is
homogeneously or heterogeneously catalyzed; thus, the iridium
nanoparticles were employed in catalysis. From now on, the
following conditions were used for the catalytic experiments: irid-
ium concentration of 100 ppm, COD concentration of 10,000 ppm
Fig. 3. Average yield of chloro(3-chloropropyl)dimethylsilane as a function of the
À1
and a rate of silane addition at 0.061 equiv. min (200 mmol/h).
COD concentration: iridium concentration of 20 ppm and rate of Me
at 0.015 equiv. min ; 100 ppm of COD was the first concentration which could be
experimentally realized, and the value at 1 ppm correlates to [COD] = 0 ppm; at
2
SiHCl addition
À1
As shown in Fig. 5, particles form immediately after [{IrCl(COD)}
exposed to Me SiHCl. Therefore, an ‘‘inverse” reaction protocol
was used to evaluate whether the iridium nanoparticles are active
in catalysis. [{IrCl(COD)} ], COD and Me SiHCl were mixed and
2
] is
2
least three catalytic runs per data point with the standard deviation as ±y error bars
2
(R
= 0.995).
2
2
allyl chloride was slowly added over time, i.e. the addition of the
substrates was inverted. With this procedure, the desired chloro(
experimental point of view, considering that the co-catalyst then
amounts to over 9% of the total reaction volume.
In summary, all three parameters investigated influence the
yield of chloro(3-chloropropyl)dimethylsilane significantly. It can
be concluded that the yield can be improved by the following:
3
-chloropropyl)dimethylsilane is obtained in slightly over 1% yield,
while the blind reaction without iridium does not yield any pro-
duct and the ‘‘normal” reaction protocol with the slow addition
1
2
3
. Increased catalyst loadings.
. Higher concentration of co-catalyst.
. Decreased rate of silane addition.
(
In cases 1 and 2 an optimum concentration can be found. With
an elevation beyond that value no further increase in yield takes
place.)
3.2. Reactions of the iridium precursor and the substrates
After investigating the main parameters influencing the cat-
alytic performance, further reactivity studies were conducted. A
reaction of [{IrCl(COD)} ] and Me SiHCl (16 equiv.) was used to
2
2
elucidate the main catalyst deactivation pathway at high silane
concentrations (Scheme 3).
2 2
The reaction of [{IrCl(COD)} ] and Me SiHCl yields about 90% of
cyclooctane in a formal hydrogenation reaction, iridium nanoparti-
cles and a mixture of different chloromethylsilanes, formed by
halide abstraction and Si–R/–X redistribution [26,27]. The ability
of the chlorinated silane to reduce the metal center and the olefin
ligand appears interesting, since it has been reported for reactions
with alkylhydrosilanes that the olefin ligands on halide-bridged
iridium complexes simply decoordinate and the metal stays in its
oxidation state [26–28]. In these reports, both the olefin and the
halide ligands are replaced by silyl and bridging silylene ligands,
Fig. 4. TEM image of iridium nanoparticles isolated from a reaction of [{IrCl
(
2 2
COD)} ] and Me SiHCl.
2 2
Scheme 3. Reaction of [{IrCl(COD)} ] and Me SiHCl.

K. Riener et al. / Journal of Catalysis 331 (2015) 203–209
207
Cl
Ir
Cl
Ir
Cl
Cl
10 : 3
Fig. 5. UV–Vis spectra of [{IrCl(COD)}
nanoparticles (solid line, first spectrum measured 3 min after silane addition) upon
reaction with Me SiHCl.
2
] (dotted line) and the formation of iridium
2
of silane to an iridium/COD/allyl compound mixture yields the
product in >99%. These results, in combination with the yield
dependence on the rate of silane addition support the assumption
that the reaction is mostly homogeneously catalyzed and the
contribution from the formed iridium nanoparticles is very minor at
best.
Fig. 6. 1H NMR spectra of [Ir(allyl)(Cl)
2
2 2
(COD)] in CD Cl at room temperature (top)
and at À60 °C (bottom, x-axis in ppm), two isomers with ratio at À60 °C in CD
2 2
Cl
depicted in the middle.
Table 1
Yield of
c-substituted product in the iridium-catalyzed hydrosilylation of allyl
After stoichiometric reactions with the silane, the catalyst
compounds; iridium concentration: 100 ppm, COD concentration: 10,000 ppm, rate of
precursor [{IrCl(COD)}
10 equiv.; Scheme 4).
The iridium dimer is opened by oxidative addition, resulting in
2
] was treated with excess allyl chloride
À1
silane addition: 0.061 equiv. min
.
(
Cl
OEt
quantitative formation of the respective allyl iridium(III)
compound. This is interesting due to reports on dimer opening
by simple coordination of the double bond rather than via C–Cl
activation [1,33]. However, it does not seem surprising since
similar reactivities have been observed for reactions with allyl
bromo/chloro derivatives [34,35]. With regard to 1H NMR spec-
Me
MeSiHCl
2
SiHCl
>99%
3%
75%
3%
2
allyl ethyl ether only results in 75% product formation. Compared
to allyl chloride (>99% selectivity), the selectivity is lower (85%).
About 12% of allyl ethyl ether are not converted, while the main
by-products are cis- and trans-isomerized starting material – i.e.
after double bond isomerization to the internal position adjacent
to the ether oxygen atom – with 7% and 2%, respectively. Similar
iridium-catalyzed isomerization processes have been reported
2
troscopy, the signals of [Ir(allyl)(Cl) (COD)] at room temperature
in CD Cl
2
2
are particularly broad but measurements at À60 °C
reveal a very defined spectrum and in combination with 2D NMR
experiments, two isomers in a ratio of 10:3 (different orientation
of the allyl ligand) can be identified (Fig. 6 and Supplementary
Material).
[
23,36]. Since internal olefins are more difficult to be transformed
in catalytic hydrosilylation [3], this might be the reason for the
lower product yield and selectivity, as the olefin then occupies a
coordination site on the metal without reacting with the silane;
thus, the silane can possibly deactivate the catalyst.
When the higher homologue MeSiHCl was employed as sub-
2
strate, only 3% of the respective dichloromethylsilanes are
obtained for both allyl chloride and allyl ethyl ether. Therefore,
3
.3. Expansion of the substrate scope
In order to understand the capabilities and limitations of the
iridium-catalyzed hydrosilylation of allyl compounds compared
to the platinum-based systems, the substrate scope was expanded
toward the higher chlorinated silane MeSiHCl
as the olefin (Table 1).
2
and allyl ethyl ether
in analogy to the reaction illustrated in Scheme 3, [{IrCl(COD)}
was exposed to an excess of an equimolar mixture of Me SiHCl/
MeSiHCl (10 equiv. each). Again, cyclooctane, iridium nanoparti-
cles and different chloromethylsilanes are yielded; however, the
consumption of MeSiHCl exceeds the one of Me SiHCl by a factor
2
]
While the hydrosilylation of allyl chloride and Me
2
SiHCl gives
2
the desired
c-substituted product in quantitative yield, the use of
2
2
2
of 1.4. This implies that a higher degree of chlorination facilitates
the reaction of the catalyst to the deactivation products, which is
in accord with the observations for the platinum-catalyzed
variant [4]. Oxidative addition of an increasingly electron-deficient
silane – i.e. higher degree of chlorination – to the olefin-bound,
2
Scheme 4. Reaction of [{IrCl(COD)} ] and allyl chloride.

2
08
K. Riener et al. / Journal of Catalysis 331 (2015) 203–209
electron-poor metal results in an extremely electron-deficient
metal species, which is prone to decomposition processes.
No chloro(3-chloropropyl)dimethylsilane is obtained, but
instead the linear/defunctionalized propylsilane is formed in
quantitative yield as well as the typical deactivation products.
These two findings (Schemes
5 and 6) prove that neither
3.4. Mechanistic aspects
[
Ir(allyl)(Cl) (COD)] nor any allyl intermediate are part of the cat-
2
alytic cycle itself and no product is formed from the allyl complex.
For the hydrosilylation of allyl ethyl ether, it appeared worth-
while studying the importance of dimer opening of the catalyst
To obtain a profound mechanistic picture for the iridium-
catalyzed hydrosilylation of allyl compounds, further experiments
were performed. In a stoichiometric experiment analogous to the
reaction shown in Scheme 4, [{IrCl(COD)}
2
] was treated with an
precursor [{IrCl(COD)}
the iridium dimer, [Ir(allyl)(Cl)
precursor in a reaction of allyl ethyl ether and Me
no difference in catalytic performance is observed with regard to
the employed iridium source, which suggests that Me SiHCl is also
able to open the dimeric compound by oxidative addition and
enables catalysis. Since the reaction is performed neatly in the allyl
compound and the silane is just added slowly, the ‘‘silane-opened”
iridium dimer is immediately coordinated by the alkene. No
additional deactivation through formal ligand hydrogenation and
iridium particle formation occurs as compared to starting from
2
]. Since allyl ethyl ether is not able to open
(COD)] was employed as catalyst
SiHCl. Almost
excess of allyl ethyl ether. However, no reaction occurs and the
dimeric iridium structure stays intact as indicated by 1H NMR
spectroscopy. This is distinctly different to the reactivity with
2
2
allyl chloride, which quantitatively yields [Ir(allyl)(Cl)
vide supra).
Due to this fact, it appeared interesting whether or not the two
2
(COD)]
2
(
substrates are transformed in mechanistically different catalytic
cycles or what the role of the allyl intermediate really is. Therefore,
deuterated allyl chloride was employed in a catalytic run with
Me
The reaction does not result in any scrambling/shift of the
deuterated methylene group which seems surprising, given the
fact that [Ir(allyl)(Cl) (COD)] is quantitatively formed after expo-
2
SiHCl (Scheme 5).
2
[Ir(allyl)(Cl) (COD)].
By combining the obtained data, the following mechanistic
picture for the iridium-catalyzed hydrosilylation of allyl
compounds is proposed (Scheme 7).
2
sure to allyl chloride (Scheme 4). Therefore, no allyl intermediate
is part of the catalytic cycle since allyl formation would lead to a
In the catalytic hydrosilylation of allyl chloride starting from
full scrambling of the d
stand whether the allyl ligand can in general lead to product
formation, isolated [Ir(allyl)(Cl) (COD)] was treated with an excess
of Me SiHCl (8.0 equiv.; Scheme 6).
2
-group in the product. To further under-
[{IrCl(COD)}
iridium dimer to [Ir(allyl)(Cl)
cleavage of the allyl ligand with two equivalents of Me
2
], the first step is the quantitative conversion of the
(COD)] (IAC), which is followed by
SiHCl to
(IIAC). The resulting [IrCl(COD)] is
coordinated by the olefin (III) and oxidative addition of the silane
IV) yields an iridium(III) intermediate. The coordinated alkene is
2
2
2
n
2 2 2
form Me SiCl Pr and Me SiCl
2
(
converted to an alkyl ligand upon hydride transfer (V) which ulti-
mately yields the product after reductive elimination (VI) and
regenerates the catalyst in its resting state ([IrCl(COD)]). For allyl
ethyl ether as substrate, only the first two reaction steps are differ-
ent: The iridium dimer is opened via oxidative addition of the
2 2
Scheme 5. Catalytic run with allyl chloride-d and Me SiHCl.
2 2
Scheme 6. Reaction of [Ir(allyl)(Cl) (COD)] and Me SiHCl.
Scheme 7. Proposed mechanistic picture for the iridium-catalyzed hydrosilylation of allyl compounds.

K. Riener et al. / Journal of Catalysis 331 (2015) 203–209
209
silane (IAE) and subsequent coordination of the alkene (IIAE) is nec-
References
essary to enter the catalytic cycle.
[
[
[
1] D. Troegel, J. Stohrer, Coord. Chem. Rev. 255 (2011) 1440–1459.
2] Y. Nakajima, S. Shimada, RSC Adv. 5 (2015) 20603–20616.
3] B. Marciniec, Hydrosilylation: A Comprehensive Review on Recent Advances,
Springer, London, 2008.
When the local silane concentration during catalysis becomes
too high, i.e. the iridium to silane ratio is too low, the respective
iridium intermediates are prone to leaving the catalytic cycle upon
reacting to iridium nanoparticles and cyclooctane. Exemplarily,
when [IrCl(COD)] reacts with silane, the silyl hydrido complex
[
[
4] P. Gigler, M. Drees, K. Riener, B. Bechlars, W.A. Herrmann, F.E. Kühn, J. Catal.
295 (2012) 1–14.
5] B. Marciniec, H. Maciejewski, W. Duczmal, R. Fiedorow, D. Kity n´ ski, Appl.
[
2
IrClH(SiMe Cl)(COD)] is formed, which is known to be an interme-
Organomet. Chem. 17 (2003) 127–134.
diate in the catalyst deactivation pathway (Scheme 3). This deacti-
vation can be prevented by application of an excess of COD (Fig. 3).
This is most probably caused by coordination and replacement of
partially hydrogenated ligand, keeping the active catalyst in
solution.
[6] M. Jankowiak, H. Maciejewski, J. Gulinski, J. Organomet. Chem. 690 (2005)
478–4487.
4
[
7] Z.V. Belyakova, E.A. Chernyshev, P.A. Storozhenko, S.P. Knyazev, G.N.
Turkel’taub, E.V. Parshina, A.V. Kisin, Russ. J. Gen. Chem. 76 (2006) 925–930.
[8] A. Wawrzy n´ czak, M. Dutkiewicz, J. Guli n´ ski, H. Maciejewski, B. Marciniec, R.
Fiedorow, Catal. Today 169 (2011) 69–74.
[
9] M. Igarashi, T. Kobayashi, K. Sato, W. Ando, T. Matsumoto, S. Shimada, M. Hara,
H. Uchida, J. Organomet. Chem. 725 (2013) 54–59.
[
[
10] M. Igarashi, T. Matsumoto, T. Kobayashi, K. Sato, W. Ando, S. Shimada, M. Hara,
H. Uchida, J. Organomet. Chem. 749 (2014) 421–427.
11] A. Bauer, O. Schäfer, M. Kriegbaum, L. Brader, B. Pachaly, V. Frey, US6388119,
2002 (to Consortium für elektrochemische Industrie GmbH).
4
. Conclusion
For the first time a comprehensive mechanistic understanding
of the iridium-catalyzed hydrosilylation of allyl compounds can
be presented. To achieve this, a combination of stoichiometric
and catalytic experiments has been used. The main reason for
the short catalyst lifetimes is found to be the formal hydrogenation
[12] Y. Tonomura, T. Kubota, M. Endo, US6359161, 2002 (to Shin-Etsu Chemical Co.,
Ltd.).
[
13] A. Bauer, L. Brader, V. Frey, M. Kriegbaum, B. Pachaly, O. Schäfer, EP1201671B1,
003 (to Consortium für elektrochemische Industrie GmbH).
[14] F. Baumann, M. Hoffmann, DE102004052424A1, 2006 (to Wacker Chemie AG).
2
[
15] Y. Tonomura, T. Kubota, M. Endo, EP1156052B1, 2006 (to Shin-Etsu Chemical
Co., Ltd.).
2
of the cyclooctadiene (COD) ligands on [{IrCl(COD)} ] and the for-
mation of iridium nanoparticles at elevated silane concentrations.
Possible solutions regarding the reaction setup are a slower rate
of silane addition and an increased concentration of cyclooctadiene
in the short term. For further fundamental improvements, we are
currently synthesizing and employing new catalyst motifs which
should pave the way toward a more generally applicable iridium-
catalyzed hydrosilylation of allyl compounds; especially focusing
on allyl chloride, the most widely used allyl compound in industry.
Since there are reports on iridium complexes bearing olefin ligands
which are not prone to hydrogenation, the concept is to mimic the
[
16] F. Baumann, DE102005030581A1, 2007 (to Wacker Chemie AG).
[17] T. Kornek, A. Bauer, D. Senden, US7208618, 2007 (to Wacker Chemie AG).
[
18] K. Ramdani, B. Vogin, N. Guennouni, EP1515977B1, 2007 (to Rhodia Chimie
SA).
[
19] K. Ramdani, N. Guennouni, B. Vogin, EP1637534B1, 2009 (to Rhodia Chimie).
[20] K. Ramdani, B. Vogin, N. Guennouni, US7655813, 2010 (to Rhodia Chimie).
[
[
[
21] K. Ramdani, B. Vogin, N. Guennount, US7659418, 2010 (to Rhodia Operations).
22] N. Guennouni, J.-C. Galland, US7884225, 2011 (to Rhodia Chimie).
23] M. Igarashi, T. Matsumoto, T. Kobayashi, K. Sato, W. Ando, S. Shimada, M. Hara,
H. Uchida, J. Organomet. Chem. 752 (2014) 141–146.
24] W.L.F. Armarego, C.L.L. Chai, Purification of Laboratory Chemicals,
Butterworth-Heinemann, Oxford, 2013.
[
[
25] 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.
electronic structure of [{IrCl(COD)}
via formal hydrogenation, i.e. hydrogen transfer from R
2
] while preventing deactivation
Si–H to the
[26] C. Cheng, M. Brookhart, J. Am. Chem. Soc. 134 (2012) 11304–11307.
3
[
27] S. Park, B.G. Kim, I. Göttker-Schnetmann, M. Brookhart, ACS Catal. 2 (2012)
07–316.
28] C. Cheng, M. Brookhart, Angew. Chem. Int. Ed. 51 (2012) 9422–9424.
alkene. Ultimately, the combination of iridium- and platinum-
catalyzed processes in the hydrosilylation of allyl compounds will
be complementary for a broad range of substrates. While the latter
process is fairly established for allyl acetate, ethers and amines,
iridium-catalysis is superior in the hydrosilylation of allyl chloride
due to the intrinsic selectivity issues of platinum catalysis.
3
[
[29] J. Dupont, G.S. Fonseca, A.P. Umpierre, P.F.P. Fichtner, S.R. Teixeira, J. Am.
Chem. Soc. 124 (2002) 4228–4229.
[
30] G.S. Fonseca, A.P. Umpierre, P.F.P. Fichtner, S.R. Teixeira, J. Dupont, Chem.-Eur.
J. 9 (2003) 3263–3269.
[31] V. Mévellec, A. Roucoux, E. Ramirez, K. Philippot, B. Chaudret, Adv. Synth.
Catal. 346 (2004) 72–76.
[
[
[
32] R. Redón, F. Ramírez-Crescencio, A.L. Fernández-Osorio, J. Nanopart. Res. 13
2011) 5959–5965.
33] G. Pannetier, P. Fougeroux, R. Bonnaire, J. Organomet. Chem. 38 (1972) 421–
25.
34] B. Bartels, C. García-Yebra, F. Rominger, G. Helmchen, Eur. J. Inorg. Chem. 2002
2002) 2569–2586.
[35] C. García-Yebra, J.P. Janssen, F. Rominger, G. Helmchen, Organometallics 23
2004) 5459–5470.
36] T. Ohmura, Y. Yamamoto, N. Miyaura, Organometallics 18 (1998) 413–416.
Acknowledgments
(
4
Support from the Fonds der Chemischen Industrie (Ph.D. grants
to T.K.M. and K.R.) and the TUM Graduate School is gratefully
acknowledged. Further, we thank Dr. Jürgen Stohrer for helpful
discussions.
(
(
[
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.2015.09.003.