α-CAM mechanisms for the hydrogenation of alkenes by cis- and trans- disilametallacyclic carbonyl complexes (M = Fe, Ru, Os): Experimental and theoretical studies

Article abstract of DOI:10.1246/bcsj.20170004

The hydrogenation of alkenes catalyzed by disilametallacyclic carbonyl complexes of iron, ruthenium or osmium was studied experimentally and theoretically. The disilaruthenacycle 2 with two CO ligands in the trans-configuration was prepared, characterized, and its ability to catalyze hydrogenation was studied. Similar to the corresponding iron analogue 1 in which the CO ligands are in the cis-configuration, 2 contains a H2MSi4 core with SiHSi SISHA (secondary interaction of silicon and hydrogen atoms) and catalyzed the hydrogenation of several alkenes under mild conditions. DFT calculations of 1 and 2 with cis- and trans-CO configurations (cis-1, trans-1, cis-2 and trans-2) revealed that the mechanism of ethylene hydrogenation comprises three catalytic cycles, and a key step involves the H-H bond of H2 being activated by an M-Si bond through oxidative hydrogen migration. These mechanisms are a variety of α-CAM (complex-assisted metathesis) mechanisms. Further calculations suggest that these catalytic cycles can apply to the catalytic hydrogenation of ethylene by osmium analogues of 1 and 2 (cis-3 and trans-3). Some of the elementary reactions in the cycles are dependent on the metal, and the osmium complexes show different performance from the iron and ruthenium analogues due to the characteristic natures of the third-row transition metals.

Full text of DOI:10.1246/bcsj.20170004

Advance Publication Cover Page

-CAM Mechanisms for the Hydrogenation of Alkenes by cis- and
trans-Disilametallacyclic Carbonyl Complexes (M = Fe, Ru, Os):
Experimental and Theoretical Studies
Konoka Hoshi, Atsushi Tahara, Yusuke Sunada, Hironori Tsutsumi, Ryoko Inoue,
Hiromasa Tanaka, Yoshihito Shiota, Kazunari Yoshizawa, and Hideo Nagashima*
Advance Publication on the web February 13, 2017
doi:10.1246/bcsj.20170004
©
2017 The Chemical Society of Japan


-CAM Mechanisms for the Hydrogenation of Alkenes by cis- and
trans-Disilametallacyclic Carbonyl Complexes (M = Fe, Ru, Os): Experimental and
Theoretical Studies
1
2,3
2,3,4
1
1
2
Konoka Hoshi, Atsushi Tahara, Yusuke Sunada,
Hironori Tsutsumi, Ryoko Inoue, Hiromasa Tanaka,
2
2
1,2,3
Yoshihito Shiota, Kazunari Yoshizawa, and Hideo Nagashima*
1
2
Graduate School of Engineering Sciences, Institute for Materials Chemistry and Engineering, Kyushu University, Kasugakoen 6-1,
Kasuga-shi, Fukuoka, 816-8580, JAPAN.
3
CREST, Japan Science and Technology Agency, Gobancho 7, Chiyoda-ku, Tokyo, 102-0076, JAPAN.
4
Present Address: Institute of Industrial Science, the University of Tokyo, Komaba 4-6-1, Meguro-ku, Tokyo, 153-8505, JAPAN.
E-mail: nagasima@cm.kyushu-u.ac.jp
Hideo Nagashima
Ph.D. Tokyo Institute of Technology (1982) supervised by Professor Jiro Tsuji
Assistant prof., senior lecturer, associate prof. of Toyohashi University of Technology (1982-1997)
Research associate, University of Wisconsin, Madison, with Professor Charles P. Casey (1985-1986)
Professor of Kyushu University (1997-present)
Abstract
The hydrogenation of alkenes catalyzed by
disilametallacyclic carbonyl complexes of iron, ruthenium or
osmium was studied experimentally and theoretically. The
disilaruthenacycle 2 with two CO ligands in the trans-
configuration was prepared, characterized, and its ability to
catalyze hydrogenation was studied. Similar to the
corresponding iron analogue 1 in which the CO ligands are in
the cis-configuration, 2 contains a H MSi core with SiHSi
2
4
SISHA (secondary interaction of silicon and hydrogen atoms)
and catalyzed the hydrogenation of several alkenes under mild
conditions. DFT calculations of 1 and 2 with cis- and trans-CO
configurations (cis-1, trans-1, cis-2 and trans-2) revealed that
the mechanism of ethylene hydrogenation comprises three
catalytic cycles, and a key step involves the H-H bond of H2
being activated by an M-Si bond through oxidative hydrogen
migration. These mechanisms are a variety of -CAM
Scheme 1. Iron and ruthenium disilametallacyclic compounds,
and 2, supported by two CO ligands. The CO ligands in 1 are
in cis-configuration, whereas those of 2 are in trans.
1
(
-complex-assisted
metathesis)
mechanisms.
Further
calculations suggest that these catalytic cycles can apply to the
catalytic hydrogenation of ethylene by osmium analogues of 1
and 2 (cis-3 and trans-3). Some of the elementary reactions in
the cycles are dependent on the metal, and the osmium
complexes show different performance from the iron and
ruthenium analogues due to the characteristic natures of the
third-row transition metals.
Scheme 2. Schematic view of hydrogenation of ethylene
catalyzed by 1 [M = Fe(CO) ]: step (a) and step (c); -bond
metathesis (oxidative hydrogen migration); step (b) insertion of
2
1
. Introduction
The catalytic hydrogenation of alkenes is one of the most
fundamental reactions in synthetic organic chemistry.
1
CH
=CH into the M-H bond and subsequent formation of
2 2
A

-agostic interaction.
number of complexes composed of noble metals and ligands
have been developed to realize this reaction, leading to the
efficient hydrogenation of various functionalized and
unfunctionalized substituted alkenes. Among these catalysts,
disilaferracyclic compound 1 was recently reported by our
research group as a unique catalyst that promotes the
hydrogenation of various alkenes even at room temperature
under 1 atm of H . A number of ruthenium complexes are
known to exhibit excellent catalytic activity towards the
3
hydrogenation of alkenes, whereas active iron catalysts are
4
rare. Examples of iron catalysts for which the structures of the
catalyst precursors are well-defined are limited to a tridentate
4
a
phosphoraborate iron complex reported by Peters, and
4
b
bis(imino)pyridine Fe(0) complexes described by Chirik. In
addition, Fe(CO) shows catalytic activity under mild
conditions when the reaction solution is continuously
2
5
2
4
g
photo-irradiated. The catalytic activity of 1 is higher than that
1

of most of these conventional iron catalysts.
searched for reasonable reaction pathways for catalytic
hydrogenation using DFT calculations. The results show that
all these compounds have similar structures, including an
H MSi core with a SiHSi SISHA (secondary interaction
of silicon and hydrogen atoms), and that the hydrogenation of
ethylene occurs via catalytic cycles involving -CAM
mechanisms. Further calculations predicted the structures of the
osmium homologues of cis-2 and trans-2, and their catalytic
activity towards the hydrogenation of alkenes was found to
differ slightly from those of the corresponding iron and
ruthenium complexes.
The good catalytic performance of 1 is ascribed to a
mechanism which differs from that of conventional catalytic
cycles. The transition metal-catalyzed hydrogenation of alkenes
is generally explained by catalytic cycles involving the
2
4
oxidative addition of H by the metal center, the insertion of a
2
C=C bond between the resulting M-H bond moieties, and the
subsequent reductive elimination of alkanes to regenerate the
1
initial species. DFT calculations clearly demonstrate the key
feature of hydrogenation catalyzed by 1, in which activation of
the H-H bond of molecular hydrogen does not occur at the iron
center but rather through the assistance of a Fe-Si bond. As
5
outlined in Scheme 2, the catalytic cycle is initiated by the
2. Experimental
disilaferracyclic intermediate X, in which both ethylene and H
are bonded to the iron center in  -coordination mode. The
2.1 General. Manipulation of air and moisture sensitive
compounds was carried out under a dry nitrogen atmosphere
using Schlenk tubes connected to a high-vacuum line or in a
glove box filled with dry nitrogen. All solvents were distilled
over appropriate drying reagents prior to use (toluene, pentane,
2
2
intermediate
X
is formed by ligand substitution of
bis(dimethylsilyl)benzene (BDSB) by ethylene and H . While
2
the oxidative addition of H2 to the iron center of X must
overcome a high energy barrier and is therefore impossible,
1
13
29
C D ; Ph CO/Na). H-, C-, and Si-NMR spectra were
6 6 2

-bond metathesis of late transition metals (oxidative hydrogen
recorded on a JEOL Lambda 400 or a Lambda 600
spectrometer at ambient temperature. H-, C-, and Si-NMR
6
2
1
13
29
migration ) easily occurs between the  -H moiety and a Fe-Si
bond to generate a Fe-H and an  -Si-H species [step (a) in
Scheme 2]. The resulting Fe-H species migrates to a carbon of
the coordinated ethylene to form a Fe(-agostic ethyl)( -Si-H)
species, which corresponds to insertion of a C=C bond between
the atoms of the Fe-H bond [step (b)]. This is followed by
Si-H/Fe-C oxidative hydrogen migration [step (c)] to form the
ethane on the iron. It is noteworthy that both the catalyst
precursor 1 and all species in the catalytic cycle have
octahedral geometry, which is advantageous in promoting steps
2
2
chemical shifts (δ values) were given in ppm relative to the
1
13
29
solvent signal ( H, C) or standard resonances ( Si: external
tetramethylsilane). Elemental analysis was performed using a
Perkin Elmer 2400II/CHN analyzer, and IR spectra were
2
recorded by
compounds,
a
JASCO FT/IR-550 spectrometer. Two
9
1,2-bis(dimethylsilyl)benzene
and
4
10
( -C H )Ru(CO) were synthesized by a method reported in
6
8
3
the literature. Alkenes were purchased from Tokyo Chemical
Industries Co., Ltd., and were used after purification by
distillation.
(a) ~ (c) by minimizing the motions of the hydrogen and
4
carbon atoms during the catalytic cycle. The strong -acceptor
2.2 Synthesis of 2. In a 50 mL schlenk tube ( -C H )Ru(CO)
6
8
3
ligand (CO) and -donor ligand (SiR ) result in all species in
the catalytic cycle being in the low spin state, similar to the
species in precious-metal-catalyzed hydrogenation.
(500 mg, 1.77 mmol) and 1,2-bis(dimethylsilyl)benzene (695
mg, 3.54 mmol) were dissolved in pentane (40 mL). After
mixing this solution for 24 h at room temperature under
irradiation with a high-pressure mercury lamp, the solvent was
evaporated in vacuo. The crude product was dissolved in
pentane (40 mL) and centrifuged to remove the small amount
of insoluble materials. The supernatant was collected,
3
Reactions catalyzed by iron have received much attention
from synthetic chemists as environmentally benign chemical
processes and as substitutes for noble metal-catalyzed
7
reactions. The design of catalyst 1 allows the catalysis of
o
organic reactions by a well-defined catalyst precursor. It should
be noted, however, that activation of molecular hydrogen by a
Fe-Si bond can be extended to that by other transition metal-Si
bonds. In 2007, Perutz and Sabo-Etienne proposed -CAM
mechanisms (-complex-assisted metathesis mechanisms), in
which H-H, Si-H, B-H, and C-H bonds could be activated by
M-E (M = various transition metals, E = main group species
such as SiR , CR , BR , or H) bonds by -bond metathesis, and
concentrated to ca. 20 mL, and cooled at –35 C to give 2 as
1
colorless crystals (400 mg, 0.74 mmol, 39%). H NMR (600
MHz, r.t., C D ):  = –7.33 (s, 2H, H-Si, with a satellite signal
6
6
2
9
due to the coupling with Si, JSi-H = 16.8 Hz), 0.76 (s, 24H,
SiMe ), 7.23-7.29 (m, 4H, C H ) 7.54-7.61 (m, 4H, C H ).
2
6
4
6
4
1
3
1
C{ H} NMR (150 MHz, r.t., C D ):  = 8.6 (s, SiMe ), 128.9,
131.5, 152.5 (s, C H ), 198.2. (CO). Si{ H} NMR (119 MHz,
6
6
2
2
9
1
6
4
r.t., C D );  = 30.1 (s, SiMe ). IR (KBr, pellet):  = 1956
3
3
2
6
6
2
H-Si
-
1
-1
a possible catalytic cycle for ethylene hydrogenation by
cm ,  = 1980 cm . Anal calcd for C H O RuSi ; C 48.34,
CO 22 34 2 4
8

-CAM mechanisms was depicted in their micro-review. Our
H 6.29; found: C 48.58, H 6.30.
hydrogenation mechanism can be categorized as a -CAM
mechanism, although the reaction pathways of the catalytic
cycle proposed by Perutz and Sabo-Etienne are somewhat
different from the cycle catalyzed by 1. In other words,
disilametallacyclic compounds analogous to 1 and containing
any transition metal might act as a hydrogenation catalyst.
Herein we report the preparation and characterization of
the ruthenium analogue of 1 and its application to the catalytic
hydrogenation of alkenes. The ligand arrangement of the
2.3 Crystallography. X-ray crystallography was performed on
Rigaku Saturn CCD area detector with graphite
a
monochromated Mo-Kα radiation (λ=0.71075 Ǻ). The data
were collected at 123(2) K using ω scan in the θ range of 3.29
≤  ≤ 27.44 deg for 2. The data obtained were processed using
the Crystal-Clear software (Rigaku) on a Pentium computer,
and were corrected for Lorentz and polarization effects. The
11
structures were solved by direct methods and expanded using
1
2
1
Fourier techniques. Hydrogen atoms except for H atom
disilaruthenacyclic complex
2
is similar to the ligand
(Ru-H) in Figure 1 were refined using the riding model,
whereas the position of H atom was determined by the
1
arrangement in 1 but the orientation of the two CO ligands is
different: the two CO ligands are in the cis configuration in 1
but in the trans configuration in 2. Nevertheless, the
hydrogenation of alkenes occurs under mild conditions, similar
to the reaction catalyzed by 1. Based on these experimental
results, we optimized the geometries of the cis- and
trans-isomers of 1 and 2 (cis-1, trans-1, cis-2, and trans-2) and
analysis of the difference Fourier map. The final cycle of
2
full-matrix least-squares refinement on F was based on 3,014
observed reflections and 137 variable parameters for 2. Neutral
1
3
atom scattering factors were taken from Cromer and Waber.
1
4
All calculations were performed using SHELXL-97. Details
of final refinement as well as the bond lengths and angles are
2

Table 1. Catalytic hydrogenation of various alkenes by 1 and 2
a
1
(ref. 2)
Yield
2
Catalyst
b
Cat
mol%)
0.25
1
1
1
5
5
5
Time
(h)
2
Cat
Time
(h)
6
6
6
6
6
4
6
Yield
(%)
>99
>99
>99
>99
>99
>99
5
Alkenes
TON
TON
(
(%)
>99
>99
>99
>99
>99
>99
20
(mol%)
c
1
-octene
400
100
100
20
20
20
1
100
20
20
20
20
20
1
methyl 10-undecenoate
styrene
1
2
6
4
6
6
5
5
5
5
5
5
stilbene
cyclohexene
ethyl 3-methyl-2-butanoate
2
,3-dimethyl-2-butene
4
a
All reactions were performed in the presence of the catalyst 2 (0.05 mmol), the alkene (1.0 mmol) and toluene (0.5 ml) at room
temperature under 1 atm of hydrogen. Determined by GC with anisole as an internal standard. 0.01 mmol of 2 was used.
b
c
summarized in Table S-1, and the numbering scheme employed
is also shown in Figures S-1, which was drawn with ORTEP at
equiv. for Ru) was photo-irradiated at room temperature for 24
h. Recrystallization of the crude sample led to isolation of the
ruthenium dicarbonyl complex 2 in 39% yield as colorless
crystals (Scheme 1).
5
0% probability ellipsoids. CCDC 1520573 (2) contains the
supplementary crystallographic data. These data can be
obtained free of charge from The Cambridge Crystallographic
Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
The molecular structure of 2 (denoted as 2cryst) was
determined by X-ray diffraction analysis. The ORTEP drawing
is shown in Figure 1 (a). Selected bond distances are shown in
Table 2 for comparison with the optimized structures obtained
by DFT calculations, and other bond distances and bond angles
are listed in Supporting Information (Figure S-1, Table S-1).
The ruthenium atom of 2 resides at a crystallographic inversion
center, and the entire molecular structure can be described as a
C -symmmetric octahedral coordination geometry with two
2
.4 Catalytic hydrogenation of alkenes by 2 (Table 1). In a
2
0 mL Schlenk tube fitted with a H -filled balloon was placed a
2
ruthenium complex [5.4 mg, 0.01 mmol (entry 1) or 27 mg,
.05 mmol (entries 2-7)], and a toluene (0.5 mL) solution of
0
alkene (1.0 mmol) was added. The atmosphere was replaced by
H , and the mixture stood with stirring at room temperature
2
under a hydrogen atmosphere for 6 h (entries 1-5, 7) or 4 h
2
v
(
entry 6). Anisole (108 L, 1.0 mmol) was added as the internal
standard, and the product yield was determined by GC analysis.
.5 DFT calculations. All of the calculations were performed
using the Gaussian 09 program to search for all intermediates
trans CO ligands located at the apical position. Analysis of the
difference Fourier map of 1 did not show hydrogen atoms
bonded to the iron center, whereas the Fourier map of 2 showed
that the two hydrogens were located between the two Si groups.
The H NMR spectrum of 2 measured at room temperature
showed only one singlet at 0.76 ppm and one high-field shifted
2
1
5
1
and transition structures on potential energy surfaces. For
1
6
optimization, the M06 functional was selected based on the
comparison of the crystal structure of 4 (see details in
Supporting Information 3-1-1). We also employed the SDD
singlet at –7.33 ppm with an integral ratio of 24:2. The former
resonance can be assigned to the four SiMe moieties whereas
2
1
7
18
(Stuttgart/Dresden pseudopotentials) and 6-31G** basis sets
the latter is derived from the two H-Si groups. The latter singlet
2
9
for Ru atoms and the other atoms, respectively [BS1]. All
stationary-point structures were found to have the appropriate
number of imaginary frequencies. An appropriate connection
between a reactant and a product was confirmed by intrinsic
reaction coordinate (IRC) and quasi-IRC (qIRC) calculations.
In the quasi-IRC calculation, the geometry of a transition state
was first shifted by perturbing the geometries very slightly
along the reaction coordinate, and then released for equilibrium
optimization. To determine the energy profile of the proposed
catalytic cycle, we performed single-point energy calculations
at the optimized geometries using the SDD (Stuttgart/ Dresden
was accompanied by a satellite signal coupled with Si (J
=
Si-H
1
2
16.8 Hz), which is smaller than the usual JSi-H of  - or  -Si-H
2
3,24
groups but is in the range of SISHA (JSi-H < 20 Hz).
Furthermore, the IR spectrum of 2 indicates relatively weak
H-Si interactions. These interactions showed strong
absorption band due to the coordinated H-Si moiety at 1956
cm , which is 184 cm lower than that of uncoordinated BDSB
(2140 cm ). A detailed discussion of the coordination modes of
Si-H groups in 2 is provided later in this paper.
1
9
a
-
1
-1
-
1
3.2 Hydrogenation of alkenes. As mentioned in the
introduction, the disilaferracyclic complex 1 showed good
2
0
2
pseudopotentials) and 6-311+G** basis sets for Ru atoms and
catalytic activity toward the hydrogenation of various alkenes.
the other atoms, respectively. Solvent effects of toluene ( =
Complex 1 has two characteristics: first, hydrogenation
catalyzed by 1 proceeds under mild reaction conditions, such as
2
.379) were evaluated using the polarizable continuum model
2
1
(PCM) [BS2]. Energy profiles of the calculated reaction
at room temperature and under 1 atm of H . Second, 1 catalyzes
2
pathways are presented as Gibbs free energy changes (G)
involving thermal corrections at 298.15 K. Spin-spin coupling
constants were computed during an NMR run, via the SpinSpin
option by using the GIAO method. All of the optimized
structures (ball-and-stick models) and optimized geometries in
XYZ file format are summarized in the Supporting
Information.
the hydrogenation of unfunctionalized sterically hindered
alkenes, including tri- and tetra-substituted alkenes.
The catalytic performance of 2 was examined for the
hydrogenation of various alkenes. Table 1 shows the results of
hydrogenation of several alkenes catalyzed by 2. For
comparison, the reactions catalyzed by 1 reported previously
are also listed in the table. It was found that 2 exhibit
comparable or somewhat lower catalytic activity than 1. The
hydrogenation of 1-octene catalyzed by 1 mol% of 2 proceeded
effectively and to completion at room temperature under 1 atm
2
2
3
. Results and Discussion
3
.1 Preparation and characterization of 2. The ruthenium
complex 2 can be synthesized by a procedure similar to that
of H (entry 1). Styrene and methyl 10-undecenoate required
2
used to prepare 1. A pentane solution of a mixture of
higher catalyst loadings for completion of the reaction, but the
reaction time never exceeded 6 h (entries 2 and 3). It should be
4
(
 -C H )Ru(CO) and 1,2-bis(dimethylsilyl)benzene (BDSB, 2
4
8
3
3

Table 2. Calculated bond distances and Meyer bond orders of disilametallacycles of iron (1), ruthenium (2) and osmium (3).
Bond distances (Å) Meyer bond orders
M-H
M-Si
.387,
2.388,
2.402,
Si-H
2.064,
2.120,
2.342,
2.361
1.882
1.971,
2.522
1.983
2.010,
2.609
2.015
M-CO
M-H
M-Si
Si-H
M-CO
2
0
.52,
0.03,
1
.526,
1.754,
1.755
cis-1opt
0.74, 0.75
0.55,
.57, 0.60
0.06, .08, 1.30, 1.31
0.11
1
.531
0
2
.404
2.446
.521,
trans-1opt
cis-2opt
1.519
1.657
1.659
1.681
1.688
1.782
1.931
1.993
1.949
1.951
0.68
0.64
0.57
0.70
0.69
0.54
0.11
1.22
0.98
1.03
0.97
1.01
2
0.55, 0.73 0.04, 0.21
0.62 0.14
0.60, 0.72 0.03, 0.18
2
.552
trans-2opt
cis-3opt
2.557
2
.553,
2
.577
trans-3opt
2.578
0.65
0.10
2
2
2
2
2
.3928(12),
.4116(13),
.4233(12),
.4240(12)
.5637(9),
.5514(10)
1.760(5),
1.762(4)
1cryst (ref. 2)
–
–
–
–
–
–
–
–
1.94(3),
2.01(4)
2cryst
1.61(3)
1.926(3)
–
–
2
2
9
1
Table 3. Observed and simulated chemical shifts of Ru-H and Si- H coupling constant of 2.
1
2
Complex
H (Ru-H) [ppm]
 JSi-H  for H [Hz]
 JSi-H  for H [Hz]
a
Experimental data of 2
–7.33
16.8
b
1
2
1
3
1
3
Simulated data of trans-2opt
–5.60(H ) / –5.60(H )
17.3(Si ) / 17.3(Si )
17.3(Si ) / 17.3(Si )
a
b
29
Condition: 600 MHz, r.t., C D . Condition: M06 / SSD (Ru), 6-31G** (C, H, O, Si), NMR=(SpinSpin). The values of H are
6
6
determined by comparison of the isotropic parameter in SCF GIAO magnetic shielding tensor (ppm) of H atoms with
tetramethylsilane (TMS) as a standard.
noted that methyl 10-undecenoate underwent smooth
hydrogenation, with the ester group remaining intact. The
internal olefins in trans-stilbene (entry 4), cyclohexene (entry
5
), and ethyl 3-methyl-2-butanoate (entry 6) can also be
hydrogenated to give the desired alkane in quantitative yield
using 5 mol% of 2 under mild reaction conditions within 6 h.
The catalytic hydrogenation of tetra-substituted alkenes
remains challenging: hydrogenation of 2,3-dimethyl-2-butene
occurred, but the yield was only approximately 5% due to
decomposition of the catalyst during hydrogenation.
Chart 1. Optimized structures of cis-1 containing the H FeSi
2
4
5
core and trans-2 containing the H RuSi core.
2
4
3
.3 Optimized structures of cis-1, trans-1, cis-2, and trans-2.
Structures for 1 and 2 compatible with the 18-electron rule are
shown in Scheme 1 as 1proposed and 2proposed. The metal is
bonded to two Si atoms to construct a disilametallacycle, and
two Si-H groups in BDSB are coordinated to the metal in an
2

-fashion. In contrast, DFT calculations of cis-1 provided the
optimized structure shown as cis-1opt (Chart 1), in which the
iron center primarily interacts with two hydrides and four
silicon atoms to form a H FeSi core, and secondary interaction
2
4
was observed between two silicon atoms and one hydride
2
3
(
SISHA). The structure of trans-2opt similarly includes a
H RuSi core with SISHA [Figure 1 (b)], in accordance with
2
4
2cryst determined by crystallography [Figure 1 (a)]. The H-Si
bond distances of 2cryst are 1.94(3) Å for Si(1)-H(1) and 2.01(4)
Å for Si(2)*-H(1), whereas those of trans-2 are both 1.983 Å.
opt
2
1
These are longer than those of the  - or  -(H-Si) moiety in
several ruthenium complexes but are in the range of SISHA
2
4
(
1
(
1.8−2.4 Å). The Ru-H bond distance in 2cryst and trans-2opt is
.61 (3) Å and 1.659 Å, respectively. The Ru-Si bond distances
2cryst: Ru(1)-Si(1) = 2.5637(9) and Ru(1)-Si(2) = 2.5514(10)
Å; trans-2opt: 2.557 Å) are somewhat longer than those in
Figure 1. The molecular structures and representative bond
distances of [(a) The ORTEP drawing obtained by
2
2
5
crystallography, (b) The optimized structure by DFT
calculations.]
previously reported disilaruthena(II)cycle complexes. Table 3
4

shows simulated J
observed in the H NMR spectrum of 2 in C D , both of which
values for trans-2opt and for 2 actually
from cis-2opt, trans-1opt and trans-2opt revealed that they
converge as intermediates at slightly higher energy levels (G
Si-H
1
o
6
6
are small enough to be in the range of SISHA (JSi-H < 20 Hz).
The H MSi core accompanied by SiHSi interactions
= +0.1 kcal/mol for cis-2”, +2.2 kcal/mol for trans-1” and
trans-2”) than the corresponding species bearing the H MSi
4
2
4
2
is generally observed by DFT calculations in other iron and
ruthenium complexes as trans-1opt and cis-2opt. Therefore, the
SiHSi SISHA moiety generally behaves as a four-electron
dianionic ligand to the Fe(II) or Ru(II) center in 1 and 2. The
calculated G values of the four optimized structures suggest
that cis-1opt is a 0.9 kcal/mol lower energy structure than
with SISHA (see details in Supporting Information 3-2-2, 3-2-3
and 3-4-1). Interconversion between the two isomers proceeds
with a G value of 0.6 ~ 4.4 kcal/mol. These calculation
‡
outcomes are consistent with the experimental results that
facile dissociation of 1 equivalent of BDSB was confirmed by
the reaction of 1 with 2,6-dithiaheptane at room temperature.
Similarly, 2 was readily converted to the disilaruthena(II)cycle
o
trans-1opt and that trans-2opt is a 4.0 kcal/mol lower energy
o
26
structure than cis-2opt. Although these differences in G
complex 4.
between the cis- and trans- isomers are not large, in particular
for 1, in considering the accuracy of DFT calculations, it is
likely that these values are consistent with the experimental
results showing that the iron complex 1 is present solely in the
cis form, whereas the ruthenium complex 2 is present solely in
the trans conformation.
As described previously, cis-1 has two isomers other than
cis-1opt that have similar energy levels. All three isomers have
the H FeSi core, and the Fe-H and Fe-Si bond distances and
2
4
bond orders are almost identical. The H atom interacts with the
two Si atoms, but each isomer has slightly different Si-H bond
distances and bond orders. Interconversion among the isomers
via intersilicon atom hydrogen fluctuation proceeds with
‡
5
essentially no energy barrier (G < 0.9 kcal/mol). Similar
barrierless intersilicon fluctuation is also observed between
trans-2opt and trans-2’, as shown in Chart 2. While the
SiHSi interaction is symmetrical in trans-2opt, it is
unsymmetrical in trans-2’. Since the convergence of similar
isomers was not observed for trans-1opt and cis-2opt in the
present calculations, these isomers probably exist and
intersilicon atom fluctuation would be generally observed in
these compounds (see details in Supporting Information 3-2-1).
Chart 3. Relative energy of proposed disilametallacycles
bearing two -(H-Si) ligands.
Chart 2. Isomerization between trans-2opt and trans-2’.

-CAMmechanisms of catalytic hydrogenation. As
described in the introduction, DFT calculations of the
hydrogenation of ethylene catalyzed by 1 support a scheme
5
involving -CAM mechanisms, as outlined in Scheme 2. The
calculations show that cis-1, cis-2, trans-1, and trans-2 all
Scheme 3. Ligand exchange from BDSB to H and C H on
2 2 4
cis-1”, cis-2”, trans-1” or trans-2” to form A1 , B1 or C1 .
M
M
M
catalyze the hydrogenation of ethylene by
a
-CAM
mechanism, although the detailed reaction pathways of the
trans-complexes are slightly different from those of the
cis-complexes.
Dissociation of BDSB from the bis[-(H-Si)] species in
cis-1”, cis-2”, trans-1”, and trans-2” is followed by the
The complexes 1 and 2 are catalyst precursors, and the
dissociation of BDSB and the coordination of ethylene and H2
results in formation of the intermediate which is involved in the
catalytic cycle. This step is considered to occur in two steps:
the precursor with a H MSi core with SISHA is first converted
coordination of ethylene and H to give disilametallacyclic
2
2
2
species bearing  -CH =CH and  -H ligands, thereby
2 2 2
M
M
initiating the catalytic cycle. Two species, A1 and B1 , are
formed from the cis-complexes, whereas C1 (M = Fe or Ru)
M
is generated from the trans-complexes. This substitution is
believed to proceed with a sufficiently low energy barrier for
the reaction to proceed at room temperature. Calculations of
2
4
to the disilametallacycle bearing two -(H-Si) ligands such as
1proposed and 2proposed, then the coordinated Si-H bonds are
substituted by H and CH =CH . These -(H-Si) intermediates
o
Fe
Fe
Fe
G from cis-1” to A1 or B1 and trans-1” to C1 , and
2
2
2
Ru
Ru
Ru
do not converge for cis-1, but single point calculation of the
those from cis-2” to A1 or B1 and trans-2” to C1 ,
expected structure suggests that 1 is 5.9 kcal/mol higher
in energy compared to cis-1opt. Geometry optimization of the
disilametallacycles bearing two -(H-Si) ligands generated
indicate that the substitution is endergonic, with the energy
difference being at most 5.9 kcal/mol.
proposed
5
M
M
M
The formed species A1 , B1 , and C1 initiate the
5

Figure 2. Three catalytic cycles catalyzed by disilametallacyclic compounds involving -CAM mechanisms (M = Fe, Ru, Os)
catalytic cycle for the hydrogenation of ethylene, which
pseudo-trigonal bipyramidal, leading to movement of the ethyl
Fe
Ru
involves the following four steps as shown in Figure 2: step
group to close to the Si-H moiety to form C3” or C3”
,
2
2
(
a), -H /-Fe-Si oxidative hydrogen migration; step (b),
respectively [step (e)]. Finally,  -Si-H/-M-C oxidative
2
Fe
Ru
insertion of CH =CH between the atoms of the Fe-H bond;
hydrogen migration occurs [step (c)] to form C4 or C4 , to
2
2
2
2
27
and step (c),  -Si-H/-Fe-C oxidative hydrogen migration.
which ethane is weakly bonded in an  -manner.
Replacement of the coordinated ethane of the product formed
Figure 3 shows energy diagrams for cycle A, cycle B, and
cycle C, where M = Fe or Ru. The activation energies for each
elementary reaction are listed in Table 4. In cycle A, the
by step (c) by ethylene and H releases ethane and regenerates
2
2
2
the disilaferracycle possessing  -H and  -CH =CH ligands.
The two initial species generated from cis-2”, A1 and B1
2
2
2
Ru
Ru
‡
,
rate-determining step is step (b), but the G values of step (a)
Fe
are components of cycle A and cycle B, respectively. After the
and step (c) are not very different. The reaction from A1 to
Ru
Ru
Fe
‡
generation of A1 or B1 , step (a) forms the ruthenium
A4 proceeds with a small energy barrier (G < 2.7 kcal/mol),
Ru Ru
Ru
Ru
hydrides A2
or B2 , respectively, which contain an
-CH=CH2 ligand and a non-classically coordinated Si-H
whereas that from A1 to A4 comprises reactions with
rather higher activation energies (G = 5.1 ~ 7.0 kcal/mol).
The rate-determining step of cycle B is in step (c); G of
2
‡

‡
moiety. Step (b) corresponds to the insertion of ethylene into
the M-H bond, but this reaction actually occurs by migration of
the hydrogen atom from the metal center to a carbon atom of
the coordinated ethylene to form an ethyl-metal complex. The
2
Fe
Fe
‡
2Ru
B3- to B4 is 4.9 kcal/mol, whereas G of B3-
to
Ru
28
B4 is considerably higher (8.5 kcal/mol). In sharp contrast,
high energy barriers exist for cleavage of the -agostic M-H
bond and for the change in geometry from octahedral to
pseudo-trigonal bipyramidal in cycle C, for which the
Ru
Ru
ethyl group of the resulting A3 or B3 is coupled with a
Ru
hydrogen atom of the coordinated Si-H moiety to form A4 or
Ru
‡
B4 . These processes are almost identical to cycle A and cycle
activation energy (G ) is 8.0 ~ 16.3 kcal/mol. Another
Ru
B of the iron-catalyzed process. A difference between A1 and
interesting difference between Fe and Ru is in step (a) + (b):
Fe
Fe
Fe
A1 is the orientation of the coordinated ethylene, which is
the reaction from C1 to C3 is slightly exergonic and
Ru
Ru
perpendicular to the plane formed by the metal center, a CO
ligand, and an M-Si -bond. During the progress of step (a),
rotation of the coordinated ethylene by 90º occurs
proceeds with no energy barrier, while that from C1 to C3
‡
requires a G of 7.1 kcal/mol.
These results show the following trends: (1) replacement
of BDSB by ethylene and H2 is easier for trans-precursors
compared to cis-precursors regardless of the metal; (2)
cis-complexes catalyze the reaction with a lower energy barrier
than the corresponding trans-compound, regardless of the
metal; (3) the energy barrier of the rate-determining step is
cis-Fe < cis-Ru, while trans-Ru < trans-Fe; and (4) the highest
activation energy obtained by these calculations was around 16
kcal/mol, which is sufficiently low for the reaction to proceed
at room temperature.
Ru
concomitantly with the H-H bond cleavage to form A2 , in
2
which the Ru-H and Ru-Si -bonds,  -ethylene moiety, and a
CO ligand are all in-plane (see Figure S-8 in the Supporting
Information).
The trans-forms of disilaferra- and disilaruthenacyclic
catalysts behave differently from the corresponding
cis-compounds in that cycle C is the sole catalytic cycle. The
reaction pathway mediated by trans-1” (Fe) is quite similar to
that mediated by trans-2” (Ru). Interestingly, splitting of the
2
Fe
Ru
H-H bond of the  -H moiety in C1 and in C1 induces the
As described previously, the octahedral geometry of cis-1
and of the species in the catalytic cycle facilitate the reactions
of steps (a) ~ (c). Similar advantages of this geometry are
observed in the hydrogenation reaction catalyzed by cis-2. The
2
Fe
concomitant formation of Si-H and C-H bonds to give C3 and
C3 , in which the ethyl group with -agostic interaction is
Ru
bonded to the metal center and the Si-H group is coordinated in
2
M
an  -manner. One hydrogen atom therefore migrates from the
initial species involved in the catalytic hydrogenation (A1 or
2
M

-H ligand to a Si atom [step (a)], and at the same moment,
2
B1 ) comprises a silyl group, an ethylene ligand, and
the other hydrogen atom migrates to a carbon of the
molecular hydrogen in a facial orientation. The H-H bond is
parallel to the M-Si -bond, and the overlap of these bonds
allows facile generation of a kite-shaped four-membered ring
transition state. This accomplishes the first oxidative hydrogen
2

-CH =CH ligand [step (b)]. The agostic C-H moiety in the
2 2
Fe
Ru
ethyl group in the resulting C3 and C3 dissociates [step (d)],
and the geometry subsequently changes from octahedral to
6

Fe
Ru
Fe
Ru
Figure 3. Relative energy diagrams for alkene hydrogenation starting from (a) A1 or A1 [Cycle A] (b) B1 or B1 [Cycle B] (c)
Fe
Ru
a
b
1
C1 or C1 [Cycle C] (G [kcal/mol]). The result of Single point calculation. The cycle A and B involve the conversion of  -SiH
coordination mode of A3- to A3- and B3- to B3- . This step is described in the Supporting Information.
1
M
1M
1M
1M
28
‡
a
Table 4. Comparison in activation energy of each elementary reactions [G ; kcal/mol]
b
b
c
d
e
Step (a)
Step (b)
TSQ2/Q3-1
+2.7
Step (d)
TSC3/C3’
Step (e)
TSC3’/C3”
Step (c)
TS
M
M
M
TSQ1/Q2
+2.3
+7.0
+7.1
+2.1
+3.1
-1.0
Fe
Ru
Os
Fe
Ru
Os
Fe
Ru
Os
-
-
-
-
-
-
-
-
-
-
+1.8
+5.1
+8.9
+4.9
+6.9
+8.0
+2.0
+1.2
Cycle A
Cycle B
Cycle C
+8.7
+9.6
+0.5
+3.4
+6.9
-
-
-0.5
+7.1
+16.3
+8.0
+15.0
+13.1
+7.4
+2.3
+0.1
+9.9
b
+1.5
+1.2
a
‡
c
Figures of bold red font are G of the rate determining step. Q = A or B, M = Fe, Ru or Os. Dissociation of -agostic hydrogen
atom. Rearrangement of geometry. Cycle A and B; TS
TSC3”/C3”’ and TSC3”’/C4
d
e
M
M
(Q = A or B). Cycle C (M = Fe, Ru); TSC3”/C4 . Cycle C (M = Os);
Q3-2/Q4
Os
Os
.
7

Figure 4. Geometries of ruthenium species in cycle A, cycle B and cycle C. Each structure is similar to that observed in the catalytic
cycles of iron and osmium.
2
migration [step (a)] to form A2 or B2 having M-H and  -Si-H
diagrams of the catalytic cycles from trans-1” and trans-2”
show that step (d) and step (e) require a relatively high G (15
and 16 kcal/mol for Fe, 8 and 13 kcal/mol for Ru) compared
with step (a) + step (b).
2
‡
moieties. It should be noted that the  -CH =CH ligand in
2
2
Ru
Fe
A1 differs from A1 in the orientation of the C-C axis,
which is perpendicular to the plane containing the ruthenium
center, one of the silicon atoms, and one of the CO ligands;
however, the  -CH =CH ligand is rotated by 90 degrees
during step (a) to form A2 . Since the formed hydride in A2
3.5 DFT calculation of the optimized structures and of
hydrogenation activity of the disilaosmacyclic homologue.
DFT calculations can predict the structure and reactivity of the
osmium homologues of the disilaferra- or disilaruthenacyclic
compounds so far discussed. Geometry optimization of the
osmium complexes cis-3 and trans-3 provides the
corresponding cis-3opt and trans-3opt shown in Figure 5. These
complexes have a structure containing the H OsSi core with
2
2
2
Ru
Fe
Ru
2
and A2 is in-plane with the C-C axis of the  -CH =CH
2
2
Fe
Ru
ligand as well as with B2 and B2 , hydrogen migration from
the metal to a carbon of the coordinated ethylene occurs
through a M-H-C-C four-membered transition state to form
1M
1M
A3-
or B3-
having a -agostic ethyl group. This
2
4
migration corresponds to step (b). After conversion of the
SISHA. The trans isomer is lower in energy than the cis-isomer
by 2.4 kcal/mol. Table 2 shows the Si-H bond distances and the
Meyer bond orders of the M-H, M-Si, and Si-H bonds in the
iron, ruthenium, and osmium complexes. Close similarity in the
M-H and M-Si bond orders between Fe, Ru, and Os complexes
indicates that all the structures contain the H MSi core with
1
1M
1M
2

-Si-H group in A3- or B3- to the  -coordination
2
mode, oxidative hydrogen migration from the  -Si-H group in
2M
2M
the resulting A3- or B3- to the -carbon of the ethyl
group takes place with minimal motion to accomplish step (c)
to form A4 or B4. Figure 4(a) and (b) show the geometry of
each isomer, using A1 -A4 and B1 -B4 as representative
structures.
2
4
Ru
Ru
Ru
Ru
SISHA.
Similar to trans-2, fluctuation of the hydrogen atom
between the two Si atoms in SiHSi SISHA in trans-3
forms the isomer trans-3’, which is 0.5 kcal/mol higher in
energy compared to trans-3opt. The Os-H and Os-Si bond
distances of the H OsSi core of trans-3’ are similar to those of
In contrast, the octahedral geometries of the
trans-complexes trans-1 and trans-2 provide both advantages
2
and disadvantages. The M-Si -bond, and the  -H-H and
2
M

-CH =CH ligands of the initial species C1 , are in the same
2 2
2
4
plane. This orientation is useful for simultaneously promoting
trans-3opt, and their interconversion occurs with no energy
barrier (see details in Supporting Information 3-5-1). Similar
isomers containing the H OsSi core did not converge for cis-3.
the two reactions step (a) and step (b) with a low energy barrier.
Fe
Fe
Indeed, no energy barrier from C1 to C3 was observed,
2
4
‡
Ru
Ru
whereas G from C1 to C3 is as small as 8 kcal/mol. In
contrast, the ligand configuration of C3 is not suitable for step
The isomer having a disilaosmacycle and two -(H-Si) ligands
converged for trans-3” but not for cis-3” (Scheme 4). G° of
‡
(
c). Two additional elementary reactions, step (d) and step (e)
trans-3” from trans-3opt is 5.3 kcal/mol, and G of this
M
M
to give C3’ and C3” , respectively, are preceded by step (c)
transformation is as low as 5.6 kcal/mol. One interesting
difference in the structure of trans-3” compared to the
corresponding Fe and Ru homologues is the coordination mode
of the two Si-H ligands. The bond order of one of the Si-H
M 27
M
to form C4 . The coordinatively unsaturated C3” is
stabilized by agostic interaction of a Si-Me group to the metal
center (see details in Supporting Information 3-6-1). Energy
8

Another feature of the osmium-catalyzed reactions is that
two of the elementary reactions are not concerted in cycle C. In
the iron- and ruthenium-catalyzed reactions, step (a) and step
Os
(b) occur simultaneously. In contrast, the intermediate C2
converges in the osmium-promoted process. The
Os
Os
Os
Os
transformation of C1 to C2 [step (a)] and C2 to C3
[
‡
step (b)] proceeds with a small energy barrier (G = 2.3 and
0
.1 kcal/mol).
Steps (a) and (b) of the iron-, ruthenium-, and
osmium-promoted reactions were studied in more detail by
Mayer bond order vs. IRC plots (Figure 6). The progress of the
Fe
Ru
reaction from C1 and C1 results in a sharp decrease of the
1
1
1
H -H interaction and an increase in the M-H bond strength
near the transition state [○ and ○ in Figure 6]. Formation of
the M-H bond is followed by an increase in the weak Si-H
2
5
1
1
interaction, whereas the M-Si bond interaction does not
1
1
change significantly. This produces a Si -M-H moiety with
2
SISHA. The M-H bond order increases in the transition state
2
2
Figure 5. The optimized structures and representative
interatomic distances [Å] of (a) cis-3opt and (b) trans-3opt
and decreases afterwards. At the same time, the C -H
interaction is increased. These observations indicate that H-H
cleavage is assisted in a concerted fashion by M-Si and C-H
bond formation to give the agostic ethyl group; this is clearly
M
M
visible in the structural change from C1 to C3 through
M
TSC1/C3 (M = Fe, Ru) shown in Figure 6.
In sharp contrast, the Mayer bond order vs. the IRC plot of
Os
Os
the reaction from C1 to C3 indicates that the reaction is
Os
stepwise and proceeds through C2 as the intermediate. Near
Os
the first transition state of step (a) [○ in Figure 6], TSC1/C2 ,
8
a sharp decrease in the H-H interaction is accompanied by an
1
2
1
increase in the bond orders of Os-H , Os-H , and Si-H ,
2
2
whereas the C -H interaction is not observed. This corresponds
1
1
2
to Os-Si /H -H oxidative hydrogen migration, which produces
Os
2
1
1
C2 containing the Os-H bond and Si -Os-H linkage with
Scheme 4. Ligand exchange from BDSB to H and C H on
cis-3 or trans-3 to form A1 , B1 or C1 .
2
2
4
Os
Os
Os
SISHA. As shown in Figure 6 (c), the optimized structure of
C2 suggests the osmium center to be Os(IV). At the second
Os
transition state corresponding to step (b) [○10 in Figure 6], the
ligands is similar to that of trans-1” and trans-2”,
indicating -Si-H coordination (0.46 for H -Si ), but the bond
2
2
2
2
2
2
Os-H interaction is decreased and the C -H interaction is
increased to complete the metal-to-carbon hydrogen atom
migration.
order of the other ligand is significantly smaller (0.14 for
1
1
H -Si ). The H-Os and Si-Os bond orders of the latter are larger
1
1
Osmium may act somewhat differently from iron and
ruthenium because osmium is a third row transition metal. It is
known that the stabilities of coordinatively unsaturated species
and metal species with a higher oxidation state follow the
order: first-row < second-row < third-row transition metals. Gº
(
(
0.70 for Os-Si and 0.77 for Os-H ) than that of the former
2
2
0.23 for Os-Si and 0.42 for Os-H ), possibly indicating that
the latter should be considered as a H-Os-Si species with
SISHA (see details in Supporting Information 3-5-2).
Os
Os
Os
The hydrogenation pathways from A1 , B1 , and C1
Os
Os
of C3’’ relative to C3 is 11.2 kcal/mol, which is larger than
that for the corresponding iron and ruthenium cases by 6.9 and
2.0 kcal/mol, respectively. Consequently, the stability of
are similar to those catalyzed by the disilaferra- or
disilaruthenacyclic compounds involving -CAM mechanisms.
Os
Os
The G° values of A1 and B1 from the catalyst precursor
Os
coordinatively unsaturated C3’’ could facilitate step (e),
cis-3opt are +7.7 and +12.6 kcal/mol, respectively, whereas that
‡
Os
Os
lowering the G of TSC3/C3’ . High stability of the Os(IV)
of C1 from trans-3opt is +9.4 kcal/mol. These values are
Os
species could explain why C2 contains a H OsSi moiety
having a Si H SISHA as the intermediate rather than as the
transition state.
.6 Summary of the calculations. DFT calculations of six
2
2
significantly larger than those of the corresponding iron- and
ruthenium complexes, indicating that the generation of
catalytically active species requires higher energy. The energy
diagram is described in the Supporting Information (Figure
1
1
2
9
3
metallacyclic complexes of cis-1 (Fe), trans-1 (Fe), cis-2 (Ru),
trans-2 (Ru), cis-3 (Os), and trans-3 (Os) revealed that all the
complexes have the H MSi with SISHA structure and promote
‡
S-27), and Table 4 shows a comparison of G in each
elementary reaction. The activation energy of the rate
2
4
determining step of cycle A [step (b)] and cycle B [step (c)]
catalytic hydrogenation of ethylene via catalytic cycles
involving -CAM mechanisms. Detailed studies of the
calculation results provided the following insights: (1) the
optimized structure of trans-2 is consistent with the
experimental results obtained by crystallography and NMR
‡
shows that G follows the order Os > Ru > Fe. Cycle C
involves dissociation of the -agostic hydrogen atom from the
metal center [step (d)] and subsequent movement of the ethyl
moiety [step (e)] between step (a) + (b) and step (c). The
rate-determining step of cycle C is either step (d) or (e), and
(
JSi-H). (2) Gº of the optimized structures showed cis-1opt
<
‡

G of the rate-determining step follows the order Fe > Ru >
Os
trans-1opt and cis-2opt > trans-2opt. This may explain the
experimental results showing that 1 is cis and 2 is trans. The
cis-isomer is more stable than the trans isomer of 1 (Fe), while
this tendency is reversed in 2 (Ru). (3) Three catalytic cycles
Os. This suggests that C1 may more efficiently promote
Fe
Ru
Os
hydrogenation than C1 and C1 , in contrast to A1 and
B1 .
Os
9

Figure 6. Plots of Mayer bond
orders along the minimum-energy
M
M
paths between C1 and C3 (a) M
Fe, (b) M = Ru, (c) M = Os with
their relative energies
G [kcal/mol]).
=
(
Quasi-IRC calculations have been
performed.
for the hydrogenation of ethylene were proposed from the
calculations: cycles A and B for the cis-complexes, and cycle C
for the trans-compounds. Replacements of BDSB in the
catalyst precursors of 1 and 2 by H and CH =CH should be
and ruthenium-catalyzed processes, were found in the reaction
catalyzed by trans-3: the facile dissociation of the -agostic
hydrogen atom, and the observation that steps (a) and (b) occur
in a stepwise manner. These two characteristic features can be
explained by features of the third-row transition metals.
2
2
2
facile, while those of 3 are more endergonic. Regardless of the
‡
metal, the reaction barrier, i.e., G of the rate-determining step,
is at most 16 kcal/mol, which is sufficiently low to promote the
4. Conclusion
reaction at room temperature. (4) Cycles A ~ C comprise the
In our previous papers we reported the synthesis and
characterization of the disilaferracyclic carbonyl complex 1, in
which the two carbonyl groups are in the cis-configuration.
2
following elementary reactions:  -H /M-Si oxidative
2
2
hydrogen migration [step (a)], M to C hydrogen migration [step
2
(
b)], and  -Si-H/M-C oxidative hydrogen migration [step (c)].
Complex 1 exhibits higher catalytic activity towards the
In cycles A and B, octahedral geometry helps facilitate these
elementary reactions. The rate determining step of cycle A and
cycle B is step (b) and step (c), respectively; in both cases, G
hydrogenation of alkenes than most other iron complexes
described in the literature. DFT calculations showed both the
4
‡
characteristic structure of 1 which contains a H FeSi core with
2
4
follows the order Fe < Ru < Os. (5) In the hydrogenation
mediated by trans-1 (Fe) and trans-2 (Ru), step (a) and step (b)
occur sequentially. Step (a) + (b) and step (c) occur with small
energy barriers. In contrast, step (d) and step (e), either of
which is the rate-determining step, require a relatively high
activation energy. (6) Two characteristic features of the
osmium-catalyzed reaction, which differ from those of iron-
SISHA, and the mechanism of catalytic hydrogenation by way
of -CAM, in which the H-H bond is not oxidatively added to
the metal center but rather is cleaved with the aid of the Fe-Si
5
-bond. In this paper, we described for the first time the
preparation of 2, the ruthenium homologue of 1, and found that
2 has a trans-configuration and exhibits catalytic activity
towards the hydrogenation of alkenes that is comparable to or
1
0

lower than that of 1. The experimental results brought about the
and 4, How to select functionals, Ball-and-stick models for all
of the species obtained from the DFT calculation which include
energy diagrams, The optimized geometries in XYZ file format.
This material is available on http://dx.doi.org/10.1246/bcsj.***.
question, of whether the H MSi with SISHA and
2
4
hydrogenation mechanisms through -CAM are generally seen
in iron, ruthenium, and osmium homologues of 1 and 2 with
cis- and trans-configurations.
The calculations suggest a reasonable reaction barrier for
the three catalytic cycles either from the trans- or the
cis-precursor, which can promote the hydrogenation reaction at
room temperature. All the catalyst precursors, intermediates,
and transition states of cycle A and cycle B contain two CO
ligands in the cis-configuration, whereas those of cycle C have
trans-CO ligands. It should be noted that cis/trans
isomerization could occur at any step during catalysis, for
example, via dissociation and re-coordination of one of the CO
ligands. The involvement of cis/trans isomerization in the
hydrogenation mechanism would allow catalytic reactions with
a lower energy barrier. For example, cis to trans isomerization
of the cis-catalyst precursors results in easier generation of
References
1.
2.
3.
J. F. Hartwig, In Organotransition Metal Chemistry:
From Bonding to Catalysis; Univ. Science Books, 2009;
Chapter 15.
Y. Sunada, H. Tsutsumi, K. Shigeta, R. Yoshida, T.
Hashimoto, H. Nagashima, Dalton Trans. 2013, 42,
16687-16692.
(a) T. Naota, H. Takaya, S. Murahashi, Chem. Rev. 1998,
98, 2599-2660. (b) J. Halpern, J. F. Harrod, B. R. James,
J. Am. Chem. Soc. 1961, 83, 753-754. (c) D. Evans, J. A.
Osborn, F. H. Jardin, G. Wilkinson, Nature 1965, 208,
1203-1204. (d) R. A. Grey, G. P. Pez, A. Wallo, J. Am.
Chem. Soc. 1980, 102, 5948-5949. (e) J. L. Graff, M. S.
Wrighton, Inorg. Chim. Acta 1982, 63, 63-70. (f) Y. Blum,
D. Czarcle, Y. Rahamlm, Y. Shvo, Organometallics 1985,
4, 1459-1461. (g) C. Pranckevicius, L. Fan, D. W.
Stephan, J. Am. Chem. Soc. 2015, 137, 5582-5589.
(a) E. J. Daida, J. C. Peters, Inorg. Chem. 2004, 43,
7474-7485. (b) S. C. Bart, E. Lobkovsky, P. J. Chirik, J.
Am. Chem. Soc. 2004, 126, 13794-13807. (c) R. P. Yu, J.
M. Darmon, J. M. Hoyt, G. W. Margulieux, Z. R. Turner,
P. J. Chirik, ACS Catal. 2012, 2, 1760-1764. (d) P. J.
Chirik, Acc. Chem. Res. 2015, 48, 1687-1695. (e) S. C.
Bart, E. J. Hawrelak, E. Lobkovsky, P. J. Chirik,
Organometallics 2005, 24, 5518-5527. (f) J. Hoyt, M.
Shevlin, G. W. Margulieux, S. W. Krska, M. T. Tudge, P.
J. Chirik, Organometallics 2014, 33, 5781-5790. (g) M. A.
Schroeder, M. S. Wrighton, J. Am. Chem. Soc. 1976, 98,
551-558. (h) E. N. Frankel, E. A. Emken, H. M. Peters, V.
L. Davison, Butterfield, R. O. J. Org. Chem. 1964, 29,
3292-3297. (i) S. Chalraborty, W. W. Brennessel, W. D.
Jones, J. Am. Chem. Soc. 2014, 136, 8564-8567. (j) S.
Chalraborty, P. O. Lagaditis, M. Förster, E. A. Bielinski,
N. Hazari, M. C. Holthausen, W. D. Jones, S. Schneider,
ACS. Catal. 2014, 4, 3004-4003.
2
2
active species bonded with  -H and  -ethylene.
2
Alternatively, trans to cis isomerization of the species involved
in step (a)+(b) in cycle C provides an opportunity to allow
catalysis through cycle A or B, which do not involve the
high-energy processes of steps (d) and (e).
4.
The results described here clearly show the remarkable
similarity in the structures and reactivities of iron, ruthenium,
and osmium species as catalyst precursors, catalytic
intermediates, and in the transition states. In particular, the
overall catalytic performance towards the hydrogenation of
alkenes is essentially independent of the metal. This is
interesting because many ruthenium complexes act as
hydrogenation catalysts, whereas iron and osmium catalysts are
3
,4,30
rare.
This commonality can be explained by the -CAM
mechanism: H is not activated at the metal center but rather at
2
the M-Si -bond. Consequently, the metal of the present
2
catalyst must be able to form the  -H species but does not
2
3
1
need to promote oxidative addition. This suggests that any
catalyst with metal-silicon bonds could show catalytic activity
towards the hydrogenation of alkenes, as long as molecular
hydrogen can coordinate to the metal center. The metal is not
necessarily limited to transition metals. It should be noted that
H activation by a metal-nitrogen bond provided the basis for
5.
6.
A. Tahara, H. Tanaka, Y. Sunada, Y. Shiota, K. Yoshizawa,
H. Nagashima, J. Org. Chem. 2016, 81, 10900-10911.
It is well-known that -bond metathesis of d -transition
2
Noyori’s famous concept of metal-ligand cooperation for the
3
2
0
catalytic asymmetric hydrogenation of ketones. Peters and
coworkers have recently reported a stoichiometric reaction in
metals proceeds through a square-shaped four membered
ring transition state, in which no interaction is found
between the metal and the atom at the diagonal position.
In contrast, -bond metathesis of late transition metals is
essentially a sequence of oxidative addition and reductive
elimination in concerted manner, and called as oxidative
hydrogen migration. The reactions proceed through a
kite-shaped transition state, in which the metal is
interacted with all three other atoms included in the
transition state. (a) ref 1, Chapter 6.5, pp285-286. (b) J.
Oxgaard, W. A. Goddard III, J. Am. Chem. Soc. 2004, 126,
442-443. (c) J. Oxgaard, R. P. Muller, W. A. Goddard III,
R. P. Periana, J. Am. Chem. Soc. 2004, 126, 352-363. (d)
W. J. Tenn III, K. J. H. Young, G. Bhalla, J. Oxgaard, W.
A. Goddard III, R. P. Periana, J. Am. Chem. Soc. 2005,
127, 14172-14173.
(a) E. Nakamura, K. Sato, Nat. Mater. 2011, 10, 158-161.
(b) C. Bolm, J. Legros, J. L. Paith, L. Zani, Chem. Rev.
2004, 104, 6217-6254. (c) B. Plietker, in Iron Catalysis in
Organic Chemistry (Ed: B. Plietker), WILEY-VCH,
Weinheim, Germany, 2008 (d) A. Furstner, ACS Cent. Sci.
2016, 2, 778-779. (e) H. Nagashima, Y. Sunada, T.
Nishikata, A. Chaiyanurakkul, In PATAI’s Chemistry of
3
3a
which a metal-boron bond activates molecular hydrogen.
Both these examples suggest that the counter element of the
metal is not necessarily limited to silicon. We believe that the
design of new hydrogenation catalysts composed of M-E (M =
any metal capable of bonding with molecular hydrogen, E =
any element) is possible by appropriate choice of the metal and
the counterpart element. The disilametallacyclic compounds
presented in this paper may be a prototype for this catalyst
design.
Acknowledgement
This work was supported in part by Grants-in-Aid (Nos.
2
4109014, 24550190, 15K13710 and 15K21222) from JSPS
and MEXT. This work also was supported by Integrated
Research Consortium on Chemical Sciences (IRCCS), Network
Joint Research Center for Materials and Devices and the Core
Research Evolutional Science and Technology (CREST)
Program of Japan Science and Technology Agency (JST),
Japan.
7.
Supporting Information
The CIF files, crystallographic data and actual charts for 2
11

Functional Groups, The Chemistry of Organoiron
Compounds (Ed: Saul Patai), WILEY-VCH, Weinheim,
Germany, 2014; Chapter 9.
D. Cremer, J. Chem. Phys. 2000, 113, 3530-3547. (c) V.
Barone, J. E. Peralta, R. H. Contreras, J. P. Snyder, J.
Phys. Chem. A, 2002, 106, 5607-5612. (d) J. E. Peralta, G.
E. Scuseria, J. R. Cheeseman, M. J. Frisch, Chem. Phys.
Lett., 2003, 375 452-458. (e) W. Deng, J. R. Cheeseman,
M. J. Frisch, J. Chem. Theory and Comput., 2006, 2,
1028-1037.
8
9
1
1
.
.
R. N. Perutz, S. Sabo-Etienne, Angew. Chem. Int. Ed.
2
007, 46, 2578-2592.
H. Nagashima, K. Tatebe, I. Ishibashi, A. Nakaoka, J.
Sakakibara, K. Itoh, Organometallics 1995, 14, 2868.
0. B. A. Sosinsky, S. A. R. Knox, F. G. A. Stone, J. Chem.
Soc. Dalton. Trans., 1975, 1633-1640.
1. SIR2008: M. C. Burla, R. Caliandro, M. Camalli, B.
Carrozzini, G. L. Cascarano, L. De Caro, C. Giacovazzo,
G. Polidori, D. Siliqi, R. Spagna, J. Appl. Crystallogr.
23. (a) F. Delpech, S. Sabo-Etienne, J. –C. Daran, B.
Chaudret, K. Hussein, C. J. Marsden and J. –C. Barthelat,
J. Am. Chem. Soc., 1999, 121, 6668-6682. (b) I. Atheaux,
F. Delpech, B. Donnadieu, S. Sabo-Etienne, B. Chaudret,
K. Hussein, J. -C. Barthelat, T. Braun, S. B. Duckett and
R. N. Perutz, Organometallics 2002, 21, 5347-5357. (c) T.
Ayed, J. -C. Barthelat, B. Tangour, C. Pradꢀre, B.
Donnadieu, M. Grellier and S. Sabo-Etienne,
Organometallics 2005, 24, 3824−3826. (d) S. Lachaize
and S. Sabo-Etienne, Eur. J. Inorg. Chem., 2006,
2115-2127. (e) M. Grellier, T. Ayed, J. –C. Barthelat, A.
Albinati, S. Mason, L. Vendier, Y. Coppel and S.
Sabo-Etienne, J. Am. Chem. Soc., 2009, 131, 7633−7640.
24. (a) J. Y. Corey, Chem. Rev. 2016, 116, 11291-11435. (b) J.
Y. Corey, Chem. Rev. 2011, 111, 863-1071. (c) J. C. Corey,
J. Braddock-Wilking, Chem. Rev. 1999, 99, 175-292.
25. (a) Y. Sunada, T. Imaoka and H. Nagashima,
Organometallics 2013, 32, 2112-2120. (b) J. J. G.
Minglana, M. Okazaki, K. Hasegawa, L. –S. Luh, N.
Yamahira, T. Komuro, H. Ogino and H. Tobita,
Organometallics 2007, 26, 5859-5866. (c) H. Tobita, K.
Hasegawa, J. J. G. Minglana, L. –S. Luh, M. Okazaki and
H. Ogino, Organometallics 1999, 18, 2058-2060.
2
007, 40, 609-613.
1
1
2. DIRDIF99: P. T. Beurskens, G. Admiraal, G. Beurskens,
W. P. Bosman, R. de Gelder, R. Israel, J. M. M. Smits,
The DIRDIF-99 program system; Technical Report of the
Crystallography Laboratory; University of Nijmegen,
Nijmegen, The Netherlands, 1999.
3. D. T. Cromer, J. T. Waber, International Tables for X-ray
Crystallography; Kynoch Press: Birmingham, U.K., 1974,
Vol. 4.
1
1
4. SHELX97: G.M. Sheldrick, Acta Cryst., 2008, A64,
1
12-122.
5. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B.
Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X.
Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J.
L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda,
J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao,
H. Nakai, T. Vreven, J. A. Montgomery Jr., J. E. Peralta, F.
Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N.
Kudin, V. N. Staroverov, T. Keith, R. Kobayashi, J.
Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S.
Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M.
Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J.
Jaramillo, R. Gomperts, R. E. Stratmann, O, Yazyev, A. J.
Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L.
Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P.
Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, O.
Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski, D. J.
Fox, Gaussian 09, Revision C.01; Gaussian, Inc.,
Wallingford, CT, 2010.
26. Preparation and characterization of 4 is described in
Supporting Information 1-2-1.
27. The M-H and Si-H bond orders (Supporting Information
M
3-6-1) indicates C3’’ may rather be described as
Si M(IV)(H)(C H ) with
a
SiH SISHA than
2
2
5
2
SiM(II){ -(H-Si)}C H . These are produced by the
2
5
M
oxidative addition of the -(H-Si) group in C3’ , which
concomitantly occurs with step (d). As a result, step (e)
rather be called as reductive elimination.
1
2
28. The  - to  - conversion occurs with a small barrier.
‡
(G < 1.1 kcal/mol), and does not affect the overall
kinetics. See details in Supporting Information 3-3-3 and
1
1
1
6. Y. Zhao, D. G. Truhlar, Theor. Chem. Acc. 2008, 120,
3-3-7.
2
15-241.
29. R. H. Crabtree, In The Organometallic Chemistry of the
th
7. D. Andrae, U. Häußermann, M. Dolg, H. Stoll, H. Preuß,
Theor. Chim. Acta 1990, 77, 123-141.
8. (a) M. S. Gordon, Chem. Phys. Lett. 1980, 76, 163−168.
Transition Metals; 6 edition, WILEY, 2014; Chapter 2,
15.
30. (a) L. Vaska, J. Am. Chem. Soc. 1964, 86, 1943-1950. (b)
L. Vaska, Inorg. Nucl. Chem. Lett. 1965, 1, 89-95. (c) R.
A. S. Delgado, A. Andriollo, N. Valencia, J. Mol. Cat.
1984, 24, 217-220. (d) R. A. S. Delgado, M. Rosales, M.
A. Esteruelas, L. A. Oro, J. Mol. Cat. A. 1995, 96,
231-243. (e) R. A. S. Delgado, M. Medina, F. L. Linares,
A. Fuentes, J. Mol. Cat. A. 1997, 116, 167-177. (f) D. F.
Schreiber, C. O’Connor, C. Grave, H. M. Bunz, R.
Scopelliti, P. J. Dyson, A. D. Phillips, Organometallics
2013, 32, 7345-7356.
(b) Pc. Harihara, J. A. Pople, Mol. Phys. 1974, 27,
2
1
09−214. (c) Pc. Harihara, J. A. Pople, Theor. Chem. Acc.
973, 28, 213−222. (d) W. J. Hehre, R. Ditchfield, J. A.
Pople, J. Chem. Phys. 1972, 56, 2257−2261. (e) R.
Ditchfield, W. J. Hehre, J. A. Pople, J. Chem. Phys. 1971,
5
4, 724−728.
1
2
9. (a) K. Fukui, J. Phys. Chem. 1970, 74, 4161-4163. (b) K.
Fukui, Acc. Chem. Res. 1981, 14, 363-368. (c) C.
Gonzalez, H. B. Schlegel, J. Phys. Chem. 1990, 94,
5
523-5527.
31. Calculations of model compounds suggest the
0. (a) R. Krishnan, J. S. Binkley, R. Seeger, J. A. Pople, J.
Chem. Phys. 1980, 72, 650-654. (b) A. D. McLean, G. S.
Chandler, J. Chem. Phys. 1980, 72, 5639-5648. (c) T.
Clark, J. Chandrasekhar, G. W. Spitznagel, P. v. R.
Schleyer, J. Comput. Chem. 1983, 4, 294-301.
“M(IV)H Si ”
species
to
be
much
less
2
2
2
thermodynamically favorable than the “M(II)( -H )Si ”
(M = Fe, Ru and Os). See details in Supporting
Information 3-6-2.
2
2
32. (a) T. Ohkuma, H. Ooka, S. Hashiguchi, T. Ikariya, R.
Noyori, J. Am. Chem. Soc. 1995, 117, 2675-2676. (b) R.
Noyori, T. Ohkuma, Angew. Chem. Int. Ed. 2001, 40,
40-73. (c) R. Noyori, Angew. Chem. Int. Ed. 2002, 41,
2008-2022. (d) R. Noyori, M. Kitamura, T. Ohkuma,
2
2
1. Miertuš, E. Scrocco, Tomasi, J. Chem. Phys. 1981, 55,
1
17−129.
2. (a) T. Helgaker, M. Watson, N. C. Handy, J. Chem. Phys.
000, 113, 9402-9409. (b) V. Sychrovsky, J. Gräfenstein,
2
1
2

Proc. Natl. Acad. Sci U.S.A 2004, 101, 5356-5362. (e) M.
Ito, T. Ikariya, Chem. Commun. 2007, 48, 5134–5142. (f)
K. Abdur-Rashid, S. E. Clapham, A. Hadzovic, J. N.
Harvey, A. J. Lough, R. H. Morris, J. Am Chem. Soc.
2
002, 124, 15104-15118. (g) R. J. Hamilton, S. H.
Bergens, J. Am Chem. Soc. 2008, 130, 11979-11987.
3. (a) H. Fong, M. -E. Moret, Y. Lee, J. C. Peters,
Organometallics 2013, 32, 3052-3062. (b) L. T. Pin, J. C.
Peters, J. Am. Chem. Soc. 2013, 135, 15310-15313. (c) L.
T. Pin, J. C. Peters, J. Am. Chem. Soc. 2014, 136,
3
1
3672-13683. (d) W. H. Harman, J. C. Peters, J. Am.
Chem. Soc. 2012, 134, 5080-5082. (e) W. H. Harman, L.
T. Pin, J. C. Peters, Angew. Chem. Int. Ed. 2014, 53,
1
081-1085.
1
3

Graphical Abstract
<

Title>
-CAM Mechanisms for the Hydrogenation of Alkenes by cis- and trans-Disilametallacyclic Carbonyl Complexes (M = Fe,
Ru, Os): Experimental and Theoretical Studies
<Authors' names>
Konoka Hoshi, Atsushi Tahara, Yusuke Sunada, Hironori Tsutsumi, Ryoko Inoue, Hiromasa Tanaka, Yoshihito Shiota, Kazunari
Yoshizawa, and Hideo Nagashima*
<Summary>
A disilaruthenacycle having two CO ligands in the trans-configuration showed catalytic performance for hydrogenation of alkenes
comparable to the iron homologue with cis-CO ligands. DFT calculations suggest that cis- and trans-isomers of dislametallacyclic
carbonyl complexes of iron, ruthenium, and osmium have similar structures containing a H MSi core with SiHSi SISHA and
2
4
catalyze the hydrogenation of alkenes via -CAM mechanisms. [59 words]
<Diagram>
1
4