Electron-rich N-heterocyclic silylene (NHSi)-iron complexes: Synthesis, structures, and catalytic ability of an isolable hydridosilylene-iron complex

Article abstract of DOI:10.1021/ja402480v

The first electron-rich N-heterocyclic silylene (NHSi)-iron(0) complexes are reported. The synthesis of the starting complex is accomplished by reaction of the electron-rich Fe⁰ precursor [(dmpe)₂Fe(PMe ₃)] 1 (dmpe =1,2-bis(dimethylphosphino)ethane) with the N-heterocyclic chlorosilylene LSiCl (L = PhC(N^tBu)₂) 2 to give, via Me₃P elimination, the corresponding iron complex [(dmpe)₂Fe(←:Si(Cl)L)] 3. Reaction of in situ generated 3 with MeLi afforded [(dmpe)₂Fe(←:Si(Me)L)] 4 under salt metathesis reaction, while its reaction with Li[BHEt₃] yielded [(dmpe) ₂Fe(←:Si(H)L)] 5, a rare example of an isolable Si^II hydride complex and the first such example for iron. All complexes were fully characterized by spectroscopic means and by single-crystal X-ray diffraction analyses. DFT calculations further characterizing the bonding situation between the Si^II and Fe⁰ centers were also carried out, whereby multiple bonding character is detected in all cases (Wiberg Bond Index >1). For the first time, the catalytic activity of a Si^II hydride complex was investigated. Complex 5 was used as a precatalyst for the hydrosilylation of a variety of ketones in the presence of (EtO)₃SiH as a hydridosilane source. In most cases excellent conversions to the corresponding alcohols were obtained after workup. The reaction pathway presumably involves a ketone-assisted 1,2-hydride transfer from the Si^II to Fe⁰ center, as a key elementary step, resulting in a betaine-like silyliumylidene intermediate. The appearance of the latter intermediate is supported by DFT calculations, and a mechanistic proposal for the catalytic process is presented.

Full text of DOI:10.1021/ja402480v

Article
pubs.acs.org/JACS
Electron-Rich N‑Heterocyclic Silylene (NHSi)−Iron Complexes:
Synthesis, Structures, and Catalytic Ability of an Isolable
Hydridosilylene−Iron Complex
Burgert Blom, Stephan Enthaler, Shigeyoshi Inoue,* Elisabeth Irran, and Matthias Driess*
Department of Chemistry: Metalorganics and Inorganic Materials, Technische Universitat Berlin, Strasse des 17. Juni 135, Sekr. C2,
̈
D-10623 Berlin, Germany
S
* Supporting Information
ABSTRACT: The first electron-rich N-heterocyclic silylene (NHSi)−iron(0)
complexes are reported. The synthesis of the starting complex is accomplished
by reaction of the electron-rich Fe0 precursor [(dmpe)2Fe(PMe3)] 1 (dmpe =1,2-
bis(dimethylphosphino)ethane) with the N-heterocyclic chlorosilylene LSiCl (L =
PhC(NtBu)2) 2 to give, via Me3P elimination, the corresponding iron complex
[(dmpe)2Fe(←:Si(Cl)L)] 3. Reaction of in situ generated 3 with MeLi afforded
[(dmpe)2Fe(←:Si(Me)L)] 4 under salt metathesis reaction, while its reaction with Li[BHEt3] yielded [(dmpe)2Fe(←:Si(H)L)]
5, a rare example of an isolable SiII hydride complex and the first such example for iron. All complexes were fully characterized by
spectroscopic means and by single-crystal X-ray diffraction analyses. DFT calculations further characterizing the bonding
situation between the SiII and Fe0 centers were also carried out, whereby multiple bonding character is detected in all cases
(Wiberg Bond Index >1). For the first time, the catalytic activity of a SiII hydride complex was investigated. Complex 5 was used
as a precatalyst for the hydrosilylation of a variety of ketones in the presence of (EtO)3SiH as a hydridosilane source. In most
cases excellent conversions to the corresponding alcohols were obtained after workup. The reaction pathway presumably involves
a ketone-assisted 1,2-hydride transfer from the SiII to Fe0 center, as a key elementary step, resulting in a betaine-like
silyliumylidene intermediate. The appearance of the latter intermediate is supported by DFT calculations, and a mechanistic
proposal for the catalytic process is presented.
Chart 1. Examples of Recently Reported NHSi Complexes
Active in Various Stoichiometric or Catalytic
Transformations
INTRODUCTION
■
Since the seminal work of Schmid and Wels in 1977, who
reported the first N-heterocyclic silylene (NHSi) complex,1 a
plethora of NHSi complexes have been reported for a large
array of the transition metal elements.2 Burgeoning interest, in
our and other groups, in these complexes has been further
precipitated by recent reports highlighting their use in a variety
of small molecule activations (NH3, H2S, hydrazines etc.; I,
Chart 1),3 interesting stoichiometric transformations, for
example hydrosilylation of alkynes (II, Chart 1),4 or silane
activation (III, Chart 1).5 Moreover they have been
demonstrated to be active in a variety of catalytic trans-
formations such as [2 + 2 + 2] cycloaddition of substituted
alkynes (IV, Chart 1),6 Suzuki cross coupling of aryl boronic
acids with bromoarenes,7 Heck coupling of styrene and
bromoacetophenone (V, Chart 1),8 and the recently reported
borylation of arenes,9 clearly demonstrating their potential and
spurring on further development.
An abundance of synthetic routes enabling access to NHSi
complexes now exist, mostly involving ligand elimination
reactions (CO,10 thf,11 cod (1,5-cycloocta-1,3-diene),12
PPh3,13 PEt3,14 or even N-heterocyclic carbenes15) from
suitable transition metal precursors upon reaction with “free”
NHSis.2 We recently showed that a PMe3 elimination strategy
can even enable facile entry to novel titanium NHSis complexes
of the type [(C5H5)2Ti(←:SiXL)2] (L = PhC(tBuN)2, X = Cl,
CH3, H), representing the first examples of group IV NHSi
complexes.16
Received: March 13, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja402480v | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Chart 2. Previouly Known Fe−NHSi Complexes
Scheme 1. Synthesis of Electron-Rich Fe−NHSi Complex 3
Scheme 2. Synthesis of 4 and 5 by in Situ Generation of 3 and Subsequent Reaction with MeLi or Li[HBEt3], Respectively
With a view of exploring the generality of this PMe3
elimination route to access NHSi complexes of other transition
metals and exploring their inherent reactivity and catalytic
activity, we became interested in exploring an analogous route
to access iron NHSis, which are surprisingly rare. To date, the
few existing iron−NHSi complexes exclusively bear the
{Fe(CO)4} fragment (Chart 2).1,17,18 The probable reason
for this is a lack of suitable iron precursors to access other
NHSi iron complexes. With this in mind, we considered the
complex [(dmpe)2Fe(PMe3)] (dmpe =1,2-bis(dimethy-
phosphino)ethane) (1) as a possible starting material
previously shown to be a suitable precursor in accessing
iron−Group 14 element (Ge and Sn) multiple bonds, by PMe3
elimination.19 We reasoned that by increasing the electron
density at the iron center with electron pushing alkyl
phosphane ligands as opposed to CO ligands, enhanced π-
back-bonding to the SiII center should result, potentially
increasing the reactivity of the overall complex, which is
desirable in a catalytic context. Surprisingly, no previous reports
exist where Fe-based silylene complexes20 have been used in
catalysis, and given the extremely low cost and abundance of
iron, we felt further warranted investigation.
Herein we report the first examples of electron-rich Fe−
NHSi complexes, obtained by a PMe3 elimination strategy, and
their reactivity. Moreover, the first catalytic investigations on
any Fe-based NHSi or SiII hydrido complex are reported,
showcasing the potential of reactive SiII hydrides in catalysis.
B
dx.doi.org/10.1021/ja402480v | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 1. Stacked variable temperature (VT) 31P{1H} NMR spectra of the hydrido-NHSi complex 5 revealing two bands at 298 K collapsing into a
broad signal at 309 K with comcomitant signal sharpening on temperature ascent due to the onset of pseudorotation (measured in C6D6 at 161.9
MHz).
temperature, at a lower measuring frequency (80 MHz),
however, shows only one very broad signal pointing to very
slow fluxional exchange of the P atoms at the lower frequency,
in accordance with expectations.25 At higher temperatures, the
31P{1H} NMR spectra of 5 reveal a broad signal at 309 K (δ
59.5 ppm, (Δν1/2 = 2236 Hz)), with concomitant signal
sharpening on increased temperature (330 K, δ 60.3 ppm,
Δν1/2 = 902 Hz) (Figure 1). The coalescence temperature for P
exchange thus lies between 298 and 309 K at this measuring
frequency (161.9 MHz) in complex 5. These findings reveal an
increase in the energy barrier associated with the Berry
pseudorotation of the dmpe ligands in complex 5, compared
to 3 and 4. This observation cannot be accounted for on steric
grounds since 5 features the smallest substituent (H) on Si
compared to 3 (Cl) and 4 (CH3). Wilson and co-workers have
documented a relationship between the energy barrier of the
pseudorotation and the π-accepting ability of the alkene in
complexes of the type [Fe(CO)4(alkene)], where a π-acidic
alkene exhibits a higher pseudorotational energy barrier.26 In
accordance with these findings, in our case the: Si(L)H ligand
may exhibit a higher π-accepting ability in 5 compared to:Si(L)
Cl (3) or:Si(L)Me (4), which was investigated by DFT studies
(see below).
RESULTS
■
Reaction of the electron-rich Fe0 precursor [(dmpe)2Fe-
(PMe3)] (1)19 with the intramolecularly N-donor-stabilized
chlorosilylene LSiCl (L = PhC(NtBu)2) (2)21 in diethyl ether
affords, with concomitant PMe3 elimination, the extremely air
sensitive orange complex [(dmpe)2Fe(←:Si(Cl)L)] (3) in 81%
yield (Scheme 1). Complex 3 is a rare example of an NHSi
complex of Fe, and can be considered much more electron-rich
than the previously reported [Fe(CO)4] complexes due to the
presence of electron pushing alkylphosphane donor ligands.
Owing to the stabilization of the reactive SiII center by the
{FeP4} fragment in complex 3, further reactions can be carried
out at the SiII center. For example, in situ generation of 3
followed by reaction with MeLi afforded, by metathetical
exchange at the silicon center, the complex [(dmpe)2Fe(←
:Si(Me)L)] (4), in 67% yield as a red solid (Scheme 2).
Furthermore, the reaction of in situ generated 3 with
Li[HBEt3] afforded by halide−hydride exchange a mixture of
[(dmpe)2Fe(←:Si(H)L)] (5) and Me3P:→BEt3 (6). The
desired complex 5 could be separated from 6 by crystallization,
affording 5 in 89% yield as a red-brown solid (Scheme 2).
Isolable SiII hydrides are very rare due to their concomitant
high reactivity,22 and generally require stabilization by a
transition metal complex fragment or intramolecular donor−
acceptor stabilization. Complex 5 is the first example of an
isolable hydrido−SiII complex of iron. Strikingly, complexes 3−
5 are all thermally very stable with decomposition temperatures
well above 110 °C, a desirable property to probe their
suitability as precatalysts for organic transformations.
Complexes 3 and 4 feature relatively broad signals (3: Δν1/2
= 58.1 Hz; 4: Δν1/2 = 48.6 Hz) in their 31P{1H} NMR spectra
at room temperature both at ca. δ 60 ppm, suggesting slow
exchange of Pax and Peq atoms on NMR time scale facilitated by
Berry pseudorotation.23,24 The broadness of the signals is likely
due to the close proximity of the exchange process to the
coalescence temperature, below which the pseudorotation
would be completely frozen out into two distinct signals.
This is clearly demonstrated in complex 5, which features two
very broad signals in the 31P{1H} NMR spectrum at room
temperature at the same measuring frequency (161.9 MHz) at
δ 51.5 (Δν1/2 = 1503 Hz) and δ 67.9 (Δν1/2 = 1503 Hz),
indicative of frozen out Berry pseudorotation and a static
structure. The 31P{1H} NMR spectrum of 5 at ambient
1
The H NMR spectra (400 MHz) of 3 and 4 at room
A
temperature exhibit two sets of diastereotopic signals (CH3
B
and CH3 ) for the methyl groups of the dmpe ligands, in
accordance with the fluxional picture observed in the 31P{1H}
NMR spectra (complex 4 exhibits an additional signal for the
Si−CH3 group). Due to the frozen out Berry pseudorotation at
ambient temperatures in 5, four diastereotopic CH3 groups
A
B
C
D
(CH3 , CH3 , CH3 , CH3 ) are observed as expected (at
measuring frequency of 400 MHz). These four signals also
coalesce into two bands at higher temperatures (T ≥ 309 K),
resulting in spectra resembling those of 3 and 4 at room
temperature, in accordance with the observations in the VT
31P{1H} spectra, also confirming pseudorotation at higher
temperatures.27 A key feature in complex 5 is the presence of
the Si-H signal, which at room temperature is not clearly
distinguishable and obscured by aromatic resonance signals.
1
The signal could unambiguously be located in the H NMR
spectrum by use of a 2D (1H, 29Si) heteronuclear shift
correlation experiment at δ 6.85 ppm, providing additional
evidence of its connectivity to Si. At higher temperatures (T ≥
C
dx.doi.org/10.1021/ja402480v | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
309 K), the Si−H signal is clearly visible as a quintet (δ 6.85
of all three compounds reveal slightly distorted trigonal-
bipyramidal coordination geometries around the iron centers
(3: τ = 0.87; 4: τ = 0.82; 5: τ = 0.95)30 with the NHSi ligand
occupying one of the equatorial coordination sites in each case.
This is the typically preferred position for π-accepting ligands in
a trigonal-bipyramidal geometry, which provides some evidence
for the π-accepting ability of the NHSi ligand. The structure
solution of complex 5 was complicated by substantial statistical
disorder over the dmpe bridges, which could, however, be
resolved in a satisfactory fashion by solving (number of
restraints = 484) for the coexistence of two stereoisomers. The
origin of this disorder is likely caused by the increased energy
barrier of pseudorotation observed spectroscopically, which
resulted during crystallization in the formation of two
coexisting stereoisomers, in a ratio of 1:1. Figures 3 − 5
3
ppm, JHP = 9.9 Hz), the quintet multiplicity deriving from
coupling to the phosphorus atoms rendered equivalent at this
temperature by the ensuing pseudorotation vida supra (Figure
2).
Figure 2. 1H NMR spectrum (400 MHz) of 5 in C6D6 at 340 K in the
aromatic region with an enlargement of the Si−H resonance signal
3
showing its quintet multiplicity derived from JPH coupling.
The 29Si{1H} NMR shifts of complexes 3−5 were also
recorded, and in each case feature a quintet multiplicity,
Figure 3. ORTEP representation of the molecular structure of 3.
Hydrogen atoms are omitted for clarity, thermal ellipsoids set at 30%
probability. Selected distances [Å] and angles [deg]: Fe1−P1
2.166(2), Fe1−P2 2.156(2), Fe1−P3 2.153(3), Fe1−P4 2.148(3);
P1−Fe1−P2 171.48(10), P1−Fe1−P3 93.22(11), P2−Fe1−P4
92.89(11), P4−Fe1−P3 122.45(11), P4−Fe1−Si1 118.46(10), P3−
Fe1−Si1 119.09(10), P2−Fe1−Si1 93.81(9), P1−Fe1−Si1 94.67(9).
2
resulting from a JSiP coupling to four equivalent P atoms
(rendered equivalent by Berry pseudorotation) on NMR time
scale for 29Si (Table 1).
Table 1. 29Si{1H} NMR Data and Hammett Constants for
Complexes 3−5
29Si{1H} NMR δ
value (ppm)
2JSiP coupling
constant (Hz)
Hammett
constant (σp)
complex, (Si−X)
3 (X = Cl)
4 (X = CH3)
5 (X = H)
43.1
102.5
63.6
12.1
16.1
23.7
0.23
−0.17
0
Inspection of the 29Si{1H} NMR chemical shifts (δ values) of
complexes 3−5 reveals a negative linear relationship between
chemical shift and Hammett constant (σp)28 of the substituents
that is evidence that inductive and resonance effects of the
substituents are a factor in the 29Si chemical shift position. The
same trend was observed for an analogous series of TiII NHSi
complexes, bearing the same substituents on Si,16 and can be
explained on the basis of the Bent rule.29 The shielding of the
silicon nucleus in 3 results from a concentration of p-character
in the Si−Cl bond, which concomitantly results in increased s-
character in the contribution of the silicon atom of the iron
silicon bond. This increase in the s-character results in a
shielding effect and the lower chemical shift value. Conversely,
in the case of 4 the exact opposite can be expected on the basis
of the Bent rule, and a lower s-character in the contribution of
the silicon atom in the silicon iron bond is predicted, resulting
in deshielding and a larger chemical shift. This was confirmed
on the basis of bonding analysis of the Si−Fe bond (vide infra).
Suitable crystals for X-ray diffraction analyses of 3−5 were
obtained by slow cooling of saturated solutions (3: Et2O; 4, 5:
n-hexane) to −30 °C for several days. The molecular structures
Figure 4. ORTEP representation of the molecular structure of 5. All
carbon atoms on dmpe ligands and H atoms omitted for clarity
(except H1), thermal ellipsoids set at 30% probability. Selected
distances [Å] and angles [deg]: Fe1−P1 2.156(9), Fe1−P2 2.123(12),
Fe1−P3 2.161(18), Fe1−P4 2.193(8); P1−Fe1−P2 88.7(3), P2−
Fe1−P3 91.7(4), P1−Fe1−P3 179.6(5), P2−Fe1−P4 113.0(4), P2−
Fe1−Si1 123.9(4), P1−Fe1−Si1 86.5(2).
show the molecular structures of complexes 3 and 5 in the solid
state (for the structure of 4 see the Supporting Information).
Figure 5 shows the resolved disorder over the dmpe bridges
resulting from the coexistence of two stereoisomers.
The Fe−Si distances for the three complexes range from
2.1634(9) Å in 3 to 2.200(2) Å in 4, with complex 5 exhibiting
a value of 2.184(2) Å (Table 2). The observed Fe−Si lengths
correlate perfectly with the Hammett constants (σp) associated
D
dx.doi.org/10.1021/ja402480v | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 5. ORTEP representations of complex 5 showing the disorder over the dmpe bridges, resolved as two stereoisomers. The two stereoisomers
were found in approximately 1:1 ratio, accounting for the disorder. Thermal ellipsoids set at 30% probability, H atoms and all substituents on Si
omitted for clarity.
a
Table 2. Selected Distances for the Complexes 3−5 )
Table 3. Comparison of Wiberg Bond Indices (WBI) Values
and the Calculated Fe−Si Distances of Complexes 3−5
a
d(Fe−Si)/Å
d(Si−X)/Å
d(Si−N1)/Å
d(Si−N2)/Å
NBO charge
3
4
5
2.1634(9)
2.200(2)
2.184(2)
2.2281(11)
1.933(9)
1.920(2)
1.954(7)
1.930(6)
1.931(2)
1.948(7)
1.915(7)
b
complex
Fe
Si
WBI
Fe−Si distances (Å)
b
1.50(2)
3
4
5
−2.205
−2.180
−2.198
+1.462
+1.531
+1.237
1.1934
1.1420
1.1585
2.1794 (2.1634)
2.2259 (2.184)
2.1976 (2.200)
a
b
X = Cl (3), CH3 (4), H (5. The H atom attached to Si in 5 could
not be located experimentally and was constrained at a distance of
1.50(2) Å.
a
Level of theory: B3LYP/6-31G(d) for C, H, N, Cl and Si; LANL2DZ
for Fe. Experimental values by X-ray diffraction analysis in
b
with the substituents on the SiII centers in analogy to the trend
observed with the respective 29Si NMR shifts of 3−5 vide supra.
This is an additional indication of the influence of the
substituent on the nature of the Fe−Si bond. Furthermore,
these bond lengths are slightly shorter when compared to the
only two existing Fe−NHSi complexes reported to date: 2.196
Å (VII) and 2.237 Å (VIII) (Chart 2). This observation is
likely due to the increased electron density at the Fe centers in
3−5 resulting from electron pushing alkyl phosphane ligands,
as opposed to good σ-donor and additionally strong π-acceptor
ability of the CO ligands in VII and VIII, facilitating stronger π-
back-donation to the SiII centers and concomitant bond length
contraction. The Fe−Si distances are in fact comparable to that
of the donor-free silylene complex [(η5-C5Me5)2Fe(CO)-
(SiMe3)(SiMes2)] (Mes =2,4,6-Me3-C6H2) by Tobita and
co-workers, where a formal double bonding interaction
between iron and silicon is observed,31 suggesting double
bond character in 3−5.
DFT Calculations. DFT calculations of the complexes 3−
5,32 including Natural Bond Orbital (NBO) analysis, were also
performed particularly to elucidate the nature of the bonding
interaction between the Si and Fe centers. A reasonable
agreement exists for the Fe−Si bond lengths found in the
geometry optimized structures with those observed by X-ray
structure determinations (3: 2.179 Å; 4: 2.226 Å; 5: 2.198 Å). A
summary of some pertinent metrical parameters for the DFT
calculations of 3−5 is presented in Table 3.
parentheses.
a
Table 4. NBO Analysis of the Fe−Si Bond
bond polarization
(%)
s-character
(%)
p-character
(%)
d-character
(%)
complex
3
Fe (33.79)
Si (66.21)
Fe (37.02)
Si (62.98)
Fe (36.54)
Si (63.46)
22.30
84.52
22.18
61.49
21.48
64.94
69.28
15.48
67.55
38.46
66.93
35.05
8.51
0.01
4
5
10.28
0.05
12.22
0.02
a
Level of theory: B3LYP/6-31G(d) for C, H, N, Cl and Si; LANL2DZ
for Fe.
donation from the Fe0 center to SiII. The calculated Wiberg
bond index (WBI)33 of all three complexes is >1, which is clear
evidence of multiple bonding character in the Fe−Si bonds.34
Figure 6 shows the distribution of the energies and shapes of
the frontier MOs of complex 3, which were found to be very
similar in complex 4.
Inspection of the calculated MOs for complexes 3 and 4
reveals that the HOMO has an asymmetric distribution of
electron density resulting from the π-back-donation from the
Fe0→SiII while the HOMO-1 and HOMO-2 MOs also reflect a
potential π-interaction between the Fe and Si centers, since
their shape reveals somewhat polarized metal d-orbitals. The
presence of the electron pushing alkylphosphane ligands
enhances the electron density at the iron center, and facilitates
π-back-bonding to silicon. This is in contrast to other examples
of NHSi complexes bearing the {Fe(CO)4} fragment, where
the π-acidity of the CO ligands should result in considerable
The calculated NBO charges on the Fe and Si atoms in
complexes 3 and 4 are comparable to each other within narrow
limits, and point to highly polarized Fe−Si bonds in both
complexes (Table 4). The NBO charge on Si in complex 5 is
slightly less positive, possibly as a result of enhanced π-back-
E
dx.doi.org/10.1021/ja402480v | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 6. Calculated boundary surfaces of some key frontier MOs and their respective energies of complex 3. Complex 4 was found to exhibit
analogous frontier MOs.
Figure 7. Calculated boundary surfaces of some pertinent frontier MOs and their energies of complex 5. The ligand facilitates better π-back-donation
from Fe0→SiII seen in the HOMO compared to that of 3 and 4.
electron depletion, although DFT calculations on these systems
have not been reported. Finally, the HOMO-5 reflects the
SiII→Fe0 σ-donation, which is also heavily polarized toward the
SiII center. The LUMO in complexes 3 and 4 is centered on the
phenyl ring of the NHSi ligand. Complex 5 on the other hand
exhibits slightly different frontier orbitals in comparison to the
other two complexes, most likely due to the small size of the H
substituent on the SiII center (Figure 7). Of particular interest is
the HOMO orbital, which resembles a more symmetric π-back-
donation from the Fe0→SiII (Figure 8). This can account for
the reduced positive charge on the SiII center, found in the
NBO analysis. Moreover, this subtle difference represented by
the HOMO in 5 can also account for the earlier observation of
a higher energy barrier of Berry pseudorotation observed
spectroscopically, vide supra. A better Fe0→SiII π-back-bond
would result in electron depletion at the Fe0 center, which
might force stronger coordination of the electron pushing
alkylphosphane ligands, accounting for their increased rigidity
in solution compared to complexes 3 and 4. As with complexes
3 and 4, the HOMO-5 is seen to exhibit the SiII→Fe0 σ-
F
dx.doi.org/10.1021/ja402480v | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
donation, which is also polarized toward the SiII center (Figure
6), with the LUMO centered on the phenyl ring of the NHSi
ligand.
Table 5. Probing the Catalytic Ability of Complex 5 for
Hydrosilylation of Ketones
Catalytic Hydrosilylation of Ketones with 5. Over the
past few years, iron catalysis has received considerable attention
as an alternative to that by precious and toxic metals (e.g., Rh,
Ir, Ru) owing to its abundance and low cost.35−38 In this regard,
excellent performances with Fe precatalysts have been
demonstrated in redox chemistry. Moreover, it has been
shown that the ligand sphere is of crucial importance for
achieving excellent catalytic activities and selectivities. Con-
sequently, we were interested in the evaluation of the catalytic
abilities of complex 5 in ketone hydrosilylation chemistry,
particularly since no studies have been reported on catalytically
active iron−silylene complexes to date. Moreover, no
previously reported catalytic studies mediated by a reactive
Si(II) hydride have been reported, further precipitating our
interest. In the presence of catalytic amounts of 5 (5 mol %), a
variety of ketones bearing different steric and electronic
properties (7−13) were converted, nearly in all cases in
excellent yields to the corresponding alcohols (7a−13a)
applying (EtO)3SiH as a hydridosilane source (Table 5). In
one case we even tested the ability of 5 to perform the catalytic
hydrosilylation of ketone 9 at ambient temperatures and found
that quantitative conversion is even observed without heating.
These results are at least comparable to some modern
benchmark systems for other iron-based catalytic ketone
hydrosilylations.37
To gain insights into the mechanism affecting these
hydrosilylation reactions, the reaction of 5 with ketone 7, in
the absence of (EtO)3SiH, was carried out on preparative scale
in the hope of isolating the adduct [(dmpe)2HFe←{:Si←[:O
C(CH3)(napthyl)]L}] (14), which is a likely first step in the
catalytic process. Despite several efforts the product could not
be isolated, so the formation of the emerging product was
monitored and analyzed by 1H and 31P{1H} NMR spectroscopy
over time. Heating of a sample of 5 with a stoichiometric
amount of ketone 7 in toluene-d8 revealed the selective
formation of a new iron hydride species, featuring a broad
multiplet signal at δ −13.94 ppm, typical of an terminal iron
hydride, which is clear evidence of H migration from the silicon
to the iron center. The thermal stability of 5 was independently
checked in the absence of ketone. In fact, no H-migration from
Si to Fe occurs even upon prolonged heating, which supports
that the hydride transfer process is induced by ketone
coordination to the silicon(II) center. Moreover, monitoring
the reaction by 31P{1H} NMR spectra over time revealed the
selective formation of two triplet signals at δ 62.2 and 75.7 ppm
(2JPP = 27 Hz) along with concomitant consumption of 5.
These data suggest nonrigid stereochemistry, or fluxionality of
the formed iron hydride species 14 in accordance with previous
observations for some cis-configured Fe−hydrido complexes.39
Attempts at obtaining a 29Si{1H} NMR spectrum of the
emerging product were unsuccessful. Nevertheless, these data
provide some evidence for the initial formation of a ketone
adduct with concomitant 1,2-H migration from Si to Fe, which
is likely the activation step in the catalytic process. Similar
observations for Ru-based systems featuring Schrock or Fischer
type silylenes had been observed very recently by Tilley and co-
workers40 and Tobita and co-workers41 for similar stoichio-
metric reactions with ketones, further lending supporting this
hypothesis.
a
Reaction conditions: precatalyst (5.0 mol %), substrate (0.16 mmol),
b
(EtO)3SiH (0.24 mmol), 70 °C, 24 h. Yield determined by GC-MS,
using n-dodecane as internal standard. Catalytic reaction repeated
with use of precatalyst (5.0 mol %), substrate (0.16 mmol), (EtO)3SiH
(0.24 mmol), 25 °C, 24 h.
c
Moreover, the appearance of the proposed intermediate 14 is
supported by DFT calculations where it was found to be a
minimum on the potential energy hyper surface (Figure 8).
The geometry optimized structure of the simplified model of
14′ reveals a distorted octahedral coordination geometry
around the iron center with a terminal hydride ligand,
originating from the Si center in 5. The oxygen atom of the
ketone is coordinated to the silicon center and features a Si−O
bond length of 1.6427 Å and a C−O bond distance of 1.4495 Å.
The latter parameter indicates substantial elongation from a
CO double bond in the free ketone to single bond character,
which is the likely source of the CO bond activation. The
Fe−Si bond distance of 2.342 Å is notably longer than is the
calculated and experimentally determined value in complex 5
(Table 6), indicating Fe−Si single bond character. Noteworthy
is the substantial increase in the NBO charge on the silicon
atom in 14′ compared to 5 (Table 6), which suggests 14 can be
G
dx.doi.org/10.1021/ja402480v | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
ketone to the silicon(II) center results in 1,2-hydride migration
to iron affording the betaine-like silyliumylidene−Fe0 hydride
complex featuring an activated CO bond of the ketone. The
reaction with the (EtO)3SiH likely follows this activation step
affording the desired silylated product and reformation of 5
(Scheme 3). The details of the last step of the proposed
sequence are currently under investigation.
SUMMARY AND CONCLUSIONS
■
A facile entry to the first electron-rich iron NHSi complexes has
been reported. This includes straightforward access to an
unprecedented SiII hydride stabilized by an electron-rich {FeP4}
fragment (5) by Cl→H exchange. All complexes show multiple
bonding character between the iron and the silicon atom of the
NHSi ligand, enhanced by the electron pushing alkylphosphane
donor ligands. This is particularly true in the case of the SiII−
hydride complex 5, featuring a classical π-back-bond. Moreover,
the first catalytic study of any iron-based NHSi complex or any
SiII hydride complex has been reported employing complex 5 as
a precatalyst. In this context 5 was found to effectively catalyze
the hydrosilylation of a variety of ketones. The NHSi ligand
appears to play a key role in this process, and the first step is
likely a ketone-mediated 1,2-H migration from SiII to Fe0.
These findings indicate for the first time that SiII hydride
complexes might emerge as a new class of catalysts, by taking
advantage of the reactive SiII center in substrate activation.
Further experimental and theoretical studies on this and other
potential catalytic processes mediated by 5 are currently
underway.
Figure 8. Geometry optimized structure of a simplified model 14′,
using the aldehyde HC(O)Ph) as a likely intermediate in the
hydrosilylation of ketones. All H atoms except the Fe−H and the H
atom on the aldehyde were omitted for clarity.
Table 6. Comparison of Wiberg Bond Indices (WBI) Values
a
and the Calculated Fe−Si Distances of Complexes 5 and
14′
NBO charge
complex
Fe
Si
WBI
Fe−Si distances (Å)
5
−2.198
−2.026
+1.237
+1.935
1.1585
0.864
2.1976
2.342
14′
a
Level of theory: B3LYP/6-31G(d) for C, H, N, Cl and Si; LANL2DZ
for Fe.
considered a betaine-like silyliumylidene−Fe0 hydride complex
bearing a cationic NHSi (silyliumylidene) ligand,42 which is
stabilized by a dative coordination from the lone pairs on the
oxygen atom of the ketone.
Based on these collective findings, a possible mechanism for
the hydrosilylation of ketones mediated by 5 could follow a
scenario presented in Scheme 3. The coordination of the
EXPERIMENTAL SECTION
■
General Considerations. All experiments and manipulations were
carried out under a dry and oxygen-free atmosphere of nitrogen or
argon, using standard Schlenk techniques or in an MBraun inert
atmosphere drybox containing an atmosphere of purified nitrogen.
Scheme 3. Proposed Catalytic Cycle for the Hydrosilylation of Ketones by Complex 5
H
dx.doi.org/10.1021/ja402480v | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Solvents were dried by standard methods and stored over LiAlH4.
Prior to use all solvents were recondensed and degassed at least once
via a freeze−pump−thaw cycle. C6D6 was stirred over excess KC8 for
48 h, recondensed, and degassed once via a freeze−pump−thaw cycle
and stored over activated molecular sieves (4 Å) for use. Complex 1
was prepared by the reduction of [FeCl2(dmpe)2]43 according to a
published method by Filippou,19 while the N-heterocylic chlorosily-
lene (2) was prepared according to the method by Roesky and co-
workers.21 MeLi was bought from Acros chemical company as a 1.6 M
solution in hexane and used as received. Free dmpe and LiBHEt3 (1 M
solution in THF) were obtained from Sigma-Aldrich and used as
received. The NMR spectra were recorded on a Bruker AV 200 or 400
spectrometer. Concentrated solutions of samples in C6D6 were sealed
off in a Young-type NMR tube for measurements. The 1H and
13C{1H} were referenced to the residual solvent signals as internal
standards (δH 7.15 ; δc 128.0 ppm for C6D6). The 29Si{1H} NMR
spectra were referenced to TMS (tetramethylsilane) as an external
standard and 31P{1H} spectra were also externally calibrated with
H3PO4 in sealed capillaries. Abbreviations: s = singlet; m = multiplet;
br = broad, quint = quintet. Unambiguous signal assignments were
made by employing a combination of 2D NMR H,H−COSYand
H,C−COSY (HMQC, HMBC) experiments. The PCH2CH2P are
diastereotopic and indicated by X or Y as superscripts in the 1H NMR
spectra. In addition, the methyl groups on the phosphorus atoms of
the dmpe ligands are also diastereotopic and indicated by the
superscripts A, B, C, or D, respectively. Melting points were
determined on an electronic “Melting point tester” device from
BSGT company and are uncorrected. For this purpose samples were
sealed off in capillaries under nitrogen and heated slowly to observed
decomposition or melting. High resolution ESI mass spectra were
recorded on an Orbitrap LTQ XL of Thermo Scientific mass
spectrometer and the raw data evaluated using the X-calibur computer
program. In all cases the isotope distribution pattern of the [M + 1]+
signal was checked against theory and the value reported is the line of
highest intensity.
Synthesis of [(dmpe)2Fe(←:Si(Me)L)] (4). A Schlenk tube was
charged with 1 (0.100 g, 0.231 mmol) and 2 (0.068 g, 0.231 mmol)
and the two solids were rapidly stirred with cooling to −30 °C.
Precooled (to −30 °C) Et2O (25 mL) was added to the stirring solid
mixture rapidly via cannula. The reaction was allowed to stir for 2 h at
room temperature generating 3 in situ. The resulting red-orange
reaction solution was cooled to −5 °C and MeLi (0.15 mL, 1.6 M
solution) was added dropwise to the solution under stirring. An
immediate color change to crimson-red is noticed. The reaction was
taken out of the cooling bath and stirred at room temperature for a
further 1 h. The solvent was removed in vacuo affording a maroon
residue, which was extracted with hexane (25 mL). The red hexane
filtrate was concentrated to ca. 2 mL and cooled to −30 °C for 2 h. A
red solid was isolated by cannula filtration and dried in vacuo at room
temperature affording 4 (0.097 g, 67%). Crystals suitable for X-ray
diffraction analysis were obtained by dissolving a small portion of 4 in
hexane and slow cooling to −30 °C for several days; mp 153 − 156 °C
dec. 1H NMR (400 MHz, C6D6, 298 K): δ 1.04 ppm (s, 3H, 1 × Si−
CH3), 1.21 (s, Δν1/2 = 1.6 Hz, 18H, 2 × NC(CH3)3), 1.41 (br s Δν1/2
A
= 4.9 Hz, 12H, 4 × CH3 ), 1.65−1.70 (br m, 8H, 4 × PCHXHY + 4 ×
PCHXHY) (signal partially obscured by that at δ 1.69 ppm), 1.69 (br s
B
Δν1/2 = 5.9 Hz, 12H, 4 × CH3 ), 6.88−6.93 (m, 1H, C4−H, Ph),
6.96−7.02 (m, 2H, C3,5−H, Ph), 7.03−7.07 (m, 1H, C2 or 6−H, Ph),
7.34−7.39 (m, 1H, C2 or 6−H, Ph). 31P{1H} NMR (161.9 MHz, C6D6,
298 K): δ 59.1 ppm (br s, Δν1/2 = 48.6 Hz, 2 × dmpe). 13C{1H} NMR
(100.6 MHz, C6D6, 298 K): δ 25.1 ppm (m, 1C, Si−CH3), 26.7 (ps
B
A
quint, 4C, 4 × CH3 ), 28.3 (ps quint, 4C, 4 × CH3 ), 32.6 (s, 6C, 2 ×
NC(CH3)3), 35.2 (m, 4C, 4 × PCHXHY), 53.6 (s, 2C, 2 × C(CH3)3),
127.10 (s, 1C, 1 × C2 or 6−H, Ph), 127.11 (s, 1C, 1 × C4−H, Ph),
127.6 (s, 1C, 1 × C3 or 5−H, Ph), 128.7 (s, 1C, 1 × C3 or 5−H, Ph),
130.4 (s, 1C, 1 × C2 or 6−H, Ph), 135.6 (s, 1C, C1, Ph), 164.6 (s, 1C, 1
× NCN). 29Si{1H} NMR (79.5 MHz, C6D6, 298 K): δ 102.5 ppm
2
(quint, JSiP = 16.1 Hz). ESI-MS, m/z: Calcd for [C28H58N2SiFeP4 +
1]+ 631.2742. Found 631.2726.
Synthesis of [(dmpe)2Fe(←:Si(H)L)] (5). A Schlenk tube was
charged with 1 (0.300 g, 0.694 mmol) and 2 (0.210 g, 0.694 mmol)
and the two solids were rapidly stirred with cooling to −30 °C.
Precooled (to −30 °C) Et2O (50 mL) was added to the stirring solid
mixture rapidly via cannula. The reaction was allowed to stir for 3 h at
room temperature generating 3 in situ. The resulting red-orange
reaction solution was cooled to 0 °C and Li[BHEt3] (0.69 mL, 1.0 M
thf solution) was added dropwise to the solution under stirring. The
reaction was stirred at 0 °C for 20 min then at room temperature for 3
h. At this point in situ 31P{1H} NMR spectroscopy of the reaction
solution showed two new signals: at δ −9.2 ppm and a very broad
signal δ 60.6 ppm. The solvent was removed in vacuo affording a red-
Synthesis of [(dmpe)2Fe(←:Si(Cl)L)] (3). A Schlenk tube was
charged with 1 (0.100 g, 0.231 mmol) and 2 (0.068 g, 0.231 mmol).
The solids were rapidly stirred in a −30 °C bath, and 20 mL of diethyl
ether which was also precooled to −30 °C was added to the mixture
via cannula. The reaction was taken out of the bath and stirred at room
temperature for 2 h during which a dark red solution with the
formation of a small amount of colorless precipitate is observed. An in
situ 31P{1H} NMR spectrum of an aliquot of the reaction solution at
this point showed complete consumption of 1, the formation of free
PMe3 (δ −60 ppm) and a new broad signal at δ 60 ppm. The reaction
solution was filtered via cannula and the red supernatant solution
concentrated to ca. 1 mL and cooled to −78 °C. The formation of an
orange precipitate was observed, and the light orange supernatant was
separated by syringe from this solid. The orange solid was dried in
vacuo at room temperature for 1 h resulting in the highly air and
moisture sensitive complex 3 (0.121 g, 81%). Crystals suitable for X-
ray diffraction analysis were obtained by dissolving a small portion of 3
in diethyl ether and slow cooling to −30 °C for several days; mp 143−
1
brown residue. A combination of H, 31P{1H}, and 11B{1H} NMR
spectroscopy on a small portion of this residue showed it to be a 1:1
mixture of the desired 5, and Me3P→BEt3 (6) resulting from a cross
reaction of liberated PMe3 in the first step of the reaction, and BEt3 in
the second step. The brown residue was extracted into 50 mL of
hexane and filtered via cannula then the red-brown solution was
concentrated to incipience. It was cooled to −30 °C and a red-brown
solid was isolated by careful decantation of the mother liquor. A
second crop can be obtained by recooling the mother liquor to −30
°C and decanting the mother liquor, affording 5 as a red-brown solid
(0.377 g, 89%). Crystals suitable for X-ray diffraction analysis were
obtained by dissolving a small portion of 4 in hexane and slow cooling
to −30 °C for several days; mp 131−133 °C dec. 1H NMR (200 MHz,
C6D6, 298 K): δ 1.27 ppm (s, Δν1/2 = 5.6 Hz, 18H, 2 × NC(CH3)3),
1
146 °C dec. H NMR (400 MHz, C6D6, 298 K): δ 1.31 ppm (br s
A
Δν1/2 = 7.1 Hz, 12H, 4 × CH3 ), 1.33 (s, Δν1/2 = 1.6 Hz, 18H, 2 ×
NC(CH3)3), 1.60 (m, 4H, 4 × PCHXHY), 1.75 (br s Δν1/2 = 7.1 Hz,
B
12H, 4 × CH3 ), 1.77 (m, 4H, 4 × PCHXHY) (this signal partially
obscured by that at δ 1.75 ppm), 6.84−6.91 (m, 1H, C4−H, Ph),
6.91−7.00 (m, 2H, C3,5−H, Ph), 7.20−7.27 (m, 2H, C2,6−H, Ph).
31P{1H} NMR (161.9 MHz, C6D6, 298 K): δ 59.8 ppm (br s, Δν1/2
=
A
1.44 (br, Δν1/2 = 18.6 Hz, 6H, 2 × CH3 ), 1.55 (br, Δν1/2 = 24.9 Hz,
58.1 Hz, 2 × dmpe). 13C{1H} NMR (100.6 MHz, C6D6, 298 K): δ
B
8H, PCH2), 1.65 (br, Δν1/2 = 24.9 Hz, 6H, 4 × CH3 ), 6.88 (br m,
B
A
2H, C4−H, Ph + 1 × Si−H signal located by 2D 29Si,H Cosy), 6.97 (br
m, 2H, C3,5−H, Ph), 7.12 (m, 1H, C2 or 6−H, Ph), 7.31 (m, 1H,
25.7 ppm (ps quint, 4C, 4 × CH3 ), 27.7 (ps quint, 4C, 4 × CH3 ),
n
32.5 (s, 6C, 2 × NC(CH3)3), 34.7 (quint, JP,C = 13.8 Hz, 4C, 4 ×
PCHXHY), 54.4 (s, 2C, 2 × C(CH3)3), 126.6 (s, 1C, 1 × C2 or 6−H,
Ph), 127.5 (s, 1C, 1 × C3 or 5−H, Ph), 127.52 (s, 1C, 1 × C4−H, Ph),
129.1(s, 1C, 1 × C3 or 5−H, Ph), 129.9 (s, 1C, 1 × C2 or 6−H, Ph),
134.4 (s, 1C, C1, Ph), 170.5 (s, 1C, 1 × NCN). 29Si{1H} NMR (79.5
MHz, C6D6, 298 K): δ 43.1 ppm (quint, 2JSiP = 12.1 Hz). ESI-MS, m/
z: Calcd for [C27H55N2SiFeP4Cl + 1]+ 651.2196. Found 651.2209.
1
C2 or 6−H, Ph). H NMR (400 MHz, C6D6, 298 K): δ 1.27 ppm (s,
A
Δν1/2 = 2.7 Hz, 18H, 2 × NC(CH3)3), 1.41 (br m, 6H, 2 × CH3 ),
B
1.46 (br m, 6H, 2 × CH3 ), 1.50−1.65 (br m, 8H, 2 × PCH2CH2P)
(signal partially obscured by that at δ 1.65 ppm), 1.65 (br m, 6H, 4 ×
C
CH3 ), 1.69 (br m, 6H, 4 × CH3D), 6.84−6.94 (m, 2H, C4−H, Ph + 1
× Si−H signal located by 2D 29Si,H COSY), 6.96−7.02 (m, 2H, C3,5−
I
dx.doi.org/10.1021/ja402480v | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
H, Ph), 7.12 (m, 1H, C2 or 6−H, Ph), 7.28−7.34 (m, 1H, C2 or 6−H,
Cartesian coordinates, NBO analyses, and optimized structures
and additional MOs of 3−5 and model complex 14′. This
material is available free of charge via the Internet at http://
pubs.acs.org.
1
Ph). H NMR (400 MHz, C6D6, 309 K): δ 1.27 ppm (s, Δν1/2 = 2.7
Hz, 18H, 2 × NC(CH3)3), 1.42 (br m, 12H, 2 × CH3A+B), 1.50−1.65
(br m, 8H, 2 × PCH2CH2P) (signal partially obscured by that at δ 1.65
ppm), 1.65 (br m, 12H, 4 × CH3C+D), 6.80−6.94 (m, 2H, C4−H, Ph +
1 × Si−H), 6.96−7.02 (m, 2H, C3,5−H, Ph), 7.12 (m, 1H, C2 or 6−H,
1
Ph), 7.28−7.34 (m, 1H, C2 or 6−H, Ph). H NMR (400 MHz, C6D6,
AUTHOR INFORMATION
■
330 K): δ 1.27 ppm (s, Δν1/2 = 2.7 Hz, 18H, 2 × NC(CH3)3), 1.42 (br
m, 12H, 2 × CH3A+B), 1.50−1.64 (br m, 8H, 2 × PCH2CH2P) (signal
partially obscured by that at δ 1.64 ppm), 1.64 (br m, 12H, 4 ×
Corresponding Author
matthias.driess@tu-berlin.de; shigeyoshi.inoue@tu-berlin.de
CH3C+D), 6.85 (quint, JHP = 9.9 Hz, 1H, Si−H), 6.88−6.96 (m, 2H,
3
Notes
C4−H, Ph), 6.98−7.04 (m, 2H, C3,5−H, Ph), 7.16 (m, 1H, C2 or 6−H,
The authors declare no competing financial interest.
1
Ph), 7.28−7.34 (m, 1H, C2 or 6−H, Ph). H NMR (400 MHz, C6D6,
340 K): δ 1.27 ppm (s, Δν1/2 = 2.7 Hz, 18H, 2 × NC(CH3)3), 1.42 (br
m, 12H, 2 × CH3A+B), 1.50−1.65 (br m, 8H, 2 × PCH2CH2P) (signal
partially obscured by that at δ 1.63 ppm), 1.63 (br m, 12H, 4 ×
ACKNOWLEDGMENTS
■
We are grateful to the Cluster of Excellence UniCat (sponsored
by the Deutsche Forschungsgemeinschaft and administered by
the TU Berlin) for financial support of this research. We thank
Dr. J. D. Epping for help concerning the NMR measurements
and Dr. S. Yao and G. Tan for help with the structure solutions
of 3 and 4.
CH3C+D), 6.84 (quint, JHP = 9.9 Hz, 1H, Si−H), 6.88−6.96 (m, 2H,
3
C4−H, Ph), 6.98−7.04 (m, 2H, C3,5−H, Ph), 7.16 (m, 1H, C2 or 6−H,
Ph), 7.28−7.36 (m, 1H, C2 or 6−H, Ph). 31P{1H} NMR (80.1 MHz,
C6D6, 298 K): δ 51.5 ppm (v br, Δν1/2 = 1567 Hz). 31P{1H} NMR
(161.9 MHz, C6D6, 298 K): δ 51.5 ppm (br s, Δν1/2 = 1503 Hz,
dmpe), 67.9 (br s, Δν1/2 = 1503 Hz, dmpe). 31P{1H} NMR (161.9
MHz, C6D6, 309 K): δ 59.5 ppm (br s, Δν1/2 = 2236 Hz, dmpe).
REFERENCES
31P{1H} NMR (161.9 MHz, C6D6, 219 K): δ 59.3 ppm (br s, Δν1/2
=
■
1297 Hz, dmpe). 31P{1H} NMR (161.9 MHz, C6D6, 330 K): δ 60.3
ppm (br s, Δν1/2 = 902 Hz, dmpe). 13C{1H} NMR (100.6 MHz, C6D6,
298 K): δ 24.0 ppm (br m, 2C, 2 × CH3D), 27.2 (br m, Δν1/2 = 122.9
Hz, 8C, 6 × CH3A+B+C), 32.9 (s, 6C, 2 × NC(CH3)3), 35.9 (br m, 4C,
(1) This was the first reported NHSi complex that is reported to be
thermolabile and no X-ray data or 29Si NMR data were reported, see:
Schmid, G.; Welz, E. Angew. Chem., Int. Ed. Engl. 1977, 16, 785.
(2) For a recent and comprehensive review on transition metal NHSi
complexes, see: Blom, B.; Stoelzel, M.; Driess, M. Chem.Eur. J. 2013,
19, 40.
4 × PCHXHY), 53.8 (s, 2C, 2 × C(CH3)3), 127.2 (s, 1C, 1 × C2 or 6
−
H, Ph), 127.2 (s, 1C, 1 × C4−H, Ph), 127.4 (s, 1C, 1 × C3 or 5−H, Ph),
128.0 (s, 1C, 1 × C3 or 5−H, Ph (obscured by solvent peak, located by
HMQC spectrum), 128.7 (s, 1C, 1 × C2 or 6−H, Ph), 136.0 (s, 1C, C1,
Ph), 164.8 (s, 1C, 1 × NCN). 29Si{1H} NMR (79.5 MHz, C6D6, 298
K): δ 63.6 ppm (quint, 2JSiP = 23.7 Hz). High resolution ESI-MS, m/z:
Calcd for [C27H56N2SiFeP4 + 1]+ 617.2585. Found 617.2581.
Procedure for the Hydrosilylation of Ketones. In a glovebox
(N2-atmosphere) a Schlenk flask was charged with complex 5 (0.036
mmol, 5.0 mol %), the corresponding ketone (0.72 mmol), and THF
(2.0 mL). The flask was removed from the glovebox and (EtO)3SiH
(0.79 mmol) was added by syringe. The reaction mixture was stirred in
a preheated oil bath at 100 °C for 24 h. The mixture was cooled in an
ice bath and was treated with dodecane (10 μL) as GC standard (for
GC analysis) and aqueous sodium hydroxide solution (1.0 mL) with
stirring. The reaction mixture was stirred for 60 min at room
temperature and was then extracted with diethyl ether (2 × 10.0 mL).
The combined organic layers were washed with brine, dried with
anhydrous Na2SO4, and filtered. n-Dodecane (internal standard) was
added and an aliquot was removed for GC analysis (30 m Rxi-5 ms
column, 40−300 °C). 7a: MS (EI) m/z 172 (6, M+), 154 (51), 129
(26), 76 (18), 63 (10), 40 (100). 8a: MS (EI) m/z 192 (65, M+), 174
(100), 159 (68), 149 (94), 133 (47), 119 (23), 105 (17), 91 (20), 77
(12), 43 (24). 9a: MS (EI) m/z 212 (100, M+), 197 (69), 169 (94),
154 (33), 138 (45), 123 (12), 109 (12), 95 (12), 77 (12), 65 (11), 42
(48). 10a: MS (EI) m/z 152 (27, M+), 137 (100), 107 (82), 94 (16),
77 (32), 65 (14), 50 (11), 43 (17). 11a: MS (EI) m/z 150 (8, M+),
107 (100), 79 (53). 12a: MS (EI) m/z 198 (41, M+), 180 (63), 165
(21), 119 (100), 105 (82), 91 (61), 77 (60), 65 (22), 51 (15). 13a:
MS (EI) m/z 128 (1, M+), 110 (19), 95 (21), 82 (65), 67 (47), 55
(58), 45 (100).
(3) (a) Meltzer, A.; Prasang, C.; Driess, M. J. Am. Chem. Soc. 2009,
̈
131, 7232. (b) Meltzer, A.; Inoue, S.; Prasang, C.; Driess, M. J. Am.
̈
Chem. Soc. 2010, 132, 3038. (c) Amoroso, D.; Haaf, M.; Yap, G. P. A.;
West, R.; Fogg, D. E. Organometallics 2002, 21, 534.
(4) Stoelzel, M.; Prasang, C.; Inoue, S.; Enthaler, S.; Driess, M.
̈
Angew. Chem., Int. Ed. 2012, 51, 399.
(5) Campion, B. K.; Heyn, R. H.; Tilley, T. D. J. Chem. Soc., Chem.
Commun. 1988, 278.
(6) Wang, W.; Inoue, S.; Enthaler, S.; Driess, M. Angew. Chem., Int.
Ed. 2012, 51, 6167.
(7) Furstner, A.; Krause, H.; Lehmann, C. W. Chem. Commun. 2001,
̈
2372.
(8) Zhang, M.; Liu, X.; Shi, C.; Ren, C.; Ding, Y.; Roesky, H. W. Z.
Anorg. Allg. Chem. 2008, 634, 1755.
(9) Bruck, A.; Gallego, D.; Wang, W.; Irran, E.; Driess, M.; Hartwig,
̈
J. F. Angew. Chem., Int. Ed. 2012, 51, 11478.
(10) See as selected examples: (a) Azhakar, R.; Ghadwal, R. S.;
Roesky, H. W.; Hey, J.; Stalke, D. Chem. Asian J. 2012, 7, 528.
(b) Schmedake, T. A.; Haaf, M.; Paradise, B. J.; Millevolte, A. J.;
Powell, D. R.; West, R. J. Organomet. Chem. 2001, 636, 17.
(c) Ramachandran, A.; Sarish, S. P.; Roesky, H. W.; Hey, J.; Stalke,
̌
D. Inorg. Chem. 2011, 50, 5039. (d) Tavcar, G.; Sen, S. S.; Azhakar, R.;
Thorn, A.; Roesky, H. W. Inorg. Chem. 2010, 49, 10199.
(11) Azhakar, R.; Ghadwal, R. S.; Roesky, H. W.; Wolf, H.; Stalke, D.
J. Am. Chem. Soc. 2012, 134, 2423.
(12) Neumann, E.; Pfaltz, A. Organometallics 2005, 24, 2008.
(13) Gehrhus, B.; Hitchcock, P. B.; Lappert, M. F.; Maciejewski, H.
Organometallics 1988, 17, 5599.
(14) (a) Perti, S. H. A.; Eikenberg, D.; Neumann, B.; Stammler, H.-
Theoretical Calculations. DFT calculations on the complexes 3,
4, and 5 were performed at B3LYP level using 6-31G(d) basis sets of
the GAUSSIAN-03 program package. Optimized structure of model
compounds were obtained without symmetry constraints.
G.; Jutzi, P. Organometallics 1999, 18, 2615.
(15) (a) Herrmann, W. A.; Harter, P.; Gstottmayr, C. W. K.; Bielert,
̈
̈
F.; Seeboth, N.; Sirsch, P. J. Organomet. Chem. 2002, 649, 141.
(b) Zeller, A.; Bielert, F.; Harter, P.; Herrmann, W. A.; Strassner, T. J.
̈
Organomet. Chem. 2005, 690, 3292.
(16) Blom, B.; Driess, M.; Gallego, D.; Inoue, S. Chem.Eur. J. 2012,
ASSOCIATED CONTENT
■
18, 13355.
S
* Supporting Information
(17) Yang, W.; Fu, H.; Wang, H.; Chen, M.; Ding, Y.; Roesky, H. W.;
Jana, A. Inorg. Chem. 2009, 48, 5058.
(18) Schmedake, T. A.; Haaf, M.; Paradise, B. J.; Millevolte, A. J.;
Details on X-ray crystal structure analyses (including a graphical
representation of 4), details of the NMR experiment of 14,
selected NMR and HR-ESI-MS spectra of 3−5 as well as
Powell, D. R.; West, R. J. Organomet. Chem. 2001, 636, 17.
J
dx.doi.org/10.1021/ja402480v | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(19) Complex 1 was first prepared in the laboratory of Prof. Dr. A. C.
Filippou (University of Bonn), by the first author of this publication in
2009. For the synthesis and full characterization of 1 and its reactivity
to germylenes and stannylenes of the type RECl (E = Ge or Sn, R =
2,6-Mes2-C6H3 or 2,6-Trip2-C6H3; Mes = 2,4,6-Me3-C6H2, Trip =
2,4,6-iPr3-C6H2) see: Blom, B., Reactivity of Ylenes at Late Transition
Metal Centers, Cuvillier Verlag, 2011, Doctoral thesis, University of
Bonn.
(20) By “silylene complexes” we refer to all types of silylene
complexes including the classical Fischer or Schrock type silylene
complexes (LnFeSiR2), base stabilized silylene complexes (LnFe
Si←:DR2) or NHSi complexes of the type LnFe←:NHSi.
(21) (a) So, C.-W.; Roesky, H. W.; Magull, J.; Oswald, R. B. Angew.
Chem., Int. Ed. 2006, 45, 3948. (b) Sen, S. S.; Roesky, H. W.; Stern, D.;
Henn, J.; Stalke, D. J. Am. Chem. Soc. 2010, 132, 1123.
(22) (a) See also ref 4. (b) Jana, A.; Leusser, D.; Objartel, I.; Roesky,
H. W.; Stalke, D. Dalton Trans. 2011, 40, 5458. (c) Rodriguez, R.;
Gau, D.; Contie, Y.; Kato, T.; Saffon-Merceron, N.; Baceiredo, A.
Angew. Chem., Int. Ed. 2011, 50, 11492 and references cited therein.
(d) Abraham, M. Y.; Wang, Y.; Xie, Y.; Wei, P.; Schaefer, H. F., III;
Schleyer, P. v. R.; Robinson, G. H. J. Am. Chem. Soc. 2011, 133, 8874.
(e) Al-Rafia, S. M. I.; Malcolm, A. C.; McDonald, R.; Ferguson, M. J.;
Rivard, E. Chem. Commun. 2012, 48, 1308 and references cited
therein.
Le Paih, J.; Zani, L. Chem. Rev. 2004, 104, 6217. (h) Junge, K.;
Schroder, K.; Beller, M. Chem. Commun. 2011, 47, 4849. (i) Bauer, E.
̈
B. Curr. Org. Synth. 2008, 12, 1341. (j) Czaplik, W. M.; Mayer, M.;
Cvengros, J.; Jacobi von Wangelin, A. ChemSusChem 2009, 2, 396.
(k) Enthaler, S.; Junge, K.; Beller, M. Angew. Chem., Int. Ed. 2008, 47,
3317.
(36) See for instance: (a) Zhang, M.; Zhang, A. Appl. Organomet.
Chem. 2010, 24, 751. (b) Darwish, M.; Wills, M. Catal. Sci. Technol.
2012, 2, 243. (c) Shaikh, N. S.; Enthaler, S.; Junge, K.; Beller, M.
Angew. Chem., Int. Ed. 2008, 47, 2497.
(37) Selected references for hydrosilylation of alkynes and ketones
with iron catalysts: (a) Bart, S. C.; Lobkovsky, E.; Chirik, P. J. J. Am.
Chem. Soc. 2004, 126, 13794. (b) Haberberger, M.; Irran, E.; Enthaler,
S. Eur. J. Inorg. Chem. 2011, 2797. (c) Enthaler, S.; Haberberger, M.;
Irran, E. Chem. Asian J. 2011, 6, 1613. (d) Belger, C.; Plietker, B.
Chem. Commun. 2012, 48, 5419. (e) For a recent review see: Junge,
K.; Schroder, K.; Beller, M. Chem. Commun. 2011, 47, 4849 and
̈
references therin. (f) Bezier, D.; Sortais, J.-D.; Darcel, C. Adv. Synth.
Catal. 2013, 355, 19 and references therin.
(38) References deoxygenation of sulfoxides with iron catalysts:
(a) Enthaler, S. ChemCatChem 2011, 3, 666. (b) Cardoso, J. M. S.;
Royo, B. Chem. Commun. 2012, 48, 4944.
(39) Meakin, P.; Muetterties, E. L.; Jesson, J. P. J. Am. Chem. Soc.
1973, 95, 75.
(40) Fasulo, M. E.; Tilley, T. D. Organometallics 2012, 31, 5049.
(41) Ochiai, M.; Hashimoto, H.; Tobita, H. Organometallics 2012, 31,
527.
(42) For a recent example of a silyliumylidene, see: (a) Xiong, Y.;
Yao, S.; Inoue, S.; Irran, E.; Driess, M. Angew. Chem., Int. Ed. 2012, 51,
10074. (b) Inoue, S.; Epping, J. D.; Irran, E.; Driess, M. J. Am. Chem.
Soc. 2011, 133, 8514.
(23) (a) Berry, R. S. J. Chem. Phys. 1960, 32, 933. (b) Ugi, I.;
Marquarding, D.; Klusacek, K.; Gillepsie, P. Acc. Chem. Res. 1971, 4,
288.
(24) See as examples for similar complexes featuring similar
phenomena: (a) Hoffman, P. R.; Miller, J. S.; Ungermann, C. B.;
Caulton, K. G. J. Am. Chem. Soc. 1973, 95, 7902. (b) Ungermann, C.
B.; Caulton, K. G. J. Organomet. Chem. 1975, 94, C9. (c) Komiya, S.;
Akita, M.; Yoza, A.; Kasuga, N.; Fukuoka, A.; Kai, Y. J. Chem. Soc.,
Chem. Commun. 1993, 787. (d) Hirano, M.; Akita, M.; Morikita, T.;
Kubo, H.; Fukuoka, A.; Komiya, S. Dalton Trans. 1997, 3453.
(e) Matos, R. M.; Verkade, J. G. J. Braz. Chem. Soc. 2003, 14, 71. (f)
Reference 19.
(43) (a) Chatt, J.; Hayter, R. G. J. Chem. Soc. 1961, 5507. (b)
Reference 19.
(25) For an explanation on how measuring frequency affects the
NMR time scale see: Bryant, R. G. J. Chem. Educ. 1983, 60, 933.
(26) Wilson, S. T.; Coville, N. J.; Shapely, J. R.; Osborn, J. A. J. Am.
Chem. Soc. 1974, 96, 4038.
(27) A selection of NMR spectra of complexes 3−5 showing these
phenomena is available in the Supporting Information, along with
crystallographic and DFT data.
(28) Hansch, C.; Leo, A.; Taft, R. W. Chem. Rev. 1991, 91, 165.
(29) Bent, H. A. Chem. Rev. 1961, 61, 275.
(30) The τ value is a reflection of the distortion between a perfect
square pyramidal geometry (τ = 0) or a perfect trigonal bipyramidal
geometry (τ = 1). It can be calculated by using the formula τ = (a −
b)/60°, where a and b are the two largest angles. See: Anderson, A.
W.; Rao, T. N.; Reedijk, J.; v. Rijn, J.; Verschoor, G. C. J. Chem. Soc.,
Dalton Trans. 1984, 1349.
(31) Tobita, H.; Matsuda, A.; Hashimoto, H.; Ueno, K.; Ogino, H.
Angew. Chem., Int. Ed. 2004, 43, 221.
(32) Basis sets employed: B3LYP/6-31G(d) for C, H, N, Cl and Si;
LANL2DZ for Fe. Performed by using Gaussian 03, Revision E.01. See
Supporting Information for more full details concerning the DFT
calculations of 3−5.
(33) Wiberg, K. B. Tetrahedron 1968, 24, 1083.
(34) For a concise discussion on bond order and the WBI index for a
series of iron silylene and carbene complexes of Fischer and Schrock
type, see: Irwin, R. P. Organometallic Chemistry: Research Perspectives,
1st ed.; Nova Science Pub Inc.: Hauppauge, NY, 2008; p 292.
(35) See as examples: (a) Morris, R. H. Chem. Soc. Rev. 2009, 38,
2282. (b) Iron Catalysis in Organic Chemistry; Plietker, B., Ed.; Wiley-
VCH: Weinheim, Germany, 2008. (c) Gaillard, S.; Renaud, J.-L.
ChemSusChem 2008, 1, 505. (d) Correa, A.; Mancheno, O. G.; Bolm,
̃
C. Chem. Soc. Rev. 2008, 37, 1108. (e) Sherry, B. D.; Furstner, A. Acc.
̈
Chem. Res. 2008, 41, 1500. (f) Bullock, R. M. Angew. Chem. 2007, 119,
7504; Angew. Chem., Int. Ed. 2007, 46, 7360. (g) Bolm, C.; Legros, J.;
K
dx.doi.org/10.1021/ja402480v | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX