Peripheral mechanism of a carbonyl hydrosilylation catalysed by an SiNSi iron pincer complex

Article abstract of DOI:10.1039/c5sc02855h

Combined experimental and theoretical analysis of the carbonyl hydrosilylation catalysed by an iron(0) pincer complex reveals an unprecedented mechanism of action. The iron(0) complex is in fact a precatalyst that is converted into an iron(ii) catalyst through oxidative addition of a hydrosilane. Neither the hydrogen atom nor the silicon atom bound to the iron(ii) centre are subsequently transferred onto the carbonyl acceptor, instead remaining at the sterically inaccessible iron(ii) atom throughout the catalytic cycle. A series of labelling, crossover and competition experiments as well as the use of a silicon-stereogenic hydrosilane as a stereochemical probe suggest that the iron(ii) site is not directly involved in the hydrosilylation. Strikingly, it is the silyl ligand attached to the iron(ii) atom that acts as a Lewis acid for carbonyl activation in this catalysis. The whole catalytic process occurs on the periphery of the transition metal. Computation of the new peripheral as well as plausible alternative inner and outer sphere mechanisms support the experimental findings.

Full text of DOI:10.1039/c5sc02855h

Chemical
Science
View Article Online
View Journal
EDGE ARTICLE
Peripheral mechanism of a carbonyl hydrosilylation
catalysed by an SiNSi iron pincer complex†
Cite this: DOI: 10.1039/c5sc02855h
a
a
b
a
*
Toni T. Metsanen,‡ Daniel Gallego,‡ Tibor Szilvasi, Matthias Driess
¨
´
a
*
and Martin Oestreich
Combined experimental and theoretical analysis of the carbonyl hydrosilylation catalysed by an iron(0)
pincer complex reveals an unprecedented mechanism of action. The iron(0) complex is in fact a
precatalyst that is converted into an iron(II) catalyst through oxidative addition of a hydrosilane. Neither
the hydrogen atom nor the silicon atom bound to the iron(II) centre are subsequently transferred onto
the carbonyl acceptor, instead remaining at the sterically inaccessible iron(II) atom throughout the
catalytic cycle. A series of labelling, crossover and competition experiments as well as the use of a
silicon-stereogenic hydrosilane as a stereochemical probe suggest that the iron(II) site is not directly
involved in the hydrosilylation. Strikingly, it is the silyl ligand attached to the iron(II) atom that acts as a
Lewis acid for carbonyl activation in this catalysis. The whole catalytic process occurs on the periphery
of the transition metal. Computation of the new peripheral as well as plausible alternative inner and
outer sphere mechanisms support the experimental findings.
Received 3rd August 2015
Accepted 11th September 2015
DOI: 10.1039/c5sc02855h
www.rsc.org/chemicalscience
where both the substrate and the hydrosilane are directly in
contact with the metal or outer sphere7,8 where only one of the
Introduction
two is in contact with the metal centre. The proposed mecha-
nisms for iron-catalysed hydrosilylations range from inner
sphere mechanisms with s-bond-metathesis-type Si–H bond
cleavage at an iron–oxygen bond3k,3l to outer sphere mechanisms
with iron acting as a Lewis acid,3h either activating the hydro-
silane or the carbonyl group.
Driess and co-workers recently introduced silylenes as
s-donor ligands in iron-based catalysis, and iron(0) complexes 1
and 2 (Fig. 1) were applied to carbonyl hydrosilylation.9 Coop-
erativity between the iron(0) atom and the silicon(II) hydride in 1
was postulated to be relevant in the catalytic cycle.9a The SiNSi
iron(0) pincer complex 2 was, in turn, believed to be a pre-
catalyst9b,10 but a detailed mechanistic analysis remained chal-
lenging. We report here the disclosure of a unique mechanism
of a transition-metal-catalysed carbonyl hydrosilylation that
takes place neither inner nor outer sphere but on the periphery of
the metal centre without its direct involvement.
Iron-catalysed carbonyl hydrosilylation can be traced back to
seminal reports by Brunner1 but these contributions were
decades ahead of their time. With few exceptions, catalyst
development was focused on complexes of rare transition
metals while limited progress had been made involving
abundant transition metals. The pressing demand for
sustainable processes nally shied iron catalysis into the
limelight,2 and several iron-based catalysts for carbonyl
hydrosilylation with different ligand designs were introduced
in recent years.3 The advent of these new catalysts immediately
poses the question whether the mechanisms are similar to
those established for rare transition-metal complexes or totally
unprecedented.4 However, little detail is known about the
mechanisms of action of iron complexes in catalytic
hydrosilylation.
Mechanisms of transition-metal-catalysed hydrosilylations
exhibit a wide variety of modes of activation.5 However, the
known mechanisms are characterised as either inner sphere6
a
¨
¨
Institut fur Chemie, Technische Universitat Berlin, Straße des 17. Juni 115, 10623
Berlin, Germany. E-mail: matthias.driess@tu-berlin.de; martin.oestreich@tu-berlin.de
bDepartment of Inorganic and Analytical Chemistry, Budapest University of Technology
´
´
and Economics, Szent Gellert ter 4, 1111 Budapest, Hungary
† Electronic supplementary information (ESI) available: Experimental procedures
and
computational
details,
characterisation,
crystallographic
and
quantum-chemical calculation data as well as NMR spectra. CCDC 1416378. For
ESI and crystallographic data in CIF or other electronic format see DOI:
10.1039/c5sc02855h
Fig. 1 Iron(0) complexes 1 (ref. 9a) and 2 (ref. 9b) introduced by Driess
and co-workers.
‡ These authors contributed equally.
This journal is © The Royal Society of Chemistry 2015
Chem. Sci.

View Article Online
Chemical Science
Edge Article
Results and discussion
Optimisation and scope
The SiNSi pincer complex 2 (2.5 mol%) was found to catalyse the
hydrosilylation of various acetophenones 3a–3i with silane 4a at
elevated temperatures11 (Table 1, entries 1–9; see the ESI† for
the optimisation of the reaction conditions). Both electron-
donating (entries 1 and 3) as well as -withdrawing (entry 6)
substituents at the aryl group were tolerated with the exception
of a Et2N group in the para position (entry 2). The reaction was,
however, sensitive toward steric hindrance. Substituents in the
ortho position signicantly lowered the yield (entries 7–9), and a
2,6-disubstituted substrate did not react (3j, entry 10). Benzo-
phenone (3k, entry 11) reacted readily while propiophenone and
isobutyrophenone (3l and 3m, entries 12 and 13) afforded 18
and 16% yield, respectively. Hydrosilylation of cyclopropyl
substituted ketone (3n, entry 14) proceeded efficiently12 (see the
ESI† for full scope).
Isolation of the active catalyst
To gain insight into the mechanism, we investigated the reac-
tion between iron(0) complex 2 and hydrosilanes 4a–c (2 /
7a–c, Scheme 1). Heating at 70 ꢀC, a new set of distinct signals
Scheme 1 Identification of the catalytically active iron(II) complexes 7
from iron(0) precatalyst 2 and molecular structure of 7b. Hydrogen
atoms except for the iron hydride are omitted for clarity.
1
appeared in the H as well as in the 29Si and 31P NMR spectra.
The 1H NMR spectrum clearly indicated the formation of an
iron hydride, and detailed NMR analysis revealed that the
hydride was likely to be trans to the apical phosphine ligand.
The silyl group was assigned by 2D NMR experiments to be in
the equatorial position trans to the pyridine ligand. We then
obtained single crystals of 7b (Si ¼ Me2PhSi) suitable for X-ray
diffraction analysis, and that conrmed the molecular structure
deduced from the NMR analysis. The structure shows a dis-
torted octahedral iron(II) coordination environment. The
hydride was located tilted toward one of the silylene donor arms
deviated from the trans coordination to the Me3P ligand (P–Fe–
H1 170.20(5)ꢀ), a situation similar to that of known iron(II)
hydride pincer complexes.3h The Fe–Si bond distances
Table 1 Carbonyl hydrosilylation with precatalyst 2a
˚
2.1509(7)/2.1715(7) A for Fe–Si1/Si2 and 2.2986(8) for Fe–Si3 are
within the range of iron silylene and silyl complexes.9a,13 The
t-Bu groups encage the iron hydride with a Si1–Fe–Si2 angle far
from the linearity, 144.54(3)ꢀ.
To validate whether the thus formed iron(II) complex 7 is the
active catalyst, we measured the kinetic proles for the hydro-
silylation of 3a with hydrosilane 4a catalysed by 2 or 7a
(Scheme 2). Conversion with 2 was only 15% aer 1 h while the
reaction had reached 74% with 7a. The reaction with 7a
continued with signicantly higher rate reaching 86% at 4 h
compared to only 53% with 2. Aer 22 h, nearly full conversion
Entry
3
R1
R2
Yield of 5b (%)
1
2
3
4
5
6
7
8
9
3a
3b
3c
3d
3e
3f
3g
3h
3i
X ¼ MeO
X ¼ Et2N
X ¼ Me
X ¼ Br
Me
Me
Me
Me
Me
Me
Me
Me
Me
>99 (5a)
40 (5b)
82 (5c)
>99 (5d)
93 (5e)
95 (5f)
70 (5g)
70 (5h)
49 (5i)
X ¼ H
X ¼ CF3
X ¼ MeO
X ¼ Me
X ¼ Cl
10
11
12
13
14
3j
3k
3l
3m
3n
Mes
Ph
Ph
Ph
c-Pr
Me
Ph
Et
i-Pr
Me
0 (5j)
60 (5k)
18 (5l)
16 (5m)
>99 (5n)
a
Reactions were performed on 0.10 mmol scale employing precatalyst 2
(2.5 mol%) and (EtO)3SiH (4a 1.5 equiv.). b Average yield from two runs
determined by GLC-MS analysis and 1H NMR spectroscopy using
anisole as internal standard.
Scheme 2 Kinetic profiles of the iron(0) and iron(II) complexes 2 and
7a in hydrosilylation.
Chem. Sci.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Edge Article
Chemical Science
is obtained for both. The greater initial rate of the catalysis with
7a strongly supports the assignment of the iron(II) complex 7 as
the active catalyst.
Stoichiometric experiments: hydride transfer
With the iron(II) hydrides 7 in hand, we had a closer look at the
hydride transfer. Maintaining 7b and deuterium-labelled
hydrosilane 4b-d1 in THF at 70 ꢀC resulted in slow H/D
exchange, visible both at the silicon and iron atoms (Scheme 3,
le). The stoichiometric reaction between iron(II) hydride 7b,
hydrosilane 4b-d1 (>95% deuteration grade), and ketone 3e was
puzzling though (Scheme 3, right). Initially, the H : D ratio at
the methine position of silyl ether 8eb is nearly 50 : 50.
However, it quickly decreases to 36 : 64 at 25% conversion
within 6 hours and then gradually increases again, returning to
50 : 50 at full conversion aer a few days. Meanwhile, the cor-
responding reaction with partially deuterated 4b-d1 (ca. 50%
deuteration grade) yielded 8eb with little deuterium incorpo-
ration at 19% conversion (H/D ¼ 90 : 10). That ratio subse-
quently decreases to 78 : 22 to reach equilibrium aer 24 hours.
These results reveal that even though the hydride at the silicon
atom in 4 is exchanging with the iron-bound hydride in 7,
hydride transfer to the carbonyl carbon atom of 3 most likely
occurs from the hydrosilane 4 and not from complex 7.14 Also,
the reaction with partially deuterated hydrosilane indicates that
the kinetic isotope effect (KIE) of the hydride transfer is
signicant. Precise value of the KIE could not be measured due
to competing H/D exchanges (vide infra).
The possible H/D exchange at the methine position of silyl
ether 8eb was veried using 8eb-d1 (Scheme 4, top). Treatment
of 8eb-d1 with equimolar amounts of the iron(II) hydride 7b
indeed led to H/D scrambling. Conversely, no erosion of the
enantiomeric purity was seen when subjecting enantiopure silyl
ether (S)-8eb to the typical protocol (precatalyst 2 and hydro-
silane 4b generate catalyst 7b, Scheme 4, bottom). The cong-
urational stability of (S)-8eb suggests that the hydride transfer
itself is irreversible, and a concerted mechanism involving
frontside attack at the asymmetrically substituted carbon atom
is needed to explain the hydrogen atom exchange between the
catalyst and the product.
Scheme 4 H/D scrambling at the methine carbon atom.
These unusual scramblings were then investigated by DFT
calculations.15 Both were found to proceed via a silylene-assis-
ted concerted mechanism (9a‡ for Si–H and 10a‡ for C–H,
Scheme 5) where the hydride on the iron atom is rst shied to
the silicon atom of the adjacent donor-stabilised silylene ligand
forming a pentacoordinate silicon atom16 while the Si–H or C–H
bond interact with the now accessible iron centre. Both transi-
tion states are paired with their corresponding isomer where
the second silylene ligand accepts the hydride. Attempts to
locate the transient intermediates between the two degenerate
conformations were not successful.
Aer the transition state, the silylene-bound hydrogen atom
migrates back to the silicon and carbon atom, respectively. The
activation barriers of the scrambling reactions (18.5 kcal molꢁ1
for Si–H and 20.1 kcal molꢁ1 for C–H) are energetically acces-
sible under the reaction conditions.
Stoichiometric experiments: phosphine dissociation
To probe the potential lability of the phosphine, we performed
phosphine crossover experiments. Slow exchange of the
phosphine ligand with (CD3)3P (6-d9) was observed (Scheme 6,
Scheme 5 Silylene-assisted H/D scrambling at the hydrosilane silicon
Scheme 3 H/D scrambling at the silicon and iron atoms and identi- atom (9a‡, left) and the methine carbon atom (10a‡, right); Gibbs free
fication of the hydride source.
energies given in parentheses in kcal molꢁ1; Si ¼ Si(OEt)3.
This journal is © The Royal Society of Chemistry 2015
Chem. Sci.

View Article Online
Chemical Science
Edge Article
As is to be expected from the above observations, the hydride
complex 7b did not produce any silyl ether 8eb when reacted
stoichiometrically with ketone 3e in the absence of a hydro-
silane (Scheme 8, top le). What was fascinating though is that
the silyl ligand in 7b also remains untouched throughout the
catalysis: 7b derived from Me2PhSiH (4b) catalyses the hydro-
silylation of 3e with MePh2SiH (4c) with hardly any incorpora-
tion of the Me2PhSi moiety into the product; silyl ether 8ec
rather than 8eb is formed almost exclusively (Scheme 8, right).
This crossover experiment unequivocally proves that iron(II)
complexes 7 are the actual catalysts, originating from oxidative
addition of hydrosilanes 3 to the iron(0) complex 2; 2 is a pre-
catalyst. During the crossover experiment no changes in the
characteristic signals of complex 7b in the 1H and 31P NMR
spectra were detected. However, when the assumed inability of
7b and 4c to exchange their silyl groups was examined with
another control experiment (Scheme 8, bottom le), we
observed slow exchange with ca. 36% conversion of 7b to 7c
aer 24 h. The Me2PhSi/MePh2Si scrambling was accompanied
with formation of phosphine-dissociated, benzene-stabilised
compounds 11b and 11c. Only traces of Me2PhSiH (4b) were
observed, indicating that the exchange (7b to 7c) is in fact a side
product of the decomposition rather than simple scrambling of
the silyl groups.
Scheme 6 Phosphine scrambling and dissociation. Double-ended
arrows in 11b-d6 show 1H, 29Si HMQC NMR correlation. Gibbs free
energies given in parentheses in kcal molꢁ1 [Si ¼ SiMe2Ph].
top le). When 7b was subjected to repeated reux/freeze–
pump–thaw cycles, we detected a new iron hydride species that
was tentatively assigned as the expected phosphine-dissoci-
ated iron silyl hydride 11b-d6 with C6D6 as an h2-ligand
(Scheme 6, top right). The generation of this new iron
Hydrosilylation with a silicon-stereogenic hydrosilane
compound was accompanied by formation of disilane 12b and, With sufficient knowledge of the active catalyst, we decided to
with longer reaction times, decomposition into a complex analyse the stereochemical course at the silicon atom of the
mixture. The role of the disilane 12b in the formation of the reacting hydrosilane (Scheme 9).18 Catalyst 7b promoted the
complex 11b remains unexplained. Unfortunately, attempts to reaction between highly enantioenriched hydrosilane (SiS)-4d
isolate 11b-d6 were unsuccessful. Its generation in the pres- (e.r. > 95 : 5) and ketone 3e but conversion was slow as expected
ence of acetophenone (3e) did not lead to the formation of the from the data obtained with achiral triorganosilane 4b (see
silyl ether 8eb (see the ESI† for details), providing further Table S1, entry 12 in the ESI†). Aer 6 days, we were able to
evidence against the role of 11 as an intermediate in the
catalytic reaction. In fact, the dissociation of the Me3P (6) is
signicantly slower in the presence of ketone 3e. The forma-
tion 11 was also investigated computationally (Scheme 6,
bottom). Phosphine dissociation from 7a gives energetically
highly unfavoured intermediate cis-14a17 that readily coordi-
nates benzene to form adduct 11a.
It must be noted here that catalysis with Guan's related
iron(II) POCOP-pincer complex is thwarted by additional Me3P
(6), indicating dissociation of one of the phosphine ligands as
part of the catalytic cycle.3h When we added 25 mol% of Me3P (6,
10 equiv./catalyst) to the reaction mixture, the reaction was
unaffected (Scheme 7, cf. Table 1, entry 5).
Scheme 7 Effect of excess Me3P on the catalytic activity.
Scheme 8 Stoichiometric control and crossover experiments.
Chem. Sci.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Edge Article
Chemical Science
iron(II) centre in 7a (grey box), we did not locate any structure
resulting from direct insertion of the ketone C]O group into the
iron hydride 7a or the silylene ligands (not shown). Instead, we
were able to nd a minimum structure for the phosphine-
dissociated complex cis-14a (Scheme 10, le). In agreement with
the experiments, cis-14a is however signicantly higher in energy
(29.2 kcal molꢁ1 relative to 7a). The intermediate cis-14a readily
coordinates THF forming the adduct 17a. This intermediate is
however a resting state if not a “mechanistic dead-end”. Ketone
coordination to the iron centre of cis-14a gives intermediate 19oa
with activated carbonyl group. The catalytic cycle is closed by an
outer sphere concerted hydrosilane addition 20oa‡ to the ketone
Scheme
probe.
9 Silicon-stereogenic hydrosilane as a stereochemical
with an activation barrier of 33.7 kcal molꢁ1
.
Isomerisation of cis-14a to trans-14a was found to be strongly
disfavoured, but again, ketone coordination to trans-14a low-
ered the energy signicantly (Scheme 10, right). The following
hydride transfer from 23oa passes through 24oa‡ (30.9
kcal molꢁ1) to afford the iron alkoxide 25oa. The silylated
alcohol is released by an inner sphere silylation through 26oa‡
with an energy barrier of 34.9 kcal molꢁ1. An alternative inner
sphere mechanism could be a reductive elimination from the
intermediate 25oa. However, the energy barrier for the transi-
tion state 27oa‡ was found to be high, and the resulting iron(0)
complex 28 is energetically disfavoured. Recoordination of
phosphine 6 gives iron(0) complex 29 which oxidatively adds to
a silane 4a to form 7a.
isolate the silyl ether (SiR)-8ed in 31% yield; diastereoselection
was poor. The enantiomeric ratio of unreacted (SiS)-4d was
found to be unaffected. Subsequent reductive cleavage of the
Si–O bond in (SiR)-8ed (known to proceed with stereoretention
at silicon atom19) liberated (SiS)-4d with overall retention of the
stereochemistry at the silicon atom (e.r. > 95 : 5). Hence, the
hydrosilylation step involves frontside attack at the silicon
atom, and that makes a mechanism involving Lewis-acid acti-
vation of the hydrosilane unlikely.18
DFT calculation of conventional inner and outer sphere
mechanisms15
Peripheral mechanism: support from DFT calculations
Based on the combined ndings, we propose that the active
iron(II) catalyst 7 is generated from the iron(0) precatalyst 2 by In addition to the high energy barriers, neither outer nor inner
oxidative addition of hydrosilane 4 to the zero-valent iron atom sphere mechanisms give satisfactory ts to the experimental
(Scheme 1). As expected due to the steric congestion around the evidence. Hence, we looked for an adduct of 7a and 3o wherein
Scheme 10 Alternative mechanisms. Gibbs free energies given in parentheses in kcal molꢁ1 [Si ¼ Si(OEt)3].
This journal is © The Royal Society of Chemistry 2015
Chem. Sci.

View Article Online
Chemical Science
Edge Article
the silyl group would act as a Lewis acid (Scheme 11).8,18,20
Coordination of the ketone to the silyl group via low energy
transition state 30oa‡ (8.1 kcal molꢁ1) led to intermediate 31oa
(2.8 kcal molꢁ1) with a pentacoordinate silicon atom, and the
C]O double bond being signicantly elongated compared to
Conclusions
The value of the present study is that it demonstrates, to our
knowledge for the rst time, an unusual case where the tran-
sition metal of a catalyst complex is not directly involved in the
catalytic process. Activation of both substrate and reagent as
well as the bond-forming and -breaking events happen in the
ligand sphere, i.e., on the periphery of the transition metal.
Coincidentally, the mechanism becomes outer sphere at
silicon.22 The more conventional inner or outer sphere mecha-
nisms do not apply to this unique catalyst.
˚
its equilibrium distance from 1.211 to 1.251 A, indicating
activation.21 Lewis pair formation is followed by coordination
of hydrosilane 4a to the carbonyl group in 31oa, and the
hydrosilylation event releases 8oa through transition state
32oa‡ with retention at the silicon atom. In accordance with
our labelling experiments (cf. Scheme 3), this is the rate-
determining step (14.3 kcal molꢁ1). To further validate this,
we conducted a competition experiment between electron-rich
3a and electron-decient 3f (Scheme 12). The para substitu-
Acknowledgements
tion in
3 exerts a pronounced electronic effect, and
This research was supported by the Cluster of Excellence UniCat
of the Deutsche Forschungsgemeinscha (EXC 314/2). T.S. is
F3C-substituted 3f was consumed signicantly faster than
MeO-substituted 3a. This reactivity pattern is not unprece-
dented, and it has been seen previously in the activation of
carbonyl compounds with silicon-based Lewis acids.22 The
carbonyl carbon atom in 3f (X ¼ CF3) is more positively
polarised accelerating the hydride transfer, than that of
donor-substituted 3a (X ¼ OMe). The reactivity is also in
agreement with the proposed KIE based on the control reac-
tions with deuterated silane 4b-d1 (Scheme 3).
´
grateful to The New Szechenyi Plan TAMOP-4.2.2/B-10/1-2010-
0009. M.O. is indebted to the Einstein Foundation Berlin for an
endowed professorship.
Notes and references
1 For the seminal work, see: H. Brunner and K. Fisch, Angew.
Chem., Int. Ed. Engl., 1990, 29, 1131–1132.
2 For reviews of iron-catalysed hydrosilylation, see: (a)
¨
K. Junge, K. Schroder and M. Beller, Chem. Commun., 2011,
47, 4849–4859; (b) M. Zhang and A. Zhang, Appl.
Organomet. Chem., 2010, 24, 751–757; (c) R. H. Morris,
Chem. Soc. Rev., 2009, 38, 2282–2291.
3 Selected reports include: (a) H. Nishiyama and A. Furuta,
Chem. Commun., 2007, 760–762; (b) N. S. Shaikh,
S. Enthaler, K. Junge and M. Beller, Angew. Chem., Int. Ed.,
2008, 47, 2497–2501; (c) B. K. Langlotz, H. Wadepohl and
L. H. Gade, Angew. Chem., Int. Ed., 2008, 47, 4670–4674; (d)
A. M. Tondreau, J. M. Darmon, B. M. Wile, S. K. Floyd,
E. Lobkovsky and P. J. Chirik, Organometallics, 2009, 28,
3928–3940; (e) T. Inagaki, L. T. Phong, A. Furuta, J.-i. Ito
and H. Nishiyama, Chem.–Eur. J., 2010, 16, 3090–3096; (f)
J. Yang and T. D. Tilley, Angew. Chem., Int. Ed., 2010, 49,
10186–10188; (g) A. P. Dieskau, J.-M. Begouin and
B. Plietker, Eur. J. Org. Chem., 2011, 5291–5296; (h)
P. Bhattacharya, J. A. Krause and H. Guan, Organometallics,
2011, 30, 4720–4729; (i) A. J. Ruddy, C. M. Kelly,
S. M. Crawford, C. A. Wheaton, O. L. Sydora, B. L. Small,
M. Stradiotto and L. Turculet, Organometallics, 2013, 32,
5581–5588; (j) Y. Sunada, H. Tsutsumi, K. Shigeta,
R. Yoshida, T. Hashimoto and H. Nagashima, Dalton
Trans., 2013, 42, 16687–16692; (k) H. Zhao, H. Sun and
X. Li, Organometallics, 2014, 33, 3535–3539; (l) Z. Zuo,
H. Sun, L. Wang and X. Li, Dalton Trans., 2014, 43, 11716–
11722; (m) T. Bleith, H. Wadepohl and L. H. Gade, J. Am.
Chem. Soc., 2015, 137, 2456–2459.
Scheme 11 Peripheral mechanism. Gibbs free energies given in
parentheses in kcal molꢁ1 [Si ¼ Si(OEt)3].
4 For a mechanistic perspective, see: S. Chakraborty and
H. Guan, Dalton Trans., 2010, 39, 7427–7436.
5 J. Y. Corey, Chem. Rev., 2011, 111, 863–1071.
6 For selected examples of inner sphere hydrosilylation, see: (a)
I. Ojima, M. Nihonyanagi, T. Kogure, M. Kumagai,
Scheme 12 Probing the electronic effect in a competition experiment.
Chem. Sci.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Edge Article
Chemical Science
´
S. Horiuchi, K. Nakatsugawa and Y. Nagai, J. Organomet. 10 K. J. Szabo and O. F. Wendt, Pincer and Pincer-Type
Chem., 1975, 94, 449–461; (b) I. Ojima, T. Kogure,
M. Kumagai, S. Horiuchi and T. Sato, J. Organomet. Chem.,
Complexes: Applications in Organic Synthesis and Catalysis,
Wiley-VCH, Weinheim, 2014.
1976, 122, 83–97; (c) G. Z. Zheng and T. H. Chan, 11 For state-of-the-art iron catalysts operating at room
Organometallics, 1995, 14, 70–79; (d) N. Schneider, temperature, see: (a) Ref. 3f; (b) Ref. 3i; (c) Ref. 3j; (d) Ref. 3m.
M. Finger, C. Haferkemper, S. Bellemin-Laponnaz, 12 No ring opening of the cyclopropyl group was observed.
P. Hofmann and L. H. Gade, Angew. Chem., Int. Ed., 2009, 13 (a) W. Yang, H. Fu, H. Wang, M. Chen, Y. Ding, H. W. Roesky
48, 1609–1613.
7 For reviews of the related outer sphere hydrogenation, see: (a) 14 O. G. Shirobokov, L. G. Kuzmina and G. I. Nikonov, J. Am.
O. Eisenstein and R. H. Crabtree, New J. Chem., 2013, 37, 21– Chem. Soc., 2011, 133, 6487–6489.
27; (b) R. M. Bullock, Chem.–Eur. J., 2004, 10, 2366–2374; (c) 15 All calculations were performed at uB97X-D/6–31G(d)[Fe:cc-
and A. Jana, Inorg. Chem., 2009, 48, 5058–5060; (b) Ref. 5.
R. Noyori, M. Yamakawa and S. Hashiguchi, J. Org. Chem.,
2001, 66, 7931–7944.
8 For outer sphere silane activation, see: (a) X.-L. Luo and
pVTZ] level of theory. Iron(II) complex 7a, triethoxysilane
(4a), and acetone (3o) were used as model substrates. See
the ESI† for full details.
R. H. Crabtree, J. Am. Chem. Soc., 1989, 111, 2527–2535; (b) 16 The calculated structures show in fact partial opening of the
V. K. Dioumaev and R. M. Bullock, Nature, 2000, 424, 530–
532; (c) E. A. Ison, E. R. Trivedi, R. A. Corbin and
M. M. Abu-Omar, J. Am. Chem. Soc., 2005, 127, 15374–
15375; (d) G. Du, P. E. Fanwick and M. M. Abu-Omar, J.
NHSi rings indicating formal oxidative addition taking place
on the silicon(II) centre. For a review on non-innocent
ligands, see: O. R. Luca and R. H. Crabtree, Chem. Soc.
Rev., 2013, 42, 1440–1459.
Am. Chem. Soc., 2007, 129, 5180–5187; (e) D. V. Gutsulyak, 17 The complex cis-14a was calculated to be diamagnetic with
S. F. Vyboishchikov and G. I. Nikonov, J. Am. Chem. Soc.,
2010, 132, 5950–5951; (f) S. Park and M. Brookhart,
vacant centre on the iron. The singlet-triplet gap was
calculated to be only 5.1 kcal molꢁ1
.
Organometallics, 2010, 29, 6057–6064; (g) R. Lalrempuia, 18 (a) S. Rendler and M. Oestreich, Angew. Chem., Int. Ed., 2008,
´
M. Iglesias, V. Polo, P. J. Sanz Miguel, F. J. Fernandez-
47, 5997–6000; (b) Ref. 8h.
´
Alvarez, J. J. Perez-Torrente and L. A. Oro, Angew. Chem., 19 M. Oestreich, G. Auer and M. Keller, Eur. J. Org. Chem., 2005,
¨
Int. Ed., 2012, 51, 12824–12827; (h) T. T. Metsanen,
184–195.
´
P. Hrobarik, H. F. T. Klare, M. Kaupp and M. Oestreich, J. 20 During the preparation of this manuscript a neutral silicon
Am. Chem. Soc., 2014, 136, 6912–6915; (i) M. C. Lipke and
T. D. Tilley, J. Am. Chem. Soc., 2014, 136, 16387–16398; (j)
Lewis acid catalysis was reported: A. L. Liberman-Martin,
R. G. Bergman and T. D. Tilley, J. Am. Chem. Soc., 2015,
137, 5328–5331.
´
M. Iglesias, F. J. Fernandez-Alvarez and L. A. Oro,
ChemCatChem, 2014, 6, 2486–2489.
21 The adduct 31 was not observed by the 1H or 31P NMR
spectroscopy, even at lower temperatures. Also attempts to
perform a Gutmann–Beckett analysis with Et3PO led to
only slight broadening of the 31P NMR resonance of the
phosphine oxide.
9 (a) B. Blom, S. Enthaler, S. Inoue, E. Irran and M. Driess, J.
Am. Chem. Soc., 2013, 135, 6703–6713; (b) D. Gallego,
S. Inoue, B. Blom and M. Driess, Organometallics, 2014, 33,
6885–6897.
¨
22 (a) K. Muther and M. Oestreich, Chem. Commun., 2011, 47,
334–336; (b) Ref. 20.
This journal is © The Royal Society of Chemistry 2015
Chem. Sci.