Reduction of ketones with hydrocarbon-soluble calcium hydride: Stoichiometric reactions and catalytic hydrosilylation

Article abstract of DOI:10.1002/ejic.200701255

Reactions of the dimeric calcium hydride complex [(DIPP-nacnac) CaH·thf]₂ {1; DIPP-nacnac = CH[(CMe)(2,6-iPr₂C ₆H₃N)]₂} with the α-hydrogen containing ketones acetophenone, acetone, dibenzylketone and

Full text of DOI:10.1002/ejic.200701255

FULL PAPER
DOI: 10.1002/ejic.200701255
Reduction of Ketones with Hydrocarbon-Soluble Calcium Hydride:
Stoichiometric Reactions and Catalytic Hydrosilylation
Jan Spielmann[a] and Sjoerd Harder*[a]
Keywords: Alkaline earth metals / Calcium / Hydride / Catalysis / Hydrosilylation
Reactions of the dimeric calcium hydride complex
[(DIPP-nacnac)CaH·thf]2 {1; DIPP-nacnac = CH[(CMe)(2,6-
iPr2C6H3N)]2} with the α-hydrogen containing ketones aceto-
phenone, acetone, dibenzylketone and 2-adamantone are
smooth. In most cases not only addition but also substantial
enolization is observed as a side reaction and in some cases
also aldol condensation was found. Despite this unselectivity,
the addition products could be isolated crystalline pure.
Crystal structures of [(DIPP-nacnac)CaOCH(Me)Ph]2 (3),
[(DIPP-nacnac)CaOCH(CH2Ph)2]2 (6) and [(DIPP-nacnac)-
Ca(2-adamantoxide)]2 (7) have been determined. The cal-
cium hydride complex 1 is an effective catalyst in the hydro-
silylation of ketones. Independent from the silane/ketone ra-
tio, a strong preference for formation of bis-alkoxy silanes
[PhSiH(OR)2] is observed. In most cases no enoxy groups
have been found in the product. This indicates that the
mechanism does not involve addition of the calcium hydride
to the ketone functionality. A concerted addition of silane to
ketone through a six-coordinate hypervalent silicon interme-
diate is proposed.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2008)
reagent by a more detailed study on the reduction of vari-
ous ketones. As nucleophilic conversion of ketones is often
plagued by competing side reactions like enolization and
aldol condensation (Scheme 1),[4] we are especially inter-
ested in: a) the selectivity of nucleophilic addition of 1 to a
variety of ketones and b) successful isolation of well-defined
products which would broaden the general scope of reagent
1 as a synthetic precursor in calcium chemistry.
Introduction
We recently published a synthetic route and crystal struc-
ture for the first well-defined calcium hydride complex (1)
[Equation (1)].[1] This dimeric complex, which readily dis-
solves in hydrocarbons like benzene, is stabilized towards
ligand exchange and concomitant precipitation of insoluble
CaH2 by the bulky β-diketiminate ligand.
As we have shown in a first screening study on the reac-
tivity of this reagent, 1 smoothly adds to unsaturated sub-
strates like conjugated alkenes, ketones, imines, and (iso)-
cyanides. In most cases well-defined calcium complexes
could be isolated and have been characterized by crystal
structure determinations.[2] Since the nucleophilic addition
to ketones is one of the most versatile reactions in organic
syntheses,[3] we now expand the scope of this new hydride
Scheme 1.
Since calcium hydride complexes have been proposed as
intermediates in the recently investigated Ca-mediated
hydrosilylation of alkenes,[5e] we are also particularly inter-
ested in the application of dimeric calcium hydride 1 in the
catalytic hydrosilylation of ketones. This would add to the
rapidly growing field of low-cost biocompatible calcium
catalysts for an increasing number of catalytic conver-
sions.[5]
[a] Anorganische Chemie, Universität Duisburg-Essen,
Universitätsstraße 5–7, 45117 Essen, Germany
E-mail: sjoerd.harder@uni-due.de
Supporting information for this article is available on the
WWW under http://www.eurjic.org or from the author.
1480
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2008, 1480–1486

Reductions with Hydrocarbon-Soluble Calcium Hydride
Table 1. Selectivity of the reaction of 1 in benzene with various
Results and Discussion
ketones[a] and subsequent quenching with Me3SiCl.
Stoichiometric Reactions
Ketone
% Addition
% Enolization % Aldol cond.
PhC(O)Ph
PhC(O)Me
MeC(O)Me
cyclohexanone
PhCH2C(O)CH2Ph
100
85
Ͻ1
58
68
100
–
–
The reaction of 1 and benzophenone in benzene is
smooth and the expected addition product could be isolated
in the form of a heteroleptic calcium alkoxide complex in
good yield (2, Scheme 2).[2] As benzophenone is not prone
15
76
35
32
Ͻ1
Ͻ1
24
7
Ͻ1
Ͻ1
to any of the side reactions summarized in Scheme 1, we adamantone
screened a variety of α-hydrogen containing ketones.
[a] A solution of 1 in benzene was stirred with the ketone at 20 °C
(16 h) and then quenched with Me3SiCl. The product distributions
have been analyzed by GC-MS.
The crystalline products 3 and 6–7 have been charac-
terized by single-crystal X-ray diffraction (Figure 1). Struc-
tures of 2 and 5 have been published earlier (the latter was
obtained in reaction of 1 with cyclohexene oxide). In all
cases a thf-free dimeric structure is observed in which β-
diketiminate ligands are bound terminally while the alk-
oxide ligands occupy the bridging positions. Likewise, in all
cases the structures are crystallographically centrosymmet-
ric. Apparently, their constitution as well as their nuclearity
and general assembly are fully independent on the alkoxide
ligand. All crystal structures exhibit surprisingly similar
Ca–ligand bond lengths (Table 2). Deviations from symmet-
rical bridging of the alkoxide ligands is only found for 2
and is supposedly correlated to intramolecular Ph(π)···Ca
interaction.[2] The crystal structure of 3 shows the centro-
symmetric (R,S)-diastereomer (Figure 1). Like in 2, the co-
ordination sphere of Ca2+ is completed by interaction with
the Ph ring of the alkoxide ligand. Whereas in 2 this contact
has been described as a Ph(π)···Ca interaction [Ca···C
3.180(2) Å],[2] the geometry of the Ph···Ca contact in 3 is
more conclusive for an agostic interaction: the Ca···C37Ј
distance is short [3.191(5) Å] but the Ca···H37Ј distance
(2.79 Å) is considerably shorter. Also the obtuse C37Ј–
H37Ј···Ca angle of 107° is more in agreement with an agos-
tic interaction. Complex 6 contains potentially donating Ph
rings, however, no Ph···Ca interactions are evident in this
structure. In 7 the steric bulk of the adamantoxide ligand
induces a slight tilt of the β-diketiminate ligand and conse-
quently some asymmetry in its chelating coordination mode
(Table 2).
Scheme 2.
Reactions of these selected ketones with 1 were generally
fast and gave full conversion. Analyses of the products after
quenching with Me3SiCl indicate a large variation in selec-
tivity (Table 1). For acetophenone 85% addition and 15%
enolization was observed. Reactions with cyclohexanone
and the readily enolizable 1,3-diphenylacetone were even
more unselective. For the substrate acetone no addition
product but only enolization and aldolcondensation could
be observed. 2-Adamantone, which is less enolizable due to
the in-plane orientation of the α-CH bonds in respect to the
ketone functionality, gave in reaction with 1 only addition
(possible enolization products are below the GC-MS detec-
tion limit).
Despite the unselectivity in the reaction of 1 with various
α-hydrogen containing ketones, crystalline pure addition
products could be obtained in most cases (3 and 5–7,
Scheme 2). The yields roughly reflect the chemoselectivities
observed in the reduction reactions. Consequently, a good
crystalline yield is obtained for 7 (the product obtained in
reaction with 2-adamantone) whereas 4 could not be ob-
tained via this route.
Table 2. A comparison of selected distances for the dimeric calcium
alkoxide complexes 2–3 and 5–7 (Å).
Complex
Ca–O
Ca–N
Ca···CaЈ
2[2]
2.261(1)
2.317(1)
2.227(1)
2.254(3)
2.226(2)
2.226(2)
2.233(1)
2.252(1)
2.247(1)
2.258(1)
2.398(1)
2.405(1)
2.376(3)
2.363(3)
2.351(3)
2.363(3)
2.371(2)
2.354(2)
2.341(1)
2.447(1)
3.5910(4)
3
3.488(1)
5[2][a]
3.4414(9)
3.4494(6)
3.4412(4)
6
7
[a] The crystal of complex 5 has two independent dimers in the
asymmetric unit. The values for the non-disordered molecule are
summarized here.
Eur. J. Inorg. Chem. 2008, 1480–1486
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
1481

J. Spielmann, S. Harder
FULL PAPER
Catalytic Hydrosilylation of Ketones
Catalytic hydrosilylation of ketones, i.e. the formal ad-
dition of a silane R3SiH to a ketone RЈ2C=O to give R3Si-
OCHRЈ2, is a convenient one-step procedure for prepara-
tion of protected alcohols.[6] The numerous catalysts for this
conversion are based on typical transition metals (Ni, Pd,
Rh) but lately also catalysts based on Sn,[7] Cu,[8] Zn,[9] Li
and Na[10] have been reported. In addition, it has been
shown that a strong Lewis acid like (C6F5)3B also efficiently
catalyzes this reaction.[11] As recently well-defined organ-
ocalcium complexes have shown efficient catalytic activity
and remarkable selectivity in alkene hydrosilylation, hydro-
silylation of ketones with calcium hydride 1 seems another
promising application of calcium complexes in catalysis.
Hydrosilylation of cyclohexanone with phenylsilane (1:1
ratio) and catalytic amounts of 1 in benzene unexpectedly
gave PhSiH(OCy)2 as the main product. Investigations on
the product distribution as a function of the silane/ketone
ratio (Table 3) show that the dialkoxy silane is the favoured
product under all circumstances. Even at high silane/ketone
ratio of 2:1 the major product is PhSiH(OCy)2. At low si-
lane/ketone ratio also small amounts of the trialkoxy prod-
uct can be obtained. Hydrosilylation of cyclohexanone gave
mainly products originating from addition of the hydride to
the ketone, however, besides the main product PhHSi(OCy)2
also small amounts of PhHSi(OCy)(OC6H9) could be de-
tected in the GC-MS spectra, i.e. a dehydrogenative product
has been formed by formal incorporation of the cyclohexen-
olate anion C6H9O–. Interestingly, the trialkoxy product
which was formed only in minor amounts consists mainly
of PhHSi(OCy)2(OC6H9). The overall alkoxy/enoxy ratio of
98:2, however, is much higher than the addition/enolization
ratio in reaction of 1 with cyclohexanone (Table 1). Thus,
hydrosilylation of cyclohexanone appears to be consider-
ably more selective than the stoichiometric addition of cal-
cium hydride 1 to this ketone.
Table 3. Product distribution in the catalytic hydrosilylation of cy-
clohexanone with PhSiH3 in benzene at 50 °C (1.25 mol-% 1).
Entry
Silane/ PhH2SiOR PhHSi(OR)2 PhSi(OR)3 Total alkoxy/
ketone
(%)
(%)
(%)
enoxy
1
2
3
4
2/1
1:1
1:2
1/3
29
20
0
69 (95:5)[a]
78 (95:5)[a]
91 (94:6)[a]
89 (91:9)[a]
2 (11:89)[b]
2 (11:89)[b]
9 (4:96)[b]
11 (1:99)[b]
98:2
98:2
96:4
98:2
0
[a] In parentheses: ratio PhHSi(OCy)2/PhHSi(OCy)(OC6H9). [b] In
parentheses: ratio PhSi(OCy)3/PhHSi(OCy)2(OC6H9).
As the dialkoxy silane is the major hydrosilylation prod-
uct, catalytic experiments for a larger variety of ketones
were carried out with a silane/ketone ratio of 1:2 (Table 4).
In all cases the dialkoxy products PhHSi(OR)2 were formed
in Ͼ90% yield. In addition, the well-defined dibenzylcal-
cium catalyst 8 was also tested and gave similar results.
Whereas for cyclohexanone also a small percentage of prod-
ucts originating from an enolate intermediate could be de-
Figure 1. Crystal structures of the dimeric calcium alkoxide com-
plexes 3, 6 and 7. In all cases the iPr substituents and hydrogens
(except for the hydrogen transferred in the addition reaction) have
been omitted for clarity.
1482
www.eurjic.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2008, 1480–1486

Reductions with Hydrocarbon-Soluble Calcium Hydride
tected, other ketone substrates exclusively gave products The addition reaction is irreversible[12] and results in forma-
with alkoxy substituents. These observations hint to a pos- tion of the calcium alkoxide complex A1. Addition of this
sible mechanism for this catalytic reaction.
alkoxide to PhSiH3 gives the hypervalent intermediate A2
which after elimination of the hydrosilylation product, the
monoalkoxy silane, generates the original catalyst (LCaH).
The observed preference for formation of a dialkoxy silane
can be explained by precoordination of product PhH2SiOR
to the intermediate A1 followed by intramolecular alkoxide/
hydride exchange [Equation (2)], i.e. A1 reacts faster with
PhH2SiOR than with PhSiH3 on account of a complex-in-
duced proximity effect (CIPE).[13] Further conversion of
PhHSi(OR)2 to PhSi(OR)3 is likely hindered for steric
reasons.
Table 4. Hydrosilylation of ketones with PhSiH3 and the calcium
catalysts 1 or 8 (1.25 mol-%); reaction conditions: benzene, 50 °C.
Results for catalyst 8 are given in square brackets.
Entry Ketone
Time / h
%
Total alkoxy/enoxy
PhHSi(OR)2
1
2
3
4
5
PhC(O)Ph
PhC(O)Me
15 [15]
34 [38]
3 [3]
96 [96]
95 [95]
91 [96]
96 [95]
95 [98]
–
100:0 [100/0]
96:4 [95/5]
100:0 [100/0]
100:0 [100/0]
cyclohexanone
PhCH2C(O)CH2Ph[a] 34 [54]
adamantone[b]
0.2 [1.5]
[a] 5 mol-% catalyst used. [b] Reactions at 20 °C.
Although a similar mechanism has been proposed for the
As for most hydrosilylation catalysts,[6–9] the mechanism Ca-mediated hydrosilylation of alkenes, there are several
could proceed through a cycle in which a hydride species is reasons to believe that the Ca-mediated hydrosilylation of
the actual catalyst (Scheme 3, cycle A). The first step in the ketones does not proceed through cycle A. i) Stoichiometric
mechanism is addition of the calcium hydride to the ketone. addition of the various ketones to 1 gave in most cases sub-
Scheme 3.
Eur. J. Inorg. Chem. 2008, 1480–1486
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
1483

J. Spielmann, S. Harder
FULL PAPER
stantial enolate formation (Table 1), whereas the overall en-
It should be noted that the intermediates A2 in cycle A
oxy content in the hydrosilylation products is low. ii) Under and B3 in cycle B are equal. This explains why heteroleptic
stoichiometric conditions the calcium hydride catalyst calcium alkoxide complexes (A1) also catalyze the hydro-
(LCaH) could not be generated by addition of PhSiH3 to a silylation of ketones (A1ǞA2 = B3ǞB1).
well-defined heteroleptic calcium alkoxide species (A1); i.e.
Hypervalent anions are the key intermediates in cycle B
the pathway A1ǞA2ǞLCaH could not be verified experi- and the species LCa+, which is likely solvated by ketones
1
mentally. The H NMR spectrum of a solution of 7 and or silyl ether products, will hardly influence the catalytic
PhSiH3 in C6D6 showed only signals for the separate species reaction. This is in agreement with the observation that cat-
and no signs for calcium hydride 1 could be observed. Ad- alysts 1 and 8 give very similar results (Table 4).
dition of a small amount of cyclohexanone, however, gave
It is noteworthy that the small amount of trisubstituted
immediate reaction and the products PhHSi(O-adamantyl)- phenylsilane formed during hydrosilylation of cyclohexa-
(OCy) and PhH2Si(OCy)2 could be detected. This suggests none is largely PhSi(OCy)2(OC6H9). As the intermediate di-
that the calcium hydride species is not the intermediate but substituted phenylsilane mainly consists of PhHSi(OCy)2,
instead a proton is transferred directly from silane to the last step involves almost exclusive introduction of the
ketone. It also indicates that heteroleptic calcium alkoxide enoxy substituent. In principal two routes could explain
species are active catalysts. The latter was verified by experi- formal incorporation of the cyclohexenolate anion. The lat-
ment.
ter anion could be formed by reaction of cyclohexanone
We propose cycle B (Scheme 3) as a catalytic cycle that with the calcium hydride (LCaH). As also substantial hy-
agrees better with the observations. In the first step the cat- dride addition would occur (Table 1), i.e. CyO– formation,
alyst (LCaH) does not react with the ketone but with this reaction seems less likely. Alternatively, PhSi(OCy)2-
PhSiH3 to give a hypervalent species (B1). 1H NMR investi- (OC6H9) could be formed in a concerted deprotonation re-
gation on a stoichiometric reaction of 1 with PhSiH3 in [D8]- action that involves a six-membered ring in the transition
THF shows that this is an equilibrium which at room tem- state [Equation (4)]. On account of steric crowding the six-
perature is fast on the NMR time scale: coalescence of the membered ring transition state for α-deprotonation would
signals for the hydride and Si–H protons is observed. Cool- be preferred over the four-membered ring transition state
ing the solution gives decoalescence of signals in those for for addition (e.g. B3).
1 and PhSiH3. This equilibrium is confirmed by the obser-
vation of rapid H/D exchange between 1 and PhSiD3 in
C6D6. In a second step, transient B1 forms a coordination
complex with a ketone (B2) that after a concerted insertion
gives a new hypervalent species (B3). Hydride transfer from
B3 to PhSiH3 gives the product and the hypervalent inter-
mediate B1.
Route B is similar to the mechanism proposed for the
sodium alkoxide catalyzed hydrosilylation of ketones with
Ph2SiH2.[10] Also in the latter reaction a high proportion of
the dialkoxy silane Ph2Si(OR)2 is formed as the product.
The observed preference for formation of a dialkoxy silane
can be explained by coordination of the ketone to B3 fol-
Conclusions
lowed by insertion and elimination of the product [Equa-
The stoichiometric reactions of the soluble heteroleptic
tion (3)]. As the silicon center in B3 is more Lewis acidic calcium hydride complex 1 with ketones are generally
than that in B1, precoordination of the ketone to B3 is pre- smooth. Apart from addition of the hydride to the ketone,
ferred over coordination to B1 thus explaining major for- α-hydrogen containing ketones also show substantial enol-
mation of the dialkoxy silane. Further conversion of B5 to ization as a side reaction. Despite this unselectivity, crystal-
a trialkoxy silane is likely hindered for steric reasons.
line pure addition products could be isolated in most cases.
The rather low yields are partly due to the poor chemoselec-
tivities. The calcium hydride complex 1 is also an effective
catalyst for the hydrosilylation of ketones. Independent of
the silane/ketone ratio, a dialkoxy silane [PhHSi(OR)2] is
found as the major product. As the overall alkoxy/enoxy
ratio in the hydrosilylation products is very high (in most
cases no enoxy substituents could be detected) it is likely
that the mechanism for Ca-mediated hydrosilylation does
not involve the addition of calcium hydride to the ketone.
Instead an intermediate with a hypervalent six-coordinate
silicon atom is proposed. In this catalytic cycle the hydride
atom is transferred from silicon to ketone in a concerted
step.
1484
www.eurjic.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2008, 1480–1486

Reductions with Hydrocarbon-Soluble Calcium Hydride
ten minutes at room temperature and then cooled to 8 °C. Over-
night colourless blocks of 7 precipitated: 114 mg (67%). H NMR
Experimental Section
1
General: All manipulations were performed under a dry and oxy-
gen-free atmosphere (argon or nitrogen) by using freshly dried sol-
vents and Schlenk line and glove box techniques. Following com-
plexes have been prepared according to literature: 1[1] and 8.[14]
(300 MHz, C6D6, 25 °C): δ = 7.21–7.10 (m, 6 H, Har), 4.64 (s, 1 H,
DIPP-nacnac bridge), 3.81 (m, 1 H, α-H adamantyl), 3.24 [sept,
3J(H,H) = 6.8 Hz, 4 H, Me2CH], 1.87–1.82 (m, 2 H, adamantyl),
1.72–1.54 (m, 6 H, adamantyl), 1.61 (s, 6 H, DIPP-nacnac Me),
1.28–1.14 (m, 6 H, adamantyl), 1.18 [d, 3J(H,H) = 6.8 Hz, 12 H,
Me2CH], 1.17 [d, 3J(H,H) = 6.8 Hz, 12 H, Me2CH] ppm. 13C NMR
(300 MHz, C6D6, 25 °C): δ = 166.7, 148.0, 141.8, 124.4, 124.0, 91.7
(aromatics, DIPP-nacnac bridge), 76.0 (α-C adamantyl), 38.3 (C-
bridge adamantyl), 37.9 (C-bridge adamantyl), 37.5 (C-bridge ada-
mantyl), 32.0 (CH2 adamamtyl), 28.5 (CH2 adamantyl), 28.3
(Me2CH), 27.7 (CH2 adamantyl), 25.7 (DIPP-nacnac Me), 25.4
(Me2CH), 24.8 (Me2CH) ppm. C39H56CaN2O (608.97): calcd. C
76.92, H 9.27; found C 76.43, H 9.16.
Synthesis of 3: A solution of 1 (250 mg; 0.24 mmol) and acetophe-
none (57 mg; 0.47 mmol) in benzene (1.8 mL) was stirred for ten
minutes at room temperature. Overnight colourless octahedral crys-
tals of 3 precipitated at room temperature in low yield: 68 mg
(25%). The moisture sensitive crystals were suitable for X-ray
analysis but are composed of both possible diastereomers in a ratio
of about 2:1. Multiple recrystallisation in benzene afforded one of
the diastereomers in sufficient purity for NMR characterization.
1H NMR (300 MHz, C6D6, 25 °C): δ = 7.35–6.98 (m, 11 H, Har),
3
4.77 (s, 1 H, DIPP-nacnac bridge), 4.75 [q, J(H,H) = 6.8 Hz, 1 H,
Reaction of 1 with Ketones and Subsequent Quenching with Me3-
SiCl: A mixture of 1 (30 mg, 0.057 mmol) and the ketone substrate
(0.057 mmol) in benzene (0.45 mL) was kept overnight at room
temperature (samples with cyclohexanone were heated at 60 °C for
three hours). The Me3SiCl used in the quench reaction, was freed
from HCl impurities by addition of an equimolar amount of dry
triethylamine to a solution of Me3SiCl in dry THF and separating
possibly formed ammonium chloride by centrifugation. The
quenching solution (10% in THF, 0.30 mL, 0.20 mmol Me3SiCl)
was added to the reaction mixture of 1 and the ketone. After stir-
ring overnight at room temperature a small amount of water was
added and the organic layer was separated and analyzed by GC-
MS.
OCHCH3Ph], 3.37 [sept, 3J(H,H) = 6.8 Hz, 2 H, CHMe2], 2.61 (br,
2 H, CHMe2), 1.70 (s, 3 H, DIPP-nacnac Me), 1.48 (s, 3 H, DIPP-
3
nacnac Me), 1.21 [d, J(H,H) = 6.8 Hz, 3 H, OCHCH3Ph], 1.21–
0.91 (24 H, CHMe2) ppm. 13C NMR (500 MHz, C6D6, 60 °C): δ
= 167.0, 151.8, 147.7, 142.1, 130.1, 128.5, 126.6, 125.1, 124.7, 124.1,
94.6 (aromatics, DIPP-nacnac bridge), 72.5 (OCHCH3Ph), 31.3
(OCHCH3Ph), 28.8 (Me2CH), 28.3 (Me2CH), 25.6 (DIPP-nacnac
Me), 25.3 (DIPP-nacnac Me), 25.2 (Me2CH), 25.1 (Me2CH), 24.8
(Me2CH), 28.8 (Me2CH) ppm. C37H50CaN2O (578.90): calcd. C
76.77, H 8.71; found C 76.32, H 8.92.
Synthesis of 5: A solution of 1 (170 mg; 0.16 mmol) and cyclohexa-
none (37 mg; 0.37 mmol) in benzene (3.0 mL) was stirred for three
hours at 60 °C. After concentrating the solution to one-third of its
original volume, the solution was cooled to 8 °C. Overnight colour-
Typical Procedure for the Catalytic Hydrosilylation of Ketones: The
catalysts 1[1] and 8[14] were used in crystalline purity. In a typical
hydrosilylation experiment a dried NMR tube was charged with the
silane (0.30 mmol) and the ketone substrate (0.60 mmol) in C6D6
(0.45 mL). After addition of the catalyst, normally 1.25 mol-% 1
(calculated on the dimer) or 8, the reaction mixture was heated to
50 °C (or in some cases allowed to stand at room temperature). At
1
less crystals of 5 precipitated: 40 mg (23%). H NMR (300 MHz,
C6D6, 25 °C): δ = 7.15–7.09 (m, 6 H, Har), 4.70 (s, 1 H, DIPP-
nacnac bridge), 3.30 (m, 1 H, α-H Cy), 3.22 [sept, 3J(H,H) =
6.8 Hz, 4 H, Me2CH], 1.67–1.58 (br m, 5 H, Cy), 1.61 (s, 6 H,
DIPP-nacnac Me), 1.52–1.38 (m, 2 H, Cy), 1.18 [d, 3J(H,H) =
3
6.8 Hz, 12 H, Me2CH], 1.07 [d, J(H,H) = 6.9 Hz, 12 H, Me2CH],
1
regular time intervals the conversion was determined by H NMR
1.01–0.87 (m, 3 H, Cy) ppm. 13C NMR (300 MHz, C6D6, 25 °C):
δ = 166.7, 147.7, 142.2, 124.9, 124.5, 93.9 (aromatics, DIPP-nacnac
bridge), 72.5 (α-C Cy), 40.1 (Cy), 29.2 (Me2CH), 26.6 (Cy), 26.4
(Cy), 25.5 (DIPP-nacnac Me), 25.3 (Me2CH), 25.2 (Me2CH) ppm.
C35H52CaN2O (556.90): calcd. C 75.49, H 9.41; found C 74.98, H
9.23.
spectroscopy. After at least 95% conversion the reaction mixture
was quenched with small amounts of water to remove all calcium
salts and the organic layer was analysed via GC-MS.
Crystal Structure Determination: All data were collected on a Sie-
mens SMART CCD diffractometer at –70 °C. The structures have
been solved by direct methods (SHELXS-97)[15] and were refined
with SHELXL-97.[16] The geometry calculations and graphics have
been performed with PLATON.[17] CCDC-660496 (for 3), -660497
(for 7) and -660498 (for 6) contain the supplementary crystallo-
graphic data for this paper (excluding structure factors). These data
can be obtained free of charge from The Cambridge Crystallo-
graphic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
Synthesis of 6: A solution of 1 (310 mg; 0.29 mmol) and 1,3-di-
phenylacetone (130 mg; 0.62 mmol) in toluene (3.5 mL) was kept
overnight at room temperature. The alkoxide 6 precipitated from
the mother liquor in form of colourless crystals and could be ob-
tained in low yields: 35 mg (9%). 1H NMR (300 MHz, C6D6,
25 °C): δ = 7.26–7.08 (m, 6 H, Har DIPP-nacnac), 6.96–6.80 [m, 10
H, OCH(CH2Ph)2], 4.60 (s, 1 H, DIPP-nacnac bridge), 4.56 [m, 1
H, OCH(CH2Ph)2], 3.24 [sept, 3J(H,H) = 6.8 Hz, 4 H, Me2CH],
Crystal Data for 3: C37H50CaN2O·(C6H6)0.5, Mr = 617.92, mono-
clinic, space group P21/n, a = 12.1570(6) Å, b = 12.8541(6) Å, c =
3
3
2.96 [d, J(H,H) = 12.6 Hz, 1 H, OCH(CH2Ph)2], 2.94 [d, J(H,H)
23.3203(13) Å, β = 94.268(4), V = 3634.1(3) Å3, Z = 4, ρcalcd.
=
3
= 13.5 Hz, 1 H, OCH(CH2Ph)2], 2.56 [d, J(H,H) = 9.4 Hz, 1 H,
1.129 Mg·m–3, F(000) = 1340, µ(Mo-Kα) = 0.204 mm–1. Of the
13419 measured reflections, 3749 were independent (Rint = 0.068,
θmax = 20.8°) and 2721 observed [IϾ2σ(I)]. The final refinement
converged to R1 = 0.0538 for IϾ2σ(I), wR2 = 0.1619 and GOF =
1.06 for all data. The final difference Fourier synthesis gave a min./
max. residual electron density of –0.30/+0.55 eÅ–3. All hydrogen
atoms have been placed on calculated positions and were refined
in a riding mode.
3
OCH(CH2Ph)2], 2.52 [d, J(H,H) = 9.4 Hz, 1 H, OCH(CH2Ph)2],
3
1.67 (s, 6 H, DIPP-nacnac Me), 1.18 [d, J(H,H) = 7.1 Hz, 12 H,
Me2CH], 1.15 [d, 3J(H,H) = 6.1 Hz, 12 H, Me2CH] ppm. 13C NMR
(300 MHz, C6D6, 25 °C): δ = 166.6, 147.6, 141.8, 139.2, 129.6,
128.0, 125.7, 124.5, 124.3, 93.8 (aromatics, DIPP-nacnac bridge),
75.4 [OCH(CH2Ph)2], 49.6 [OCH(CH2Ph)2], 28.5 (Me2CH), 25.3
(DIPP-nacnac Me), 25.0 (Me2CH), 24.8 (Me2CH) ppm.
C44H56CaN2O (669.03): calcd. C 78.99, H 8.44; found C 79.23, H
8.21.
Crystal Data for 6: C44H56CaN2O·(C6H6)2.5, Mr = 864.26, triclinic,
¯
Synthesis of 7: A solution of 1 (150 mg; 0.14 mmol) and 2-ada-
space group P1, a = 13.5135(6) Å, b = 13.6938(6) Å, c =
mantone (42 mg; 0.28 mmol) in benzene (2.5 mL) was stirred for
15.8605(8) Å, α = 82.900(3), β = 69.930(3), γ = 69.320(3), V =
Eur. J. Inorg. Chem. 2008, 1480–1486
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
1485

J. Spielmann, S. Harder
FULL PAPER
2579.1(2) Å3, Z = 2, ρcalcd. = 1.113 Mg·m–3, F(000) = 934, µ(Mo-
Kα) = 0.162 mm–1. Of the 32045 measured reflections, 13110 were
independent (Rint = 0.050, θmax = 28.9°) and 8689 observed
[IϾ2σ(I)]. The final refinement converged to R1 = 0.0538 for
IϾ2σ(I), wR2 = 0.1480 and GOF = 1.05 for all data. The final
difference Fourier synthesis gave a min./max. residual electron den-
sity of –0.30/+0.31 eÅ–3. Most hydrogen atoms have been placed
on calculated positions and were refined in a riding mode. The
hydrogen transferred by the reduction reaction (Cα) has been lo-
cated in the Fourier-difference map and was refined isotropically.
369; i) O. Riant, J. Hannedouche, Org. Biomol. Chem. 2007, 5,
873.
[4]
[5]
a) T. Imamoto, Y. Sugiura, N. Takiyama, Tetrahedron Lett.
1984, 38, 4233; b) T. Imamoto, N. Takiyama, K. Nakamura,
T. Hatajima, Y. Kamiya, J. Am. Chem. Soc. 1989, 111, 4392;
c) M. Hatano, T. Matsumura, K. Ishihara, Org. Lett. 2005, 7,
573; d) A. Krasovskiy, F. Kopp, P. Knochel, Angew. Chem.
2006, 118, 511; Angew. Chem. Int. Ed. 2006, 45, 497.
a) S. Harder, F. Feil, K. Knoll, Angew. Chem. 2001, 113, 4391;
Angew. Chem. Int. Ed. 2001, 40, 4261; b) M. H. Chisholm, J.
Gallucci, K. Phomphrai, Chem. Commun. 2003, 48; c) Z.
Zhong, S. Schneiderbauer, P. J. Dijkstra, M. Westerhausen, J.
Feijen, J. Polym. Bull. 2003, 51, 175; d) M. R. Crimmin, I. J.
Casely, M. S. Hill, J. Am. Chem. Soc. 2005, 127, 2042; e) F.
Buch, J. Brettar, S. Harder, Angew. Chem. 2006, 118, 2807; An-
gew. Chem. Int. Ed. 2006, 45, 2741; f) M. R. Crimmin, A. G.
Barrett, M. S. Hill, P. A. Procopiou, Org. Lett. 2007, 9, 331; g)
S. Datta, P. W. Roesky, S. Blechert, Organometallics 2007, 26,
4392; h) F. Buch, S. Harder, Organometallics 2007, 26, 5132.
a) I. Ojima, The Chemistry of Organic Silicon Compounds (Eds.:
S. Patai, Z. Rappoport), Wiley, New York, 1989; b) B. Marcin-
iec, Comprehensive Handbook on Hydrosilation, Pergamon,
New York, 1992; c) H. Brunner, Henri, Transition Metals for
Organic Synthesis (Eds.: M. Beller, C. Bolm), Wiley-VCH,
1998; d) J. Jun, S. Buchwald, J. Am. Chem. Soc. 1999, 121,
5640; e) A. C. Fernandes, R. Fernandes, C. C. Romao, B. Royo,
Chem. Commun. 2005, 213; f) O. Riant, N. Mostefai, J. Courm-
arcel, Synthesis 2004, 18, 2943.
Crystal Data for 7: C39H56CaN2O·(C6H6), Mr = 687.05, triclinic,
¯
space group P1, a = 12.6423(6) Å, b = 13.2216(6) Å, c =
13.9195(6) Å, α = 112.101(2), β = 98.680(2), γ = 104.757(2), V =
2003.98(17) Å3, Z = 2, ρcalcd. = 1.139 Mgm–3, F(000) = 748, µ(Mo-
Kα) = 0.191 mm–1. Of the 93670 measured reflections, 10533 were
independent (Rint = 0.044, θmax = 29.0°) and 8819 observed
[IϾ2σ(I)]. The final refinement converged to R1 = 0.0403 for
IϾ2σ(I), wR2 = 0.1126 and GOF = 1.02 for all data. The final
difference Fourier synthesis gave a min./max. residual electron den-
sity of –0.25/+0.33 eÅ–3. Some hydrogen atoms have been located
in the Fourier-difference map and were refined isotropically, others
have been placed on calculated positions and were refined in a ri-
ding mode.
[6]
Supporting Information (see also the footnote on the first page of
this article): Experimental details on the stoichiometric quench re-
actions and the catalytic hydrosilylation reactions (including GC-
MS and or NMR analyses of the products). Also a catalytic experi-
ment on a preparative scale is described.
[7]
[8]
N. J. Lawrence, S. M. Bushell, Tetrahedron Lett. 2000, 41, 4507.
a) H. Brunner, W. Miehling, J. Organomet. Chem. 1984, 275,
C17; b) B. H. Lipshutz, W. Chrisman, K. Noson, J. Organomet.
Chem. 2001, 624, 367; c) S. Rendler, M. Oestreich, Angew.
Chem. 2007, 119, 504; Angew. Chem. Int. Ed. 2007, 46, 498.
H. Mimoun, J. Org. Chem. 1999, 64, 2582.
L. Gan, M. A. Brook, Organometallics 2007, 26, 945.
a) D. J. Parks, W. E. Piers, J. Am. Chem. Soc. 1996, 118, 9440;
b) D. J. Parks, J. M. Blackwell, W. E. Piers, J. Org. Chem. 2000,
65, 3090.
[9]
[10]
[11]
Acknowledgments
Prof. Dr. R. Boese and D. Bläser (University of Duisburg-Essen)
are thanked for collection of X-ray diffraction data.
[12]
[13]
The reversibility of the reaction has been tested by addition of
benzophenone to a solution of 7 in C6D6. Even under reflux
conditions no signs of formation of 2 could be observed.
B. Chauder, L. Green, V. Snieckus, Pure Appl. Chem. 1999, 71,
1521.
S. Harder, F. Feil, A. Weeber, Organometallics 2001, 20, 1044.
G. M. Sheldrick, SHELXS-97, Program for Crystal Structure
Solution, 1997, University of Göttingen, Germany.
G. M. Sheldrick, SHELXL-97, Program for Crystal Structure
Refinement, 1997, University of Göttingen, Germany.
A. L. Spek, PLATON, A Multipurpose Crystallographic Tool
2000, Utrecht University, Utrecht, The Netherlands.
Received: November 22, 2007
[1] S. Harder, J. Brettar, Angew. Chem. 2006, 118, 3554; Angew.
Chem. Int. Ed. 2006, 45, 3474.
[2] J. Spielmann, S. Harder, Chem. Eur. J. 2007, 13, 8928.
[3] a) B. J. Wakefield, Organomagnesium Methods in Organic
Chemistry Academic Press, San Diego, 1995; b) H. G.
Richey Jr, Grignard Reagents: New Developments, Wiley,
Chichester, 2000; c) Y.-H. Lai, Synthesis 1981, 585; d) J. J.
Eisch, Organometallics 2002, 21, 5439; e) P. Knochel, W. Dohle,
N. Gommermann, F. F. Kneisel, F. Kopp, T. Korn, I. Sa-
pountzis, V. A. Vu, Angew. Chem. Int. Ed. 2003, 42, 4302; f) A.
Inoue, H. Shinokubo, K. Oshima, J. Synth. Org. Chem. Jpn.
2003, 61, 25; g) J. Christoffers, A. Baro, Adv. Synth. Catal.
2005, 347, 1473; h) B. M. Trost, C. Jiang, Synthesis 2006, 3,
[14]
[15]
[16]
[17]
Published Online: February 1, 2008
1486
www.eurjic.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2008, 1480–1486