Copper-catalyzed hydrosilylation with a bowl-shaped phosphane ligand: Preferential reduction of a bulky ketone in the presence of an aldehyde

Article abstract of DOI:10.1002/anie.200906348

Chemical Equation Presented Hollywood bowl: A highly active copper catalyst with a bowl-shaped phosphane (bsp) ligand was used in the hydrosilylation reaction of bulky ketones. The preferential reduction of a bulky ketone in the presence of an unprotected aldehyde is unprecedented.

Full text of DOI:10.1002/anie.200906348

Communications
DOI: 10.1002/anie.200906348
Homogeneous Catalysis
Copper-Catalyzed Hydrosilylation with a Bowl-Shaped Phosphane
Ligand: Preferential Reduction of a Bulky Ketone in the Presence of an
Aldehyde**
Tetsuaki Fujihara, Kazuhiko Semba, Jun Terao, and Yasushi Tsuji*
Table 1: Effect of phosphanes on the copper-catalyzed hydrosilylation of
4a.[a]
Ligands play a crucial role in the efficiency and selectivity of
homogeneous catalysts;[1] therefore, much effort has been
given to their development. Bowl-shaped tris(3,5-diarylphe-
nyl)phosphanes (1–3, Figure 1a)[2] have a unique structure
Entry
Ligand
Yield [%][b]
after 20 min
after 5 h
1
2
3
4
5
6
7
8
PPh3
P(o-Tol)3
P(Mes)3
P(tBu)3
1
2
3
bdp
<1
<1
<1
9
7
43
2
<1
2
11
25
92
99
4
99
<1
<1
<1
9
(S)-dtbm-segphos
2
1
5
10[c]
Figure 1. a) Bowl-shaped phosphanes 1–3. b) An optimized structure
of 2 calculated by the HF/6-31G(d)-CONFLEX/MM3 method.[2]
[a] 4a (2.0 mmol), Ph2SiH2 (2.4 mmol), CuCl (0.02 mmol, 1.0 mol%),
phosphane (0.02 mmol, 1.0 mol%), tBuONa (0.12 mmol, 6.0 mol%),
toluene (2.0 mL), RT. [b] Yield of alcohol determined using a GC internal
standard. [c] [{RhCl(C2H4)2}2] (0.01 mmol, 1.0 mol%) and 2 (0.02 mmol,
1.0 mol%) were used as the catalyst. Tol=methylphenyl, Mes=mesityl,
bdp=1,2-bis(diphenylphosphino)benzene, dtbm-segphos=(4,4’-bi-1,3-
benzodioxole)-5,5’-diylbis(3,5-di-tert-butyl-4-methoxyphenyl)phosphane.
that consists of a bulky periphery (the rim of the bowl has a
diameter of 2.0–2.6 nm) and substantial empty space around
the phosphorus atom (Figure 1b). We have recently reported
that 2 and 3 are particularly effective ligands in the rhodium-
catalyzed hydrosilylation reaction of ketones[2a,b] and the
palladium-catalyzed Suzuki–Miyaura coupling of aryl chlor-
ides.[2c] Herein, we report the use of extremely active copper
catalysts[3] that contain bowl-shaped phosphane (bsp) ligands
(2 or 3) in the hydrosilylation reaction of ketones. These
catalysts afford the unprecedented preferential reduction of a
bulky ketone substrate in the presence of an unprotected
aldehyde.
PPh3, P(o-Tol)3,[5] P(Mes)3,[5] and P(tBu)3, afforded the
corresponding alcohol after hydrolysis in only 2%, < 1%,
2%, and 11% yields, respectively (5 h; Table 1, entries 1–4).
In contrast, 2 was remarkably effective in this reaction,
affording the product in 92% after 5 h (Table 1, entry 6).
Furthermore, 3 was a more efficient ligand, providing the
alcohol in quantitative yield after only 20 min (Table 1,
entry 7). Bidentate phosphane ligands, such as bdp[5] and
(S)-dtbm-segphos,[5] which are very effective ligands in
copper-catalyzed hydrosilylation reactions,[6] were unsuccess-
ful (Table 1, entries 8 and 9). Under the same reaction
conditions as in Table 1, entry 6 (5 h), other silanes, such as
Et3SiH, MePh2SiH, and Me(OEt)2SiH, afforded the product
in trace, 2%, and 51% yields, respectively. An excess
(5 equiv) of silane only slightly improved the yield, namely
trace amounts with Et3SiH and 8% using MePh2SiH. Nolan
et al. reported that N-heterocyclic carbene (NHC) ligands,
such as ICy (N,N’-bis(cyclohexyl)imidazol-2-ylidene), IMes
(N,N’-bis(2,4,6-trimethylphenyl)imidazol-2-ylidene), and IPr
First, the hydrosilylation of the poorly reactive ketone[4]
2,2,4,4-tetramethyl-3-pentanone (4a) with Ph2SiH2 was inves-
tigated at room temperature (Table 1). A CuCl/tBuONa
catalyst system with conventional phosphane ligands, such as
[*] Prof. Dr. T. Fujihara, K. Semba, Prof. Dr. J. Terao, Prof. Dr. Y. Tsuji
Department of Energy and Hydrocarbon Chemistry
Graduate School of Engineering, Kyoto University
Kyoto 615-8510 (Japan)
Fax: (+81)75-383-2514
E-mail: ytsuji@scl.kyoto-u.ac.jp
Homepage: http://twww.ehcc.kyoto-u.ac.jp/
[**] This work was supported by a Grant-in-Aid for Scientific Research on
Priority Areas (“Synergy of Elements” and “Chemistry of Concerto
Catalysis”) from the Ministry of Education, Culture, Sports, Science,
and Technology (Japan).
(N,N’-bis(2,6-diisopropylphenyl)imidazol-2-ylidene),
were
highly efficient in the copper-catalyzed hydrosilylation reac-
tions of bulky ketones.[4c,7] Under the same reaction con-
ditions as in Table 1, entry 7 (20 min), the ICy ligand afforded
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200906348.
1472
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 1472 –1476

Angewandte
Chemie
the alcohol in a poor yield (6%), whereas IMes and IPr
afforded high yields (84% and 94%, respectively). Con-
versely, a rhodium catalyst with bsp ([{RhCl(C2H4)2}2] with 2;
1.0 mol% rhodium, P/Rh = 1)[2a,b] was not as effective for the
bulky ketone (Table 1, entry 10).
The efficacy of the bsp ligands was further confirmed
using various ketones (Table 2). The results using PPh3 were
also listed for comparison, because PPh3 represents the core
by the Cu–bsp catalyst systems (Table 2, entries 10,11, and
14). Aromatic ketones, such as cyclohexyl phenyl ketone (4g),
tert-butyl phenyl ketone (4h), and 2-acetylthiophene (4i),
were smoothly reduced in high yields (Table 2, entries 17, 19,
20, and 22), whereas the attempted hydrosilylation reactions
of methyl 2,4,6-trimethylphenylketone was unsuccessful using
ligands 1–3.
Bulky ketones bearing multiple substituents (4j–4m)
were converted into their corresponding alcohols in 15 min
at room temperature whilst retaining their functionalities in
excellent to good isolated yields (Scheme 1). Furthermore,
Table 2: Hydrosilylation of various ketones.[a]
Entry
Substrates
Ligand
t [min]
Yield [%][b]
1
2
3
1
3
PPh3
10
98
68
23
4
5
6
2
3
PPh3
360
20
28[c]
98[d]
<1
7
8
9
2
3
PPh3
99
Scheme 1. Reduction of ketones 4j–4m to alcohols. Reaction condi-
tions: a) CuCl (1.0 mol%), 3 (1.0 mol%), tBuONa (6.0 mol%),
Ph2SiH2 (2.4 mmol), toluene, RT, 15 min; b) HCl/MeOH.
99 (83)[e]
8
10
11
12
2
3
PPh3
40
90
60
10
79
95 (89)[e]
3
the Cu–bsp catalyzed hydrosilylation of estrone derivative
(4n), which contains an allyl ether functionality, proceeded
chemoselectively toward the carbonyl moiety, affording the
isolated allyloxy product in 79% yield (Scheme 2). In
13
14
15
2
3
PPh3
53
83 (75)[e]
4
16
17
18
2
3
PPh3
77
(93)[e]
2
19
20
21
2
3
PPh3
95
99 (87)[e]
9
22
23
24
1
3
PPh3
20
84 (69)[e]
24
5
Scheme 2. Reduction of ketone 4n containing an allyl ether. Reaction
conditions: a) Catalyst system A or B, Ph2SiH2 (1.2 mmol), toluene, RT,
8 h; b) K2CO3/MeOH.
[a] Ketone 4 (2.0 mmol), Ph2SiH2 (2.4 mmol), CuCl (0.02 mmol, 1.0
mol%), phosphane (0.02 mmol, 1.0 mol%), tBuONa (0.12 mmol, 6.0
mol%), toluene (2.0 mL), RT. [b] Yield of alcohol determined using a GC
internal standard. [c] cis/trans=44:56. [d] cis/trans=58:42. [e] Yield of
isolated product.
contrast, the rhodium catalyst system[2a,b] with a bsp ligand
provided a mixture of products from the hydrosilylation of
both the allyl and the carbonyl functional groups.
part of all the bsp ligands (1–3). The less-bulky cyclohexanone
(4b) was easily reduced in 10 min to cyclohexanol in 98% and
68% yields using ligands 1 and 3, respectively (Table 2,
entries 1 and 2), whilst the yield with PPh3 was only 23%
(Table 2, entry 3). The more bulky 2-tert-butylcyclohexanone
(4c) afforded its corresponding alcohol in 98% yield (ligand
3; Table 2, entry 5). The highly congested 2,2,6,6-tetrame-
thylcyclohexanone (4d) readily afforded the corresponding
alcohol in almost quantitative yield in 20 min with either 2 or
3 as the ligand (Table 2, entries 7 and 8, respectively).
Dicyclohexyl ketone (4e) and 1-adamantyl methyl ketone
(4 f) were also readily reduced to their corresponding alcohols
Initial rates of reaction for the hydrosilylations of 4h,
isopropyl phenyl ketone (4o), and acetophenone (4p) with
Ph2SiH2 were found to be 1.1, 4.9 ꢀ 10À1, and 8.5 ꢀ
10À2 molLÀ1 hÀ1, respectively,[5] which indicates that the
reaction was much faster with a more bulky substrate. This
intriguing difference between the rates of reaction was
investigated using a competition reaction between equimolar
amounts of two ketones that have substituents of different
bulkiness (A: more bulky, B: less bulky) in the presence of
half an equivalent of Ph2SiH2 at room temperature (Table 3).
In the competition reaction between 4h and 4p, the more
Angew. Chem. Int. Ed. 2010, 49, 1472 –1476
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1473

Communications
Table 3: Competitive hydrosilylation between two substrates (A and B) of
different bulkiness.[a]
Table 4: Hydrosilylation of bulky ketones in the presence of aldehydes.[a]
Entry
Substrates
t [h]
Yield [%][b]
A C
Entry
Substrates
T [8C]
t [h]
Yield [%][b]
A
C
A
B
A
B
1
3
90
8
41
56
55
5
86
46
41
43
1
RT
0.2
88
87
2
3
2[c]
3[d]
4[e]
5[f]
2[c]
3
RT
0.3
89
21
11
79
4[d]
6
2.5
92
6
7
8
12
3
81
87
12
12
5
RT
RT
0.5
0.2
3
71
99
87
78
17
11
13
25
C6H13CHO
6[e]
7[e]
8[e]
6b
C6H13CHO
9
10
11
2
3
1
92
72
62
9
21
28
À40
À40
6b
C6H13CHO
C6H13CHO
1
6b
6b
[a] Bulky substrate A (2.0 mmol), less bulky substrate B (2.0 mmol),
Ph2SiH2 (2.0 mmol), CuCl (0.02 mmol), (0.02 mmol), tBuONa
(0.12 mmol), toluene (2.0 mL). [b] Yield of alcohol determined using a
GC internal standard. [c] 2 was used instead of 3. [d] PPh3 was used
instead of 2. [e] Toluene (4.0 mL).
3
[a] Bulky ketone
A
(2.0 mmol), aldehyde (2.0 mmol), Ph2SiH2
(2.0 mmol), CuCl (0.02 mmol), 3 (0.02 mmol), tBuONa (0.12 mmol),
toluene (4.0 mL) at À408C. [b] Yield of alcohol determined using a GC
internal standard. [c] PPh3 was used instead of 3. [d] ICy was used
instead of 3. [e] IMes was used instead of 3. [f] IPr was used instead of 3.
bulky 4h was preferentially reduced using either 3 or 2 as the
ligand (Table 3, entries 1 and 2). Similarly, the more bulky 4d
was reduced preferentially over 4b using ligand 3 (Table 3,
entry 3). However, when PPh3 was used as the ligand, the less
bulky 4b was preferentially reduced, as expected (Table 3,
entry 4). Higher reactivity of the more bulky 4a over the less
bulky 5-nonanone (4q) was also observed (Table 3, entry 5).
In the competition reaction between 4o and 4p (isopropyl
ketone versus methyl ketone), the alcohol from the isopropyl
ketone substrate was obtained as the major product (Table 3,
entry 6). Even when two competing aldehydes were
employed, the more bulky substrates (6a and 6c) were
predominantly reduced over heptanal (6b; Table 3, entries 7
and 8).
The most remarkable feature of this bsp catalyst system is
the preferential reduction of a ketone (4) in the presence of an
aldehyde (6). The hydrosilylation reaction of an equimolar
mixture of 4h, 2,4,6-trimethylbenzaldehyde (6d), and
Ph2SiH2, using 3 as the ligand, afforded the corresponding
alcohol from the ketone in 90% yield and the corresponding
alcohol from the aldehyde in only 5% yield (Table 4, entry 1).
When PPh3 was used instead of 3, the alcohol from the
aldehyde was predominantly obtained (Table 4, entry 2).
NHC ligands, such as ICy, IMes, and IPr, were highly efficient
in the copper-catalyzed hydrosilylation of bulky ketones.[4c,7]
However, when these ligands were used, no selectivity in
hydrosilylation was observed between ketone and aldehyde
substrates (Table 4, entries 3–5). In the reactions of 4h with 4-
methoxybenzaldehyde (6e), benzaldehyde (6 f), and even 6b,
the alcohol from 4h was obtained preferentially in 92%,
81%, and 87% yields, respectively (Table 4, entries 6–8).
Surprisingly, even highly congested ketones, such as 4d and
4a, were preferentially reduced in the presence of aldehydes
(6b or 6d) to their corresponding alcohols in 92%, 72%, and
62% yields, respectively (Table 4, entries 9–11). To the best of
our knowledge, there have been only six precedents for the
preferential reduction of a ketone in the presence of an
aldehyde.[10] However, all of these previous reactions neces-
sitated the prior in situ protection of the more reactive
aldehyde, followed by the reduction of the unprotected
ketone and successive deprotection of the aldehyde during
work-up. However, in these reactions, the in situ protections
were significantly affected by subtle changes in the reaction
conditions, thus making these methods unreliable.
The hydrosilylation of diketone 7 was carried out as
shown in Scheme 3. After 3 h, almost all of 7 had been
consumed, and the reaction mixture contained 8 as the major
1474
www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 1472 –1476

Angewandte
Chemie
A possible catalytic cycle for this hydrosilylation reaction
is shown in Figure 3. The key intermediates in the cycle are
the bsp-bearing copper hydride species 11 and alkoxide 12.
Scheme 3. Reduction of diketone 7. Reaction conditions: a) CuCl
(2.0 mol%), 3 (2.0 mol%), tBuONa (12.0 mol%), Ph2SiH2 (1.0 mmol),
toluene, RT, 3 h; b) HCl/MeOH.
product (78% yield) with a small amount of the correspond-
ing diol from the reduction of the both keto groups (5%
yield). Thus, the more bulky ketone functionality of 7 was
preferentially reduced during the reaction. In the hydro-
silylation reaction of 9, which has both ketone and the formyl
substituents (Scheme 4), 9 was fully consumed within 17 h and
Figure 3. Plausible reaction mechanism.
Typically, copper hydrides prefer to aggregate; indeed, the
copper hydride complex with PPh3 was isolated as the
hexamer [(Ph3PCuH)6].[12] In this cycle, the complexation of
11 with a ketone or an aldehyde would reversibly afford 12
(step a). Here, 12 from bulky ketones could be of much lower
nuclearity (possibly, mono) owing to the bulkiness of the bsp
and bulky alkoxide moieties. Such a highly unsaturated
complex 12 would be extremely reactive for s-bond meta-
thesis[13] with a silane to afford the product and regenerate 11
(step b). Conversely, following complexation with a less bulky
ketone or aldehyde, 12 might be susceptible to aggregation
owing to the smaller alkoxide moieties, thus their relative
reactivity in step b would be low.
In conclusion, we have developed a highly active copper
catalyst using a bsp ligand for the hydrosilylation reaction of
ketones, and these reactions proceed faster with more bulky
ketone substrates. It is noteworthy that the present catalysts
afford the unprecedented preferential reduction of a bulky
ketone in the presence of an unprotected aldehyde. Further
studies on the reaction mechanism and characterization of the
catalytic species are underway.
Scheme 4. Reduction of ketoaldehyde 9. Reaction conditions: a) CuCl
(2.0 mol%), 3 (2.0 mol%), tBuONa (12.0 mol%), Ph2SiH2 (1.0 mmol),
toluene/CH2Cl2, À40 8C, 17 h; b) K2CO3/MeOH.
the resulting reaction mixture contained 10 as the major
product, with a small amount of the corresponding diol from
the reduction of both the keto and the formyl moieties (3%
yield) and the mono alcohol bearing a keto functionality from
the reduction of only the formyl moiety (3% yield). The pure
compound 10 was isolated in 69% yield from the reaction
mixture, thereby indicating that the keto functionality had
been preferentially reduced.
It is well-known that copper complexes are easy to
aggregate.[11] Indeed, it was reported that the copper tetramer
with a cubane structure, [{CuCl(PPh3)}4], was obtained from
the reaction of an equimolar mixture of CuCl and PPh3.[12] In
contrast, when the similar reaction of an equimolar mixture of
CuCl and 2 was carried out, a lower-nuclearity complex, the
copper dimer [{CuCl(2)}2] was obtained in 54% yield; the
structure was confirmed by X-ray crystallography
(Figure 2).[8] These results suggest that the unique steric
bulk of 2 could suppress the aggregation of copper centers.
Received: November 11, 2009
Revised: December 3, 2009
Published online: January 25, 2010
Keywords: aldehydes · copper · hydrosilylation · ketones ·
.
P ligands
[1] a) P. W. N. M. van Leeuwen, Homogeneous Catalysis, Kluwer
Academic, Dordrecht, 2004; b) Comprehensive Organometallic
Chemistry III, Vol. 10 and 11 (Eds.: I. Ojima, T. Hiyama),
Elsevier Science, Amsterdam, 2006.
[2] For catalytic reactions using bsp as the ligand, see: a) O.
Niyomura, M. Tokunaga, Y. Obora, T. Iwasawa, Y. Tsuji,
Angew. Chem. 2003, 115, 1325 – 1327; Angew. Chem. Int. Ed.
2003, 42, 1287 – 1289; b) O. Niyomura, T. Iwasawa, N. Sawada,
M. Tokunaga, Y. Obora, Y. Tsuji, Organometallics 2005, 24,
Figure 2. Crystal structure of [{CuCl(2)}2].
Angew. Chem. Int. Ed. 2010, 49, 1472 –1476
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1475

Communications
3468 – 3475; c) H. Ohta, M. Tokunaga, Y. Obora, T. Iwai, T.
data for this paper. These data can be obtained free of charge
from the Cambridge Crystallographic Data Centre via www.
ccdc.cam.ac.uk/data_request/cif. For further details, see the
Supporting Information.
Iwasawa, T. Fujihara, Y. Tsuji, Org. Lett. 2007, 9, 89 – 92.
[3] a) C. Deutsch, N. Krause, B. H. Lipshutz, Chem. Rev. 2008, 108,
2916 – 2927; b) S. Rendler, M. Oestreich, Angew. Chem. 2007,
119, 504 – 510; Angew. Chem. Int. Ed. 2007, 46, 498 – 504.
[4] To date, only two catalyst systems have been reported for the
hydrosilylation of 4a: a) G. Hamasaka, A. Ochida, K. Hara, M.
Sawamura, Angew. Chem. 2007, 119, 5477 – 5479; Angew. Chem.
Int. Ed. 2007, 46, 5381 – 5383; b) G. Hamasaka, S. Kawamorita,
A. Ochida, R. Akiyama, K. Hara, A. Fukuoka, K. Asakura, W. J.
Chun, H. Ohmiya, M. Sawamura, Organometallics 2008, 27,
6495 – 6506; c) S. Dꢁez-Gonzꢂlez, H. Kaur, F. K. Zinn, E. D.
Stevens, S. P. Nolan, J. Org. Chem. 2005, 70, 4784 – 4796.
[5] See the Supporting Information.
[6] a) B. A. Baker, Z. V. Boskovic, B. H. Lipshutz, Org. Lett. 2008,
10, 289 – 292; b) B. H. Lipshutz, J. M. Servesko, B. R. Taft, J. Am.
Chem. Soc. 2004, 126, 8352 – 8353; c) B. H. Lipshutz, K. Noson,
W. Chrisman, A. Lower, J. Am. Chem. Soc. 2003, 125, 8779 –
8789.
[7] a) S. Dꢁez-Gonzꢂlez, E. D. Stevens, N. M. Scott, J. L. Petersen,
S. P. Nolan, Chem. Eur. J. 2008, 14, 158 – 168; b) S. Dꢁez-
Gonzꢂlez, N. M. Scott, S. P. Nolan, Organometallics 2006, 25,
2355 – 2358; c) H. Kaur, F. K. Zinn, E. D. Stevens, S. P. Nolan,
Organometallics 2004, 23, 1157 – 1160.
[9] a) J.-L. Luche, A. L. Gemal, J. Am. Chem. Soc. 1979, 101, 5848 –
5849; b) A. L. Gemal, J.-L. Luche, J. Org. Chem. 1979, 44, 4187 –
4189; c) M. P. Paradisi, G. P. Zecchini, G. Ortar, Tetrahedron
Lett. 1980, 21, 5085 – 5088; d) T. Chihara, T. Wakabayashi, K.
Taya, H. Ogawa, Can. J. Chem. 1990, 68, 720 – 724; e) J. H. An,
T. B. Sim, J. Choi, N. M. Yoon, Bull. Korean Chem. Soc. 1997, 18,
111 – 113; f) A. Clerici, N. Pastori, O. Porta, Eur. J. Org. Chem.
2002, 3326 – 3335.
[10] G. van Koten, J. G. Noltes in Comprehensive Organometallic
Chemistry, Vol. 2 (Eds.: G. Wilkinson, F. G. A. Stone, E. W.
Abel), Pergamon, Oxford, 1982; pp. 709 – 763.
[11] a) G. Costa, E. Reisenhofer, L. Stefani, J. Inorg. Nucl. Chem.
1965, 27, 2581 – 2584; b) M. R. Churchill, K. Kalra, Inorg. Chem.
1974, 13, 1065 – 1071.
[12] a) W. S. Mahoney, D. M. Brestensky, J. M. Stryker, J. Am. Chem.
Soc. 1988, 110, 291 – 293; b) W. S. Mahoney, J. M. Stryker, J. Am.
Chem. Soc. 1989, 111, 8818 – 8823.
[13] a) T. Gathy, D. Peeters, T. Leyssens, J. Organomet. Chem. 2009,
694, 3943 – 3950; b) S. Rendler, O. Plefka, B. Karatas, G. Auer, R.
Frꢃchlich, C. Mꢄck-Lichtenfeld, S. Grimme, M. Oestreich,
Chem. Eur. J. 2008, 14, 11512 – 11528.
[8] CCDC 753994 ([CuCl(2)]2·1.5(C4H10O)·0.5(CH2Cl2)), 753995
(8), and 753996 (10) contain the supplementary crystallographic
1476
www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 1472 –1476