Angewandte
Chemie
Table 1: Asymmetric hydrosilane reduction with 2a or Zn salts in the
is involved. It may also be possible that a hydride is directly
transferred from the hydrosilane.
presence of various additives.[a]
Other hydrosilanes, including (EtO)3SiH, Ph3SiH, and
Ph2SiH2, were tested with the catalyst 2c and exhibited similar
activities, giving 65–71% ee with the same absolute config-
uration (S) as that obtained with (EtO)2MeSiH (Table 2).
Thus, the observed effect of a change in the absolute
configuration of the product was not influenced by the
hydrosilanes.
Entry Cat., Additive (mol%)
t
Yield [%]
ee [%]
[h] (recov. 3 [%])
1
2
3
4
2a, AgOAc (9)
2a, AgBF4 (9)
2a, NaOAc (9)
2a, NaOtBu (9)
2a, Cu (10)
24 n.r.
24 n.r.
24 97
–
–
57 (R)
55 (R)
24 99
5
48 n.r.
24 n.r.
24 92
24 60 (40)
48 97
6
2a, Mn (10)
7
2a, Mg (10)
15 (R)
44 (S)
41 (S)
8
2a, Zn (6)
Table 2: Asymmetric hydrosilane reduction with various hydrosilanes.[a]
9
2a, Zn (6)
10
11
12
13[b]
14[b]
15[b]
16
17
18
2a, Zn (6), 1a (7)
2a, ZnEt2 (5)
2a, ZnCl2 (4.5)
–, ZnCl2 (5)
–, ZnCl2 (5), 1a (7)
–, Zn(OAc)2 (5), 1a (7)
–, Fe(OAc)2 (5), 1a (6), Zn (8) 48 83 (17)
2b, Zn (6)
2c, Zn (6)
48 n.r.
24 64 (36)
24 n.r.
24 97
48 n.r.
48 96 (3)
33 (S)
–
21 (R)
23 (R)
21 (S)
65 (S)
Entry
Hydrosilane
Yield [%]
(recov. 3 [%])
ee [%]
48 67 (32)
48 98 (2)
1
2
3
(EtO)2MeSiH
(EtO)3SiH
Ph3SiH
Ph2SiH2
Ph2MeSiH
98 (2)
99
97
92 (2)
n.r.
65 (S)
71 (S)
67 (S)
70 (S)
–
[a] Reaction conditions: Cat. 2a (5 mol%), 3 (0.5 mmol), (EtO)2MeSiH
(1 mmol), THF (1.5 mL), 658C, then H3O+. All reported yields are of the
isolated product. [b] 3 (1 mmol), THF (3 mL), 658C.
4[b]
5
[a] Reaction conditions: Cat. 2c (5 mol%), 3 (0.5 mmol), hydrosilane
(1 mmol), THF (1.5 mL), 658C, 48 h, then H3O+. All reported yields are
of the isolated product. [b] 72 h.
(Table 1, entry 8), and surprisingly the product alcohol 4 had
an absolute configuration of S (44% ee). The reaction that
was run for 48 hours produced 4 in 97% yield and 41% ee
(Table 1, entry 9), and the addition of extra 1a (7 mol%)
negated the effect of the zinc metal (Table 1, entry 10). When
diethylzinc (5 mol%) was used instead of Zn, it activated the
complex 2a, giving predominantly the S enantiomer with
33% ee (Table 1, entry 11). However, using ZnCl2
(4.5 mol%) as an additive showed no activation (Table 1,
entry 12). Although ZnCl2 itself was found to promote the
reaction, giving the alcohol in 97% yield but as a racemic
mixture (Table 1, entry 13),[15,16] the combination of 1a and
ZnCl2 did not work as a catalyst (Table 1, entry 14). However,
the combination of Zn(OAc)2 and 1a did promote the
reaction, giving 96% yield of the alcohol 4 with an R confi-
guration in 21% ee (Table 1, entry 15). The addition of Zn
powder to the catalyst generated in situ from Fe(OAc)2 and
1a decreased the enantioselectivity to 23% ee compared to
61% ee obtained without the Zn powder (Scheme 1a versus
Table 1, entry 16). The use of other complexes, such as 2b and
2c, in combination with zinc powder (6 mol%) were also
effective, giving the S enantiomer in 21% ee and 65% ee,
respectively (Table 1, entries 17 and 18).
We have successfully activated the Fe complexes 2 by the
addition of a small amount of zinc powder at 658C. Not only
does the catalyst combination promote hydrosilylation of the
ketone but it also results in a change in the absolute
configuration of the products. The experiments shown in
entries 10 and 12–14 in Table 1 ruled out the possibility that
only the zinc bearing the chiral ligand was involved in the
asymmetric induction. These findings imply that a combined
Fe/Zn complex may serve as the catalyst or that the Fe and Zn
atoms take part in the reaction simultaneously. At this point,
we cannot specify which hydride metal species, Fe–H or Z–H,
The reduction of other ketones was carried out using two
different methods (Methods A and B) so as to compare the
resulting enantioselectivity (Table 3); for the results of
Method B, some previous data are cited. Methyl ketones
bearing substituted phenyl groups resulted in the formation of
the corresponding S-configured secondary alcohols in high
yields (Table 3, entries 1–6). Naphthalenyl ketones 5g and 5h
were reduced with 75% ee (S) and 82% ee (S), respectively
(Table 3, entries 7 and 8). Tetralone derivatives 5i and 5j also
gave the S as the absolute configuration with 80% ee and
83% ee, respectively (Table 3, entries 9 and 10). Interestingly,
the substituted indanone derivatives 5k–5n were reduced to
the S-configured product with up to 95% ee (Table 3,
entries 11–14). Methyl phenethyl ketone (5o) was also
converted into an S-configured secondary alcohol with
33% ee (Table 3, entry 15). Thus, by using Method A, all
ketones were reduced to the corresponding alcohols as
S enantiomers, which is the opposite configuration to that
obtained by using Method B. In the case of benzalacetone 5q,
a 1,2-reduction preferentially proceeded to give the corre-
sponding secondary alcohol in 87% yield with 60% ee
(Table 3, entry 17). The reduction of 2,4,6-trimethylphenyl
methyl ketone as a bulky ketone did not proceed with the iron
complex 2c. Although the reduction of cyclopropyl phenyl
ketone (5r) is very slow, probably a result if the steric
hindrance, it gives 40% of the corresponding secondary
alcohol and no ring-opening product is obtained (Table 3,
entry 18). This fact indicates that the reduction did not
proceed by a radical mechanism.[17] The exceptions to the
trend were ketones 5g, 5o, and 5r, which resulted in the
Angew. Chem. Int. Ed. 2010, 49, 9384 –9387
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
9385