yl group in 4b and 5b had some negative effects on the re-
activity (entries 3 and 4). Moreover, this also caused a slight
decrease in the enantioselectivity of 5b compared to 5a
(entry 4 vs 2).
Astonishing solvent effects were observed with the pres-
ent catalytic system. When toluene was replaced with either
dichloromethane or chloroform, which are the best solvents
for the previously reported Lewis base catalysts for similar
transformations,[5–8] a dramatic decrease in both reactivity
and enantioselectivity was observed in the 5a-catalyzed re-
duction (entries 5 and 6). Interestingly, CCl4 proved to be as
an excellent solvent as toluene in terms of reactivity and
enantioselectivity (entry 9).
When the catalyst loading of 5a was reduced from 20 to
10 mol%, only a marginal effect on either the reactivity or
enantioselectivity (entry 10 vs 2) was observed, whereas fur-
ther lowering of the catalyst loading to 5 mol% led to a de-
crease of the enantioselectivity from 96 to 92% ee (entry 11
vs 10). A decrease of the enantioselectivity from 96 to 88%
ee was observed when the reaction temperature was lowered
from À20 to À408C (entry 12 vs 10). When the reaction was
performed at 08C, a more practical and preferable reaction
temperature, excellent reactivity and enantioselectivity in
both toluene and CCl4 remained (entries 13 and 14).
After the reaction conditions had been optimized, we set
out to examine the substrate spectrum of the present cata-
lyst system. A wide variety of aromatic N-alkyl ketimines
(6a–x) was reduced in the presence of catalyst 5a (Table 2).
Generally, the desired amine were obtained in high yields
and excellent enantioselectivities with toluene as the solvent
(method A).[10] In particular, ee values up to 99.6% were
achieved in the reduction of the p-nitrophenyl methyl
ketone derived N-benzyl imine 6h (entry 8). To the best of
our knowledge, this is the highest enantioselectivity ach-
ieved for an organocatalytic reduction of imines.
Figure 3. Catalysts evaluated in this study.
Starting from l-proline, compounds 4a and 4b and their
diastereomers 5a and 5b were easily synthesized as a mix-
ture at a diastereomeric ratio close to 1:1 and purified by
column chromatography (see Supporting Information for
the experimental details and analytic data). The stereochem-
istry of the chiral sulfur centers in 4a and 5a was deter-
mined by single-crystal x-ray diffraction analysis.[9] For 4b
and 5b, the stereochemistry on the sulfur atom was estab-
1
lished by a clear analogy of their H NMR profiles to those
of 4a and 5a.
Initial attempts to evaluate 4a and 5a as Lewis basic cata-
lysts (20 mol%) for the reduction of N-benzyl ketimine 6a
by HSiCl3 in toluene at À208C gave very good results: the
desired amine 7a was obtained in high yield in the presence
of both diastereomeric catalysts in 24 h (entries 1 and 2,
Table 1); catalyst 5a with an R-configuration on the sulfur
atom afforded a high ee up to 97%. Interestingly, a dramati-
cally lower ee was obtained with catalyst 4a bearing an S-
configured sulfur atom, suggesting that a stereochemistry
match between the chiral carbon and sulfur centers is criti-
cal for the stereocontrol in the course of asymmetric induc-
tion.
Replacement of the electron-rich 3,5-dimethylphenyl
group in 4a and 5a with electron-deficient 3,4-difluorophen-
Table 1. Asymmetric reduction of ketimine 6a under various condi-
tions.[a]
Although a few substrates gave only moderate results (en-
tries 10, 12, and 20–22), change of the solvent to CCl4
(method B) still gave high enantioselectivities. It should be
noted that the ee values of the amine products do not corre-
spond to the E/Z ratio of the starting ketimines,[11] for which
we have no explanation at present.
Entry Catalyst
(mol%)[b]
Solvent
T
[8C]
Yield
[%][c]
ee
[%][d,e]
1
2
3
4
5
6
7
8
4a (20)
5a (20)
4b (20)
5b (20)
5a (20)
5a (20)
5a (20)
5a (20)
5a (20)
5a (10)
5a (5)
toluene
toluene
toluene
toluene
CH2Cl2
CHCl3
À20
À20
À20
À20
À20
À20
95
96
80
86
76
55
90
91
93
95
95
90
98
90
22
97
37
94
22
32
47
55
97
96
92
88
96
97
In summary, a highly enantioselective catalytic method
has been developed for the reduction of aromatic N-alkyl
ketimines by trichlorosilane under mild conditions using the
newly designed Lewis base organocatalyst 5a that incorpo-
rates C- and S-chirality. Excellent enantioselectivities of up
to 99.6% ee and high yields were obtained for a wide range
of substrates. Further work is in progress to clarify the
mechanism of the transformation and explore the full appli-
cation scope of the present catalyst system.
ClCH2CH2Cl À20
THF
À20
À20
À20
À20
À40
0
9
CCl4
10
11
12
13
14
toluene
toluene
toluene
toluene
CCl4
5a (10)
5a (10)
5a (10)
0
Experimental Section
[a] Reactions were carried out on a 0.1 mmol scale with 2.0 equiv of
HSiCl3 in 0.5 mL of solvent for 24 h. [b] The molar percentage is based
on imine. [c] Isolated yield based on imine. [d] The ee values were deter-
mined using chiral HPLC. [e] Product 7a was R configured in all cases,
as revealed by comparison of the optical rotation with the literature data.
General procedure for the asymmetric reduction: Under an argon atmos-
phere, trichlorosilane (20 mL, 0.20 mmol) was added dropwise to a stirred
solution of imine 6 (0.10 mmol) and catalyst (0.01 mmol) in anhydrous
toluene (method A) or CCl4 (method B) at 08C. The mixture was al-
8790
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 8789 – 8792