Scheme 1. Chemoselective Borohydride Reduction of Aromatic vs. Aliphatic Ketones
converted into the corresponding optically active 2-substituted-
aromatic alcohol was 67% ee. When only 0.25 equiv of the
premodified borohydride was used, high enantioselectivity
was realized with high chemo- and diastereoselectivity.
These observations suggested that the excess hydride in
the catalytic system caused noncatalytic reduction, resulting
in low selectivities. To maintain the initial reaction condi-
tions, therefore, five portions of the 0.1 equiv of the
premodified borohydride were successively added at 1 h
intervals to the reaction to obtain the 3-aryl-3-hydroxyketones
in 43% yield with 97% chemoselectivity, 99% anti-selectiv-
ity, and 94% enantiomeric excess (Scheme 2). The enantio-
meric excesses of the 2-methyl-1,3-diketone remaining after
the kinetic resolution were determined by HPLC. Since
racemization of 2-substituted-1,3-diketones gradually pro-
ceeded at room temperature, the reaction mixture was directly
injected into an HPLC chiral column (Daicel chiralpak AD,
5% 2-propanol in n-hexane) to determine the ee of 2,4-
dimethyl-1-phenyl-1,3-pentanedione to be 99% ee. These
observations indicated that the cobalt complex catalyzed
reduction selectively afforded one isomer among the possible
eight isomers and that the kinetic resolution was excellent.
After the detailed optimization,7 the present kinetic resolu-
tion system was successfully applied to enantioselective
reduction of various 2-substituted-1-alkyl-3-aryl-1,3-dike-
tones to optically active 2-substituted-3-hydroxyketones
(Table 1). The 1,3-diketones having 2-methyl, 2-ethyl, and
2-allyl groups were converted into the corresponding 3-aryl-
3-hydroxyketones with high chemo-, diastereo-, and enan-
tioselectivities (entries 1-3). Kinetic resolution in the
enantioselective reduction of the substrate containing a tert-
butyl ketone (entry 4) showed that the corresponding reduced
product was obtained in 48% yield and indicated 99%
chemoselectivity, 99% anti-selectivity, and 97% ee. The
present highly selective kinetic resolution could also be
applied to substrates having primary alkyl ketones, such as
n-nonyl ketone, isobutyl ketone, and benzyl ketone, to obtain
the corresponding anti-hydroxyketones with high selectivities
(entries 5-7).
3-hydroxyketones with excellent diastereoselectivity and
enantiomeric excess with high catalytic efficiency catalyzed
by the optically active â-ketoiminato cobalt complexes.5 In
this Letter, we would like to describe that a highly chemo-,
diastereo-, and enantioselective borohydride reduction of
unsymmetrical 2-substituted-1,3-diketones was achieved in
the presence of the optically active â-ketoiminato cobalt
complexes to afford optically active anti-2-substituted-3-
hydroxyketones.
A preliminary examination of the borohydride reduction
using a catalytic amount of the cobalt complexes revealed
that aromatic ketones were preferentially reduced in the
presence of aliphatic ketones. As shown in Scheme 1, to the
methanol solution of 0.5 mmol of 2-undecanone and 0.5
mmol of 2-acetonaphthone was added 0.1 mmol of sodium
borohydride. After 8 h, 2-undecanone, an alkyl ketone, was
reduced to the corresponding alcohol in 25% yield and
2-acetonaphthone, an aromatic ketone, in 11% yield, respec-
tively. The chemoselectivity for the reduction of the aliphatic
ketone was about 70%. In contrast, in the presence of 0.05
mmol of â-ketoiminato cobalt complex 1, the chemoselec-
tivity was completely reversed. By treatment of the pre-
modified borohydride,6 an aromatic ketone, 2-acetonaphtho-
ne, was selectively reduced to 1-(2-naphthyl)-1-ethanol in
45% yield while the aliphatic ketone was reduced in only
5% yield. The chemoselectivity of the aromatic ketone was
90%. These observations encouraged us to apply the cobalt-
catalyzed reduction to 1-alkyl-3-aryl-1,3-diketones to prepare
the corresponding 1-alkyl-3-aryl-3-hydroxyketones.
As the unsymmetrical 2-substituted-1,3-diketone model for
chemo-, diastereo-, and enantioselective reduction, 2,4-
dimethyl-1-phenyl-1,3-pentanedione was adopted. Since
kinetic resolution must be considered for the model substrate,
0.5 equiv of the premodified borohydride was employed in
the presence of 5 mol % of the optically active â-ketoiminato
cobalt complex catalyst 2. After 24 h, the reaction was
quenched to afford the corresponding hydroxyketones in 44%
yield with 88% aromatic vs 12% aliphatic alcohol. Though
the diastereoselectivity in the aromatic alcohol was deter-
mined to be 93% anti, the enantioselectivity of the anti-
The excellent stereoselectivity in the present catalytic
reduction system can be explained as follows: In the
presence of (R,R)-cobalt catalyst, the enantioselective reduc-
(5) Ohtsuka, Y.; Koyasu, K.; Ikeno,; T. Yamada, T. Org. Lett. 2001, 3,
2543.
(6) Sugi, K. D.; Nagata, T.; Yamada, T.; Mukaiyama, T. Chem. Lett.
1996, 1081.
(7) To reduce the excess hydride in the catalytic system and avoid
noncatalytic reduction further, four portions of the 0.1 equiv of the
premodified borohydride were successively added at 2 h intervals to the
reaction mixture.
3422
Org. Lett., Vol. 3, No. 21, 2001