Noyori-Type Catalysts in Aqueous Media
COMMUNICATION
Table 3. Optimization experiments for substrates 17 and 18, resistant to
standard conditions.
thanide and bismuth
reactions with imines.
when silver and Lewis acid activation were used together
Table 3, entries 7–12). Increasing the size of the cation in
ACHTUNGTRENUNNG( III) triflates are known to accelerate
[
a]
[21]
The best results were obtained
[
b]
[c]
[d]
Entry Substrate Additive
Conv.
%]
ee
[%]
Configuration
[
(
[
e]
3+
3+
3+
1
2
3
4
5
6
7
8
9
1
1
1
18
18
18
17
17
17
17
17
17
17
17
17
none
AgSbF
none, cat. 16
none
40
90
88
5
27
20
31
60
80
72
54
87
ND
S
S
S
–
S
–
S
S
S
S
S
S
the series Sc <Y <La
afforded better conversions
[
[
f]
f]
6
98.5
99
[.22]
(
compare Table 3, entries 7–9);
the use of bismuth also
[
e]
gave good results (Table 3, entry 12) and good enantioselec-
tivities.
ND
97
ND
90
92
94
AgSbF
La(OTf)
6
[
e]
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
3
The results of the application of the double activation
method to the reduction of other aromatic/benzylic sub-
strates are presented in Table 4 (entries 1–13). To summarize
the results (compare Tables 1, 2, and 3), imines with alkyl
substituents could be reduced without any additive, but for
aromatic substituents both Lewis acid activation and catalyst
AgSbF
AgSbF
AgSbF
AgSbF
AgSbF
AgSbF
6
6
6
6
6
6
/Sc ACHTUNGTRENNUNG( OTf)
3
/Y
/La
/Ce
/Yb
/Bi(OTf)
A
H
U
G
R
N
U
G
3
A
H
U
G
R
N
U
G
3
3
0
1
2
A
H
U
G
R
N
U
G
93.5
94.5
94
A
T
N
T
E
N
N
3
A
T
N
T
E
N
N
3
[
(
a] Reaction time: 16 h. See the Experimental Section for conditions.
0.125 mmol scale, 1.3 mol% catalyst, 2.4 mol% AgSbF ) [b] Conv.=con-
activation (AgSbF ) were indispensable. The benzyl-substi-
6
6
tuted substrates and the fused iminium salts represent inter-
mediate cases in which only the catalyst needs to be activat-
ed with silver salts.
version, as determined by crude NMR spectroscopy. [c] The ee value was
determined by HPLC (Daicel-OD column). [d] Hydrogenation products
are levorotary. [e] ND=not determined. [f] Isolated yield.
Although the conversions with the surfactant method
were typically excellent, purification of the products from
the surfactant residues was more problematic with aromatic
substrates (especially 17 and 20a). In addition, thiophenyl-
substituted 20b was surprisingly unreactive (Table 4, en-
tries 7 and 8). For these cases, we found that aqueous meth-
Table 4. Optimization of the Ag-Lewis acid activated transfer hydrogen-
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
ations in aqueous conditions and comparison with H
2
O/MeOH condi-
[
a]
At ions.
C
H
T
U
N
G
T
R
E
N
N
U
N
G
[
d,e]
Entry Substrate Conditions
Additive
Yield
ee
[
b]
A
H
U
G
R
N
U
G
anol (2:1 H O/MeOH) could be used (Table 4, entries 9–13)
2
[23]
as a simple alternative to the surfactant conditions. Of the
solvent mixtures tested (mixtures of water with methanol,
ethanol, acetonitrile, and THF), aqueous methanol (2:1
1
2
3
4
5
6
7
8
9
1
1
1
1
18
19
19
20a
20a
20a
20b
20b
8a
H
H
H
H
H
H
H
H
H
H
H
H
H
2
2
2
2
2
2
2
2
2
2
2
2
2
O/CTAB AgSbF
O/CTAB AgSbF
O/CTAB AgSbF
O/CTAB AgSbF
O/CTAB AgSbF
O/CTAB AgSbF
O/CTAB AgSbF
O/CTAB AgSbF
O/MeOH AgSbF
O/MeOH AgSbF
O/MeOH AgSbF
O/MeOH AgSbF
O/MeOH AgSbF
6
6
6
6
6
6
6
6
6
6
6
6
6
90 (99)
98
99
99
87
70
60
90
94
99
95
93
98
94
[
[
f]
f]
[c]
/La
A
T
N
R
N
U
G
3
(94)
[c]
/Bi(OTf)
A
H
N
T
E
N
N
(99)
[
[
[
b]
c]
c]
(40)
(99)
(99)
(27)
(22)
H O/MeOH) afforded the highest rates. This protocol was
2
/La
/Bi(OTf)
/La(OTf)
/Bi(OTf)
A
H
U
G
R
N
U
G
especially useful for lipophilic substrates with aromatic sub-
stituents, such as 17, 19, and 20b. The optimized procedures
A
H
U
T
E
U
A
H
U
T
E
U
G
A
H
U
T
E
U
are summarized in Table 4. The H O/MeOH system allows
2
59 (99)
90 (99)
80 (99)
78 (99)
50 (54)
for easier purification and this procedure is more convenient
to scale up. However, for the alkyl-substituted imines, such
as 8a (Table 4, entry 9) and most polycyclic iminium salts
(see Table 2), the CTAB method afforded better results
(only the reduction of 13b proceeded with a higher yield in
0
1
2
3
13b
17
19
/La
/La
/La
A
H
U
G
R
N
N
ACHTUNGTRENNUNG
20b
ACHTUNGTRENNUNG
[
[
a] Reaction time: 16 h. See the Experimental Section for conditions.
b] Conv.=conversion, as determined by NMR spectroscopy. [c] The iso-
[24]
aqueous methanol).
lated yields in these cases were generally 20–30% lower due to difficul-
ties in removing the CTAB residues. [d] The ee value was determined by
HPLC (Daicel-OD column). [e] All hydrogenation products are levorota-
ry and are likely to have the same configuration. [f] Reaction time 40 h.
Although it has been established that the reductions pro-
ceed through protonated substrates, it is interesting to note
that the reductions of alkyl-substituted imines do proceed
under our (slightly basic) aqueous conditions. Blackmond
et al. have suggested that reductions of imines with Rh–di-
ly suggest the in situ generation of cationic intermediate 16.
Interestingly, the less bulky BF4 and PF6 salts were not ef-
fective and even decreased the catalytic activity.
Although benzyl-substituted substrate 18 (Scheme 6) was
readily reduced under the micelle conditions, substrates with
aryl substituents conjugated to the imine system (e.g., 17,
ACHTUNGTRENNUaGN mine catalysts might proceed via the unprotonated
À
À
[25]
imine.
However, to explain the observed rate enhance-
ment effect of metal triflates, we suggest that these reduc-
tions proceed by the ionic “anti” mechanism proposed by
[25b]
Wills et al.
Although the NÀH bond of the catalyst al-
ready activates the imine, the Lewis acid may provide addi-
tional activation (Scheme 7). In most cases this extra activa-
tion does not appear to disturb the enantioselectivity of the
catalyst; however, in the case of furan 20a, the enantioselec-
tivity was clearly eroded (compare entries 4–6, Table 4), pos-
sibly due to complexation of the furan ring with lanthanum.
In summary, we have developed two alternative modifica-
tions to the original Noyori reduction procedure by per-
forming the reaction under aqueous conditions and using
19, 20a, and 20b; Scheme 6) turned out to be more chal-
lenging. For the asymmetric reduction of 17, a work-around
based on the reduction of N-benzyl iminium derivatives has
[12b]
been proposed.
It has previously been established that under the original
Noyori conditions the reduction proceeds via a protonated
[20]
iminium species. We hypothesized that activation of the
imine with water-soluble Lewis acids might be possible. Lan-
Chem. Eur. J. 2009, 15, 12963 – 12967
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
12965