790
D. Guijarro et al. / Tetrahedron Letters 52 (2011) 789–791
transformed into amine 2a at room temperature with an 84% ee
using simple 2-aminoethanol as a ligand in the ruthenium catalyst,
but an excess (3 equiv) of the ligand had to be used and the yield
was very low (11%, Table 1, entry 1).5a Later on, we determined
the optimum reaction conditions for the reductions in isopropanol
with the ruthenium complex prepared with (1S,2R)-1-amino-
2-indanol as a chiral ligand, which involved the use of 4 Å molec-
ular sieves, ButOK as a base and a reaction temperature of 40 °C.5
We decided to try these new conditions in the reaction catalysed
by the complex bearing 2-aminoethanol using 50 mol % of the
ligand and we were pleased to see that the reaction time was
reduced to 6 h and the yield of 2a improved to 89%, maintaining
the enantioselectivity of 84% (Table 1, entry 2). This result encour-
aged us to test some other achiral b-amino alcohols L2–L8 as
ligands for this process. All of these ligands were either commer-
cially available (L1–L6 and L8) or prepared in one step (L7).9 The
obtained results are collected in Table 1. From these results, it
was clear that a b-amino alcohol structure with a primary amino
group in the ligand was needed for the reductions to give good
yields and diastereoselectivities. The extension of the carbon chain
separating the amino and the hydroxy groups of the ligand (L2) or
the methylation of the nitrogen atom (L3) led to incomplete reac-
tions after 22 h with diminished ee’s of around 70% in both cases
(Table 1, entries 3 and 4). The introduction of a bulkier group on
the nitrogen atom had an even more detrimental effect, the yield
of the amine 2a being only 15% (Table 1, entry 5). Having estab-
lished that a 2-aminoethanol skeleton was necessary in the ligand,
we next introduced substituents in both carbon atoms. Ligand L5,
bearing two methyl groups on C2, gave, after desulfinylation, a 94%
yield of amine 2a with a very high enantiomeric excess (97%, Table
1, entry 6). Ligand L6, in which the carbon bearing the amino group
belongs to a cyclopentane ring, gave a 97% yield of the reduction
product with an ee slightly lower than the one obtained with the
dimethyl-substituted ligand L5 (compare entries 6 and 7 in Table
1). The effect of the substituents on the carbinol carbon was stud-
ied with ligands L7 and L8, which gave amine 2a in very good
yields (Table 1, entries 8 and 9), but the stereoselectivities were
lower than the one obtained with the ligand L5. As it was the case
with ligands L5 and L6, the ee slightly decreased with the cyclic
ligand L8 in comparison with the dimethyl-substituted amino
alcohol L7. From these results, it seems clear that the increase of
the steric hindrance close to the carbon of the ligand bearing the
amino group contributes more to the improvement of the ee than
the introduction of substituents at the carbinol site. We arrived at
the conclusion that the ligand of choice was L5, not only because of
the very high yield and ee obtained with it, but also due to its low
price.
With this ligand in hand, we tried to optimize the reaction con-
ditions using imine 1a as a model substrate. First, the amount of
the ligand L5 was reduced keeping the loading of the ruthenium
source [RuCl2(p-cymene)]2 and ButOK constant at values of 5 and
25 mol %, respectively. Fortunately, no detriment in either the yield
or the enantioselectivity was observed when the amount of the
ligand was gradually reduced up to 10 mol % (Table 2, entries 1–3).
The proportion Ru-dimer: L5:ButOK was then 1:2:5, which is the
same that we found as the optimum one in our first report on
the asymmetric transfer hydrogenation of sulfinylimines.5 A mini-
mum ratio L5:ButOK of 2:5 seems to be crucial for the reaction to
work, since the change of this ratio to 2:2.5 completely prevented
the reduction process, the unaltered imine being recovered (Table
2, entry 4). Maintaining the proportion Ru-dimer:L5:ButOK = 1:2:5,
we tried to reduce the catalyst loading and we were pleased to see
that 2.5 mol % of the Ru-dimer was enough to achieve the prepara-
tion of amine 2a in 90% yield without loss of enantiomeric purity
(Table 2, entry 5). In this case, the reaction time had to be extended
to 5 h in order to get full conversion of the imine. However, this
reaction time could be reduced to 2 h by performing the transfer
hydrogenation reaction at 50 °C, which afforded the expected
amine in 97% yield and 97% ee (Table 2, entry 6). Encouraged by
this result, we tried to further reduce the catalyst loading to
1 mol % of the Ru-dimer: the ee was slightly lower but the yield
drastically fell to 50% (Table 2, entry 7). With the idea of trying
to improve the yield, we repeated the reaction at 60 °C but this
led to an even more pronounced reduction of the yield with some
loss of optical purity (Table 2, entry 8). After all of these experi-
ments, we chose the conditions of entry 6 as the optimum one. It
is worth noting that these conditions represent the reduction of
the catalyst loading to half of the one used in our previous
report on the asymmetric transfer hydrogenation of N-(tert-
butylsulfinyl)imines.5
After performing the optimization of the reaction conditions,
some other imines 1b–d10 were used as substrates (Scheme 2,
Table 2, entries 9–11).11 Imine 1b, derived from propiophenone,
gave an excellent yield and enantioselectivity in a reaction time
of only 2 h (Table 2, entry 9). Our methodology is equally efficient
for the reduction of imines derived from phenones bearing either
an electron-releasing or an electron-withdrawing group on the
aromatic ring. Very high yields and ee’s were obtained irrespective
of the electronic nature of the substituent on the aromatic ring of
the imine (Table 2, entries 10 and 11). We also performed the
reduction of the (S)-N-(tert-butylsulfinyl)benzaldimine ent-1a
and the expected enantiomeric amine ent-2a was obtained, after
desulfinylation, in very good yield with the same enantiomeric
purity as for the preparation of the (R)-amine 2a (compare entries
6 and 12). This result confirms our assumption that the diastere-
oselectivity of the reduction process is mainly controlled by the
tert-butylsulfinyl group.
Table 1
Test of several achiral b-amino alcohols as ligands for the ruthenium-catalysed
transfer hydrogenation of imine 1aa
Entry
Ligand
T (°C)
tb (h)
Yieldc (%)
eed (%)
1e
2
3
4
5
6
7
8
9
L1
L1
L2
L3
L4
L5
L6
L7
L8
25
40
40
40
40
40
40
40
40
15
6
22
22
22
4
4
3
3
11f
89
84
84
70
71
44g
31g
15g
94
h
–
97
94
89
80
97
96
96
a
The solution of imine 1a [0.9 mmol in PriOH (9 mL)] and ButOK (2.25 mL of a
0.1 M solution in PriOH) were successively added to a solution of the ruthenium
complex [prepared by refluxing a mixture of [RuCl2(p-cymene)]2 (0.045 mmol), the
ligand (0.45 mmol) and 4 Å molecular sieves (0.5 g) in PriOH (2 mL)] at the tem-
perature indicated and the reaction was stirred at the same temperature for the
time indicated.
b
Time for the transfer hydrogenation reaction.
Isolated yield of amine 2a after acid–base extraction based on the starting imine
c
In conclusion, we have presented here a very efficient proce-
dure to prepare highly optically enriched primary amines through
the asymmetric transfer hydrogenation of N-(tert-butylsulfi-
nyl)imines in isopropanol catalysed by a ruthenium complex bear-
ing a commercially available and inexpensive achiral b-amino
alcohol as a ligand. Simple 2-amino-2-methyl-1-propanol as a
ligand for the ruthenium catalyst presents some advantages over
(1S,2R)-1-amino-2-indanol that we previously used:5 the first
1a. The isolated compound 2a was always P 95% pure (300 MHz 1H NMR).
d
Determined for the corresponding benzamide by HPLC using a ChiralCel OD-H
column. The (R)-enantiomer was the major one in all cases.
e
The reaction was performed with 3 equiv of the ligand L1, in the absence of 4 Å
molecular sieves and using KOH (50 mol %) as a base instead of ButOK (see Ref. 5a).
f
Acetophenone, ButSONH2 and 1-phenylethanol were also formed in the
reduction reaction.
g
Some unaltered imine was also observed in the crude reaction mixture.
Not determined.
h