.
Angewandte
Communications
molecular sieves (MS; see examples in Table 1, entries 1 and
2).
acid additives, even though all known amination reactions of
alcohols on the basis of borrowing hydrogen methodology
were carried out under basic or neutral conditions.[6–9] As
shown by the representative results listed in entries 3–5 of
Table 1, whereas acid additives did not improve the reaction
catalyzed by Ru complex 4 (entry 3), the combination of an
acid (such as p-toluenesulfonic acid or phosphoric acid) and
the Ir complex (R,R)-5 successfully led to the formation of 3a,
although the level of efficiency and selectivity remained
disappointingly low (entries 4,5). Chiral phosphoric acids
were then tested. The identity of the 3,3’-substituents in the
chiral backbone had a dramatic effect on the reaction
outcome (Table 1, entries 6–9), whereby 10e provided the
highest level of enantioselectivity and promising efficiency.
As two chiral sources were used, the combination of 10e with
the enantiomeric Ir complex (S,S)-5 was tested. This pair
provided significantly higher selectivity (87% ee), but with
diminished 32% conversion (Table 1, entry 10). The matched/
mismatched relationship of the two chiral sources is signifi-
cant, whereby the chirality of the Ir complex is the determin-
ing factor.
We then switched our focus to the modification of the Ir
complex. Variation of the substitution on the sulfonamide
moiety as well as the chiral-diamine backbone led to the
identification of complex (S,S)-6 as the catalyst that provided
the optimal combination of reactivity and selectivity (Table 1,
entries 10–14). Mesylate (S,S)-8 and cyclohexanediamine-
derived (S,S)-9 were surprisingly not effective at all. Further
fine-tuning of various reaction parameters showed that
a higher loading of the alcohol substrate (3 equiv or 5 equiv
rather than 1.5 equiv; Table 1, entries 15 and 16 versus
entry 11) led to an increase in enantioselectivity to
> 90% ee. In an effort to identify an alternative to this
higher alcohol loading, which is detrimental to the overall
efficiency of the process, we wondered whether the beneficial
effect comes from a higher concentration of the hydrogen
donor or simply an increase in the polarity of the reaction
medium. tert-Amyl alcohol, which is not a hydrogen donor but
a much more polar solvent, was then tested as the solvent.
Gratifyingly, this change of solvent led to an increase in the
ee value of the product to 89% (Table 1, entry 17 versus
entry 11). Finally, a higher loading of 10e proved beneficial.
Under the optimal conditions, under which the alcohol was
used in slight excess (1.5 equiv), 3a was produced in 90%
yield with 93% ee (Table 2, entry 1).
Careful examination of the crude reaction mixture
indicated the formation of a small amount of the correspond-
ing ketone, which led us to speculate that the imine
condensation was the problematic step. Inspired by previous
studies on reductive amination by the research groups of
MacMillan[4g] and Xiao,[5e] in which chiral phosphoric acids
proved effective in promoting the condensation of ketones
and anilines even at 35–508C, we decided to test the effect of
Table 1: Optimization of the enantioselective amination of 1a.
Entry [M]
Additive
Solvent
Conv. ee
[%][a]
[%][b]
1
2
3
(R,R)-4
(R,R)-5
–
–
toluene
toluene
<5
<5
<5
32
N.D.
N.D.
N.D.
7
(R,R)-4 TfOH (5 mol%) toluene
(R,R)-5 TfOH (5 mol%) toluene
4
5
6
7
8
(R,R)-5 10a (5 mol%)
(R,R)-5 10b (5 mol%)
(R,R)-5 10c (5 mol%)
(R,R)-5 10d (5 mol%)
(R,R)-5 10e (5 mol%)
(S,S)-5 10e (5 mol%)
(S,S)-6 10e (5 mol%)
(S,S)-7 10e (5 mol%)
(S,S)-8 10e (5 mol%)
(S,S)-9 10e (5 mol%)
(S,S)-6 10e (5 mol%)
(S,S)-6 10e (5 mol%)
(S,S)-6 10e (5 mol%)
(S,S)-6 10e (10 mol%)
toluene
toluene
toluene
toluene
toluene
toluene
toluene
toluene
toluene
toluene
toluene
toluene
tert-amyl alcohol
tert-amyl alcohol
33
60
<5
83
50
32
>99
96
<5
<5
>99
>99
95
12
9
N.D.
À17
À46
87
81
77
N.D.
N.D.
90
9
10
11
12
13
14
15[c]
16[d]
17
18
92
89
93
>99
[a] Conversion was determined by NMR spectroscopy with 4,4’-di-tert-
butylbiphenyl as an internal standard. [b] The ee value was determined by
HPLC on a chiral stationary phase. [c] The reaction was carried out with
3 equivalents of the alcohol. [d] The reaction was carried out with
5 equivalents of the alcohol. N.D.: not determined; Tf=trifluorometh-
anesulfonyl. See the Supporting Information for details.
The same set of conditions can be applied to a wide range
of alcohol substrates (Table 2). Alcohols bearing linear alkyl
substituents (substrates 1a–c) were transformed into 3 with
high enantioselectivity (Table 2, entries 1–3). The reactions of
substrates bearing a-branched substituents were particularly
successful (up to 97% ee; Table 2, entries 4–6). Functional-
ities such as benzyl ether and silyl ether groups were well-
tolerated (products 3i and 3j), even though the reaction
involved reductive and acidic conditions (Table 2, entries 9
and 10). Aryl–alkyl-substituted products were also generally
obtained with good to excellent enantioselectivity (Table 2,
entries 11–17). The electronic properties of the substrates
seemed to play a role, whereby products bearing electron-
donating substituents were obtained with higher ee values
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1399 –1403