COMMUNICATION
with diethylzinc.[14,10b] The choice of bis(trifluorometh-
ylsulfonyl)imide (Tf2N) as counter ion was dictated by solu-
bility (very low in water) and stability reasons.
With ligands 1c–e in our hands, we started testing the
simple addition of diethylzinc to benzaldehyde as the bench-
mark reaction in different RTILs. The results are reported
in Table 1.
The efficiency of ion-tagged ligands was then compared to
that of simple diphenylprolinol (1a) and N-methyl-diphenyl-
prolinol (1b).[15] The former required longer reaction times
to ensure a good conversion but with only a 38% ee
(Table 1, entry 11). The latter was less efficient both in
terms of yield and selectivity (Table 1, entry 12) with respect
to ligand 1d.
Using the protocol of entry 4, we attempted the reaction
adding only
a slightly excess amount of diethylzinc
Table 1. Addition of diethylzinc to benzaldehyde in different RTILs cata-
(1.1 equiv), and we were delighted to obtain essentially the
same results, both in terms of yield and selectivity (87%,
90% ee), with respect to the use of a larger excess of the or-
ganometallic reagent.
Having defined entry 4 as an optimized protocol, process
refinement required to focus on catalyst recycling. This task
was efficiently addressed when a very simple workup was
developed to easily remove the inorganic zinc salts formed
during the aqueous workup and to quantitatively recover
the alkylation product.
In detail, the reaction mixture was quenched with a basic
aqueous solution of sodium salt of ethylenediamine tetra-
acetic acid (EDTA), washed with water and extracted with
Et2O. The ligand-containing IL phase was dried by heating
for 2–3 h at 708C under reduced pressure (~0.1 mmHg), and
eventually reloaded with diethylzinc and the appropriate al-
dehyde. Results obtained in ten consecutive cycles are re-
ported in Table 2.
lyzed by ligands 1.
Entry RTIL
1
T [8C] t [h] Yield [%][a] ee [%][b]
1
2
–
A
1d
1d
1d
0
0
0
0
0
3
3
3
3
3
90
92
88
92
90
23
30
69
89
87
89
75
95
72
80
38
78
ACHTUNGTRENNUNG
3
4
U
U
R
5
6
N
G
1d
1d
[NTf2][e] 1d
0
20
4.5 80
2
2
3
3
7
N
U
85
91
92
90
91
75
8
N
U
1d[f] 20
9
N
U
1c
1e
1a
1b
0
0
0
0
10
11
12
N
G
24
3
[a] Yield of isolated product. [b] The absolute configuration was assigned
by comparison of the optical rotation with that of the known product.
The ee was determined by HPLC analysis (chiralcel OD column).
[c] bpy=1-butylpyridinium.
[d] bmpy=1-butyl-1-methylpyrrolidinium.
[e] 5 mol% of ligand was used. [f] 20 mol% of ligand was used.
Table 2. Recycling experiments.
We first compared the effect of the RTIL structure on re-
activity and selectivity using 10 mol% of ligand 1d at 08C
(Table 1, entries 1–5). Yields were independent of the RTIL
used (88–92%), while the enantiomeric excesses (eeꢁs) were
strongly affected by its structure. As a matter of fact, the
electrostatic environment associated to the RTIL seems es-
sential for selectivity, since the reaction in n-hexane
(Table 1, entry 1) delivered a 23% ee. Using 1-butylpyridini-
um RTILs, the effect of the counter anion on selectivity was
apparent, since tetrafluoroborate gave a 30% ee (entry 2),
Cycle
Yield [%][a]
ee [%][b]
Cycle
Yield [%][a]
ee [%][b]
1
2
3
4
5
96
94
95
98
96
89
89
90
89
89
6
7
8
9
10
96
98
95
98
96
89
88
90
88
90
[a] Yield of isolated product. [b] Determined by HPLC analysis (chiralcel
OD column).
À
while NTf2 ensured a 40% gain (entry 3). The best enantio-
selectivity was achieved by combining the pyrrolidinium
À
cation and NTf2 anion (entry 4) or triflate (entry 5).
An outstanding reproducibility was displayed and no loss
of activity and stereoselectivity was apparent up to the tenth
cycle.
Finally, we performed an aldehyde screen to explore the
scope of this catalytic system, by applying the same opti-
mized protocol reported for the recycling experiments. Re-
sults are summarized in Table 3.
Very good yields and enantioselectivities were obtained
with aromatic aldehydes substituted in the para position
with electron-withdrawing or donating groups (Table 3, en-
tries 1–7), while 2-substituted aromatic aldehydes (Table 3,
entries 8, 9), a,b-unsaturated aldehydes (Table 3, entries 10,
11) and aliphatic aldehydes (Table 3, entry 12) afforded
lower eeꢁs.
Since [bmpy]ACHTUNGTRENNUNG[NTf2] is much more easily synthesized and
handled with respect to the corresponding triflate, it was
elected as the solvent of choice in the next study.
When the catalyst loading was decreased to 5 mol%,
80% isolated yield after 4.5 h at 08C was obtained, with
identical ee with respect to the use of 10 mol% ligand
(Table 1, entry 6). When the reaction was catalyzed by 1d at
208C, we recorded a 10% loss in ee (Table 1, entry 7), but
using the ligand in 20 mol%, a 95% ee and 91% isolated
yield was obtained within 2 h (Table 1, entry 8).
The different length of the alkyl spacer (1c,e) did not
affect the high yields, but reflected in lower enantioselectivi-
ties with respect to the use of 1d (Table 1, entries 9, 10).
Chem. Eur. J. 2008, 14, 11288 – 11291
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
11289