on development of a novel polymeric platform for catalyst
immobilization, featuring inverted solubility pattern: the
catalyst soluble in a non-polar and insoluble in a polar
medium.
In our previous work, the solid-supported catalysts 8 were
prepared by constructing an ether link between the phenolic
group in 6 and a suitable group on the polymer, namely P-
C6H4-CH2Cl or P-C6H4-CH2OH, using the Williamson and
Mitsunobu conditions, respectively.[6] In the present study,
we have adopted a different strategy, which aimed at the
preparation of the anchored catalyst by copolymerization of
methacrylates 10 and 11 (Scheme 2).
copolymers 9a and 9b was ~7800 and 6000 gmolÀ1, respec-
tively, as revealed by gel permeation chromatography
(GPC). Their poly-dispersity index (PDI) varied from 1.35
(9a) to 1.18 (9b), indicating a fairly narrow distribution of
molecular masses for the polymers.
The reactivity of our immobilized catalysts in the reduc-
tion of imines was investigated by using the same reaction
conditions as those employed for 3–7 (1 equiv of imine,
2 equiv of Cl3SiH, and catalyst 9 at room temperature). Re-
duction of imine 1a in toluene (an optimized solvent for ho-
mogeneous conditions[4,5]), catalyzed by 9a, afforded amine
(S)-2a in 85% ee (Table 1, entry 3). Catalyst 9b, with the
higher content of the active component, turned out to
be an optimum in terms of its preparation and perfor-
mance, since high conversions and enantioselectivities
were attained (86–88% ee; entries 4–6) even with grad-
ually decreasing catalyst loading (from 7 to 1 mol%).
The practicality of the protocol was demonstrated by
more than 25-fold scale-up experiment (entry 7). Signif-
icantly, enantioselectivity of the supported catalyst 9b
remained at the same level as that observed for its
monomeric congeners 4 and 6 (entries 1 and 2). When
the reaction was complete, as indicated by TLC, the re-
action mixture (in toluene) was poured into an excess
of a vigorously stirred MeOH, resulting in the precipi-
tation of ꢁ95% of the catalyst, which was then filtered
off.[13]
The scope of the homogeneous catalysts 3–5 has been
shown by us to be quite broad, spanning from a range of ar-
omatic to heteroaromatic and to some aliphatic substrates,
while tolerating various functionalities.[4,5] Therefore, only a
small set of imines, namely 1b–j, was explored in the pres-
ent study (Table 1). In agreement with the previous observa-
tions, high enantioselectivities were attained for imines de-
rived from aromatic ketones 1b–d with both electron-with-
drawing and electron-donating groups (entries 8–10). Imine
1e with the b-thiophene nucleus (entry 11) exhibited lower
enantioselectivity, consistent with the general behavior of ar-
omatic heterocycles.[4,5] By contrast, the presence of heteroa-
toms in the alkyl part (1 f,g) did not have any adverse effect
(entries 12 and 13), reaching the maximum of 91% ee. The
cyclopropyl and cyclobutyl derivatives 1h,i still exhibited
acceptable enantioselectivities (entries 14 and 15), whereas a
larger decrease was observed for the bulkier cyclohexyl de-
rivative 1j (entry 16).
Scheme 2. Synthesis of the supported catalyst.
Treatment of the phenolic derivative 6[6] with methacrylo-
yl chloride afforded the monomer 10 (83%), which was co-
polymerized with benzyl methacrylate 11 using atom trans-
fer radical polymerization (ATRP) methodology,[12] namely
by heating a mixture of 10 and 11 (1:99 to 10:90) in the pres-
ence of CuCl (1 mol% with respect to 11), 2,2’-bipyridine
(3 mol%), and ethyl 2-bromo-isobutyrate (1 mol%) in
DMSO at 908C for 4 h. The resulting copolymer 9 was pre-
cipitated by pouring the cooled mixture into an excess of
MeOH.
In the initial polymerizations, the amount of 10 in the
mixture was varied in the range of 1, 5, and 10 mol% with
respect to 11, whereas the amount of CuCl was kept con-
stant (1 mol%). Co-polymerization of 10 (1 mol%) with 11
produced 9a in 46% yield, which contained 0.13 mmolgÀ1
of the active moiety, as revealed by elemental analysis.
However, when the amount of 10 in the initial mixture was
increased to 5 mol%, the polymerization afforded the copo-
lymer in only 9% yield and a further decrease (to 4%) was
observed with 10 mol% of 10. Clearly, 10 had a negative
effect on the activity of the catalyst. Therefore, in the subse-
quent experiments, the amount of CuCl was increased to
match that of 10, which proved to have a beneficial effect.
In a scaled-up, optimized experiment, carried out with a
5:95 mixture of 10 and 11 in the presence of CuCl
(5 mol%), 2,2’-bipyridine (15 mol%), and ethyl 2-bromo-
isobutyrate (1 mol%), copolymer 9b (0.19 mmolgÀ1) was
obtained in 48% yield. Further increase in the content of 10
in the mixture did not lead to any significant increase in its
incorporation, showing that the copolymerization reached
its saturation point. The average molecular weight (Mn) of
The solid-supported, insoluble catalysts 8 retained their
activity when re-used;[6] in fact, we have observed an in-
crease in selectivity in the second run by ~5% ee, which was
maintained in the subsequent runs; this behavior was attrib-
uted to a “conditioning” effect.[6] By contrast, the soluble
catalyst 9b (Table 2) exhibited the same activity in runs 1–5;
hence, no “conditioning” was taking place here, which is ap-
parently confined to the heterogeneous protocol.
In conclusion, a new soluble polymeric platform for im-
mobilization of organocatalysts has been developed, which
may, a priori, be applied to a wide variety of other catalytic
systems. The asymmetric reduction of imines 1a–j with
9652
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 9651 – 9654