diethylzinc to aromatic aldehydes have been reported.7,13 This
positive nonlinear effect (NLE) has been explained by
predominant formation of thermodynamically favored, het-
erochiral zinc complexes in solution. The organozinc adduct
of the minor enantiomer is thus trapped in a catalytically
inactive form while the surplus of the organozinc adduct of
the major enantiomer is available to catalyze the reaction.
Because heterochiral recognition and association generates
an enantiomerically enriched or enantiopure monomeric
catalytic species, the enantiomeric excess of the reaction
product can exceed the ee of the precatalyst employed.14
Positive and negative nonlinear effects have been observed
in a variety of asymmetric reactions and are generally
explained with Kagan’s MLn model, which is based on the
assumption that a scalemic catalyst is in equilibrium with
homochiral and heterochiral adducts exhibiting different
catalytic activity.15
a mixture containing (+)-1 in 50% ee in toluene and hexanes
(1:4 v/v) for 12 h. The precipitate was isolated, and the
enantiomeric composition of 1 in the solid state and in
solution was determined.
Unfortunately, screening of several HPLC columns did
not provide a chromatographic method for direct enantio-
separation of 1. We therefore decided to subject the bis-
oxazolidine mixtures to hydrolysis and derivatization to
afford N-t-Boc-aminoindanol 7, which can be resolved by
chiral HPLC on Chiralcel OD (Scheme 2). Gravimetric and
Scheme 2. Hydrolysis and Derivatization of Ligand 1 for
Chiral HPLC Analysis
Analysis of the enantioselective alkylation of benzaldehyde
with diethylzinc in the presence of scalemic mixtures of 1
revealed a positive nonlinear effect (Figure 2). For example,
chromatographic analysis showed that the crystalline material
was almost racemic and corresponded to 47% of the
originally used bisoxazolidine. By contrast, the supernatant
contained the remaining 53% of 1 in 89% ee.16 The same
results were obtained when the experiment was conducted
in the presence of 150 equiv of diethylzinc. Upon mixing of
solutions containing a large excess of Et2Zn and either
enantiopure (+)-1 or (-)-1 to afford 50% ee, a precipitate
formed which contained nearly racemic 1 (4.5% ee) but no
trace of Et2Zn. NMR analysis of mixtures of 1 and Me2Zn
revealed that complexation occurs without concomitant
deprotonation. This shows that 1 is not a precatalyst and
unlike amino alcohols does not require 1 equiv of R2Zn to
form a catalytically active species.17 Accordingly, alkylation
of 2a proceeds with high yield and enantioselectivity when
equimolar amounts of 1 and Et2Zn are used; see the
Supporting Information.
Figure 2. Nonlinear effects in the bisoxazolidine-catalyzed enan-
tioselective ethylation, methylation, and alkynylation of benzalde-
hyde with organozinc reagents (from left to right). See the
Supporting Information for reaction conditions.
Apparently, the positive nonlinear effect observed in the
bisoxazolidine-catalyzed reaction between benzaldehyde and
diethylzinc is a result of the inherently low solubility of
racemic 1 in apolar organic solvents. This unexpected solid-
liquid phase behavior of scalemic 1 selectively increases the
purity of the major enantiomer in solution and gives rise to
a positive NLE that cannot be described by Kagan’s MLn
model.18 In our case, chiral amplification originates from
thermodynamically controlled crystallization of racemic 1,
whereas the enantiomeric surplus remains in solution and is
available for asymmetric catalysis (Scheme 3). As expected,
no NLE was observed when the reaction was carried out in
toluene and dichloromethane (1:1 v/v) because racemic 1 is
soluble in this solvent mixture (Figure 2).
1-phenylpropanol is obtained in 94% yield and 96% ee in
the presence of 2 mol % of 1 having 50% ee. During our
studies, we noticed that the solubility of the ligand signifi-
cantly decreases as the enantiopurity of the bisoxazolidine
is reduced. When scalemic 1 was used, the catalyst could
not be fully dissolved, and we suspected that this might
account for the NLE observed. We found that the solubility
of enantiopure 1 in toluene and hexanes (1:4 v/v) is 10.5
mg/mL, whereas only 1.4 mg/mL of racemic 1 dissolves in
the same solvent mixture. To prove our hypothesis, we stirred
(13) (a) Oguni, N.; Matsuda, Y.; Kaneko, T. J. Am. Chem. Soc. 1988,
110, 7877-7878. (b) Noyori, R.; Kitamura, M. Angew. Chem., Int. Ed.
Engl. 1991, 30, 49-69.
(14) Kagan, H. B. AdV. Synth. Catal. 2001, 343, 227-233.
(15) (a) Puchot, C.; Samuel, O.; Dunach, E.; Zhao, S.; Agami, C.; Kagan,
H. B. J. Am. Chem. Soc. 1986, 108, 2353-2357. (b) Mikami, K.; Motoyama,
Y.; Terada, M. J. Am. Chem. Soc. 1994, 116, 2812-2820. (c) Evans, D.
A.; Kozlowski, M. C.; Murry, J. A.; Burgey, C. S.; Campos, K. R.; Connell,
B. T.; Staples, R. J. J. Am. Chem. Soc. 1999, 121, 669-685. (d) Chen, Y.
K.; Costa, A. M.; Walsh, P. J. J. Am. Chem. Soc. 2001, 123, 5378-5379.
(e) Yuan, Y.; Li, X.; Sun, J.; Ding, K. J. Am. Chem. Soc. 2002, 124, 14866-
14867. (f) Qin, Y.-C.; Liu, L.; Pu, L. Org. Lett. 2005, 7, 2381-2383. (g)
Reetz, M. T.; Meiswinkel, A.; Mehler, G.; Angermund, K.; Graf, M.; Thiel,
W.; Mynott, R.; Blackmond, D. G. J. Am. Chem. Soc. 2005, 127, 10305-
10313.
To date, only a few examples of enantioenrichment of
chiral catalysts and chiral amplification based on solid-
(16) The ee of the precipitate was determined as 6%. We found that the
ee of the solid formed decreases further when the precipitate is washed
more thoroughly, but less material is recovered in these cases.
(17) Organozinc reagents have been reported to not deprotonate secondary
amino groups. See ref 4c.
(18) (a) Puchot, C.; Samuel, O.; Dunach, E.; Zhao, S.; Agami, C.; Kagan,
H. B. J. Am. Chem. Soc. 1986, 108, 2353-2357. (b) Girard, C.; Kagan, H.
B. Angew. Chem. Int. Ed. 1998, 37, 2922-2959.
Org. Lett., Vol. 9, No. 16, 2007
2967