Substrate dependence of nonlinear effects: Mechanistic probe and practical applications [16]

Full text of DOI:10.1021/ja0158004

5378
J. Am. Chem. Soc. 2001, 123, 5378-5379
Because nonlinear effects are so easily detected,3 many
asymmetric catalysts have now been tested. However, we are
unaware of studies of nonlinear effects that focus on the substrate
dependency of this behavior. The substrate dependency of
nonlinear effects has important implications for two primary
reasons: (1) in the optimization of asymmetric processes it is
beneficial to determine the ee of the ligand necessary to obtain a
product of the desired ee and (2) substrate dependency of
nonlinear effects can be used to probe the mechanism of
asymmetric reactions. In this Communication we present a study
of the substrate dependency of nonlinear effects using the MIB
ligand of Nugent (eq 1).14 This ligand is closely related to the
DAIB ligand 1 that has been extensively studied by Noyori and
co-workers.9,15-18 We find that simply modifying the electronic
properties of benzaldehyde derivatives results in a change in the
product ee (eep) of over 30% in the asymmetric addition (eq 1)
with 10% ee of MIB. This effect is even more pronounced with
aliphatic aldehydes. Equally important, the current model for the
mechanism of the asymmetric addition reaction (eq 1) with DAIB
is not consistent with the observed substrate dependency of the
nonlinear effect with MIB.
We chose to employ the MIB ligand because of its ease of
synthesis and its stability on long-term storage.14 Asymmetric
addition reactions (eq 1) were conducted using 4 mol % MIB of
10, 20, and 100% ee of the ligand at 0 °C (Table 1). These
additions were performed by combining the ligand and aldehyde
followed by addition of the diethylzinc over 1 min (Method A).
Under these conditions, no precipitate formed over the course of
the reaction (Table 1). Reactions were also performed at room
temperature by mixing the ligand and diethylzinc (Method B,
Supporting Information). After 1 h the aldehyde was added and
the reactions were sampled between 10 and 20% conversion
except for m-trifluoromethyl benzaldehyde which was 96%
complete after 2 min (Table S2, Supporting Information). No
precipitate was present at these conversions; however, solid did
form in some reactions at later times. Importantly, both methods
exhibited the same trend in nonlinear behavior. A plot of the ee
of the MIB ligand versus the eep for the reaction conducted at 0
°C clearly shows that nonlinear effects are substrate dependent
in this system (Figure 1). The data in Figure 1 demonstrate that
benzaldehyde derivatives with electron-donating substituents
exhibit greater nonlinear behaVior than analogues with electron-
withdrawing substituents. Changing substrates form p-methyl-
benzaldehyde to o-methylbenzaldehyde resulted in no change in
the nonlinear behavior (compare entries 2 and 7, Table 1).
However, 2,4,6-trimethylbenzaldehyde exhibited a markedly
smaller nonlinear effect (entry 8). Aliphatic aldehydes were found
to show greater variation in nonlinear behavior than aromatic
aldehydes (Table 1). The linear chain nonal gave a stronger
nonlinear effect than cyclohexanecarboxaldehyde. Increasing the
steric hindrance around the carbonyl carbon leads to a decrease
in the nonlinear effect (entries 8 and 10).
Substrate Dependence of Nonlinear Effects:
Mechanistic Probe and Practical Applications
Young K. Chen, Anna M. Costa, and Patrick J. Walsh*
P. Roy and Diane T. Vagelos Laboratories
Department of Chemistry, UniVersity of PennsylVania
231 South 34th Street, Philadelphia, PennsylVania 19104-6323
ReceiVed March 12, 2001
ReVised Manuscript ReceiVed April 24, 2001
Enantioselective catalysis has witnessed explosive growth in
the last two decades as it has become the most versatile and
efficient method for the preparation of molecules of high
enantiomeric excess.1 Of the numerous contributions that have
shaped our understanding of catalytic asymmetric reactions, few
have had such a profound impact as the experimental and
theoretical description of nonlinear effects by Kagan and co-
workers.2-5 Before this seminal work,2 it was generally believed
that a strictly linear correlation existed between the ee of the
catalyst and the ee of the product. However, Kagan demonstrated
that this assumption was not valid2 and many systems have since
been shown to exhibit nonlinear behavior.3,4
The consequences of strong positive nonlinear effects are
remarkable.3,4,6-8 For example, Noyori used the DAIB ligand (eq
1) of only 15% ee in the asymmetric addition of alkyl groups to
aldehydes from which a product of 95% ee was generated. When
enantiopure DAIB was used in this reaction, the product was
generated in 98% ee.9 Soai has demonstrated that ligands with
very low ee, or even traces of chiral material, can be used in an
autocatalytic asymmetric process to generate product in high
ee.10-13 The only drawback to using partially resolved catalysts
exhibiting strong positive nonlinear behavior such as DAIB in
production of enantioenriched material is that they display a lower
overall rate than when they are enantiomerically pure.7
(1) Jacobsen, E. N.; Pfaltz, A.; Yamamoto, H. ComprehensiVe Asymmetric
Catalysis; Springer: Berlin, 1999; Vols. 1-3.
(2) Puchot, C.; Samuel, O.; Dunach, E.; Zhao, S. H.; Agami, C.; Kagan,
H. B. J. Am. Chem. Soc. 1986, 108, 2353-2357.
(3) Girard, C.; Kagan, H. B. Angew. Chem., Int. Ed. Engl. 1998, 37, 2922-
Variation of the substituent on benzaldehyde also affects the
overall rate using nonenantiopure MIB in eq 1. With 10% ee MIB,
the reaction of the m-(trifluoromethyl)benzaldehyde was complete
in under 30 min while the p-methoxybenzaldehyde was only 23%
complete after 2 h. The higher overall rate with electron-
withdrawing aldehydes is consistent with the proposed rate
determining addition of the alkyl group to the carbonyl.16 Control
experiments were conducted to estimate the rate of the background
2959.
(4) Avalos, M.; Babiano, R.; Cintas, P.; Jime´nez, J. L.; Palacios, J. C.
Tetrahedron: Asymmetry 1997, 8, 2997-3017.
(5) Mikami, K.; Terada, M.; Korenaga, T.; Matsumoto, Y.; Ueki, M.;
Angelaud, R. Angew. Chem., Int. Ed. 2000, 39, 3532-3556.
(6) Blackmond, D. G. Acc. Chem. Res. 2000, 33, 402-411.
(7) Blackmond, D. G. J. Am. Chem. Soc. 1997, 119, 12934-12939.
(8) Blackmond, D. G. J. Am. Chem. Soc. 1998, 120, 13349-13353.
(9) Kitamura, M.; Okada, S.; Suga, S.; Noyori, R. J. Am. Chem. Soc. 1989,
111, 4028-4036.
(10) Soai, K.; Shibata, T.; Morioka, H.; Choji, K. Nature 1995, 378, 767-
768.
(14) Nugent, W. A. J. Chem. Soc., Chem. Commun. 1999, 1369-1370.
(15) Noyori, R.; Suga, S.; Kawai, K.; Okada, S.; Kitamura, M. Pure Appl.
Chem. 1988, 60, 1597-1606.
(11) Shibata, T.; Morioka, H.; Hayase, T.; Choji, K.; Soai, K. J. Am. Chem.
Soc. 1996, 118, 471-472.
(12) Shibata, T.; Yamamoto, J.; Matsumoto, N.; Yonekubo, S.; Osanai,
S.; Soai, K. J. Am. Chem. Soc. 1998, 120, 12157-12158.
(13) Soai, K.; Osanai, S.; Kadowaki, K.; Yonekubo, S.; Shibata, T.; Sato,
I. J. Am. Chem. Soc. 1999, 121, 11235-11236.
(16) Kitamura, M.; Suga, S.; Oka, H.; Noyori, R. J. Am. Chem. Soc. 1998,
120, 9800-9809.
(17) Kitamura, M.; Oka, H.; Noyori, R. Tetrahedron 1999, 55, 3605-3614.
(18) Yamakawa, M.; Noyori, R. Organometallics 1999, 18, 128-133.
10.1021/ja0158004 CCC: $20.00 © 2001 American Chemical Society
Published on Web 05/10/2001

Communications to the Editor
J. Am. Chem. Soc., Vol. 123, No. 22, 2001 5379
Scheme 1. The Noyori Mechanisma
Table 1. Product ee’s with 10, 20, and 100% ee MIB Determined
by Using Method Ah
a See Supporting Information for structures.
reaction using DAIB is shown in Scheme 1. Using Noyori’s
abbreviations, S and R are ligand adducts (LS)Zn(Et) and (LR)-
Zn(Et), Sub is the aldehyde substrate, Rea is the reagent ZnEt2,
and P is the product. Noyori and co-workers have shown that the
nonlinear effects in the DAIB system arise from a reservoir effect.
Convincing evidence has been presented that the active catalysts
are derived from the monomeric species (LS)Zn(Et) and (LR)Zn-
(Et) and that the resting states for the catalyst are the heterochiral
and homochiral dimers (S-R, S-S, and R-R in Scheme 1). The
heterochiral dimer has been shown to be significantly more stable
than the homochiral dimers.16 It is proposed that the monomers
(L)Zn(Et) bind aldehyde and ZnEt2 to give ZnEt2‚(L)(Et)Zn-Sub.
This complex reacts irreversibly to give the alkoxide product.
When nonenantiopure ligand is employed, the eep is dependent
on the concentrations of the aldehyde and diethylzinc and the
equilibrium constants Khomo, Khetero, and Kassoc (Scheme 1). Of these
equilibrium constants, Khomo and Khetero are independent of the
substrate and only Kassoc will be affected by the binding strength
of the aldehyde. Aldehydes that bind more strongly will increase
the concentration of the ZnEt2‚(L)(Et)Zn-Sub at the expense of
the dimers. This model predicts that aldehydes that bind more
tightly will result in a decrease in eep. In the extreme case of
very high Kassoc, where only monomers are formed, there will be
no nonlinear effects. This model of the mechanism is inconsistent
with the observed substrate dependency of the nonlinear effects
measured here in which more tightly binding substrates exhibit
greater nonlinear behavior. We are currently investigating whether
the data can be explained by reaction of aldehydes with the
diastereomeric homo- and heterochiral dimers through different
pathways.
a ee measured after completion of the reaction, except where
otherwise stated. b 92% conversion after 23 h. c 14% conversion after
96 h. d 33% conversion after 96 h. e 40% conversion after 24 h. f 94%
conversion after 24 h. g 94% conversion after 16 h. h Data for Method
B are given in the Supporting Information.
In conclusion, we have shown that nonlinear effects are
substrate dependent. This observation has important ramifications
in the practical application of nonenantiopure ligands to synthesis
of products with high enantiomeric excess. We have also
demonstrated that investigation of substrate-dependent nonlinear
effects has the potential to provide important information about
equilibria that control eep. We are currently attempting to modify
the Noyori model to fit the previously reported mechanistic data
as well as the data reported here. These investigations will be
reported in due course.
Figure 1. Ee of product vs ee of MIB for substituted benzaldehyde
derivatives.
reaction, that is the reaction that takes place without the
participation of the chiral ligand. Reaction of benzaldehyde in
eq 1 without the amino alcohol ligand exhibited 3% conversion
after 24 h. The more reactive p-trifluoromethyl derivative was
3% complete in 2 h and 22% complete after 24 h. These results
indicate that the background reaction is significantly slower than
the ligand accelerated pathway.19
Because of the similarity of the MIB and the DAIB ligands, it
is reasonable to assume that the reaction mechanisms in eq 1 are
the same.20 The proposed mechanism for the asymmetric addition
Acknowledgment. Dedicated to Professor H. B. Kagan, pioneer of
nonlinear effects. This work was supported by the National Science
Foundation (CHE-9733274) and the National Institute of Health
(GM58101). We thank Prof. Donna Blackmond for helpful discussions.
Supporting Information Available: Experimental conditions and
results of addition reactions at 0 (Table S1) and 23 °C (Table S2) (PDF).
This material is available free of charge via the Internet at http://pubs.acs.org.
JA0158004
(19) Berrisford, D. J.; Bolm, C.; Sharpless, K. B. Angew. Chem., Int. Ed.
Engl. 1995, 34, 1059-1070.
(20) Rosner, T.; Sears, P. J.; Nugent, W. A.; Blackmond, D. G. Org. Lett.
2000, 2, 2511-2513.