Catalyst versus Substrate Induced Selectivity
adopt an equatorial position in the cyclic intermediate.
Therefore, the reaction rate of the (R)-enantiomer can com-
pete in that case with the (S)-enantiomer.
configurational outcome confirms recent DFT studies that
planar chiral palladacycles operate similar to achiral Pd-
salts[9] via (half)chair-like intermediates (and transition
states). Our results also suggest that additional less favora-
ble reaction pathways via boat or twist transition states ar-
guably contribute to product formation.
Larger substituents R3 increase the selectivity (s up to 278
in entry 10) and accelerate the rearrangement thus permit-
ting the use of lower catalyst loadings (compare entries 3
and 9 or 1, 10, and 11). Substrates with unbranched N-alkyl
substituents R1 need lower catalyst loadings as compared to
substrates with branched moieties (compare entries 1 and 3
or 9, 10, 12, and 13).
Experimental Section
See the Supporting Information.
Yields determined by HPLC with addition of internal
standard were found to be practically identical to the yields
of isolated products (entries 5,10–11). The products can sub-
sequently be transformed into the corresponding secondary
amines by reduction with NaBH4 as shown for 3e and 3k
(see graphic in Table 4). On the other hand, imidates re-
maining from the resolution could be transformed to the
(S)-configured products 3 by PdCl2 catalyzed rearrangement
without loss of enantiopurity (see graphic in Table 4).
The selectivity factors s given in Table 4 were determined
based on the product data (yield and ee). It has to be men-
tioned though that they are higher than the s-factors calcu-
lated from the data of the starting material. Usually these
data should be identical in kinetic resolutions.[6] The differ-
ences in the present case might be partly attributed to the
fact that concomitant decomposition is observed. Moreover,
a new stereocenter is formed while another one is sacrificed
whereas in a classical kinetic resolution—for example, the
hydrolysis of a racemic ester—all bonds to the stereocenter
remain intact and a new center is not formed. Because chir-
ality transfer using a chiral catalyst is not complete (see
Table 1), competitive mechanistic pathways arguably exist,
for example, via a boat-like (or twist-like) transition state
leading to intermediate 9 (Scheme 5). By that, the (R)-con-
figured starting material, that is, the mismatched enantio-
mer, can for example, also lead to (R,E)-configured product.
This might explain the poor chirality transfer for entry 2 in
Table 1.
Acknowledgements
This work was financially supported by F. Hoffmann-La Roche. We
thank Corinna Schelzel and Marcel Weiss for skilful experimental contri-
butions during research internships in our laboratory and Priv.-Doz. Dr.
Martin Karpf (F. Hoffmann-La Roche) for critically reading this manu-
script.
Keywords: allylic amines · aza-Claisen rearrangement ·
chirality transfer · kinetic resolution · palladium
576, 290; b) L. E. Overman, N. E. Carpenter, Org. React. 2005, 66, 1.
[2] Trichloroacetimidates: a) C. E. Anderson, L. E. Overman, J. Am.
[3] Trifluoroacetimidates: see Ref. [2e]) and: a) L. E. Overman, C. E.
Prasad, C. E. Anderson, C. J. Richards, L. E. Overman, Organome-
Weiss, D. F. Fischer, Z.-q. Xin, S. Jautze, W. B. Schweizer, R. Peters,
1430; i) Z.-q. Xin, D. F. Fischer, R. Peters, Synlett 2008, 1495.
In conclusion, we have described the first kinetic resolu-
tion of allylic trifluoroacetimidates by planar chiral pallada-
cycles which is the result of a matched/mismatched combi-
nation of catalyst face selectivity for olefin coordination and
the inherent substrate selectivity. Other examples for kinetic
resolutions by palladacycles[14] are exceptionally rare.[15] The
[7] a) A. H. Hoveyda, M. T. Didiuk, Curr. Org. Chem. 1998, 2, 489;
343, 5; c) for kinetic resolution by amination of p-allyl complexes,
see: D. C. Vrieze, G. S. Hoge, P. Z. Hoerter, J. T. Van Haitsma, B. M.
[8] The (E)-configured counterpart of 2a also bearing an N-PMP group
is thermally less stable than the (Z)-isomer and thus not suitable for
the investigation of a catalytic process (20% thermal rearrangement
at RT after 18 h).
Scheme 5. Proposed competitive mechanistic pathway via a boat-like in-
termediate 9 explaining the incomplete chirality transfer with the mis-
matched enantiomer of the starting material.
Chem. Asian J. 2010, 5, 1770 – 1774
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1773