I have carried out simplified analysis of some recent papers on organocatalysis. The first example comes from a paper from the Jacobsen group (Raheem et al. J. Am. Chem. Soc. 2007, 13404, 10.1021/ja076179w) which describes a “H-bond donor” catalyst that effects an enantioselective Pictect-Spengler type reaction. Jacobsen attributes the selectivity to the substrate-halide-catalyst complex shown on the left. However, leaving the halide out of the complex yields an ion pair or perhaps a covalently bound intermediate as shown on the right. This brings the catalyst and substrate into closer contact than the halide-containing complex.
A ground-state model of the des-halide intermediate appears to work in explaining the enantioselectivity of the reaction as shown in the pictures below. In intermediate I, which leads to the minor enantiomer, the indole ring is not able to cyclize onto the acyliminium ion since it is being blocked by an appendage projecting from the catalyst. In contrast when the catalyst is bound to the other face of the acyliminium ion, as in intermediate II, the indole is free to cyclize which leads to the observed dominant enantiomer. Although Jacobsen indicates that some of their experimental work on the mechanism of catalysis makes the direct interaction of the catalyst with the acyliminium less likely, this intermediate is consistent, at least qualitatively, with the stereoselectivity reported for the paper. I am eager to see more details concerning the mechanism to see if this my interpretation holds up.