Remarkably, the enoate group lies in a environment relative to the Lewis acid (Fig. 2 ) ; the substituents at the chirality center make no significant contribution to the diastereoface-differentiation.
How does this interpretation fit into the general picture? For complexes of "classical" chiral acrylates (prototype: menthyl acrylate) with only one center capable of coordination (carbonyl-0), reactive conformers 4a with antiplanar enoate conformation are postulated on the basis of the configuration of their addition products[',81. Consistent with this postulation is the structure of a 2 : I-methyl cinnamate-SnC1, complex in the crystal" la]. This structure, as well as two additional structures of achiral ester complexes["], further suggest that o-coordination, with the position of the Lewis acid as given in 4, is preferred for nonchelate complexes. The synplanar conformation of the enoate group in the complex 4b could consequently be destabilized due to repulsive interactions"21. Other than in the case of 3, the o-coordinated Lewis acid does not contribute to the diastereodifferentiation.
Complexes of the type 4 might also play a role for the reactions of the lactate 1. We assume that 3 is in equilibrium with complexes 5 (MX,=TiCI,). For these, preferred Re-attack of the diene (to give 2b) is to be expected. Since the formation of 5b is favored by high concentration of TiCl,, the stereoselectivity of the addition reaction is lowered if the ratio TiCl, : 1 is > 1 (Fig. 1). Assumption of alternatives 3 and 5 also explains the stereodichotomic effect: with Tic],, predominantly the chelate complex 3 reacting to give 2a is formed, whereas with tetracoordinating Lewis acids such as EtAICI, or BF3 the complexes 5 predominate, from which 2b is then formed selectively.