enforce an increase in the charge separation. However, whether an 0x0 ligand is sufficiently voluminous to have this steric effect is debatable, and calculations are currently underway to resolve this question." ' 1 Although the greater thermodynamic stability of the trans alcohol with respect to that of the cis isomer (ca. 1.5 kcal mol-') could lead to a relative lowering of the transition state free energy (invoking the Hammond postulate), it would be insufficient to account for these effects. Interestingly, in the thermal gas phase rearrangements of ally1 esters,[l2] in which a similar mechanism involving an allyl migration process is postulated, the cis product was also found to be formed at a considerably lower rate than the trans isomer.
Overall these results are in good agreement with the model proposed by Charbardes et al., which, however, must be modulated by some rather subtle effects in the charge distribution in the cyclic transition state and its effects on selectivity. It appears that the type of bond formed between the allyl fragment and the perrhenium fragment in the transition state may vary greatly from ionic-covalent to nearly totally ionic." 31 Finally, given the very high activity of these rhenium catalysts under mild conditions, urther modification of the ligand environment, for example by replacing 0x0 ligands with imido groups, would appear to offer the prospect of both greater catalyst control and selectivity in rearrangements of this type.