The anionic [3 + 2] cycloaddition of allyl anions or
allyllithium compounds to double or triple
bonds is an elegant route both to carbocyclic and to heterocyclic
five-membered rings. The mechanism of
such reactions has not yet been established conclusively. In this
computational study, the concerted 4πs +
2πs mechanism, expected on the basis of the
Woodward−Hoffmann rules, is found to be less favorable
than
two-step pathways for the cycloadditions of ethylene to the allyl,
2-borylallyl, and 2-azaallyl anions and their
lithiated counterparts at Becke3LYP/6-311+G** and
MP2(fc)/6-31+G* levels of theory. Except for
allyllithium,
the 4πs + 2πs cycloadditions (in
C
s
symmetry) are not concerted, since
only second-order saddle points, rather
than true transition structures, are involved. The anisotropy of
the reactant polarizabilities is responsible.
Instead, two-step cycloaddition pathways are followed by all three
model systems. In accord with experimental
experience, 2-borylallyl and 2-azaallyl compounds are found to undergo
this type of reaction more readily
than the unsubstituted allyl anion or allyllithium. The second,
ring-closing step is facilitated by the anion-stabilizing effect of nitrogen and the boryl substituent.