Epoxidation of acyclic chiral allylic alcohols with peroxy acids: spiro or planar butterfly transition structures? A computational DFT answer

The mechanism of the epoxidation of two chiral allylic alcohols, i.e., 3-methyl-3-buten-2-ol and (Z)-3-penten-2-ol, with peroxyformic acid has been investigated by locating 20 transition structures with the B3LYP/6-31G* method and by evaluating their electronic energy also at the B3LYP/6-311+G**@B3LYP/6-31G* theory level. Relative stability of TSs, as far as electronic energy is concerned, is basis set dependent; moreover, it also depends on entropy and solvent effects. Free enthalpies, calculated by using electronic energy at the higher theory level and with inclusion of solvent effects, indicates that syn, exo TSs, where the olefinic OH group hydrogen bonds the peroxy oxygens of the peroxy acid, outweigh syn, endo TSs, where the peroxy acid carbonyl oxygen is involved in hydrogen bonding. In the former TSs the peroxy acid moiety maintains its planar geometry while in the latter ones a strong out-of-plane distortion of peroxy acid is observed. This distortion makes it viable an unprecedented 1,2-H shift, as a possible alternative to the 1,4-H shift, for the peroxy acid hydrogen. In fact, for one syn, endo TS IRC analysis demonstrated that the 1,2-H shift mechanism is actually operative. The geometry of all TSs substantially conforms to a spiro (i.e., with the peroxy acid plane almost perpendicular to the C=C bond axis) butterfly orientation of the reactants while no TS resembles, even loosely, the planar butterfly structure. Theoretical threo/erythro epoxide ratios are in fair accord with experimental data. Calculations indicate that threo epoxides derive mostly from TSs in which the olefinic OH assumes an outside conformation while erythro epoxides originate from TSs with the OH group in an inside position. Computational findings do not support the qualitative TS models recently proposed for these reactions.