Singlet-triplet splittings and barriers to Wolff rearrangement for carbonyl carbenes.

High-level ab initio calculations at the G3(MP2)//B3-LYP level have been used to study carbomethoxychlorocarbene and related halogenocarbenes and carbonyl carbenes. Initial calculations at the more accurate W1' level on the subset CH(2), HCCl, HCF, CCl(2), and CF(2) provide support for the reliability of G3(MP2)//B3-LYP for this type of problem. The W1' calculations also suggest that the experimental S-T splitting is slightly underestimated for HCCl and CF(2) and substantially underestimated for CCl(2), in keeping with other recent high-level studies. Whereas the parent carbonyl carbenes, namely formylcarbene, carbohydroxycarbene, and carbomethoxycarbene, are all predicted to have triplet ground states, their chloro and fluoro derivatives are predicted to have singlet ground states. In particular, carbomethoxychlorocarbene is predicted to have a singlet ground state, with the singlet-triplet splitting estimated as -16.0 kJ mol(-)(1). The barriers to Wolff rearrangement of the singlet carbonyl carbenes generally (but not always) correlate with the exothermicity accompanying the production of ketenes. In the case of the parent carbonyl carbenes, for which the rearrangement reaction is most exothermic, the barriers lie between about 10 and 30 kJ mol(-)(1), whereas for the less exothermic rearrangements of the chloro- and fluoro-substituted carbonyl carbenes, the Wolff rearrangement barriers increase significantly to between 58 and 75 kJ mol(-)(1). The calculated barrier for carbomethoxychlorocarbene is 58.2 kJ mol(-)(1).