Calculations which include dynamic electron correlation beyond the CASSCF level have been carried out for the interconversion of homocub-1(9)-ene (1) and homocub-9-ylidene (2). The geometry of the transition state was also located at the (4/4)CASSCF/6-31G* level and was found to di er signi® cantly from the previously published (2/2)CASSCF/3-21G geometry. In agreement with experiment, calculations at the (2/2)CASSDCI+ Q/6-31G* level ® nd 1 and 2 to have free energies at 298 K that di er by only~0 . 1 kcal mol -1 . However, the activation energy for the rearrangement of 1 to 2 of E a = 12 . 9 kcal mol -1 , computed at this level of theory, is considerably higher than the experimental estimate of E a < 5 kcal mol -1 . Additional calculations in which the size of the basis set was increased or in which the active space was expanded to 4 electrons in 4 orbitals suggest that the calculated E a may be reduced to as low as 8 kcal mol -1 . However, there is no indication that higher level calculations would give a value of E a as low as 5 kcal mol -1 . Possible reasons for this apparent disagreement between the calculations and the experiments on the rate of equilibration of 1 and 2 are discussed.
IntroductionRearrangements of carbenes to ole® ns usually are irreversible reactions [1]. Nevertheless, bridgehead ole-® ns with su cient torsional strain are capable of undergoing reversion to carbenes [2]. The best studied example of the latter process was discovered by Eaton and Ho mann in the rearrangement of 9-phenylhomocub-1(9)-ene (1a) to 1-phenylhomocub-9-ylidene (2a) [3]. Subsequent investigations in the desphenyl series demonstrated that the parent ole® n (1b) also can be formed from the carbene (2b) [4], and experiments by Jones, Platz and coworkers have found the equilibrium constant for this reaction to be approximately unity at room temperature [5].Jones, Platz, and coworkers started with di erent precursors, which led initially either to 1b or to 2b. Independent of the precursor used, the same mixture of CH 3 OD trapping products of 1b and 2b was formed, provided that the concentration of CH 3 OD was less than 1 m. Therefore, under these conditions equilibration of 1b and 2b appears to be faster than trapping with CH 3 OD.Flash photolysis studies indicated that both the ole® n and carbene react with CH 3 OD within an order of magnitude of the di usion controlled rate. Taken together, these two observations indicate that the rate constant for equilibration of 1b and 2b at room temperature