MECHANISTIC ANALYSIS OF DOUBLE HYDROGEN DYOTROPY IN SYN-SESQUINORBORNENE DISULFONES - A COMBINED KINETIC AND THEORETICAL EVALUATION OF PRIMARY DEUTERIUM-ISOTOPE EFFECTS

Two sets of syn-sesquinorbornene disulfones have been prepared that ca rry either one or two deuterium atoms ct to the sulfonyl substituents. The rates of dyotropic rearrangement were measured in these systems i n order to assess the deuterium isotope effects associated with single and double H/D transfer. The temperature dependence of the deuterium isotope effects in the biscyclopropane series 3b-5b [k(HH)/k(HD) = 5.4 (0 degrees C) and 2.9 (100 degrees C); k(HH)/k(DD) = 33 (0 degrees C) and 8.5 (100 degrees C)] is not as steep as that when only one cyclop ropane ring is present as in 3a-5a [k(HH)/k(HD) = 3.8 (0 degrees C) an d 2.1 (100 degrees C); k(HH)/k(DD) = 200 (0 degrees C) and 11.2 (100 d egrees C)]. Only in the b series is the Rule of Geometric Mean obeyed. In an effort to gain detailed insight into the relationship between s train energy changes and rates of these reactions, the isomerization r ates for a significant number of syn-sesquinorbornene disulfones were first modeled by means of the empirical MM3 force field. Good agreemen t was found based upon a concerted model. Transition structures for se ven concerted dyotropic hydrogen transfers of syn-sesquinorbornene ana logs were located with ab initio 3-21G calculations. The potential ene rgy paths for concerted and stepwise hydrogen transfer were also evalu ated at the CASSCF level of theory using minimum STO-3G and 3-21G basi s sets. The concerted path is predicted to be favored in the absence o f tunneling as in the b series. To take the tunneling effect into cons ideration, a more detailed dynamic treatment of the one- or two-dimens ional barrier issue was next implemented, This model supports-tunnelin g by a stepwise mechanism for both 3a-5a and 3b-5b, but the empirical energy parameters do not agree well with the best CASSCF calculations. Elaborate direct dynamics calculations on model compounds identical w ith 3a-5a except for the phenylsulfonyl groups were Carried out. Altho ugh they yielded results consistent with the two-oscillator calculatio ns, the resulting activation energies are too low and leave room for s ome uncertainty about the mechanism, especially for compounds 3b-5b.