We have explored the potential energy surfaces of the title reactions by density functional theory (DFT) and ab initio methods. The DFT calculations were based on the local density approximation (LDA) as well as the more sophisticated approach, NL-SCF, in which nonlocal corrections are included self consistently. The ah initio methods made use of the Hartree-Fock (HF) scheme as well as up to fourth-order Moller-Plesset perturbation theory (MP4). We have systematically characterized the geometries, frequencies, and energies for the reactants, ion-dipole complexes, and the transition states. Our study shows that the DFT methods offer overall better geometries and frequencies than the HF and MP2 schemes in comparison with the experimental results. In predicting the C-X bond energies of the reactants, CH(3)X, the NL-SCF scheme is superior to ah other methods applied in this study. The NL-SCF and MP4 complexation energies are similar and in good agreement with the experimental results for all but the fluorine system, for which the NL-SCF value is about 6 kcal/mol larger than the MP4 estimate. For the transition state energies, i.e., the barrier heights, the ab initio and DFT results turn out to be qualitatively different in the order HF >> MP2 > MP4 >> NL-SCF >> LDA. The experimental data seem to fall into the region with the MP4 and NL-SCF values as the upper and lower bounds, respectively. Within the DFT approaches, the relativistic effects on the geometries, frequencies, and energies were discussed, and the intrinsic reaction coordinate (IRC) method was utilized to provide further information about the potential energy surfaces, and to rationalize the reaction mechanism. We finally carried out bond energy decomposition and population analyses on the X-C bonds formed or broken during the reaction processes studied here.
