Theoretical studies of the potential surface for the F+H2→HF+H reaction

Abstract
The F+H2→HF+H potential energy hypersurface has been studied in the saddle‐point and entrance channel regions. Using a large [5s 5p 3d 2f 1g/4s 3p 2d] atomic natural orbital basis set, we obtain a classical barrier height of 1.86 kcal/mol at the CASSCF/multireference CI level (MRCI) after correcting for basis set superposition error and including a Davidson correction (+Q) for higher excitations. Based upon an analysis of the computed results, the true classical barrier is estimated to be about 1.4 kcal/mol. We also compute the location of the bottleneck on the lowest vibrationally adiabatic potential curve, and determine the translational energy threshold from a one‐dimensional tunneling calculation. Using the difference between the calculated and experimental threshold to adjust the classical barrier height on the computed surface yields a classical barrier in the range of 1.0–1.5 kcal/mol. Combining the results of our direct estimates of the classical barrier height with the empirical values obtained from our approximate calculations of the dynamical threshold, we predict that the true classical barrier height is 1.4±0.4 kcal/mol. Arguments are presented in favor of including the relatively large (≈1 kcal/mol)+Q correction obtained when nine electrons are correlated at the CASSCF/MRCI level.