A neural network potential energy surface for the F + H2O ↔ HF + OH reaction and quantum dynamics study of the isotopic effect
Abstract
A global potential energy surface for the F + H2O ↔ HF + OH reaction has been constructed using the neural network method based on ∼24 000 ab initio energies calculated at the all-electron CCSD(T)-F12a/cc-pCVTZ-F12 level of theory. The correction term accounting for the influence of spin–orbit couplings has also been included with a hierarchical scheme. The isotopic effect on the total reaction probabilities of the reaction was investigated using the time-dependent wave packet method.
- This article is part of the themed collection: Emerging AI Approaches in Physical Chemistry