Potential energy surfaces: Δ-machine learning from analytical functional forms

Delta-machine learning (Δ-ML) is a highly cost-effective approach to developing high-level potential energy surfaces (PES) from a large number of low-level configurations. In particular, the high flexibility of the analytical PES-2008 is exploited to efficiently sample points from the low-level data set and, using information from the highly accurate PIP-NN surface, construct the Δ-ML PES. This approach is applied to the well-known H + CH4 hydrogen abstraction reaction. In order to test the validity and accuracy of the Δ-ML approach to describe this polyatomic system, kinetic studies using the variational transition state with multidimensional tunneling corrections and dynamic studies on the deuterated reaction, H + CD4, reaction using quasiclassical trajectory calculations were performed on the three surfaces, PES-2008 (low level), PIP-NN (high-level) and Δ-ML. The delta-machine learning approach reproduces the kinetics and dynamics information of the high-level surface, showing its efficiency in describing multidimensional polyatomic systems.