Breaking the Coupled Cluster Barrier for Machine Learned Potentials of Large Molecules: The Case of 15-atom Acetylacetone

Abstract

Machine-learned potential energy surfaces (PESs) for molecules with more than 10 atoms are typically forced to use lower-level electronic structure methods such as density functional theory and second-order Møller-Plesset perturbation theory (MP2). While these are efficient and realistic, they fall short of the accuracy of the "gold standard" coupled-cluster method [CCSD(T)], especially with respect to reaction and isomerization barriers. We report a major step forward in using a ∆-machine learning method to the challenging case of acetlyacetone, whose MP2 barrier height for H-atom transfer is low by roughly 1.5 kcal/mol relative to the CCSD(T) barrier of 3.5 kcal/mol. From a database of 2151 LCCSD(T) energies, and training with as few as 430 energies, we obtain a new PES with a barrier in agreement with the LCCSD(T) one. Tunneling splittings due to H-atom transfer are calculated using this new PES, providing improved estimates over previous ones obtained using an MP2-based PES.