Multidimensional Neural Network Interatomic Potentials for CO on NaCl(100)

The advent of machine learning (ML) models has unlocked new possibilities in the realm of interatomic potentials. We use an equivariant graph Neural Network (NN) to construct interatomic potentials for a versatile system, CO on a NaCl(100) surface, and mediate efficient large-scale atomistic simulations with ab initio molecular dynamics accuracy. We report two NN potentials, one trained on equilibrium configurations at finite temperatures (T = 30, 300 K), and the other additionally trained upon nonequilibrium trajectories of pre-excited CO adsorbates. We demonstrate first applications of the ML potentials for (i) adsorption energies and barriers for reactions, (ii) potential energy landscapes for submonolayer and monolayer coverages, (iii) vibrational spectra at finite temperatures, and (iv) vibrational relaxation dynamics. Further possible applications are discussed.