Machine Learning Accelerated Nonadiabatic Dynamics at Metal Surfaces

In this chapter, we will introduce the fundamental ideas and concepts behind the simulation of gas-surface reactions at metal surfaces along with a set of well-established theoretical models to describe it. The widely assumed Born-Oppenheimer Approximation breaks down for chemical reactions at metallic surfaces and nanoparticles and an explicit description of nonadiabatic effects becomes necessary to describe energy dissipation effects between molecular adsorbate degrees of freedom and electronic excitations in the substrate. We will discuss existing methods to perform nonadiabatic molecular dynamics at surfaces and the crucial role of machine learning methods to enable such simulations.