Dynamics in Heterogeneous and Single-Site Catalysis

Sen Lin,Hua Guo
DOI: https://doi.org/10.1016/b978-0-12-821978-2.00008-8
2022-01-01
Abstract:We survey recent advances in theoretical approaches to the dynamics of molecule-surface interaction on solid surfaces relevant to catalysis. These new developments involve quantum and classical methods for exploring scattering, adsorption/desorption, diffusion, and reaction of molecules at the gas-surface interface. Such dynamics calculations require an accurate description of the electronic structure of the combined system within the Born-Oppenheimer approximation, which is provided commonly by density functional theory (DFT). For many such extended systems, the dynamics can be accurately and conveniently represented by ab initio molecular dynamics (AIMD), in which energies and forces are computed on the fly by DFT. The bottleneck in AIMD calculations is often the expensive on-the-fly DFT calculations along trajectories. A recent breakthrough is to replace the DFT calculations by training global high-dimensional potential energy surfaces (PESs) from energies and gradients using machine learning methods such as neural networks, as the resulting analytic PESs can significantly accelerate dynamical calculations. Further corrections beyond the Born-Oppenheimer approximation on metal surfaces can be included in dynamical simulations by using the friction model. These powerful theoretical tools have revealed valuable insights in these important processes, which help to understand the role played by dynamics in heterogeneous and single-site catalysis.
What problem does this paper attempt to address?