Physics-separating artificial neural networks for predicting initial stages of Al sputtering and thin film deposition in Ar plasma discharges

Abstract Simulations of Al thin film sputter depositions rely on accurate plasma and surface interaction models. Establishing the latter commonly requires a higher level abstraction means to dismiss fundamental atomic fidelity. Previous works sputtering processes addressed this issue by establishing machine learning surrogate models, which include basic state (i.e. stoichiometry) as static input. In work, an evolving defect structure are introduced jointly describe growth with physics-separating artificial neural networks. The data describing plasma–surface interactions (PSIs) stem from hybrid reactive molecular dynamics/time-stamped force bias Monte Carlo simulations neutrals Ar + ions impinging onto Al(001) surfaces. It is demonstrated that comprehensively described taking well into account. Hence, PSI model established resolves inherent kinetics high physical resulting not restricted input modeling simulation, but may similarly be applied experimental data.