Molecular dynamics-based ion-surface interaction models for ionized physical vapor deposition feature scale simulations

A procedure is presented for incorporating the results of atomistic simulations of ion–surface interactions into integrated circuit topographic simulations of ionized physical vapor deposition (PVD). Energy and angular dependent sticking probabilities for energetic Cu atoms impacting a {111} Cu surface, obtained from molecular dynamics simulations, were implemented in a simple Monte Carlo flux model. The resulting flux-averaged Cu sticking probability was found to vary significantly with position within submicron features and with the feature geometry. This illustrates the shortcomings of a constant (energy and angle independent) sticking probability model for ionized PVD.