Parameter Optimization of Tersoff Interatomic Potentials Using a Genetic Algorithm

Abstract

A method that gives the parameters of advanced Tersoff interatomic potentials for describing nonequilibrium atomic structures has been developed. This method uses a genetic algorithm to optimize the Tersoff potential parameters fitted to first-principles-calculated cohesive energies of various carbon systems, including bulk systems with atomic defects and amorphous, surface, or cluster systems under stress. These optimized parameters converge towards a set of Tersoff potential parameters that well describes not only crystals but also amorphous systems.