Prediction of the mixing enthalpy by improving the MEAM potential function for Cu-based binary systems

Abstract

In this study, the melting points and standard formation enthalpies of pure metals (aluminum, iron, cobalt, nickel, copper, silver) are verified using the second-order nearest-neighbor modified embedding atom method (2NN-MEAM), confirming the physical rationality of their basic parameters. Regarding the “failure of solid-state parameters in the liquid state” problem existing in the existing Cu-X (X = Al, Fe, Co, Ni, Ag) system MEAM potential, an optimization strategy for potential parameters based on energy gradient analysis and liquid-state configuration sensitivity is proposed. By reconstructing the dynamic correlation mechanism between the multi-body interaction parameters and the local environmental response, the optimized potentials significantly improve the accuracy of the liquid state mixing enthalpy prediction. Moreover, by comparing with the experimental data, it is proved that the optimized MEAM potential significantly improves the accuracy of the mixed enthalpy prediction for the alloy systems.