A New Embedded-Atom Method Approach Based On the p-th Moment Approximation

TL;DR

This paper introduces a novel approach to constructing embedded-atom method potentials using the p-th moment approximation, reducing the number of fitting parameters while maintaining accuracy for large-scale atomistic simulations.

Contribution

It proposes modifications to the EAM potential construction and employs a smaller reference data set to enhance robustness and reduce parameters without sacrificing precision.

Findings

Reduced FPs in EAM potentials while maintaining accuracy

Enhanced potential robustness and scalability

Efficient potential construction with smaller data sets

Abstract

Large scale atomistic simulations with suitable interatomic potentials are widely employed by scientists or engineers of different areas. Quick generation of high-quality interatomic potentials is of urgent need under present circumstances, which largely relies on the developments of potential construction methods and algorithms in this area. Quantities of interatomic potential models have been proposed and parameterized with various methods, such as analytic method, force-matching approach and multi-object optimization method, in order to make the potentials more transferable. Without apparently lowing precisions for describing the target system, potentials of fewer fitting parameters (FPs) are somewhat more physically reasonable. Thus, studying methods of reducing FP number is helpful to understand the underline physics of simulated systems and generalize the construction methods to other similar systems. However, few reported works concentrate on methods of reducing the number of FPs without affecting precisions. In this work, the methods of reducing the FP number while keeping the precisions are discussed from two aspects. Firstly, the physical ideas of constructions of the embedded-atom method (EAM) potential model are modified to make the potential more robust, flexible and scalable without introducing too many FPs.Secondly, the smaller reference data set (SRD), compared with the target reference database for describing the mechanical behaviors of aluminum, is employed in our construction procedures.