Accurate representation of formation energies of crystalline alloys with many components

TL;DR

This paper introduces a new model for accurately predicting formation energies of multicomponent crystalline alloys, especially effective with small datasets, and comparable or better than existing methods.

Contribution

A novel, simple model with two parameters that achieves high accuracy in representing alloy formation energies, outperforming cluster expansion in small data regimes.

Findings

Comparable accuracy to cluster expansion for small displacements

Better accuracy with small datasets

Minimal tunable parameters

Abstract

In this paper I propose a new model for representing the formation energies of multicomponent crystalline alloys as a function of atom types. In the cases when displacements of atoms from their equilibrium positions are not large, the proposed method has a similar accuracy as the state-of-the-art cluster expansion method, and a better accuracy when the fitting dataset size is small. The proposed model has only two tunable parameters - one for the interaction range and one for the interaction complexity.