Solute diffusion calculation in Fe-Si and Fe-Cr-Si multicomponent alloys

TL;DR

This paper develops efficient computational methods to estimate solute diffusion coefficients in multicomponent alloys, focusing on silicon in Fe-based systems, by using bond potentials and thermodynamic simulations instead of costly kinetic models.

Contribution

It introduces a computationally efficient approach to estimate silicon diffusion in Fe-based alloys using bond potentials and thermodynamic simulations, bypassing expensive kinetic calculations.

Findings

Silicon jump frequency can be estimated from thermodynamic simulations.

The silicon correlation factor varies slightly with concentration.

Bond potentials can replace machine learning potentials for diffusion calculations.

Abstract

Diffusion plays a key role in microstructure evolution at multicomponent alloys: diffusion controls the kinetics of phase transformations and alloy homogenization. This study aims at developing computationally efficient approaches to estimate the solute diffusion coefficients in two-component systems. We consider silicon as the solute example because it is highly used in industrial steels. We demonstrate that the silicon jump frequency may be calculated with the bond potential instead of the more computationally expensive machine learning potential in Fe-Si and Fe-Cr-Si alloys. We show that the silicon jump frequency can be estimated from thermodynamic simulations for the bond potential without kinetic simulations. The silicon correlation factor slightly depends on silicon concentration and can be approximately estimated by the analytical nine-frequency model.