Segregation to grain boundaries in disordered systems: an application to a Ni-based multi-component alloy

TL;DR

This paper investigates how compositional disorder in Ni-based multi-component alloys affects segregation energies at grain boundaries, providing a statistical framework to predict segregation behavior and its thermodynamic parameters.

Contribution

It introduces a method to explicitly account for compositional disorder in predicting segregation energies in complex alloys, extending previous models.

Findings

Segregation energies vary significantly with alloy composition.

A statistical approach can extract segregation enthalpy and entropy.

Results align with experimental observations for complex alloys.

Abstract

Segregation to defects, in particular to grain boundaries (GBs), is an unavoidable phenomenon leading to changed material behavior over time. With the increase of available computational power, unbiased quantum-mechanical predictions of segregation energies, which feed classical thermodynamics models of segregation (e.g., McLean isotherm), become available. In recent years, huge progress towards predictions closely resembling experimental observations was made by considering the statistical nature of the segregation process due to competing segregation sites at a single GB and/or many different types of co-existing GBs. In the present work, we further expand this field by explicitly showing how compositional disorder, present in real engineering alloys (e.g. steels or Ni-based superalloys), gives rise to a spectrum of segregation energies. With the example of a $\Sigma 5$ GB in a Ni-based model alloy (Ni-Co-Cr-Ti-Al), we show that the segregation energies of Fe, Mn, W, Nb, and Zr are significantly different from those predicted for pure elemental Ni. We further use the predicted segregation energy spectra in a statistical evaluation of GB enrichment, which allows for extracting segregation enthalpy and segregation entropy terms related to the chemical complexity in multi-component alloys.