Modeling magnetic evolution and exchange hardening in disordered magnets: The example of Mn$_{1-x}$Fe$_x$Ru$_2$Sn Heusler alloys

TL;DR

This paper uses a first-principles statistical mechanics approach to model magnetic evolution and exchange hardening in disordered Mn$_{1-x}$Fe$_x$Ru$_2$Sn Heusler alloys, revealing local magnetic interactions and temperature effects.

Contribution

It develops a mixed-basis chemical and magnetic cluster expansion to predict magnetic properties and exchange hardening in disordered alloys from first principles.

Findings

Reproduces experimental magnetic transition temperatures and magnetization.

Predicts local antiferromagnetic regions within ferromagnetic matrix.

Explains temperature-dependent decrease in exchange hardening.

Abstract

We demonstrate how exchange hardening can arise in a chemically-disordered solid solution from a first-principles statistical mechanics approach. A general mixed-basis chemical and magnetic cluster expansion has been developed, and applied to the Mn$_{1-x}$Fe$_x$Ru$_2$Sn Heusler alloy system; single-phase solid solutions between antiferromagnetic \ch{MnRu2Sn} and ferromagnetic \ch{FeRu2Sn} with disorder on the Mn/Fe sublattice that exhibit unexpected exchange hardening. Monte Carlo simulations applied to the cluster expansion are able to reproduce the experimentally measured magnetic transition temperatures and the bulk magnetization as a function of composition. The magnetic ordering around a site is shown to be dependent not only on bulk composition, but also on the identity of the site and the local composition around that site. The simulations predict that local antiferromagnetic orderings form inside a bulk ferromagnetic region at intermediate compositions that drives the exchange hardening. Furthermore, the antiferromagnetic regions disorder at a lower temperature than the ferromagnetic regions, providing an atomistic explanation for the experimentally-observed decrease in exchange hardening with increasing temperature. These effects occur on a length scale too small to be resolved with previously-used characterization techniques.