Implementation of the magnetic force theorem for large-scale calculations of magnon bands: application to yttrium iron garnet

TL;DR

This paper introduces an efficient, high-throughput compatible method to compute magnon band structures directly from Bloch states without localized orbital mapping, demonstrated on yttrium iron garnet with excellent experimental agreement.

Contribution

The paper presents a novel implementation of the magnetic force theorem for large-scale magnon calculations directly in q-space, bypassing the need for localized orbitals and enabling high-throughput applications.

Findings

Accurate magnon dispersion matching experiments

Identification of multiple exchange pathways

Validation of the method for complex magnetic materials

Abstract

We present an efficient implementation of the magnetic force theorem which allows for direct evaluation of exchange parameters in q-space. The exchange parameters are calculated directly from Bloch states and the implementation does not rely on any mapping onto localized orbitals. This renders the approach well suited for high-throughput computations, where the construction of a localized basis set (for example Wannier functions) often is impractical. We demonstrate the versatility of the method by applying it to yttrium iron garnet, where we obtain excellent agreement with the experimental magnon dispersion and Curie temperature without any prior assumptions of important exchange pathways. In particular, the calculations reveal the existence of several inequivalent exchange pathways associated with the same interatomic distances. Performing such calculations in q-space fully accounts for long-range exchange interactions and provides a convenient route for validating models obtained by fitting to inelastic neutron scattering data.