Magnetocrystalline anisotropy in metallic systems: fast and stable estimation in Green`s functions formalism

TL;DR

This paper introduces a stable and efficient Green's functions-based method to calculate magnetocrystalline anisotropy in metallic systems, improving numerical stability and convergence, validated on models and ab initio calculations.

Contribution

The work presents a novel analytical and numerical approach for calculating MCA energies that enhances stability and convergence in metallic systems using Green's functions formalism.

Findings

Reciprocal space resolution improves numerical stability.

Analytical replacement simplifies atomic summation.

Validated on models and ab initio calculations.

Abstract

In this work we suggest a theoretical approach, that allows to study the effects of magnetocrystalline anisotropy (MCA) in metallic systems using the Green`s functions formalism. We demonstrate that employment of the reciprocal space resolution instead of its reduction in the inter-site variant essentially improves the numerical stability of MCA energy by means of Monkhorst-Pack grid density and spatial convergence. The latter problem is able to be completely removed due to rigorous analytical replacement of pairwise atomic summation by simple composition of sublattices contributions, calculated as a whole. The approach is validated on the effective model of single atom, which nevertheless inherits the qualitative MCA picture of Co monolayer and Au/Co/Au sandwiched material. The numerical convergence is confirmed using the model of atomic chain in the strong metallic regime. For cobalt monoxide, described by ab initio calculations using GGA+U, the MCA energy angular profile reveals the prevailing role of ferromagnetically aligned Co sublattices in forming of the easy axis.