GPU Accelerated Atomistic Energy Barrier Calculations of Skyrmion Annihilations

TL;DR

This paper introduces GPU-accelerated atomistic simulations to compute the energy barriers of magnetic skyrmion annihilations, crucial for their stability in storage and logic applications.

Contribution

It applies GPU acceleration to atomistic spin models to efficiently calculate skyrmion annihilation energies considering multiple magnetic interactions.

Findings

Boundary annihilations have slightly lower energy barriers than Bloch point annihilations.

Both annihilation processes exhibit similar behavior across various interaction energies.

GPU acceleration enables faster and more detailed simulations of skyrmion stability.

Abstract

We present GPU accelerated simulations to calculate the annihilation energy of magnetic skyrmions in an atomistic spin model considering dipole-dipole, exchange, uniaxial-anisotropy and Dzyaloshinskii-Moriya interactions using the simplified string method. The skyrmion annihilation energy is directly related to its thermal stability and is a key measure for the applicability of magnetic skyrmions to storage and logic devices. We investigate annihilations mediated by Bloch points as well as annihilations via boundaries for various interaction energies. Both processes show similar behaviour, with boundary annihilations resulting in slightly smaller energy barriers than Bloch point annihilations.