A semi-analytical model to simulate the spin-diode effect and accelerate its use in neuromorphic computing

TL;DR

This paper introduces a semi-analytical, data-driven model called DD-TEA that accurately and rapidly simulates the spin-diode effect in spin-torque vortex oscillators, with potential applications in neuromorphic computing.

Contribution

The paper presents the novel DD-TEA model that combines analytical equations with micromagnetic data to efficiently simulate non-linear spin-torque vortex dynamics.

Findings

The DD-TEA model accurately predicts experimental spin-diode behavior.

Reversal of the spin-diode effect depends on vortex chirality.

The model demonstrates potential for neuromorphic computing applications.

Abstract

The spin-diode effect is studied both experimentally and with our original semi-analytical method. The latter is based on an improved version of the Thiele equation approach (TEA) that we combine to micromagnetic simulation data to accurately model the non-linear dynamics of spin-torque vortex oscillator (STVO). This original method, called data-driven Thiele equation approach (DD-TEA), absorbs the difference between the analytical model and micromagnetic simulations to provide a both ultra-fast and quantitative model. The DD-TEA model predictions also agree very well with the experimental data. The reversal of the spin-diode effect with the chirality of the vortex, the impact of the input current and the origin of a variation at half of the STVO frequency are presented as well as the ability of the model to reproduce the experimental behavior. Finally, the spin-diode effect and its simulation using the DD-TEA model are discussed as a promising perspective in the framework of neuromorphic computing.