Enhanced Field-Free Perpendicular Magnetization Switching via spin splitting torque in Altermagnetic RuO2-based Heterostructures

TL;DR

This paper demonstrates a novel approach to achieve field-free, low-power perpendicular magnetization switching in spintronic devices by leveraging altermagnetic RuO2 and Pt layers, enhancing efficiency without external magnetic fields.

Contribution

The study introduces the use of altermagnetic RuO2 combined with Pt to enable deterministic, field-free magnetization switching with optimized layer thickness and current direction, advancing spintronic technology.

Findings

Enhanced field-free switching achieved at 1.5 nm Pt thickness.

Low current density of 2.56e11 A/m2 used for switching.

Effective out-of-plane magnetic field maximized by tuning layer parameters.

Abstract

Current-induced spin-orbit torque (SOT) has emerged as a promising method for achieving energy-efficient magnetization switching in advanced spintronic devices. However, technological advancement has been inadequate because an external in-plane magnetic field is required to attain deterministic switching. Several approaches have been explored to address these challenges. In this work, we explored the potential of a newly emerged altermagnetic material RuO2 in combination with a Pt layer to achieve both field-free and low-power switching concurrently. We leveraged out-of-plane (OOP) spin polarization via the spin-splitter effect (SSE) in RuO2 for field-free switching (FFS) and in-plane spin polarization combined with spin Hall effect (SHE) in Pt for enhanced SOT efficiency. We revealed that the effective OOP magnetic field and FFS can be maximized by tuning the nominal thickness of the Pt underlayer and the direction of the applied current. We observed a significant enhancement in FFS at an optimized Pt thickness of 1.5 nm for an applied current density as low as 2.56e11 A/m2 at a crystal angle of 90 deg. Our study paves the way for energy-efficient spintronics devices for non-volatile memory, logic circuits, and neuromorphic computing.