Manipulating transient SOT-MRAM switching dynamics for efficiency improvement and probabilistic switching - Scientific Reports

Manipulating Transient SOT-MRAM Switching Dynamics for Efficiency Improvement and Probabilistic Switching

This study examines transient dynamics in the switching process of spin-orbit torque magnetic random-access memory (SOT-MRAM) devices stabilized by in-plane uniaxial magnetocrystalline anisotropy. A theoretical framework is developed to describe the interaction between spin torques and effective fields throughout the magnetization write trajectory, identifying regions of both failed and successful switching.

Quasi-Stochastic Regime in Switching Dynamics

The research highlights a "quasi-stochastic" regime situated between deterministic regions of failure and success. This regime arises from the interplay between torque-driven and precession-driven magnetization during switching, which introduces complexity in the device operation.

Enhancing Switching Efficiency

By implementing minor changes to device geometry, material properties, and electrical inputs, the study leverages transient effects to reduce the switching barrier. This approach enables SOT-MRAM switching with notably lower current requirements and faster write speeds compared to conventional designs.

Probabilistic Switching at Elevated Temperatures

At higher temperatures, the initially unpredictable stochastic regime transitions into a well-defined probabilistic "transition band" featuring monotonic and adjustable regions of probabilistic behavior. This controlled probabilistic operation is essential for emerging computing paradigms.

“Through this addition of control mechanisms through electrical inputs, our framework paves the way for the creation of a fast, efficient probabilistic bit (p-bit) for the field of probabilistic computing.”

Author's summary: The study demonstrates how transient effects in SOT-MRAM switching can be controlled to achieve lower power, faster operation, and enable tunable probabilistic behavior suitable for next-generation computing.

more

Nature Nature — 2025-11-01