Self-consistent calculation of spin transport and magnetization dynamics

TL;DR

This paper reviews the importance of self-consistent calculations in spin transport and magnetization dynamics, emphasizing feedback mechanisms that influence magnetic behaviors and improve agreement with experimental results.

Contribution

It introduces a comprehensive approach to self-consistently model coupled spin transport and magnetization dynamics, highlighting non-local feedback effects in various magnetic systems.

Findings

Feedback mechanisms generate non-local effective interactions.

Self-consistent calculations align better with experimental data.

Feedback reduces oscillation linewidth in spin valves.

Abstract

A spin-polarized current transfers its spin-angular momentum to a local magnetization, exciting current-induced magnetization dynamics. So far, most studies in this field have focused on the direct effect of spin transport on magnetization dynamics, but ignored the feedback from the magnetization dynamics to the spin transport and back to the magnetization dynamics. Although the feedback is usually weak, there are situations when it can play an important role in the dynamics. In such situations, self-consistent calculations of the magnetization dynamics and the spin transport can accurately describe the feedback. This review describes in detail the feedback mechanisms, and presents recent progress in self-consistent calculations of the coupled dynamics. We pay special attention to three representative examples, where the feedback generates non-local effective interactions for the magnetization. Possibly the most dramatic feedback example is the dynamic instability in magnetic nanopillars with a single magnetic layer. This instability does not occur without non-local feedback. We demonstrate that full self-consistent calculations generate simulation results in much better agreement with experiments than previous calculations that addressed the feedback effect approximately. The next example is for more typical spin valve nanopillars. Although the effect of feedback is less dramatic because even without feedback the current can induce magnetization oscillation, the feedback can still have important consequences. For instance, we show that the feedback can reduce the linewidth of oscillations, in agreement with experimental observations. Finally, we consider nonadiabatic electron transport in narrow domain walls. The non-local feedback in these systems leads to a significant renormalization of the effective nonadiabatic spin transfer torque.