Theory of spin-transfer torque in the current-in-plane geometries

Wessely, O.; Umerski, A. and Mathon, J. (2009). Theory of spin-transfer torque in the current-in-plane geometries. Physical Review B, 80(1) 014419-1.

DOI: https://doi.org/10.1103/PhysRevB.80.014419

Abstract

Two alternative current-induced switching geometries, in which the current flows parallel to the magnet/nonmagnet interface, are investigated theoretically using the nonequilibrium Keldysh theory. In the first geometry, the current is perpendicular to the polarizing magnet/nonmagnet interface but parallel to the nonmagnet/switching magnet interface (CPIP). In the second geometry, the current is parallel to both polarizing magnet/nonmagnet and nonmagnet/switching magnet interfaces (CIP). Calculations for a single-orbital tight-binding model indicate that the spin current flowing parallel to the switching magnet/nonmagnet interface can be absorbed by a lateral switching magnet as efficiently as in the traditional current-perpendicular-to-plane (CPP) geometry. The results of the model calculations are shown to be valid also for experimentally relevant Co/Cu CPIP system described by fully realistic tight-binding bands fitted to an ab initio band structure. It is shown that almost complete absorption of the incident spin current by a lateral switching magnet occurs when the lateral dimensions of the switching magnet are of the order of 50–100 interatomic distances, i.e., about 20 nm and its height as small as a few atomic planes. It is also demonstrated that strong spin current absorption in the CPIP/CIP geometry is not spoiled by the presence of a rough interface between the switching magnet and nonmagnetic spacer. Polarization achieved using a lateral magnet in the CIP geometry is found to be about 25% of that in the traditional CPP geometry. The present CPIP calculations of the spin-transfer torque are also relevant to the so-called pure-spin-current-induced magnetization switching that had been recently observed.

Viewing alternatives

Metrics

Public Attention

Altmetrics from Altmetric

Number of Citations

Citations from Dimensions

Item Actions

Export

About

Recommendations