Time-resolved measurements of spin-transfer-induced (STI) magnetization reversal were made in current-perpendicular spin-valve nanomagnetic junctions subject to a pulsed current bias. These results can be understood within the framework of a Landau-Lifshitz-Gilbert equation that includes STI corrections and a Langevin random field for finite temperature. Comparison of these measurements with model calculations demonstrates that spin-transfer induced excitation is responsible for the observed magnetic reversal in these samples.

Current-induced magnetization dynamics in Co/Cu/Co trilayer nanopillars (∼100 nm in diameter) have been studied experimentally at low temperatures for large applied fields perpendicular to the layers. At 4.2 K an abrupt and hysteretic increase in resistance is observed at high current densities for one polarity of the current, comparable to the giant magnetoresistance (GMR) effect observed at low fields. A micromagnetic model, that includes a spin-transfer torque, suggests that the current induces a complete reversal of the thin Co layer to alignment antiparallel to the applied field-that is, to a state of maximum magnetic energy.

A study of spin relaxation on spin transfer switching of a noncollinear magnetic multilayer structure J. Appl. Phys.

Current-induced magnetic excitations in Cu/Co/Cu single layer nanopillars (∼50 nm in diameter) have been studied experimentally as a function of Co layer thickness at low temperatures for large applied fields perpendicular to the layers. For asymmetric junctions current induced excitations are observed at high current densities for only one polarity of the current and are absent at the same current densities in symmetric junctions. These observations confirm recent predictions of spintransfer torque induced spin wave excitations in single layer junctions with a strong asymmetry in the spin accumulation in the leads.

Numerical solutions of the micromagnetic Landau-Lifshitz-Gilbert equations provide valuable information at low temperatures (T), but produce egregious errors at higher T. For example, Curie temperatures are often overestimated by an order of magnitude. We show that these errors result from the use of block or coarse-grained variables, without a concomitant renormalization of the system parameters to account for the block size. Renormalization solves the problem of the Curie-point anomaly and improves the accuracy of more complicated micromagnetic simulations, even at low T.