Analysis of switching times statistical distributions for perpendicular magnetic memories

Abstract

Abstract Magnetization switching by injected spin-polarized current or applied field pulses is considered for perpendicular magnetic nanosystems with uniaxial symmetry and subject to thermal fluctuations. Such systems can model both perpendicular spin-transfer-torque magnetic random access memories and magnetic grains for state-of-the-art perpendicular recording. It is shown that, in the limit of high energy barriers and large current/field amplitudes, the probability and cumulative distribution functions of switching times can be analytically computed as function of material, geometrical parameters and thermal noise level. The theoretical predictions are confirmed by macrospin and full micromagnetic simulations. By using the derived formulas, it is shown that analysis of Write-Error Rates (WERs) at typical levels of reliability required by magnetic recording technology can be easily performed.