A new approach to micromagnetic simulation of thermal magnetic fluctuation noise in magnetoresistive read sensors

Thermal magnetic fluctuation noise forms the ultimate application limit of small magnetoresistive devices for magnetic data storage. The noise analysis of these devices becomes increasingly important for high-density recording. Such noise analyses by micromagnetic simulation are, however, computationally very intensive and require enormous amounts of simulation time. This paper presents a faster micromagnetic method to arrive at the noise and small signal dynamics of these sensors. It uses, for every cell in the simulation, a behavioral analog (i.e., an "analog computer" model) described by the same equations as the basic magnetic simulation cell (the Landau-Lifshitz-Gilbert equation, Slonczewski's spin torque addition and the equipartition principle). The sensor, as a network of exchange and demagnetization coupled cells, is then solved with a network simulator (such as PSpice). This process gives a drastic reduction in computing time and yet leads to high resolution spectra with very little residual uncertainty. The paper further presents a large number of simulation results for uniform sensors as well as for sensors with a nonuniform magnetic bias and a nonuniform electrical bias. It addresses the spatial distribution of the noise (standing spin waves in the noise) and the correlation of the noise in various parts of the sensors. Finally, as a further example of this method, the paper shows the effect of spin torque transfer on the noise and the small signal dynamics of a current perpendicular to plane giant magnetoresistive (CPP GMR) sensor.