Computer simulation study of rare events in magnetic nanosystems: error rates and energy barriers in STT-MRAM cells

点击翻译按钮获取中文摘要

The main goal of the project is the theoretical study using computer simulations (with experimental support) of rare events in physical systems with many metastable states, taking as an example an emerging generation of novel computer memory: spin-transfer-torque magnetic random access memory (STT-MRAM). According to this general task, partial goals of this proposal are: (1) To achieve thorough physical understanding of switching processes in STT-MRAM cells (using numerical studies and supporting experimental measurements) with the purpose to identify typical switching scenarios and to predict the dependence of the switching current on the MRAM cell parameters and on the environmental conditions.(2) To calculate Write and Read Error Rates of STT-MRAM cells using full-scale massively parallel micromagnetic simulations of MRAM cells at finite temperatures (and to verify these calculations by large-scale experiments on MRAM cell arrays) in a broad range of temperatures and times.(3) To predict the information life time in STT-MRAM cells using reliable numerical methods (to be developed in frames of this project) for the evaluation of energy barriers.In order to achieve these partial goals, the theory group of General Numerics Research Lab intends to develop:(i) new GPU-based numerical methods for massively parallel Langevin-dynamics modeling of switching processes in multilayer nanoelements, and (ii) advanced methods for the direct determination of energy barriers in systems with multiply metastable states, which would also account for the possibility of a transition between two given states via several intermediate local energy minimaExpected project results should greatly improve our understanding of the long-time magnetization dynamics in magnetic systems with many metastable states in general, and obtain deep physical insights into the switching processes in the new generation of magnetic memory (STT-MRAM) in particular.