A data retention model for phase-change memory by the Monte Carlo approach

Data retention is an important issue in the phase-change memories (PCMs) development, which is induced by the spontaneous thermal-driven crystallization for amorphous phase of chalcogenide material in PCMs. In this work, a data retention model for PCMs induced by spontaneous crystallization is presented. In this model, the spontaneous crystallization mechanism is modeled by nucleation and growth (N/G) theory, and Monte-Carlo approach is introduced to model the N/G process due to the stochastic nature of nucleation. According to this model, the resistance evolution with time is calculated. The model is calibrated by the experimental data, and results fit experimental data well. By defining the failure resistance, the statistical distributions of retention failure times are calculated, and the retention characteristics of PCMs are predicted.