An integrated multi-physics approach to the modeling of a phase-change memory device

This paper presents a novel approach for the modeling and the simulation of a phase-change memory cell. The proposed multi-physics model, which describes the electrical, the thermal, and the phase change behavior of the memory cell, can be implemented in a finite-element simulator, with no need for any external algorithm to compute the phase evolution. The goal of this work is to determine, from a macroscopic point of view, the morphological evolution of the phase distribution during the programming pulses. This is useful to investigate the programming conditions that lead, for example, to a series-type or to a parallel-type phase distribution. The proposed model has been validated by comparing the simulated and the experimental SET to RESET programming curve. Finally, the model allowed us to evaluate the different phase distributions obtained after the same programming pulse when starting from different initial RESET conditions.