Investigation of forming process for metal oxide-based resistive switching memory by stochastic simulation

Forming process of metal oxide-based resistive switching memory (RRAM) is investigated by a self-consistent stochastic simulator based on the percolation theory and resistor network. The model capabilities allow reproducing Forming voltage relation with the thickness of oxide layer, statistical distribution of Forming voltage and the evolutions of oxygen vacancies´ distribution. Results indicate that the final filament prefers to locate in the district with low activation energy, thin thickness and high percentage of pre-exist oxygen vacancies, such as grain boundaries. The results can help to further understand the physical mechanism of resistive switching effect and optimize the performance of metal oxide-based RRAM.