Percolation theory based statistical resistance model for resistive random access memory

Abstract

A comprehensive understanding of the disorder-induced transport characteristics in resistive random-access memory (RRAM) is critical for its thermal stability analysis and analog switching for the coming neuromorphic computing application. Superior to the previous transport mechanisms which are only valid within their respective ranges of temperatures, we propose a unified physics-based model that can accurately predict the transport dependence on all temperature ranges up to 300 K. By utilizing percolation theory and the Fermi Golden Rule, the probability distributions for both the tunnel junction energy barrier and gap distance based statistical resistance model are described. It is found that different programming cycles and resistance states contribute to transition behavior between various low-temperature transport mechanisms. Moreover, the model can also investigate the dependence of electrical characteristics on defect generation like radiation damage. Therefore, it quantitatively relates the thermal stability and percolation effects to the structural disorders in RRAM. The good agreement between the simulation and experimental results indicates that our physics-based model can provide an accurate prediction of temperature and disorder dependent effects in RRAMs.