TCAD Simulation of Resistive Switching Devices: Impact of ReRAM Configuration on Neuromorphic Computing.

This paper presents a method for modeling ReRAM in TCAD and validating its accuracy for neuromorphic systems. The data obtained from TCAD are used to analyze the accuracy of the neuromorphic system. The switching behaviors of ReRAM are implemented using the kinetic Monte Carlo (KMC) approach. Realistic ReRAM characteristics are obtained through the use of the trap-assisted tunneling (TAT) model and thermal equations. HfO-AlO-based ReRAM offers improved switching behaviors compared to HfO-based ReRAM. The variation in conductance depends on the structure of the ReRAM. The conductance extracted from TCAD is validated in the neuromorphic system using the MNIST (Modified National Institute of Standards and Technology) dataset.