Temperature-dependent resistive switching statistics and mechanisms in nanoscale graphene–SiO2–graphene memristors

Abstract

The development of memristors presents a transformative opportunity to revolutionize electronic devices and computing systems by enabling non-volatile memory and neuromorphic computing. Silicon oxide memristors are particularly promising due to their potential for low cost, high integration and compatibility with existing manufacturing processes. In this study, we statistically investigate the switching mechanisms of a nanoscale (sub-2 nm) silicon oxide memristor at different temperatures. As a unipolar memristor, the average set voltage (switching from a high resistive state to a low resistive state) rises with a temperature drop, while the average reset voltage (switching from a low restive state to a high state) drops slightly with the temperature drop. Standard deviation of these values increases as temperature decreases. These behaviors are analyzed based on the Weibull distribution. Statistical results suggest that the set process involves the formation of Si conducting filaments promoted by the diffusion of oxygen ions from oxygen vacancies, while the reset process involves Joule heat-driven conductive filament rupture and silicon–oxygen recombination, requiring intensified heating at higher environmental temperatures to counteract extended oxygen ion migration. Beyond general resistive switching mechanisms involving only the formation and rupture of Si conductive filaments, our insights provide a novel understanding of the stochastic mechanisms of the switching process at the atomic level, with significant implications for future neuromorphic computing applications.