Modeling and device parameter design to improve reset time in binary-oxide memristors

A simulation-based analysis is conducted to study the set and reset times of TiO₂-based memristor device. This analysis uses nonlinear device model that captures the effects of large electric field inside memristor devices. Previous studies report strong asymmetry between reset and set times with reset time being several orders of magnitude higher than set time. The aim of this work was to investigate the effect of the device length and oxygen vacancies profile on the switching time. Our results show that a device model with a length of 10 nm and accurate parameters can result in more realistic device characteristics. Also, oxygen vacancies profile can be tuned to improve the reset time. MATLAB is used to simulate a 10-nm device, and the initial vacancy profile is chosen to look like inverted parabola. The results show reduced ratio between reset and set times from several orders of magnitude presented in the literature to not more than 1.5×. Thus, the proposed oxygen vacancies profile and quantitative results derived from it can be used to suggest a physical memristor device. In addition, the tuned parameters and model that matches actual device behavior can be factored in memristor fabrication.