Time series modeling of the cycle-to-cycle variability in h-BN based memristors

Abstract

We have characterized and modeled memristor devices based on the Au/Ti/multilayer h-BN/Au/Ti stack. Resistive switching (RS) operation has been analysed by extracting the reset and set voltages and currents. The evolution of the set and reset parameters along a RS series was mathematically modeled in a cycle-to-cycle (CTC) basis by means of the Time Series Analysis (TSA). To do so, the Autocorrelation Functions (ACF) and the Partial Autocorrelation Functions (PACF) have been calculated. These tools help to perform a comprehensive variability study and to obtain the corresponding analytical models within the TSA context. Finally, we have included this modeling procedure in a complete compact model such as the Stanford to be able to account for this variability at the circuit level. Experimental current versus voltage (I-V) curves have been correctly fitted with the model.