Analytic model of endurance degradation and its practical applications for operation scheme optimization in metal oxide based RRAM

An analytic model for the endurance degradation of metal oxide based RRAM is presented for the first time. The endurance degradation behaviors under various operation modes can be predicted by the model, which were verified by the measured data in different devices. Furthermore, the 10⁶ endurance for all 4-level resistance states in HfO_X-based RRAM is demonstrated by using the proposed optimization operation scheme for multi-level data storage based on the model prediction. Guided by the model, a dynamic self-recovery operation scheme is developed to achieve more than 2 orders endurance enhancement at a high switching speed (10 ns).