Statistical approach to the RESET switching of the HfO2-based solid electrolyte memory

The RESET switching of an oxide-based bipolar solid electrolyte memory with the Cu/HfO₂/Pt structure is studied in this work. The parameters of the RESET point defined at the maximum of the RESET current evolution, RESET voltage (V_RESET) and RESET current (I_RESET), are analyzed with a statistical method. The experimental raw data show a U-shape relation between VRESET and the ON-state resistance (R_ON). After data correction by a series resistance (R_S), V_RESET is roughly constant and the RESET current (IRESET) is inversely proportional to R_ON. These behaviors are in agreement with the thermal dissolution model of RESET. The further statistical analyses in terms of Weibull distributions show that the R_ON distribution has a strong influence on the I_RESET and V_RESET distributions. By using a “resistance screening” method, the I_RESET and V_RESET distributions are found to be compatible with a Weibull model. The Weibull slopes of the I_RESET and V_RESET are independent on R_ON, indicating that the RESET point corresponds to the initial phase of CF dissolution, according to our cell-based RESET model for the unipolar VCM (valence change mechanism) device. On the other hand, the scale factor of the V_RESET distribution (V_RESET63%) is roughly constant, while the scale factor of the I_RESET (I_RESET63%) is inversely proportional to R_ON, which is consistent with the thermal dissolution model of RESET and our cell-based RESET model. The RESET switching of the studied oxide-based bipolar solid electrolyte memory is dominated by the thermal dissolution mechanism.