Mechanism discrimination in multilevel RRAM using state-resolved impedance evolution

Resistive random access memory (RRAM) requires reliable mechanism discrimination in multilevel operation because distinct switching pathways produce different state evolution and frequency responses. In this study, state-resolved impedance evolution is employed to analyze and distinguish filamentary-type and interfacial-type RRAM devices across multiple programmed resistance states. By tracking impedance spectra at each state, the evolution of resistive and capacitive contributions is systematically examined and linked to the underlying switching mechanism. Nyquist and Bode analyses reveal clearly different state-dependent responses for filamentary and interfacial switching. To support this interpretation, equivalent circuit models were constructed and implemented in SPICE. The simulations reproduce the measured impedance trends across different resistance states, and the extracted parameter evolution remains consistent with the underlying physical mechanisms. This combined experimental and modeling approach establishes state-resolved impedance evolution as a practical framework for mechanism analysis and discrimination in multilevel RRAM.