Stochastic Preisach Modeling of Ferroelectric Switching Kinetics

Abstract

Although ferroelectric thin films are expected to enable nanointegrable nonvolatile memory devices, system-level evaluation in research firms remains challenging because academic experiments are typically conducted in macroscale test environments. Compact models can bridge this gap, yet existing approaches only partially satisfy practical requirements. This work proposes a stochastic Preisach model that combines the complementary strengths of the quasi-static Preisach and nucleation-limited-switching formalisms by employing a hysteresis-shape model and transient switching dynamics within a unified framework based on the hysteron concept. The model was experimentally validated using three different Hf0.5Zr0.5O2 capacitor stacks, including devices exhibiting large coercive voltages and shifted/asymmetric polarization−voltage loops. Finally, the model was embedded in a lumped one-transistor−one-capacitor ferroelectric random-access memory circuit, which demonstrated stable operation under bidirectional and sequential pulse trains. The simulation revealed a trade-off governed by serial bit-line capacitance, predicting an optimal capacitance ratio that maximizes the read sensing margin.