Atomistic Modeling of Valence Change Memory Devices: What Can We Learn from Simulations?

Abstract Resistive switching devices based on the valence change effect have shown promise for applications in emerging in‐memory and neuromorphic computing architectures. To support the development of such valence change memory (VCM) cells and enhance their figures of merit, a deep understanding of their operational principle is essential. Atomistic simulations can provide important insight to reach these objectives. Indeed, advanced models have been used to link the atomic structure and functionality of VCMs, as well as to simulate the key processes occurring in them. In particular, their electro‐forming and cycling behavior can be revealed through accurate simulations, while optimized device stacks with improved performance have been suggested and demonstrated experimentally. In this review, we therefore discuss how different atomistic modeling techniques can be combined and leveraged to shed light on the operation of VCM cells. Specifically, we outline which properties can be extracted with each method and give examples to illustrate the strengths and weaknesses of the approaches considered.