Abstract
Resistive-switching memory (RSM) is one of the promising candidates for nextgeneration edge-computing devices due to its excellent device properties. Currently, several experimental and modeling studies have been reported to understand the conduction behaviors. However, a complete physical model used for describing the conduction behavior is still missing. Here, we propose an accurate conduction model which could not only fully accounts for the conduction processes of RSM devices in a simple framework, but also provides critical insight for devices of continued design, optimization, and application. A single physical model that is able to describe both the conduction and switching behaviors is achieved. The proposed model reveals the role of temperature, mobility of electrons and the depth of traps, and provides accurate prediction of the various set and reset processes using standard device parameters. In addition, the reproducibility of resistive random-access memory (RSM) devices based on cycle-to-cycle and device-to-device consistency are discussed by experiments and simulation of the devices sampled with different compliance currents. The finite element modeling was applied to reveal the change of internal temperature and to explain potential hard breakdown under normal operation conditions. The degradation of the RSM devices is also investigated by analysis of mean time to failure (MTF). Furthermore, neuron spiking simulation via heat signals in nano-size devices was investigated. These considerations of mechanisms, reproducibility and degradation of resistive switching and neuromorphic computing in non-volatile memory could JV