Dynamical Model of Threshold Switching Process in OTS Selector

Ovonic threshold switch (OTS) selector is crucial for developing high-density nonvolatile memory (NVM) with crossbar arrays. Its operational speed has been of great concern. This work presents a multiphysics dynamic model of threshold switching in OTS that accurately reproduces the time-resolved electrical property changes during pulse operations. The model focuses on the coupling between electrothermal and structural dynamics at device level rather than the carrier motion at the atomic level. Simulation results show parameter-dependent characteristics consistent with actual devices under varying OTS structure parameters and temperatures. In addition, we simulate the intrinsic stochastic characteristics of OTS using the TS dynamics model by introducing a physical random term, which captures the randomness in delay times for device turn on and turn off. Results indicate that both turn-on delay time and turn-off holding time follow Weibull distributions similar to real devices. This work aids in optimizing the speed performance of OTS devices for NVM and neuromorphic applications.