Forming-Free Multilevel Resistive Switching in a ZnO@β-SiC Composite for Neuromorphic Computing

Recent attention has been focused on developing artificial synaptic devices for in-memory computing, with the aim of long-term device stability, low power consumption, and high performance. Oxides and perovskites have been explored in this context to address limitations of the Von Neumann architecture. However, these materials face individual constraints that hinder their full potential. This study introduces a unique approach using a ZnO@beta-SiC composite for low-power, high-performance, and forming-free bipolar resistive switching devices. Notably, these devices exhibit switching from high to low resistance states at a very low voltage of similar to 100 mV with a fast response times of similar to 40 ns and 50 ns for positive and negative pulses, respectively, and consume a very low power of similar to 100 mu W. Chemical and microstructure analyses reveal Zn2SiO4 nanocrystals embedded in an amorphous layer, and it is found to be suitable for enhancing device stability over 104 cycles with similar to 104 s retention. The phenomenon is explained by the formation and dissolution of oxygen vacancy and metal cation-driven conductive filaments. Moreover, the devices effectively replicate versatile synaptic functions such as excitatory postsynaptic current, pair pulse facilitation, potentiation/depression, long-term memory/short-term memory, and learning/forgetting behavior. This work thus presents a promising avenue for the sustainable development of artificial intelligence through in-memory neuromorphic computing.