Ar+ ions irradiation induced memristive behavior and neuromorphic computing in monolithic LiNbO3 thin films

Abstract Recently, memristors have attracted considerable attention because of their potential applications in artificial neural networks which will promote the future development of artificial intelligence. In this work, the analogue memristive and related synaptic behavior of memristors based on single crystalline LiNbO3 thin films have been studied. Low energy Ar+ ions irradiation was applied to locally dope the LiNbO3 thin films by controllably introducing oxygen vacancies acting as donors. The resistive switching performance and synaptic plasticity can be tuned by changing the size or the number of the irradiated regions below the top electrode. Linear regression, an important fundamental function belonging to the machine learning in artificial intelligence, was emulated using memristors with different synaptic plasticity. It has been shown that the local doping method significantly influences the linear regression process.