Reliable resistive switching and synaptic plasticity in Ar+-irradiated single-crystalline LiNbO3 memristor
Jiejun Wang,Xinqiang Pan,Qiang Wang,Wenbo Luo,Yao Shuai,Qin Xie,Huizhong Zeng,Gang Niu,Chuangui Wu,Wanli Zhang
DOI: https://doi.org/10.1016/j.apsusc.2022.153653
IF: 6.7
2022-09-15
Applied Surface Science
Abstract:Typical filament-type memristors suffer from temporal and spatial variations in the resistive switching due to stochastic filament formation, which hinders the implementation of memristive synapses in neuromorphic computing. In this work, the memristor based on Ar+-irradiated single-crystalline LiNbO3 (SC-LNO) thin film is reported. The high-quality SC-LNO thin films contribute to the formation of more concentrated and robust oxygen vacancy channels (OVCs) and lead to reliable self-rectifying resistive switching behaviors with an ultra-low device-to-device variability of ∼2.80% and cycle-to-cycle variability of ∼2.25%, as well as long retention time at multilevel resistance states (>3 × 104 s), stable endurance performance (>105 cycles), excellent environmental stability (>6 months), and good analog switching linearity. Furthermore, abundant synaptic plasticity characteristics including paired-pulse facilitation (PPF), post-tetanic potentiation (PTP), long-term potentiation (LTP), long-term depression (LTD), spike-timing-dependent plasticity (STDP), and associative learning, are also successfully emulated using this device as an electronic synapse. The spontaneous decay process right after applied voltage pulses and the long-term nonvolatility characteristics in Ar+-irradiated SC-LNO memristor indicate the existence of residual OVCs, which serves to guide the reconstruction of OVCs during repeated switching and to further enhances the uniformity of device switching characteristics.
chemistry, physical,physics, applied, condensed matter,materials science, coatings & films