摘要
基于可作为突触的非易失性忆阻器构建神经元,有望将构建突触和神经元统一到一种器件上,但基于非易失性忆阻器构建神经元的关键是需要完成发放后的置回,而目前缺乏针对忆阻置回方法的研究。针对具有良好忆阻特性的单晶LiNbO_(3)薄膜忆阻器,开展了置回特性研究。研究了等幅电压脉冲序列作用下的置回特性,发现其难以满足置回要求。在此基础上,研究了幅值渐变的电压脉冲序列的置回作用,重点研究了最大脉冲幅值和脉冲宽度的影响,优化了幅值渐变电压脉冲序列参数,厘清了置回作用与积分作用的平衡关系,采用最大脉冲幅值为7 V和脉冲宽度为40 ms的幅值渐变电压脉冲序列将单晶LiNbO_(3)薄膜忆阻器的电阻值置回至200 kΩ以下,并在20次积分-置回循环中具有稳定的阻值变化范围。
Constructing neurons using non-volatile memristors,which can also function as synapses,enables the unification of the hardware implementation of the synapses and neurons into one device.A key challenge in building neurons based on non-volatile memristors lies in achieving post-firing reset.However,the research on memristor reset methods is rarely reported.In this study,the reset characteristics of memristors based on single-crystal LiNbO_(3) thin films with excellent memristive properties were investigated.First,the reset characteristics under a sequence of the voltage pulses with the same amplitude were studied,and it was difficult to achieve the desired reset effect.Based on this,the reset process using a sequence of voltage pulses with gradually changing amplitudes was further explored.The effects of the maximum pulse amplitude and pulse width were specifically analyzed,and the parameters of the voltage pulse sequence with gradually changing amplitudes were optimized.This process clarified the balance between the reset effect and integration operation,and a reliable reset operation was achieved.The single-crystal LiNbO_(3) thin-film memristor was reset to below 200 kΩ using a voltage pulse sequence with a maximum pulse amplitude of 7 V and a pulse width of 40 ms.The device exhibited a stable resistance variation range over 20 reset-integration cycles.
作者
佟俊德
潘忻强
帅垚
罗文博
吴传贵
TONG Junde;PAN Xinqiang;SHUAI Yao;LUO Wenbo;WU Chuangui(School of Integrated Circuit Science and Engineering,University of Electronic Science and Technology of China,Chengdu 611731,China;State Key Laboratory of Electronic Thin Films and Integrated Device,Chengdu 611731,China)
出处
《电子元件与材料》
北大核心
2025年第5期524-530,共7页
Electronic Components And Materials
基金
国家自然科学基金(52073041)
四川省自然科学基金(2024NSFJQ0014)。
