电导
记忆电阻器
锡
材料科学
航程(航空)
MNIST数据库
电阻率和电导率
光电子学
导电体
纳米技术
电气工程
凝聚态物理
人工神经网络
复合材料
计算机科学
物理
冶金
人工智能
工程类
作者
Yujian Zhang,Kailiang Zhang,Xuanyu Zhao,Zheng Sun,Qiaozhen Zhou,Xin Shan,Xin Lin,Kunming Liu,Zexia Ma,Ke Shan,Fang Wang
标识
DOI:10.1002/pssr.202200199
摘要
Enhancement of the conductance range of memristors used in synaptic devices is essential for achieving high‐performance neural networks. Herein, a memristor based on the stack structure of TiN/AlO x /AlO y /ITO is designed to enhance the conductance range. The AlO x /AlO y devices exhibit pseudointerface switching characteristics with higher switching ratios and reliability under a compliance current of 1 mA. The high‐resistance state/low‐resistance state ratio of the AlO x /AlO y devices increases from 11.4 to 128.6 compared with the AlO x devices. Accompanied by high‐conductance linearity, the conductance range increases from 36–202 to 25–280 μS simultaneously. Based on the related electrical properties and microstructure analysis, the regulation mechanism of the formation and rupture of conductive filaments by the oxygen concentration gradient is demonstrated. Simulations using the Modified National Institute of Standards and Technology (MNIST) handwritten recognition data set prove that the AlO x /AlO y memristor can operate with a learning accuracy of 91.07%.
科研通智能强力驱动
Strongly Powered by AbleSci AI