共 50 条
Resistive switching performance and synaptic behavior of La-doped HfO2 thin film
被引:4
|作者:
Zhang, Wu-Wen-Bo
[1
]
Jiang, Yan-Ping
[1
]
Tang, Xin-Gui
[1
]
Liu, Qiu-Xiang
[1
]
Tang, Zhenhua
[1
]
Zhou, Yi-Chun
[2
,3
]
机构:
[1] Guangdong Univ Technol, Guangzhou Higher Educ Mega Ctr, Sch Phys & Optoelect Engn, Guangzhou 510006, Peoples R China
[2] Xidian Univ, Sch Adv Mat & Nanotechnol, Xian 710126, Peoples R China
[3] Xidian Univ, Acad Adv Interdisciplinary Res, Frontier Res Ctr Thin Films & Coatings Device Appl, Xian 710126, Peoples R China
来源:
基金:
中国国家自然科学基金;
关键词:
Hafnium dioxide;
Lanthanum;
Doping;
Thin film;
Resistive switching;
Memristor;
Synapse behavior;
Sol-gel deposition;
MEMRISTOR;
RRAM;
DEVICE;
D O I:
10.1016/j.tsf.2023.139842
中图分类号:
T [工业技术];
学科分类号:
08 ;
摘要:
Resistive random-access memory has great application prospects in developing nonvolatile memory and artificial synapse devices. In this work, lanthanum (La) doped hafnium dioxide (HfO2) (La: HfO2) thin film is fabricated by sol-gel deposition. Using gold (Au) and lanthanum nickelate (LaNiO3) as electrodes, the resistive switching performance and synaptic behavior of memristor based on La: HfO2 thin films with different doping concen-trations as resistive switching layers are investigated. Current-voltage characteristics show that La-doped devices all exhibit bipolar resistive switching performance. When La doping concentration is 10%, the device shows reliable switching ratio (similar to 10(3)) under 110 consecutive cycles. And the conduction mechanism of the device in high resistance state and low resistance state is analyzed by ohmic conduction and space charge limited con-duction. In addition, the analog resistive switching can be obtained by increasing the voltage scanning and the result reveals that the device can be used to mimic the potentiation and depression behavior of biological synapses. Au/10% La: HfO2/LaNiO3/Si memristor effectively simulates the synaptic properties, such as long-term potentiation/depression, short-term potentiation/depression, paired-pulse facilitation and spike-time-dependent-plasticity learning rule. Furthermore, the information is classified and processed in an 8x8 pixel array. These findings will open up a unique way for the future development of nonvolatile memory and neural morphology computation.
引用
收藏
页数:9
相关论文