神经形态工程学
材料科学
突触可塑性
神经促进
记忆电阻器
长时程增强
突触
兴奋性突触后电位
电阻随机存取存储器
计算机科学
可塑性
光电子学
非突触性可塑性
峰值时间相关塑性
突触后电位
变质塑性
神经科学
电压
记忆
感知
突触重量
纳米技术
突触后电流
促进
电子工程
人工神经网络
人工智能
作者
Jianping Lan,Zhanchuan Cai,Fengxia Yang,Yan Li
标识
DOI:10.1021/acs.jpclett.5c02625
摘要
Emerging near-infrared (NIR) optoelectronic synaptic devices, which enable parallel perception and memorization of information, play a pivotal role in developing high-efficiency neuromorphic computing systems with visual perception and complex learning capabilities. Here, CuSbS2 is used in an innovative method to achieve an artificial NIR optoelectronic synapse. This CuSbS2 memristor not only exhibits stable nonvolatile resistive switching behaviors, featuring low operating voltages (−0.69/0.68 V) with variations [ΔV(σ)/μ] both less than 13% and long retention time exceeding 104 s, but also realizes diverse synaptic plasticity triggered by both electric signals and NIR light, with the achievement of comprehensive synaptic functionalities, including excitatory postsynaptic current (EPSC), paired-pulse facilitation (PPF), short-term to long-term potentiation (STP/LTP), short-term and long-term memory (STM/LTM), as well as spike-timing-dependent plasticity (STDP). Moreover, utilizing the excellent optoelectric performance of CuSbS2, neuromorphic functions of designing a high-accuracy NIR imaging system are implemented, further verifying its practical application potential. The results confirm that CuSbS2 is a highly promising candidate material for constructing NIR optoelectronic artificial synapses for advanced synaptic applications, paving a solid way for the future development of neuromorphic systems integrating sensing, memory, and processing capabilities.
科研通智能强力驱动
Strongly Powered by AbleSci AI