神经形态工程学
材料科学
异质结
光电子学
计算机科学
突触
突触可塑性
纳米技术
神经科学
人工神经网络
化学
人工智能
生物化学
受体
生物
作者
Xiaorui Han,Youhan Mo,Yanbo Wang,Xiaoqin Yang,Jianyuan Wang,Bingcheng Luo
摘要
Neuromorphic computing can transcend the existing performance-limiting conventional computing architectures (i.e., von Neumann paradigm) by designing biologically inspired devices with intelligence capabilities. Optically stimulated synaptic devices have recently emerged as an alluring direction in neuromorphic computing due to their combined sensing, memory, and computing functions. This study introduces a new two-terminal optically stimulated synaptic device based on an amorphous Ga2O3/GaN heterostructure. This III-nitride-compatible device exhibits persistent photoconductivity at broad ultraviolet wavelengths, enabling it to mimic essential biological synapse functions, including pair-pulse facilitation, short-term plasticity (STP), long-term plasticity (LTP), and STP-to-LTP transition. Furthermore, memory and learning capabilities are achieved in this device, with an ultralow energy consumption of 0.05 pJ per synaptic event, comparable to the human brain. This work presents a CMOS-compatible nitride synapse, suggesting potential applications in low-power computing.
科研通智能强力驱动
Strongly Powered by AbleSci AI