神经形态工程学
记忆电阻器
电导
人工神经网络
计算机科学
冯·诺依曼建筑
水准点(测量)
突触重量
材料科学
电子工程
实现(概率)
人工智能
匹配(统计)
突触
突触可塑性
工作(物理)
纳米技术
调制(音乐)
能量(信号处理)
可塑性
物理神经网络
神经科学
突触标度
物理
光电子学
生物系统
脉搏(音乐)
作者
Li Shi,Yun-Tao Ding,X D Liu,Shun‐Da Wu,Wenjing Sun,Bing Sun,Yi‐Sheng Wei,Zheng Zhang,Liu S,Lichuan Chen,Dong Xiang,Yamin Zhang,Zitong Liu,Hao‐Li Zhang
摘要
Neuromorphic computing, particularly memristor-based architectures, offers a promising route to overcome the von Neumann bottleneck. Single-molecule devices, with their high integration density and low energy consumption, represent an emerging platform for next-generation computing. Here, we report the first optoelectronic volatile single-molecule memristor based on the organic photovoltaic material Y6. The device exhibits reproducible conductance switching driven by electric-field-induced structural relaxation, enabling gradual and linear conductance modulation that mimics synaptic behavior. Under red-light illumination, the Y6 junction shows a remarkable 457% increase in conductance and a significantly reduced switching threshold, demonstrating strong photoresponsivity and low-power operation. Furthermore, frequency-dependent pulse tests clearly reproduce short-term synaptic plasticity (STP), demonstrating the memristor's ability to emulate dynamic synaptic functions. When the experimentally measured current response of the Y6 single-molecule memristor is incorporated into an artificial neural network (ANN) model, the system achieves a speech-recognition accuracy of 71.50%, closely matching that of the benchmark ANN (74.90%). This work pioneers the realization of synaptic functionality in a single-molecule memristor and validates its application within an artificial neural network. It provides a new strategy for developing highly integrated molecular-scale neuromorphic computing devices.
科研通智能强力驱动
Strongly Powered by AbleSci AI