<p>Owing to the inherent discrepancy between the transient photoresponse speeds of photodetectors and artificial synapses, the integration of these elements into a single monolithic device poses notable challenges. Herein, we propose a bifunctional monolithic transparent device based on lifted-off (In,Ga)N nanowires. This device not only shows omnidirectional self-driven photodetection but also effectively functions as an artificial synapse in neuromorphic computing. The proposed device exhibits persistent photoconductivity under bias voltage and fast photoresponse without bias voltage, corresponding to the operational modes of artificial synapses and self-driven photodetectors, respectively. As a photodetector, the device exhibits omnidirectional detection, particularly owing to its transparency. As an artificial synapse, the device realizes various synaptic behaviors with ultrahigh paired-pulse-facilitation index and low energy consumption. Furthermore, an artificial neural network assembled using the proposed device exhibits a pattern recognition accuracy of 92%. The findings of the present study contribute to the advancement of omnidirectional self-driven detection and efficient neuromorphic computing.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Bifunctional monolithic transparent device for both neuromorphic computing and omnidirectional self-driven photodetection

  • Yanyan Chang,
  • Min Jiang,
  • Lian Ji,
  • Yukun Zhao

摘要

Owing to the inherent discrepancy between the transient photoresponse speeds of photodetectors and artificial synapses, the integration of these elements into a single monolithic device poses notable challenges. Herein, we propose a bifunctional monolithic transparent device based on lifted-off (In,Ga)N nanowires. This device not only shows omnidirectional self-driven photodetection but also effectively functions as an artificial synapse in neuromorphic computing. The proposed device exhibits persistent photoconductivity under bias voltage and fast photoresponse without bias voltage, corresponding to the operational modes of artificial synapses and self-driven photodetectors, respectively. As a photodetector, the device exhibits omnidirectional detection, particularly owing to its transparency. As an artificial synapse, the device realizes various synaptic behaviors with ultrahigh paired-pulse-facilitation index and low energy consumption. Furthermore, an artificial neural network assembled using the proposed device exhibits a pattern recognition accuracy of 92%. The findings of the present study contribute to the advancement of omnidirectional self-driven detection and efficient neuromorphic computing.