Scholarlink

<p>Neuromorphic devices that emulate sensory processing and pattern recognition are increasingly important in advancing artificial vision systems. Wide-bandgap materials like ZnO, known for their ability to absorb UV light and generate detectable photocurrent under an external bias, offer promising capabilities for such devices. In this study, a bilayer structure of defect-induced and nearly stoichiometric ZnO was used, and the ability of the device to recognise images was investigated. The set and reset voltages of the bilayer device was found to be lower than that of the single layer device, indicating the easy formation and rupture of the conductive filament. This device showed UV light-induced synaptic plasticity, exhibiting a short-term to long-term memory transition by optical stimuli similar to the Atkinson-Shiffrin multistore model. The bilayer device exhibited good optical PPF and high photosensitivity of 3208% for a single pulse of 1&#xa0;s width. The neural network simulation using the characteristic curve of the device showed better learning accuracy of more than 95% with a matrix of less confusion. These measurements provide promising results for an artificial visual system, demonstrating the device’s efficiency in capturing and processing light information, potentially mimicking the capabilities of natural vision systems. We have further demonstrated a ZnO bilayer transparent flexible memristor using ITO as both the bottom and top electrodes on a PET substrate.</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

Ultraviolet pulse-driven neuromorphic device for pattern recognition

  • 期刊论文

  • 发表日期：

    2025年7月14日

查看全文
  • Anjala Jayaraj,
  • P. S. Subin,
  • G. Arun,
  • K. J. Saji,
  • Aldrin Antony

摘要

Neuromorphic devices that emulate sensory processing and pattern recognition are increasingly important in advancing artificial vision systems. Wide-bandgap materials like ZnO, known for their ability to absorb UV light and generate detectable photocurrent under an external bias, offer promising capabilities for such devices. In this study, a bilayer structure of defect-induced and nearly stoichiometric ZnO was used, and the ability of the device to recognise images was investigated. The set and reset voltages of the bilayer device was found to be lower than that of the single layer device, indicating the easy formation and rupture of the conductive filament. This device showed UV light-induced synaptic plasticity, exhibiting a short-term to long-term memory transition by optical stimuli similar to the Atkinson-Shiffrin multistore model. The bilayer device exhibited good optical PPF and high photosensitivity of 3208% for a single pulse of 1 s width. The neural network simulation using the characteristic curve of the device showed better learning accuracy of more than 95% with a matrix of less confusion. These measurements provide promising results for an artificial visual system, demonstrating the device’s efficiency in capturing and processing light information, potentially mimicking the capabilities of natural vision systems. We have further demonstrated a ZnO bilayer transparent flexible memristor using ITO as both the bottom and top electrodes on a PET substrate.