Return
Ultraviolet pulse-driven neuromorphic device for pattern recognition
DOI:10.1007/s10854-025-15318-5.png)
Abstract
En 中文
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.
Journal
J
IF:
2.8
Papers:
2.4K
Citations:
5.2W
Organization
Cited Papers
Redox-Based Resistive Switching Memories - Nanoionic Mechanisms, Prospects, and Challenges
ADVANCED MATERIALS
IF26.8
Memristor-based neural networks: a bridge from device to artificial intelligence
NANOSCALE HORIZONS
IF6.6

