光探测
材料科学
响应度
光电子学
光电探测器
异质结
神经形态工程学
钙钛矿(结构)
计算机科学
人工神经网络
人工智能
化学工程
工程类
作者
Zijin Zhao,Zijun Hu,Ming Deng,Enliu Hong,Peixi Wang,Ziqing Li,Xiaosheng Fang
标识
DOI:10.1002/adma.202416033
摘要
Abstract Optoelectronic devices with imaging and recognition capabilities are crucial for developing artificial visual system (AVS). Bias‐switchable photodetection and photosynaptic devices have been developed using 2D perovskite oxide/organic heterojunctions. This unique structure allows for modulated carrier dynamics under varied bias conditions, enabling the devices to function as photodetectors without bias and as photosynapses with bias. At zero bias, the device achieves high responsivity (≈0.36 A W −1 at 320 nm) and rapid response speed (0.57 s). Under a −0.5 V bias, it exhibits persistent photoconductivity (PPC), resulting in neuromorphic synaptic behaviors with a paired‐pulse facilitation (PPF) index exceeding 300%. Moreover, an 8 × 8 sensor array demonstrates image sensing and memory capabilities, showing in situ enhanced imaging when switching the bias from 0 to −0.5 V, and over 200 s of image memory. The image processing and recognition abilities are further explored by constructing an AVS using a 28 × 28 device array combined with an artificial neural network (ANN). The adjustable synaptic weight under different reverse biases allowed for optimized simulated recognition, achieving an accuracy of 92% after 160 training epochs. This work presents a novel method for creating dual‐functional photodetection and photosynaptic devices, paving the way for a more integrated and efficient AVS.
科研通智能强力驱动
Strongly Powered by AbleSci AI