等离激元调制的硒化钼超低能耗类视网膜光电器件

光电神经突触和视觉成像器件的研发备受关注，然而，生物兼容的ms级运算速度和pJ级功耗的类视网膜器件研发仍未取得突破，在材料体系和器件性能优化方面亟待寻求解决方案。目 前，基于二维半导体材料的类视网膜器件都受限于材料自身有限的光吸收和载流子迁移率，在性能和功耗改善方面仍需新的技术优化方案。本项目将以“金颗粒-二维半导体材料”为结构“基元 ”，利用模板压印和改进的化学气相沉积技术，实现三基色光响应的“序构”阵列；结合瞬态 /稳态光谱表征，揭示等离激元的局域电磁场增强、热电子掺杂和原子替位现象对器件性能提升的物理本质；利用光电测试平台系统研究神经突触器件的“学习-记忆-遗忘”等仿生功能， 并阐明等离激元效应对器件极限性能突破的影响；进一步优化软件算法，实现万级像素元视觉图像识别与神经网络记忆。本项目的开展，将促进等离激元效应提升光电器件性能的深入理 解，为智能视觉感知系统设计与应用提供重要依据。.

The development of optoelectronic synapses and visual imaging devices have attracted much attention. However, the research of biocompatible retinal-like devices with ms-level operation speed and pJ-level power consumption have not yet made a breakthrough. It is urgent to seek solutions in finding new material system and improving device performances. At present, retina-like devices based on two-dimensional semiconductors are limited by their weak light absorption and poor carrier mobility. New techniques are still needed to explore to improve the performance and power consumption. In this project, the material system of "Au nanoparticle-two dimensional semiconductor" is selected as the "unit" structure, and "ordered" devices are fabricated by using template imprinting method and improved chemical vapor deposition technology. Combined with transient/steady-state spectral characterizations, the physical essences of plasmon-induced local electromagnetic field enhancement, hot electron doping and atomic substitution phenomenon in improving the device performance should be revealed. The "learning-memorizing-forgetting" bionic functions of synaptic devices are studied using optoelectronic measurement platform, and the influence of plasmon-effect on the breakthrough of device performance will be clarified. We further optimize the software algorithm to realize the visual image recognition and neural network memory of 10000-level pixel. This project can promote the in-depth understanding of plasmon-effect in improving the performance of optoelectronic devices, and provides a valuable guidance for designing and fabricating intelligent visual perception systems.