Donor‐redox covalent organic framework‐based memristors for visual neuromorphic system
Qiongshan Zhang, Qiang Che, Fuzhen Xuan, Bin Zhang
Abstract
Abstract Artificial visual neural systems have emerged as promising candidates for overcoming the von Neumann bottleneck via integrating image perception, storage, and computation. Existing photoelectric memristors are limited by the need for specific wavelengths or long input times to maintain stable behavior. Here, we introduce a benzothiophene‐modified covalent organic framework, enhancing the photoelectric response of methyl trinuclear copper for low‐voltage (0.2 V) redox processes. The material enables the modulation of 50 conductive states via light and electrical signals, improving recognition accuracy in low light, dense fog, and high‐frequency motion. The ITO/BTT‐Cu 3 /ITO device's accuracy increases from 7.1% with 2 states to 87.1% after training. This construction strategy and the synergistic effect of photoelectric interactions offer a new pathway for the development of photoelectric neuromorphic computing elements capable of processing environmental information in situ. image