Tailoring Wavelength-Adaptive Visual Neuroplasticity Transitions of Synaptic Transistors Comprising Rod-Coil Block Copolymers for Dual-Mode Photoswitchable Learning/Forgetting Neural Functions.

Abstract

The vision-inspired artificial neural network based on optical synapses has drawn a tremendous amount of attention for emulating biological senses. Although photoexcitation-induced synaptic functionalities have been widely studied, optical habituation via the photoinhibitory pathway is yet to be demonstrated for sophisticated biomimetic visual adaptive systems. Here, the first optical neuromorphic block copolymer (BCP) phototransistor is demonstrated as an all-optical operation responding to various wavelengths, fulfilling photoassisted dynamic learning/forgetting cycles via optical potentiation without gate bias. The polyfluorene BCPs were precisely designed to enable wavelength-adaptive responses, benefiting from interfacial semiconductor/electret morphology and the crystallinity/electron affinity of the BCPs. Notably, this is the first work to simultaneously exhibit fully light-controlled short- and long-term memory based on organic material systems. The device presents a high current contrast above 100-fold and long-term retention over 104 s. As a proof-of-concept for neural networks, a 6 × 6 array of photosynapses performed outstanding visual pattern learning/forgetting with high accuracy. This study exploits the design strategy of a conjugated BCP electret to unleash the full potential of wavelength-adaptive visual neuroplasticity transitions. It provides an effective architecture for designing high-performance and high-storage capacity required applications in next-generation neuromorphic systems.