バックボウルベースの有機単結晶フォトンシナプス：凹型アーキテクチャが画像認識の高精度を誘発

Organic photosynapses are crucial for mimicking biological visual perception, advancing neuromorphic computing, and enabling wearable and biomedical applications. While device engineering has been the primary strategy for performance optimization, the relationship between molecular structure and synaptic behavior remains poorly understood. Here, we report a novel organic single-crystal photosynapse transistors (OSPSTs) based on hetero-buckybowl 3S-2Me, which exploits a persistent photoconductivity (PPC) effect induced by defect energy levels from ambient oxygen (O₂) adsorption. The bowl-shaped structure of 3S-2Me endows it with a remarkable advantage in O₂ adsorption, significantly amplifying PPC. Consequently, our devices exhibit long-term non-volatile current retention exceeding 4000s, typical synaptic plasticity, and linearly programmable multi-level conductance states. Notably, they achieve a high classification accuracy of 94.8% for 28 × 28 digit recognition, highlighting their potential in neuromorphic computing. This work provides a molecular engineering strategy for enhancing OSPST performance and offers insights into the structure-property relationship of organic photosynapses.