Self-Powered Ultraviolet Photonic Synapses for Neuromorphic Vision

Inspired by biological neuromorphic systems, optoelectronic synapses have garnered significant attention for neuromorphic visual information processing owing to their high propagation speed, broad bandwidth, and low power consumption. To address the complexity and high bias voltage of ultraviolet (UV) synapses based on traditional wide-bandgap semiconductors, we developed a solution-processed, self-powered UV photonic synaptic device based on a MoS₂ thin film decorated with CsPbBr₃ quantum dots (QDs). Benefiting from its favorable Type-II band alignment between MoS₂ and CsPbBr₃, the device achieves efficient carrier separation and sustained synaptic responses under the zero-bias condition. This photonic synapse not only emulates essential synaptic behaviors but also demonstrates advanced image preprocessing and face recognition functions. Integration with artificial neural networks further improves the recognition accuracy and accelerates training processes. These results highlight the potential of the CsPbBr₃ QDs/MoS₂ thin film for energy efficient UV neuromorphic vision systems.