中文版本

Opportunity  

The development of artificial intelligence (AI) and the Internet of Things (IoT) has intensified the demand for advanced sensory systems, particularly in artificial vision, autonomous driving, and robotics. However, existing artificial visual systems face significant challenges, including high power consumption, limited bandwidth, and difficulties in large-scale integration. Traditional systems rely on complex multilayer heterojunctions or unstable perovskite materials, which suffer from poor air stability and intricate fabrication processes. Additionally, most current systems cannot simultaneously achieve both positive and negative photoconductivity (PPC/NPC) across multiple wavelengths, limiting their ability to mimic the human visual system’s versatility. This patent addresses these gaps by proposing a scalable, low-power solution using organic molecule-nanowire heterojunctions to enable tunable photoconductivity and retina-like synaptic behaviors.  

Technology  

The patent introduces an innovative artificial photosynaptic device based on organic molecule-nanowire heterojunctions. The core innovation involves:  

1. Heterojunction Design: Two type-I heterojunctions are formed by wrapping p-type C8-BTBT or n-type PC₆₁BM organic thin films around InGaAs nanowire arrays. These configurations enable tunable photoconductivity—NPC (negative) for C8-BTBT/InGaAs and PPC (positive) for PC₆₁BM/InGaAs—due to carrier injection differences.  
2. Synaptic Mimicry: The devices replicate biological synaptic behaviors (e.g., excitatory/inhibitory postsynaptic currents, paired-pulse facilitation) under optical stimuli (UV to visible light). Persistent photoconductivity allows long-term memory (LTM) and multi-state storage.  
3. Scalability: The use of printed nanowire arrays and solution-processed organic films enables large-area integration, overcoming limitations of traditional 2D/perovskite materials.  
4. Hardware Kernel: A 4×4 device array demonstrates optical memory functions, while NPC/PPC devices form a "hardware kernel" for visual processing, achieving 100% recognition accuracy in neural networks.  
    

Advantages  

• Tunable Photoconductivity: Simultaneous NPC/PPC responses across solar-blind to visible wavelengths.  
• Low Power Consumption: Synaptic behaviors achieved with femtojoule-level energy.  
• Scalability: Printed nanowire arrays and solution-processed organic films enable large-area fabrication.  
• Stability: Superior air stability compared to perovskite-based devices.  
• Versatility: Mimics human visual functions (color recognition, memory, and adaptive learning).  
• High Recognition Accuracy: Hardware kernels boost neural network accuracy to 100%.  

Applications  

• Artificial Vision Systems: Autonomous vehicles, robotics, and augmented reality.  
• Neuromorphic Computing: Energy-efficient AI hardware for edge devices.  
• Optical Memory Devices: Multi-state storage with photoelectric modulation.  
• Biomedical Sensors: Retina prosthetics or low-power medical imaging.  
• IoT Sensory Nodes: Distributed vision networks with low energy demands.