Multiwavelength photo-responsive RRAM: a quantum dot-based approach for artificial vision

Artificial vision systems face efficient challenges due to separated sensors, memories, and processors, resulting in redundant data generation and high energy consumption. We present an innovative optical-controlled two-terminal RRAM device that mimics the human eye's efficiency through multi-level photo-responsive capabilities. In device structure, ZnO nanoparticles protect quantum dot efficiency during electrode deposition and help to achieve multiple conductance states using three different wavelengths of light. Electrical characterization reveals distinct photo-responsive memory behavior under multiple wavelengths. Under 405nm illumination (1mW), the device shows reproducible switching between conductance states with an ON/OFF ratio of ~10 at -0.5V, transitioning from ohmic to Schottky conduction in HRS. At 365nm, it demonstrates short-term memory effects and persistent photoconductive effects. The device exhibits paired-pulse facilitation with 10s intervals, effectively mimicking biological synaptic behavior.