Neuromorphic synaptic applications of oxygen-deficient BaSnO3 thin films

Abstract

Cubic barium stannate (BaSnO3) has emerged as a promising platform for optoelectronic devices due to its remarkable room-temperature electron mobility, high transparency, excellent thermal stability, and flexible doping control. In this work, a photonic synaptic device—an image sensor integrating memory and processing—was proposed based on the persistent photoconductivity of oxygen-deficient BaSnO3 thin films. The device demonstrated a light-tunable response, allowing it to replicate key functions of biological synapses. Importantly, this device can be utilized as a reservoir layer in a reservoir computing system, achieving a recognition accuracy of over 90% when identifying handwritten digit images. This work underscores the potential of BaSnO3 for retinomorphic computing applications.