Dual‐Mode Optoelectronic Synapse Based on Ultrathin In x Ti y O Films for In‐Sensor Neuromorphic Computing and Physical Image Encoding

ABSTRACT In this work, we propose a dual‐mode optoelectronic synaptic device based on an ultrathin In x Ti y O channel and demonstrate conductance‐state‐based image encoding. The ultrathin In x Ti y O channel was fabricated via ALD super‐cycle engineering, enabling precise control of the channel characteristics for optimized synaptic operation. The device operates through an interfacial charge‐trapping mechanism and exhibits both electrical and optical synaptic behaviors. Notably, nearly linear conductance modulation was achieved through the accumulation of photogenerated carriers under optical stimulation, indicating its suitability for precise synaptic weight control in neuromorphic learning systems. Furthermore, multilevel conductance states were achieved under different electrical stimulation conditions and utilized to encode image information into current maps. Even when multiple patterns overlap, the encoded information can be effectively represented through the linear superposition of current values. These results demonstrate the potential of the device to integrate sensing, memory, and security functions within a single platform. Overall, the proposed In x Ti y O‐based optoelectronic synaptic device provides a promising platform for functionally integrated in‐sensor neuromorphic systems and next‐generation AI hardware.