A Bioinspired Low-Power Optoelectronic Synaptic Transistor for Artificial Visual Recognition and Multilevel Optical Storage

Inspired by the biological neuromorphic system with efficient information perception and processing capabilities, transistor-based synaptic devices are expected to address the von Neumann limitation and promote the evolution of a neuromorphic computing paradigm. However, achieving a wide spectrum light response and low power consumption using a simple process remains challenging for artificial visual perception electronics. Here, an optoelectronic synaptic transistor (OST) capable of both a wide light wavelength response and multilevel optical storage is verified to emulate the functions of the avian optic nerve, using a low-temperature, solution-driven organic semiconductor and a biodegradable PVA electret. The organic OST successfully exhibits diverse synaptic characteristics such as postsynaptic current, short-term plasticity, neural facilitation, long-term potentiation, and depression (LTP/D). These behaviors enable the OST as a photoreceptor to imitate the avian retina, respond to UV and RGB light, and concurrently simulate avian synapses in the visual cortex of the brain with learning-experience behavior. The OST can individually respond to gate pulse stimulation without light illumination and mimic the function of bird feathers. MNIST image recognition was performed by constructing an artificial neural network (ANN), and the recognition rate is 90.8% in the UV range, indicating the robust visual perception ability. Furthermore, Ebbinghaus’s biological memory behaviors were simulated based on four types of light wavelength. More importantly, the multilevel optical storage with 300 conductance states was confirmed using a relatively weak light intensity of 1 μW/cm 2, and the OST device shows a low power consumption of 137 pJ/pulse. Therefore, the proposed OST highlights the huge potential to design an intelligent highly efficient machine vision device. Our work should provide opportunities for developing a future neuromorphic system.