Sericin-Au nanoparticles composite-based artificial synapse for brain-inspired neuromorphic computing

Abstract Developing appropriate materials for memristor-based artificial synapses is necessary for advancing brain-inspired neuromorphic computing. Using biomaterials as the resistive switching layer offers promise due to their biocompatibility and biodegradability. This study presents a memristor fabricated using a sericin-Au nanoparticles (NPs) composite. While pure sericin-based devices exhibit digital-type switching, the incorporation of Au NPs enables analog-type resistive switching by controlling the compliance current. This behavior effectively emulates various synaptic functions. Additionally, a simulated artificial neural network utilizing the bio-memristor achieved a learning accuracy of 92.8% in handwritten digit recognition. These findings highlight the potential of the sericin-Au NPs composite for neuromorphic applications.