Memory Characteristic of Organic/Inorganic Hybrid Synaptic Transistor with Silk Fibroin Gate Insulator

According to the development of the 4^thindustry such as Artificial Intelligence (AI) and Internet of Things (IoT), needs to process large amounts of data is exploded. The conventional Von Neumann computing system is facing challenges of a bottleneck between data processing and memory unit. To overcome this issue, a neuromorphic computing system inspired by the human brain using parallel processing has appeared. Therefore, neuromorphic devices have been introduced based on various materials. Also, flexible and biocompatible organic synapse devices is studied a lot for wearable artificial intelligence systems.<br/>In this study, we introduce organic/inorganic hybrid synaptic transistor with Silk fibroin from <i>Bombyx mori</i> cocoon as gate insulator. Silk fibroin is an electronic biomaterial that has light transmittance, good bio-compatibility and high mechanical flexibility. The dielectric constant (k) of silk fibroin is around 6, which is higher than SiO₂'s. The potentiation and depression characteristics of hybrid synaptic transistor were investigated based on various pulse scheme. Artificial neural network (ANN) simulation with experimental results of silk fibroin synaptic device shows high recognition accuracy (&gt;95%) for handwritten digit database in MNIST. It will be discussed working mechanism of synaptic transistor with silk fibroin and the possibility of the application of silk fibroin for synaptic devices.