Conventional semiconductor devices are developed based on von Neumann computing system. However, due to the inherent problems with von Neumann's architecture, Moore's Law reached its limitation. On the other hand, human brain has huge advantages in handling complex problems. For example, our brain has good energy efficiency compared to von Neumann’s architecture and it can handle logic and memory in parallel. Because of these reasons, brain-like system called ‘Neuromorphic System’ is considered to be a new way to create next generation computing system. To implement a neuromorphic system, artificial neuron and synaptic devices which mimic brain are needed. Our group conduct researches about these artificial devices, especially on Synaptic Thin Film Transistors (Synaptic TFTs). Consider a Synaptic TFT with gate as a pre-synapse and source, drain, active layer as post-synapse and consider the conductance of the channel as synaptic weight. Synaptic TFTs operate in such a way that when pulse voltage is applied to the electrode, the conductance of the channel changes. In this way, TFT mimics the synaptic plasticities that occur in the real human brain, which become a mechanism of both memory and logic. Our team has advanced techniques in designing and fabricating TFTs with Oxide semiconductor. We have applied these advanced techniques to the Display field and now we are trying to combine them with semiconductor field to implement low energy consuming neuromorphic device with various operating mechanisms.

[1] S. -I. Cho et al. "Synaptic transistors with human brain-like fJ energy consumption via double oxide semiconductors engineering for neuromorphic electronics", Journal of Materials Chemistry C, 2021