CMOS-Compatible Ferroelectric Artificial Dendrite for Spatiotemporal Integration

Abstract In biological neural systems, dendrites play an essential role in spatiotemporal information processing, which enable the brain to manage complex tasks with minimal energy. In this study, for the first time, we propose a CMOS-compatible dual-gate ferroelectric thin film transistor (FeTFT) for emulating spatiotemporal information processing in biological dendrites. Owing to the utilization of CMOS-compatible materials and fabrication processes, the FeTFT exhibits exceptional retention characteristics (>10 6 s) and endurance characteristics (>10 12 ). The following experimental results show that the dendritic spatial and temporal integration can be replicated in FeTFT device with varying pulse sequences. These superior characteristics, combined with its unique spatiotemporal processing capability, position this novel device as a highly promising candidate for advanced large-scale neuromorphic computing systems, particularly in applications requiring complex signal integration and parallel processing.