A Novel Ferroelectric Tunnel FET-Based Synapse With Complementary Hebbian Plasticity Enhancing Learning of Spiking Neural Networks

In this work, for the first time, a novel ferroelectric (FE) tunnel FET (FeTFET) based synapse device with complementary Hebbian plasticity is proposed and experimentally demonstrated. The device leverages FE polarization switching dynamics to achieve spike-timing-dependent plasticity (STDP) and utilizes FE-modulated ambipolar band-to-band tunneling (BTBT) transports with opposite monotonicity for complementary synaptic weight updates. The fabricated FeTFET based on the 14nm FinFET platform enables configurable anti-STDP and STDP modes with single transistor of high CMOS compatibility, facilitating advanced supervised and reinforcement learning. Moreover, benefiting from the transition of BTBT windows, the FeTFET synapse exhibits historical activity-aware plasticity, enhancing the stability of the learning process and demonstrating its potential for highly integrated neuromorphic computing systems.