Four-Terminal Ferroelectric Schottky Barrier Field Effect Transistors as Artificial Synapses for Neuromorphic Applications

In this paper, artificial synapses based on four terminal ferroelectric Schottky barrier field effect transistors (FE-SBFETs) are experimentally demonstrated. The polarization switching dynamics gradually modulate the barriers, thus programming device conductance by applying negative or postive pulses to imitate excitation and inhibition behaviors of biological synapse. excitatory post-synaptic current can be modulated back-gate bias, enabling reconfiguration weight profile with high speed 20 ns low energy (< 1 fJ/spike) consumption. Besides, tunable long term potentiation depression show endurance very small cycle-to-cycle variations. Based good linearity, symmetricity large dynamic range synaptic updates, a recognition accuracy (92.6%) is achieved for handwritten digits multilayer perceptron neural networks. These findings demonstrate FE-SBFET has potential as an ideal component future intelligent neuromorphic network.