The study investigates the use of LiNbO, a perovskite ferroelectric material, in creating a new type of thin-film transistor (FeTFT) that can mimic synaptic behavior in neuromorphic devices.
The FeTFT is constructed with layers of LiNbO and LiAlO, enhancing its performance by reducing depolarization fields and leakage current, leading to impressive device metrics such as high mobility and good memory retention.
The FeTFT not only demonstrates effective emulation of synaptic plasticity but also achieves successful results in artificial neural network training while maintaining low energy consumption.