A Ta2o5/Zno Synaptic Se-Fet For Supervised Learning In A Crossbar

The asynchronous supervised-learning ability of a Ta2O5/ZnO synaptic thin film transistor, capable of write operation in the off-state, is demonstrated. The Poisson drift-diffusion model of the device is enhanced to reproduce the Faradaic type charge storage mechanism in its gate current characteristics. Training as well as implementation of OR and AND operations in a 2x2 crossbar array are simulated using the device measured conductance characteristics.