The Amorphous Oxide Semiconductor Thin Film Transistors a-GIZO TFT Device Simulation, Neural Modeling and Validation of Model

An a-IGZO TFT is analyzed in this paper. The Universal Organic TFT (UOTFT) equivalent circuit model is used to model the a-IGZO TFT equivalent circuit. Physical simulation is performed to examine the channel width effect (W) on its current–voltage characteristics Ids, gate-to-source capacitance Cgs, and gate-to-drain capacitance Cgd. Moreover, analytical modeling is put into the study of the source current Ids. A multilayer perceptron (MLP) neural modeling method is applied and tested with different numbers of neurons to elaborate an equivalent mathematical model of the used a-IGZO TFT output characteristic Ids, as well as to develop an equivalent circuit model of the proposed a-IGZO TFT gate-to-source capacitance Cgs, and gate-to-drain capacitance Cgd. Results prove that the multilayer perceptron (MLP) neural modeling and more precisely MLP3 [4 5 1] and MLP3 [4 4 1] represents the most optimal solution compared with the analytical model for Ids and capacitances (Cgs, Cgd), respectively. They also prove that the used Universal Organic TFT (UOTFT) Model is a good candidate for a-IGZO TFT equivalent circuit modeling.