Detection of Traps in Thin-Film Transistors using Evolutionary Algorithms

Abstract

In this work, we present a novel approach to analyzing the current-related characteristics of thin-film transistors (TFTs). We introduce a method to detect and quantify different types of trapped charges from current-voltage curves exhibiting hysteresis, as well as to track the evolution of charge density over time during experiments. To achieve this, we use a previously developed compact model for TFTs that accounts for contact effects and includes a time-dependent threshold voltage. This model is combined with an evolutionary parameter extraction procedure for trap detection. We demonstrate that our time-dependent threshold voltage model is highly adaptable to varying conditions. In fact, our method, which has been successfully applied to detect traps induced by hysteresis, is also capable of identifying unexpected traps from environmental factors. While our evolutionary procedure is slower than traditional methods, which typically rely on extracting constant values for the threshold voltage and sub-threshold swing, it offers a distinct advantage in that it can differentiate between the effects of various traps from a single current-voltage curve and allows continuous monitoring of trapped charge density throughout the experiment. To validate our approach, we conduct an experiment involving the measured output and transfer characteristics of poly(3-hexylthiophene) (P3HT) transistors with varying channel lengths, tested in a room-temperature environment.