Prediction of AC conductivity for organic semiconductors based on artificial neural network ANN model

Abstract

This article presents a theoretical study conducted to model the experimental data of alternating–current (AC) conductivity for different organic semiconductors using an artificial neural network (ANN) model. The experimental data were utilized as inputs to an ANN model. ANNs were constructed to obtain the best performance. Mean squared errors (MSE) were calculated and they were found to present small values indicating excellent matching between the ANN results and the targets. A new equation was obtained to describe the behavior of AC conductivity of organic semiconductors as a function of frequency and temperature. This equation was applied to eight different samples of organic semiconductors and excellent results were obtained. Simulation and prediction for known values of experimental data were carried out as a testing step and the results show excellent agreement with the target values. The prediction of values that were not involved in the experimental data range is the main aim of this research. The predictions of these unknown values were carried out successfully and provided excellent results. Thus, the equation obtained in this work could be considered as a generalized tool for predicting the AC conductivity for any other organic material. This article opens up a new field of collaboration between experimental and theoretical researchers.