Modeling of reflectance properties of ZnO film using artificial neural networks

Abstract

A ZnO thin film was prepared on a p-Si (100) substrate by using a pulsed filtered cathodic vacuum arc deposition system (PFCVAD). Specular reflectance, a nondestructive technique, can be used to measure thickness, refractive index of thin films grown on reflecting substrate and their dependancy on reflecting angle. In this study, the effects of reflectance angle on specular reflectance measurements of ZnO thin film is modeled by Artificial Neural Networks (ANN) utilizing "Multi-Layer Perceptron (MLP)", Back propagation Algorithms Levenberg Marqued that is learning rule on incident angle range of 30-60 degrees. Also it is shown that reliable high precision measurements can be obtained without using expensive high precision hardware.