Artificial Neural Network Modelling on TiN Coating Parameters in Sputtering Process

Sputtering is a Physical Vapor Deposition (PVD) vacuum process used to deposit very thin films onto a substrate for a wide variety of commercial and scientific purposes. Due to this coating process the material performance will be improved. The objective of this study is to evaluate the hardness of titanium nitride thin film layers by utilizing multi layer perceptron (MLP) artificial neural networks (ANN). For determining the influences of the various sputtering parameters (voltage, work pressure, ion bombard time and Sub-layer temperature) on the hardness of TiN thin films, the Taguchi approach has been used. 50 experiments were performed, varying the PVD parameters and the resulting hardness of the film was measured. From these experiments 42 were used for the training process and 8 have been utilized for validation process. A very good agreement between the ANN predictions and the experimental results was achieved, with a 99.754% correlation between the trained ANN result and the experimental measurements.