Modelling of the hysteresis effect of target voltage in reactive magnetron sputtering process by using neural networks

Reactive sputtering process is very non-linear and usually exhibits hysteresis behaviour with respect to the reactive gas flow. Most of the problems encountered in the preparation of non-stoichiometric compound films by reactive sputtering are due to the hysteresis effect. Therefore, a considerable amount of effort has been devoted to find means for its elimination or ensuring a stable sputtering in the transition mode. This paper presents a new approach based on Artificial Neural Networks (ANNs) for modelling of the hysteresis effect of target voltage at different target power levels and reactive gas flow rates in reactive sputtering. Based on this model, it is possible to predict the target voltage in reactive magnetron sputtering processes, when the target power level, reactive gas flow rate and its direction are used as inputs to the artificial neural network (ANN). The proposed ANN is trained in different structures with the use of learning algorithms to obtain better performance and faster error convergence. Broyden–Fletcher–Goldfarb Shanno (BFGS) algorithm gives the best result among other learning algorithms used in the analysis. The training and test data required to develop the ANN model are obtained from the experimental studies. Both the training and the test results are in very good agreement with the experimental results obtained in this work.