Parametric optimization and minimization of corrosion rate of electroless Ni–P coating using Box-Behnken design and Artificial Neural Network

Abstract

In this study, parametric optimization has been conducted to determine the optimal combination of parameters, considering the corrosion rate of an electroless Ni–P coating onto a copper substrate as a response, and a predictive model has been prepared using Artificial Neural Network (ANN). The concentration of nickel sulphate, sodium hypophosphite, and bath temperature—the three design factors—have been considered as the process parameters. 20 g/L of nickel sulphate, 20 g/L of sodium hypophosphite, and a bath temperature of 85 °C have been found to be the optimum conditions using the Box-Behnken Design (BBD) of the experiment for the deposition of Ni–P coating. The corrosion rate of the as-deposited coat-ed sample in the optimum condition is 1.06μm/Y, which is almost half of that of the original copper substrate. Further, an analysis of variance (ANOVA) has been performed in order to visualize and investigate the impact of the process parameters. A predictive model, the Artificial Neural Network, is employed, and ANN 3-10-1 is found to be the best fit model to predict the corrosion rate of Ni–P, where 3 indicates 3 input variables, 10 indicates 10 hidden neurons, and 1 indicates 1 output. The surface morphology of an electroless Ni–P-coated copper substrate has been studied using a scanning electron microscope (SEM), while the elemental study has been carried out using energy dispersive X-ray analysis (EDX).