Estimation of surface roughness upon electroless Ni-Fe-P coatings: experiments, characterization, modelling and optimization

The present work aims to study the surface roughness of an electroless Ni-Fe-P coating on a copper substrate by using an optimum set of concentrations of Nickel Sulphate, Ferrous Sulphate, and Sodium Hypophosphite in the electroless bath using different modelling and optimization methods namely the Taguchi Method, the Genetic Algorithm and the Fuzzy Logic. At first, three important electroless bath component parameters were identified, which influenced the surface roughness of the electroless coating formed, with other parameters and conditions kept consistent. Using the Taguchi Method with an L-27 array, 27 different combinations (with all three bath components varied across 3 discrete levels each) were obtained, and the surface roughness of the 27 electroless-coated samples was experimentally obtained. Now, considering the surface roughness as an output variable and the three electroless bath component parameters considered as input, a statistical linear regression was performed, obtaining a linear equation involving all three parameters. The contribution of the electroless bath components was then statistically studied using ANOVA on the regression equation obtained. The same regression equation was then consistently used in the genetic algorithm to find the concentrations (within their respective ranges) of the three electroless bath components, that yielded the lowest (considered as optimal) surface roughness value of the coating. Similarly, the bath component parameters that influenced the surface roughness were identified using fuzzy modelling. Finally, the results of surface roughness are then validated by experimentally measuring it for a Ni-Fe-P coated copper substrate, coated under the optimal concentrations of the electroless bath components and its surface morphology studied with Scanning Electron Microscopy and Optical Microscopy, followed by the study of weighted elemental composition using Energy Dispersive X-Ray analysis.