Regression-BPNN modelling of surfactant concentration effects in electroless Ni[sbnd]B coating and optimization using genetic algorithm

Abstract

Critical micelle concentration (CMC) is an important factor to avoid the formation of micelles from monomeric surfactant molecules. Trisodium citrate stabilized electroless NiB (ENi-B) coating on aluminium alloy (Al7075-T6) is attempted with the addition of amphoteric surfactant, 3-(N, N-Dimethylmyristylammonio) propanesulfonate (3-DMAPS), to enhance the surface finish (Ra) of the coatings. The main aim of the study is to investigate the influence of surfactant concentration on average surface roughness in ENi-B bath and determine the CMC of 3-DMAPS at minimum Ra. Mathematical models relating the concentration of amphoteric surfactant (0–0.162 g/L) as an independent variable and Ra as a dependent variable are developed using univariate regression analysis (Linear, quadratic, power, and exponential models) and back propagation neural network (BPNN) algorithm. The coefficient of determination (R2) is used to evaluate the goodness of fit between the models, and the BPNN model is found to be the best fit (R2 > 0.98). The minimum Ra of 0.171 ± 0.001 μm was achieved at the CMC of 0.049 g/L (0.135 mM) from the genetic algorithm (GA) using the validated models developed by quadratic regression analysis and BPNN as fitness functions. SEM, XRD and AFM techniques were carried out for the characterization of ENi-B coatings with and without surfactant at CMC.