An effective Application of response surface methodology combined-artificial neural network for bisphenol-A(BPA) treatment using synthesized CNT-MnO2 composite

Bisphenol-A is one of the emerging pollutants,which easily escapes conventional treatment techniques. It requires application of novel composite materials along with mathematical modelling for optimization and evaluation of treatment process. In present study, manganese oxide nanoparticles (MnO2) were doped on the surface of multi-walled carbon nanotubes (MWCNT) to develop an adsorptive-oxidative composite .Composite was characterized using transmission electronmicroscope,X-raydiffraction,RAMANspectroscopy,X-rayphotoelectron sectroscopy,fourier-transform infrared spectroscopyand Surface area analysistoconfirm composite formation and study its properties. Conventional optimization of PH(4-10)BPA initialconcentration(10-50 mg/L), contact time(0-60mins)was carried out and found to be fitting well with Freundlich isotherm model (R2 > 0.99) and followed pseudo-second-order kinetics reaction. Central composite design (CCD) model was applied using Response surface methodology (RSM) to study individual parameters and their interaction effects in order to enhance the process efficiency. Further, the experimental data sets and their responses from RSM were analyzed using Artificial Neural Network (ANN). (80%) of Random experimental sets of which (10%)each to Train,Validate and Test were selected to analyze the variance of models for higher efficiency using Levenberg-Marquardt Backpropagation (LM-BP) algorithm. Additionally, BPA spiked simulated pharmaceutical wastewater was treated with composite to explore its treatment potential. This systematic experimental and computational approach aids in optimizing the treatment efficiency for real-time application.