Optimization and modeling of simultaneous ultrasound-assisted adsorption of ternary dyes using copper oxide nanoparticles immobilized on activated carbon using response surface methodology and artificial neural network.

The present study examines simultaneous adsorption of ternary dyes such as rose bengal (RB), safranin O (SO) and malachite green (MG) from aqueous media on copper oxide nanoparticles immobilized on activated carbon (CuO-NPs-AC) in a batch system. To forecast and optimize the adsorption, artificial neural network (ANN) and response surface methodology (RSM) were utilized. The effect of various factors, e.g. dye concentration, sonication time, adsorbent dosage and pH on the adsorption process were evaluated through five level six factor central composite design (CCD) using RSM. Maximum removal efficiency of MG, SO and RB dyes were seen 94.26%, 71% and 76% under optimal operating conditions. The suggested quadratic models revealed good fit with the actual data. To testing the data, the coefficients of determination (R2) of 0.9976, 0.9971 and 0.9952 and Fisher F-values of 2048.92, 1660.95 and 926.84 were obtained for MG, SO and RB dyes, respectively. The same data were utilized to construct the ANN models. The results revealed that both models yielded high R2 values, while the RSM models were slightly more accurate in predictions as compared to ANN models for MG, SO and RB dyes removal. The equilibrium data followed the Langmuir isotherm model, although the rate of the adsorption process well fitted to pseudo-second-order kinetics. The maximum adsorption capacity of the CuO-NPs-AC for MG, SO and RB were found to be 212.79, 149.25 and 172.42 mg/g, respectively.