Optimization of Tartrazine Adsorption onto Polypyrrole/SrFe12O19/Graphene Oxide Nanocomposite Using Central Composite Design and Bat Inspired Algorithm with the Aid of Artificial Neural Networks

This project was aimed to focus on the application of bat inspired algorithm with the aid of artificial neural networks (ANN-BA) as a novel metaheuristic algorithm in chemistry and environmental sciences for optimization of tartrazine dye adsorption onto the polypyrrole/SrFe12O19/graphene oxide (PPy/SrM/GO) nanocomposite from aqueous solutions. The PPy/SrM/GO nanocomposite was fabricated by an in situ polymerization process and its structural and magnetic properties were studied by means of several instrumental techniques. Four factors affecting adsorption process were optimized in a batch system by ANN-BA and central composite design (CCD). In comparison to the CCD, the ANN- BA model obtained through levenberg marquardt back propagation methodology, gave higher percentage removal (94 %) about 6 %. Under optimal conditions obtained by ANN-BA, the values of four factors including initial concentration, adsorbent dosage, pH, and shaking rate were 15 mg/l, 0.02 g, 6.5, and 297 rpm, respectively. In the above conditions, the experimental results were fitted well to the pseudo-second-order kinetic model with the rate constant (k(2)) of 0.038 g/mg/min and the Langmuir adsorption isotherm with monolayer maximum capacity (q(m)) of 123.5 mg/g with determination coefficients (R-2) of 0.9986 and 0.9989, respectively. Thermodynamic studies revealed that tartrazine adsorption was spontaneous in all temperatures (Delta G< 0), endothermic (Delta H=30.816 kJ/mol), and feasible process with slight increase of entropy (Delta S=0.116 kJ/ mol/K). Moreover, the adsorbent application in wastewater and its regeneration studies depicted that the nanocomposite can be applied as an effective adsorbent (R%>89), magnetic separable and reusable adsorbent (R%>50 after the sixth regeneration cycle) in environmental cleanup.