Validity of zinc oxide nanoparticles biosynthesized in food wastes extract in treating real samples of printing ink wastewater; Prediction models using feed-forward neural network (FFNN).

ABSTRACT

In the present study, biosynthesised ZnO nanoparticles in food wastewater extract (FWEZnO NPs) was used in the photocatalytic degradation of real samples of printing ink wastewater. FWEZnO NPs were prepared using green synthesis methods using a composite food waste sample (2 kg) consisted of rice 30%, bread 20 %, fruits 10 %, chicken 10 %, lamb 10%, and vegetable 20%. The photocatalysis process was optimised using response surface methodology (RSM) as a function of time (15-180 min), pH 2-10 and FWEZnO NP (20-120 mg / 100 mL), while the print ink effluent after each treatment process was evaluated using UV-Vis-spectrophotometer. The behaviour of printing ink wastewater samples for photocatalytic degradation and responses for independent factors were simulated using feed-forward neural network (FFNN). FWEZnO NPs having 62.48 % of the purity with size between 18 and 25 nm semicrystalline nature. The main functional groups were -CH, CH2, and -OH, while lipid, carbon-hydrogen stretching, and amino acids were the main component in FWEZnO NP, which contributed to the adsorption of ink in the initial stage of photocatalysis. The optimal conditions for printing ink wastewater were recorded after 17 min, at pH 9 and with 20 mg / 100 mL of FWEZnO NPs, at which the decolorization was 85.62 vs. 82.13% of the predicted and actual results, respectively, with R2 of 0.7777. The most significant factor in the photocatalytic degradation was time and FWEZnO NPs. The FFNN models revealed that FWEZnO NPs exhibit consistency in the next generation of data (large-scale application) with an low errors (R2 0.8693 with accuracy of 82.89%). The findings showing a small amount of catalyst is needed for effective breakdown of dyes in real samples of printing ink wastewater.