Sustainable Photodegradation of Amoxicillin in Wastewater with a Nickel Aluminate and ZnO Heterosystem Oxides: Experimental and Gaussian Process Regression Modeling Studies

The wastewater generated by the pharmaceutical industry poses a risk to the environment due to undesirable characteristics such as low biodegradability, high levels of contaminants, and the presence of suspended solids, in addition to the high load of organic matter due to the presence of drugs and other emerging products in the effluent. This study aims to reduce the impact of wastewater pollution by removing amoxicillin (AMO) antibiotics as an organic pollutant. In this concept, two synthesized catalysts, NiAl2O4 and ZnO, are sensitive oxides to light energy. The prepared materials were then characterized using X-ray diffraction, UV–vis solid reflectance diffuse, Raman spectroscopy, scanning electron microscopy, BET, and ATR-FTIR spectroscopy. The effects of principal operating parameters under sunlight, namely, the percentage of the mixture of NiAl2O4 and ZnO, the pH of the medium, and the initial concentration of the antibiotic were studied experimentally to determine the optimal conditions for achieving a high degradation rate. The results showed that photodegradation is higher at a pH of 6, with a weight percentage of the mixture of 50% for both catalysts in 1 g/L of the total catalyst dose. Then, the effect of the initial concentration of AMO on the photodegradation reaction showed an important influence on the photodegradation process; as the degradation rate decreases, the initial AMO concentration increases. A high degradation rate of 92% was obtained for an initial AMO concentration of 10 mg/L and a pH of 6. The kinetic study of degradation established that the first-order model and the Langmuir–Hinshelwood (LH) mechanism fit the experimental data perfectly. The study showed the success of using heterosystem photocatalysts and sustainable energy for effective pharmaceutical removal, which can be extended to treat wastewater with other organic emerging pollutants. On the other hand, modeling was introduced using Gaussian process regression (GPR) to predict the degradation rate of AMO under sunlight in the presence of heterogeneous ZnO and NiAl2O4 systems. The model evaluation criteria of GPR in terms of statistical coefficients and errors show very interesting results and the performance of the model used. Where statistical coefficients were close to one (R = 0.9981), statistical errors were very small (RMSE = 0.1943 and MAE = 0.0518). The results suggest that the model has a strong predictive power and can be used to optimize the process of AMO removal from wastewater.