Photo-Assisted Degradation, Toxicological Assessment, and Modeling Using Artificial Neural Networks of Reactive Gray BF-2R Dye

Abstract

This work investigates the degradation of Reactive Gray BF-2R dye (a blend of reactive yellow 145, reactive orange 122 and reactive black 5 dyes) using UV/H2O2, Fenton, and photo-Fenton-advanced oxidative processes, with artificial sunlight and UV-C radiations. The photo-Fenton process employing UV-C radiation was the most efficient under the conditions studied. The ideal conditions for the degradation of the dye, determined using a factorial design 23 and a study of the concentration of hydrogen peroxide ([H2O2]), were [H2O2] equal to 40 mg L−1, iron concentration [Fe] of 1 mg L−1, and pH between 3 and 4. The Chan and Chu non-linear kinetic model predicted the kinetic data with a degradation of over 98% for color and 68% for aromatics after 60 min. The behavior of the chemical oxygen demand fitted the first-order kinetic model well, with a degradation of 64% after 60 min. The Multilayer Perceptron 7-11-2 artificial neural network model enabled to model the degradation process of the aromatics and accurately predict the experimental data. Toxicity tests indicated that the post-treatment samples were non-toxic for Escherichia coli bacteria, and Portulaca grandiflora and Basil sabory seeds. However, they inhibited the growth of Lactuca sativa seeds and Salmonella enteritidis bacteria. The photo-Fenton process with UV-C radiation degraded the dye studied efficiently and the degradation percentages were, on average, 7% and 5% higher for color than those observed when employing the Fenton and UV/H2O2 processes, respectively. With the aromatic, however, they were 84% and 62% higher, thus justifying the use of this process.