CWPO Degradation of Methyl Orange at Circumneutral pH: Multi-Response Statistical Optimization, Main Intermediates and by-Products

Abstract

The catalytic wet peroxide oxidation (CWPO) of the industrial azo-dye methyl orange (MO) activated by an Al/Fe-pillared clay catalyst was optimized by the Response-Surface Methodology (RSM). Three sequential sets of factorial 2^k central composite experiments were required for the full optimization of the process; catalyst loading and stoichiometric dose of hydrogen peroxide were the experimental factors studied through different times of reaction by means of all, Dissolved Organic Carbon (DOC) removal, Total Nitrogen (TN) removal, reacted fraction of hydrogen peroxide, and decolorization as experimental responses to be maximized. The resulting single-response RSM optimums were combined in a multi-response Desirability function ruling out the differential effect of adsorption on the catalyst's surface by defining all responses per gram of clay catalyst. Former two statistical sets of experiments (DOE-1 and DOE-2) showed the CWPO degradation of MO to get favored at increasing both catalyst loading and time of reaction (up to 180 min). Afterwards, third final design of experiments (DOE-3) displayed 75% of DOC removal, 78% of TN removal, 97% of reacted H₂O₂, and 95% of decolorization by using a catalyst loading of 5.0 g/L of Al/Fe-PILC together with just 50% of the stoichiometric amount of H₂O₂. The multi-response optimum conditions based on the Desirability function showed excellent fitting explaining at least 99.3% of the optimal overall responses at 95% of confidence. A further analysis revealed that no one of the non-controllable variables under real conditions of industrial wastewater treatment (covariates): starting total organic carbon (TOC) (2.0-20 mg/L), temperature (5.0-25°C) or circumneutral pH (6.0-9.0), exhibited statistically significant effect (P > 0.05), suggesting the system to be almost insensitive against them within studied range of close to ambient conditions in the tropic. Finally, HPLC/PDA and GC/FID measurements identified phenol, cyclohexa-2,5-diene-1,4-dione, phenylamine, N-methylaniline and N,N-dimethylaniline in very low concentrations as main intermediates in the CWPO degradation of MO, which nevertheless disappeared over 90 min of treatment. Meanwhile, 4-aminobenzenesulfonic and oxalic acids were recorded as unique by-products at final time of reaction, but both of them fairly less toxic than the starting azo-dye.

Keywords: Al/Fe-PILC; catalytic wet peroxide oxidation; methyl orange; response surface methodology; statistical optimization.