Multivariate experimental design for the photocatalytic degradation of imipramine Determination of the reaction pathway and identification of intermediate products

Multivariate experimental design was applied to the degradation of imipramine solutions under simulated solar irradiation. The efficiency of photocatalytic degradation was determined from the analysis of the following parameters: H2O2, Fe-(II) and TiO2 concentrations. Experimental data were then fitted using artificial neural networks (ANNs). The findings indicate that ANN provided excellent predictive performance while the influence of each parameter on the variable studied was assessed, TiO2 being the most significant factor, followed by H2O2 and Fe-(II). TOC profile shows an initial sharp decrease within 4 h of irradiation (almost 75% of the organic carbon was mineralized) and until 24h complete mineralization is achieved. Nitrogen is mainly transformed into ammonium ions (almost 90% of the stoichiometric amount) and in a negligible extent into nitrate ions. GC/MS and LC/MS were brought to bear in assessing the temporal course of the photocatalyzed process. A first pathway involves the hydroxylation, that is confined to the dibenzodiazepine moiety or the methyl groups. Another route proceeds through the oxidation of the dibenz[azepine] moiety, with the formation of the ketoderivative. Afterwards, the detachment of the aminoalkylic chain occurs. A parallel transformation involves aminoalkylic chain, with the partial or total detachment of the propylic chain. Microtox bioassay (Vibrio fischeri) was also employed in evaluating the ecotoxicity of solutions treated by photocatalysis. Results clearly demonstrate the efficiency of the photocatalytic process in the detoxification of the irradiated solution. (C) 2008 Elsevier B.V. All rights reserved.