Semiempirical Modeling for the Photocatalytic Reaction of EDTA over TiO2 applying Artificial Neural Networks

C.A. EMILIO; M.I. LITTER; J.F. MAGALLANES

HELVETICA CHIMICA ACTA

WILEY-V C H VERLAG GMBH

Lugar: Weinheim; Año: 2001 vol. 84 p. 799 - 813

0018-019X

The use of chemometric techniques, full factorial and Doehlert experimental designs, multivariate analysis by MANOVA (multiple-way analysis of the variance), and artificial neural networks (ANNs) for the photocatalytic reaction of ethylenediaminetetraacetic acid (EDTA) over TiO2 in aqueous solution is described. Based on the previous knowledge of the photocatalytic system, variables such as EDTA concentration, photocatalyst concentration, pH, and irradiation time were chosen to build a set of experiments for the analysis. By means of MANOVA, the statistical significance of the individual variables and the inspection of interactions between them were analyzed. By the use of ANNs, correlation plots among variables may help to build a semiempirical modeling for understanding and prediction the behavior of the system, optimizing parameters valuable for further technological applications.