Simultaneous electroanalytical determination of luteolin and rutin using artificial neural networks

Abstract

In this study, we propose an electroanalytical method to quantify simultaneously luteolin and rutin, two flavonoids which are present in a pharmaceutical formulation. The methodology is based on square wave voltammetry at glassy carbon electrodes modified with multiwalled carbon nanotubes dispersed in polyethylenimine. Both flavonoids show quasi-reversible surface redox couples in 10% ethanol + 1 mol L−1 HClO4 aqueous solutions, which are defined in potential regions very close to each other. The adsorption process of flavonoids on the modified electrode surface was carried out using an accumulation potential of 0.55 V (vs. Ag/AgCl, 3 mol L−1 KCl), and an accumulation time of 20 min. Considering that luteolin and rutin electrochemical responses show a high degree of overlapping, we processed the electrochemical signals using artificial neural networks. We used a supervised network, feed-forward network with Levenberg-Marquardt back propagation training. Values of 92.6 ± 0.4 and 92 ± 1 mg per tablet were determined by the artificial neural networks methodology for luteolin and rutin, respectively. According to values declared by the manufacturer, differences of 7.4 and 8.0% were calculated for luteolin and rutin, respectively. Results obtained with electroanalytical methodologies were in very good agreement with those obtained by HPLC.