Multivariate optimization of a liquid–liquid extraction of the EPA-PAHs from natural contaminated waters prior to determination by liquid chromatography with fluorescence detection

This paper reports the multivariate optimization of a liquid–liquid extraction procedure for the determination of 15 EPA-polycyclic aromatic hydrocarbons (PAHs) by high-performance liquid chromatography with fluorescence detection. A Doehlert design was used to find optimum conditions for the procedure through Response Surface Methodology. Three variables (total volume of hexane, number of extraction steps and duration of such steps) were elected as factors in the optimization study. A principal component analysis (PCA) was run with optimized data, resulting in four groups of PAHs, ordered according to their molecular weight. Final working conditions were established in order to achieve a more robust methodology in relation to all fifteen PAHs under study. Best results could be observed when 77 mL of hexane were divided in four consecutive extraction steps with 18 min each. These experimental conditions were applied in the analysis of a spiked river water sample, and the recoveries varied between 80.9 and 106%, with an average value of 97.1 ± 6.8%. The application of the methodology to river water showed that the method has a good average precision for the studied PAHs.