LAUSR

This research shows the development of a miniaturized solid-phase extraction method with UV-Vis detection for simultaneous determination of dibenzofuran, fluoranthene and phenanthrene using chemometrics approaches. After synthesis of Fe3O4 nanoparticles (Fe3O4 NPs), the surface of the nanoparticles was modified by Triton X100 coating. The influence of extraction solvent and volume, concentration of Triton X100, extraction time, and sample pH were studied and optimized. Due to high spectral overlapping, resolving ternary mixtures for simultaneous determination of targets with classical analytical methods is impossible. Therefore, the recorded UV-Vis spectra were transformed using continuous wavelet transform and then subjected to artificial neural networks (ANNs). The Db4 mother wavelet was used as the better mother wavelet. For simultaneous detection of analytes, a comparison of feed-forward back-propagation and radial basis function networks was accomplished. The calibration graphs showed linearity in the ranges of 2.4–250 ng mL−1, 50–3750 ng mL−1, and 48–5000 ng mL−1 with a limit of detection of 0.58, 9.5 ng mL−1, and 12.5 ng mL−1 under optimal conditions for phenanthrene, fluoranthene, and dibenzofuran, respectively. The limit of quantitation was achieved at 3.52 ng mL−1, 16.35 ng mL−1, and 31.3 ng mL−1 for phenanthrene, fluoranthene and dibenzofuran, respectively. The method involving TX-100-coated Fe3O4 NPs in a liquid sample phase for analyte extraction, followed by ethanol desorption and UV-Vis detection, was successfully applied for the determination of polycyclic aromatic hydrocarbons in oil-field water and algae samples.