Petroleum contamination characterization and quantification using fluorescence emission-excitation matrices (EEMs) and parallel factor analysis (PARAFAC)

摘要

This paper introduces a novel approach to characterize and semi-quantify common petroleum contaminants (natural gas condensate, gasoline, diesel, flare pit residue, and heavy crude oil) and their underlying aromatic hydrocarbon components in solutions based on their fluorescence spectral signatures. The method uses fluorescence excitation-emission matrices (EEMs) combined with multivariate statistical procedures: parallel factor analysis (PARAFAC) and soft independent method of class analogy (SIMCA) to identify the petroleum products. Quantitatively, fluorescence intensities of EEMs of analyzed petroleum products at different concentrations are used to establish standard calibration curves that can be employed to estimate unknown concentrations of similar petroleum products in solutions. As well, underlying aromatic hydrocarbon component concentrations are estimated by performing customized PARAFAC analysis. This approach provides fingerprints for different petroleum products along with estimates of their concentrations in non-fluorescing solvents. Concentrations of predicted PARAFAC components were validated by laboratory chemical analytical results of the same petroleum products.