Matching fluorescence spectra of oil spills with spectra from suspect sources

Abstract

Fluorescence spectroscopy is ideally suited to the analysis of oil spills as it allows chemical information of polycyclic aromatic hydrocarbons to be acquired quickly, sensitively and selectively. Unlike infrared spectra which have detailed peak information, many fluorescence spectra have only a few broad peaks. Nine different samples of crude and diesel oils were used for testing point-to-point matching across the spectral range. Five of them were discriminated by point-to-point matching algorithms and the other four very similar samples were not. Principal components analysis (PCA) did successfully discriminate among all similar samples. PCA could also distinguish the extent of weathering of different samples, an important factor in matching environmental spills.