Oil spill in northeastern Brazil: Application of fluorescence spectroscopy and PARAFAC in the analysis of oil-related compounds

Between November 2019 and February 2020, 53 water samples were collected along 430 km of coastline in northeastern Brazil, which was the location of an oil spill that occurred in August 2019. Synchronous fluorescence matrices (SFMs) were acquired to avoid regions affected by Raman Stokes scatterings and second harmonic signals, and then, the SFMs were converted into excitation-emission matrices (EEM) by shear transformation. The matrix coupled with parallel factor analysis (PARAFAC) was used in the study of fluorescent components present in the collected waters. A sample collected before the oil spill and another from Florianopolis-SC, 2000 km from the incident, were used as references for nonimpacted waters. In the postspill samples, 4 components were determined, with component 1 (lambda(exc) = 225 nm, lambda(em) , = 475 nm) being associated with humic-like organic matter (terrestrial), component 2 (lambda(exc) = 230 nm, lambda(em), = 390 nm) being associated with humic-like organic matter (marine), component 3 (lambda(exc) = 225/295 nm, lambda(em) = 345 nm) being associated with dibenzothiophene-like components also observed in tests with crude oil samples, and component 4 (lambda(exc) = 220/280 nm, lambda(em) = 340 nm) being associated with a naphthalene-like substance. Principal component analysis (PCA) was performed on the PARAFAC scores. The distribution of samples along the 4 components was observed and compared with the reference samples. (C) 2020 Elsevier Ltd. All rights reserved.

Automated summary

Water samples were collected along the coastline in northeastern Brazil between November 2019 and February 2020, following an oil spill in August 2019. Different fluorescent components were identified in the postspill samples, associated with terrestrial and marine organic matter, dibenzothiophene-like substances, and a naphthalene-like substance. Principal component analysis showed the distribution of samples along these components, compared with reference samples.