Comparison of oil spillages using mid-IR indexes and 3-way procrustes rotation, matrix-augmented principal components analysis and parallel factor analysis

Three different approaches for 3-way analyses, namely, Procrustes rotation, parallel factor analysis (PARAFAC) and matrix-augmented principal component analysis (MA-PCA), have been used to compare six different oil spillages made under controlled conditions (one of them corresponding to the heavy oil released after the sunk of the Prestige tanker off the Galician coast–NW Spain on November 2002). Each spillage was monitored during three and a half months by attenuated total reflectance (ATR) mid-IR spectroscopy. Ten characteristic band ratios were defined. Results showed that the three 3-way chemometric techniques lead to essentially the same conclusions, where from it was concluded that the most relevant pattern defining the oil weathering was related to ‘total aromaticity’, i.e., the total number of CC bonds in the molecules which form the products. In addition, weathering of the samples got clearly characterized by a steady evolution on the scores (sample weights), with a clear increase after 11–14 days. Differentiation of the products (slices of the data cube) was also possible due to their intrinsic characteristics as, in general, heavy products oppose to the lightest ones.