Brazilian basins characterization based on the distributions of elements in desalted crude oils using classical multivariate analysis and kohonen self-organizing map

The classification of petroleum regarding origin is usually carried out by Ni, V, and S correlations because these elements are essentially present in the oil phase. In contrast, other elements can be partitioned with the water phase. In this work, extraction of the water-soluble elements was performed on 70 crude oil samples from six Brazilian Basins. Both original and washed oils were analyzed for their metal concentrations by inductively coupled plasma mass spectrometry (ICP-MS) after microwave-assisted acid decomposition of the matrix. Then, Spearman's correlation between the original and washed oils showed element partitioning. The association of elements in the washed oils was used to avoid repetitive information by reducing the number of variables. Principal components analysis, which is a commonly employed multivariate technique, was applied to the reduced matrix, but Basin's separation was not possible because of the poor contribution (50%) of PC1 and PC2 to the variance. Then, the Kohonen self-organizing map (KSOM) artificial neural network was performed on the same matrix, with a 5x5 hexagonal topology based on Euclidean distance, resulting in a more robust analysis. The proposed data treatment allowed the association of Mo, Re, and V with marine anoxic origin, while Co, Ni, and Ce concentrations increased in petroleum from terrestrial oxic origin, and Basins were separated regarding the occurrence of these elements. Other metals such as Cu, Fe, Mn, and Zn were not significant for the classi-fication of Basins possibly due to other geochemical processes such as the formation and oxidation of sulfides.