Raman Spectroscopy Integrated with Machine Learning as a Tool for Maturity Assessment of Organic Matter: A Case Study in Santos Basin, Brazil

The thermal maturation assessment of organic matter (OM) is crucial for understanding the hydrocarbon generation and expulsion processes in oil-bearing rocks. Vitrinite, a maceral found in coal samples, has its reflectance (%Ro) measurement widely used as a paleotemperature indicator in organic-rich sedimentary rocks. However, alternative methods are necessary to determine paleotemperature in sedimentary basins where vitrinite is scarce or absent, such as in presalt formations. This work used Raman spectroscopy to determine the reflectance equivalent in OM present in carbonate rocks from a presalt petroleum system (Santos Basin, Brazil). Vitrinite fragments present in palynofacies sections from a well of the Potiguar Basin, Rio Grande do Norte, Brazil, collected in different depths, with %Ro obtained in the range from 0.46 to 2.72%, were used as reference material to model calibration using their Raman spectral parameters, assisted by the Machine Learning Least Absolute Shrinkage and Selection Operator (LASSO) algorithm. The strategy was employed to analyze 30 samples of presalt carbonate rocks containing solid bitumen, whose Raman spectral parameters were used to determine their equivalent reflectance values (%R_Raman). Calibration with vitrinite samples was done using excitation radiation with wavelengths at 532 and 632.8 nm. The %R_Raman values were determined for all rock-containing bitumen samples using the exciting radiations. The methodology has proven to be an effective way to determine the OM thermal maturity.