APPI (+) FT-ICR MS AND MACHINE LEARNING FOR THE DIFFERENTIATION OF WEATHERED CRUDE OILS FROM DIFFERENT ORIGINS

This study investigates the differentiation of marine and lacustrine crude oils before and after simulated photo-oxidative weathering, mimicking a type of compositional alteration that affects crude oil during oil spills. Ultra-high resolution mass spectrometry with positive-ion atmospheric pressure photoionization (APPI (+) FT-ICR MS) was combined with chemometric modeling using partial least squares discriminant analysis (PLS-DA) and variable selection through the ordered predictors selection (OPS) algorithm to investigate the viability of predicting the origin of crude oil. The results demonstrated good predictive capability even after weathering processes. Additionally, the combination of APPI (+) FT-ICR MS data with variable selection showed the potential to establish a link with molecular compounds and thereby enabling the identification of possible molecular markers by ultra-high resolution mass spectrometry.