Biodiesel undergoes oxidative degradation, losing stability upon exposure to heat, oxygen, and metallic traces. To investigate these alterations, six commercial B100 samples and one laboratory-synthesized sample (hereafter referred to as “originals”) were subjected to two standardized oxidation assays: RapidOxy (ASTM D7545) and Rancimat (EN 14112); and subsequently analyzed using ultrahigh-performance liquid chromatography coupled to high-resolution mass spectrometry (UHPLC-HRMS). An untargeted analytical approach initially detected 7,722 molecular features. A Python-based script was employed to remove redundancies, resulting in 3,131 unique variables. Following data standardization, supervised modeling was conducted using Partial Least Squares Discriminant Analysis (PLS-DA) and Orthogonal Partial Least Squares Discriminant Analysis (OPLS-DA). These models, validated through 200 permutation tests, exhibited determination coefficients (R2) close to 1.0 and group-based cross-validated accuracy (Q2) exceeding 0.8, effectively distinguishing nonoxidized from oxidized fuels. Features with Variable Importance in Projection (VIP) scores greater than 2.0 in PLS-DA and above 1.5 in OPLS-DA were subjected to the Wilcoxon–Mann–Whitney test. Resulting p-values were adjusted using the Benjamini–Hochberg procedure and combined with a |log2 fold-change| threshold of ≥2.0, yielding 12 statistically robust candidate molecular features. Individual Receiver Operating Characteristic (ROC) curves showed area under the curve (AUC) values ranging from 0.80 to 1.00, confirming their strong discriminatory power. Their chemical compositions clearly distinguish degradation pathways: the RapidOxy assay produced hydroperoxy-epoxydienoates and furan-fatty acids, whereas the Rancimat test yielded hydroxy-stearates, dihydroxypalmitates, and the lactone curvulalide. These findings support the differentiation between oxidation profiles and may offer insights into the predominant pathways involved in biodiesel oxidation.
Muniz et al. (2026) studied this question.