Sunflower oil authentication using green FT-IR-based fingerprinting strategies

This study evaluates green FT-IR fingerprinting strategies as alternatives to the official method for the authentication of sunflower oils, based on attenuated total reflectance Fourier transform mid-infrared (ATR-FTIR) and near-infrared (FT-NIR) spectroscopy combined with chemometric tools. Using NIR and ATR-FTIR fingerprinting, a natural clustering of sunflower oils based on their oleic acid content was observed through principal component analysis (PCA). A direct comparison between both FT-IR approaches was enabled by using the same full training and prediction blocks. Classification/discrimination chemometric models were developed using NIR and ATR-FTIR fingerprints to evaluate the behaviour of commercial medium oleic sunflower oils (MOSFO) as prediction set. Both approaches successfully differentiated high-oleic from non-high-oleic sunflower oils; while revealing samples whose spectral fingerprints were more consistent with the HOSFO category than expected from their commercial label. To quantify the oleic acid content in sunflower oils, a previously developed and validated Partial Least Squares Regression (PLS-R) model based on Raman-SORS fingerprints was employed. The NIR-based regression model exhibited satisfactory performances in the determination of oleic acid in commercial sunflower oils (RER = 9.25; RPD = 23.21) while the ATR-FTIR-based one showed greater predictive power (RER = 16.65; RPD = 40.57). Bland–Altman analysis demonstrated a higher degree of agreement between ATR-FTIR and the GC-FID reference method than for NIR, while Raman-SORS showed comparable agreement to ATR-FTIR. The proposed FT-IR fingerprinting strategies achieved high sustainability scores (AGREENIR = 0.94 and AGREEATR-FTIR = 0.88), offering rapid, non-destructive and solvent-free alternatives for sunflower oil authentication and quality control.