Energy-dispersive X-ray fluorescence combined with chemometric tools applied to tomato and sweet pepper classification

Consumers' interest in organic foods is increasing, and robust analytical tools are necessary to authenticate them. This study investigated the potential of Energy-Dispersive X-Ray Fluorescence (EDXRF) under two measurement conditions combined with chemometric tools to classify tomato and sweet pepper samples according to the agronomic production mode and geographical area cultivation. Initially, the Common Dimension Analysis (ComDim) was applied to 15 kV (Na-Sc) and 50 kV (Ti-U) EDXRF spectra to extract the common information generated by both spectral measurement conditions. According to the ComDim analysis, it was possible to verify that the data generated by both instrumental conditions are complementary. In contrast, it was impossible to cluster tomato or sweet pepper samples effectively according to the agronomic production mode or geographical origin. On the other hand, the junction between EDXRF data and Partial Least Squares with Discriminant Analysis (PLS-DA) proved to be a perfect match, reaching optimal prediction accuracy values for the classifi-cations between 98.8 and 100.0%. These results suggest that EDXRF associated with chemometrics can be an excellent technique to verify the authenticity of tomato and sweet pepper. Besides, it does not require complex sample preparation steps and is fast and relatively inexpensive compared to some chromatographic or spectro-metric analytical techniques.