FT-IR extra virgin olive oil classification based on ethyl ester content

According to the Regulation (EU) 2016/2095, extra virgin olive oils (EVOO) must contain a maximum of 35 mg kg−1 of fatty acid ethyl esters (FAEE). The official method for FAEE quantification is time-consuming and requires the use of a large amount of toxic solvents. Thus, the aim of this work was the application of FT-IR spectroscopy for the development of classification models (i.e. Linear Discriminant Analysis, LDA; Soft Independent Modelling of Class Analogy, SIMCA) able to discriminate EVOO from non-EVOO based on FAEE content. To the aim, 113 EVOO and 46 non-EVOO samples were analysed. Since the Principal Component Analysis revealed that the whole FT-IR spectral range (both raw or pre-treated) was not promising in EVOO and non-EVOO distinction, a variable selection strategy was applied (i.e. SELECT algorithm). All the classification models were validated both by cross validation and with three different external test sets. The best and more robust LDA model was obtained with the raw FT-IR selected variables, reaching 96–100% of correct classification in prediction. SIMCA models resulted less reliable. In particular, the low specificity values (40–67%) revealed that there is a high probability of assignment of non-EVOO to the EVOO class. In conclusion, FT-IR spectroscopy coupled with a discriminant classification approach is a useful tool for a rapid and fast discrimination of EVOO and non-EVOO based on FAEE content. Since the variable selection strategy was effective, the development of simplified and cheap instruments can boost the FT-IR spectroscopy application also in small enterprises, giving the opportunity to acquire many important information about olive oils.