Detection of evening primrose oil adulterated with soybean oil: differential scanning calorimetry and chemometrics

Adulteration is performed by adding inferior products imperceptible by consumers. Consequently, the objective of this work was to develop and use analytical methods to detect adulterants in Evening Primrose Oil (EPO) through differential scanning calorimetry (DSC) with a controlled cooling/heating system and chemometrics. For such a purpose, binary mixtures were prepared using soybean oil as adulterant. It was weighed and prepared in a ratio ranging from 5 to 95% (m/m). Samples were submitted to DSC based on the following parameters: dynamic atmosphere of N₂ (50 mL min⁻¹); temperatures ranging from 283.15 to 198.15 K and from 198.15 to 283.15 K; cooling/heating rate of 275.15 K min⁻¹; about 25 mg of sample in an aluminum crucible. Chemometric models were constructed from the DSC heating curves and normalized by the respective initial masses of samples. Data were pre-processed, normalized by their respective standard deviations and mean center. Results of multivariate analyzes were also compared with univariate calibration using T_onset data (referring to the EPO melting point). Chemometric models were successfully constructed to quantify the level of adulteration, showing RMSE (mean squared errors) of 2.23 and 3.27% m/m for PLS and iPLS (Partial Least Squares and interval Partial Least Squares), respectively. The univariate model demonstrated linearity ranging from 15 to 80% m/m and reached RMSE of 1.06% m/m, thus being an outstanding alternative for a preliminary analysis aimed to reveal the presence of soybean oil in EPO. Therefore, the strategy of coupling DSC with chemometrics for detecting soybean oil in EPO proved to be satisfactory.