This work aimed to predict C18:1 TFA isomers as well as other groups of fatty acids (saturated, monounsaturated, polyunsaturated, and total TFA) in chocolates by ATR-FTIR and partial least square regression. The quantification of fatty acids in representative samples (white, dark, and milk chocolates) was rapid (<30 s) and did not require derivatization. The optimized models showed satisfactory linear correlations compared to a reference gas chromatographic method. Coefficients of correlation for prediction considering C18:1 positional isomers were 0.973 (trans 6-8), 0.991 (trans 9), 0.991 (trans 10), 0.988 (trans 11), and 0.998 (trans 12). When considering fatty acids groups, these coefficients ranged from 0.965 to 0.999. The obtained results indicate that this straightforward procedure is suitable for chocolate analysis, for determining its general lipid composition and TFA isomeric profile, which would be of great interest for quality control programs in the face of the new TFA regulations.
Keywords: fatty acid composition; food quality control; infrared spectroscopy; multivariate calibration; trans fats.