Quantificação de ácidos graxos em óleos vegetais por espectroscopia no infravermelho associada à calibração multivariada por mínimos quadrados parciais

Abstract

The fatty acids are chemical structures of elevated nutritional value for the human health. These substances are present, mostly, in vegetable oils, composing one of the main energy sources. Currently, the quantification of fatty acids is performed especially using the gas chromatography technique with a flame ionization detector (GCFID), which requires sample preparation, involving a long time to accomplish, as it requires organic solvents, which implies disadvantages to the environment due to the waste generated. In view of this, the objective of this work was to develop a simple and effective analytical methodology for the determination of these compounds in complex matrices, from the technique of midinfrared (MIR) spectroscopy by regression of partial quadratic minimums per interval (iPLS) for determination of the palmitic, stearic, oleic, linoleic and linolenic fatty acids. A total of 80 and 19 samples composed the calibration and prediction sets, respectively, and the division was performed by the KennerdStone algorithm. The lowest square root values of the mean quadratic crossvalidation error (RMSECV) in the intervals were in the region of 600 to 3550 cm1 for palmitic, stearic, oleic and linolenic fatty acids. For linoleic the best range was 2707 to 3128 cm1.