A rapid and simple approach to identify different sunflower oil types by means of near-infrared reflectance spectroscopy

The potential of near-infrared reflectance spectroscopy (NIRS) to perform a n easy and rapid classification of different sunflower oil types was invest igated. A total of 118 oil samples showing large variation in their fatty a cid compositions were analyzed by both NIRS and gas-liquid chromatography ( GLC). They were classified into five classes, characterized by (i) high pal mitic acid content (>29%), (ii) high palmitic acid in high oleic acid backg round (>27 and >51%, respectively), (iii) high stearic acid content (>22%), (iv) high oleic acid content (>83%), and (v) standard oil type. Second-der ivative transformation and scatter corrections were applied to the original log (1/R) spectra, and the correlation coefficients between NIRS spectral information and GLC fatty acid values were studied to identify the waveleng ths with the best discriminating ability. The use of the spectral data at 2 134 nm permitted all the samples with high levels of total saturated fatty acids (>29%, classes i, ii, and iii) to be discriminated from the samples w ith standard levels (<22%, classes iv and v). The use of a second wavelengt h, 2192 nm, led to a further separation of the samples with high C-18:1 con tent within each group (classes ii and iv, respectively). Therefore, an acc urate discrimination of four of the five sunflower oil types was achieved b y using the spectral information at two wavelengths exclusively. The oil sa mples belonging to classes i and iii could not be separated with this appro ach, which was explained on the basis of the small spectral differences obs erved between the two classes.