Chemometric characterization of sunflower seeds

Abstract

The spectroscopic characterization of different varieties of sunflower seeds based on their oleic acid content is proposed. One hundred fifty samples of sunflower seeds from different places of Argentina were analyzed by near-infrared diffuse reflectance spectroscopy (NIRDRS). Seed samples were grounded and sieved without chemical treatment previous to the analysis. For the characterization, the used multivariate methods were: principal component analysis (PCA), cluster analysis (CA), linear discriminant analysis (LDA), and partial least square discriminant analysis (PLS-DA). By using PCA, CA, and LDA, and from the point of view of varieties of sunflower seeds, 2 groups were differentiated, based on the concentration of oleic acid: a low oleic group, which ranged from 15% to 25% w/w oleic acid; and the other one (mid-high oleic varieties) which ranged from 26% to 90% w/w oleic acid. However, by using the PLS-DA, 3 groups were correctly differentiated based on the concentration of oleic acid: low oleic (from 15% to 25% w/w oleic acid); mid oleic (26% to 76% w/w oleic acid); and high oleic (≥ than 77% w/w oleic acid), demonstrating the high classification ability of this method. This multivariate characterization of sunflower seed varieties did not require chromatographic analysis to generate the matrix of concentrations, and only direct measures of NIRDRS spectra were required. This characterization can be useful to quickly know the variety of sunflower seed in the grain market. Practical Applications: This manuscript describes a method to determine 3 varieties of sunflower seeds (high, mid, and low oleic) The advantage of this method is to avoid the use of techniques that require long-time analysis.