Hyperspectral images: a qualitative approach to evaluate the chemical profile distribution of Ca, K, Mg, Na and P in edible seeds employing laser-induced breakdown spectroscopy

Abstract

In the present study, laser-induced breakdown spectroscopy (LIBS) combined with chemometric tools was used to investigate the metal composition in nine seed samples. The samples were directly analyzed, and a matrix with 9 rows and 9 columns (81 points) and 10 consecutive pulses were analyzed in each point. A total of 810 emission spectra were collected from 186 to 1042 nm from the surface and bulk of the sample. The dataset was normalized by Euclidian norm and principal component analysis (PCA) was used for the initial exploratory investigation. Calcium, Mg, Na, K and P were mainly identified in all samples; however, the distribution of metals in these samples is not completely homogeneous, i.e., the composition of the elements change from one layer to another. This fact can be probably related to the absorption capability of nutrients resulting from different factors such as soil characteristics, physiology of the plant, water source composition and fertilizers which can influence the distribution of the elements in different seeds. To confirm the elements observed by LIBS, the samples were digested using microwave-assisted digestion, and Ca, K, Mg, Na and P were determined by inductively coupled plasma-optical emission spectrometry (ICP-OES). In addition, some minor nutrients such as S and Zn were also investigated and the relationships between elements were observed through the Pearson correlation graph, and some of them, such as Mg and Na, P and Na, S and P, S and Zn, are extremely correlated; it means that, for example, when the concentration of Mg increases, that of Na also increases.