Determination of the geographic origin of acacia honey by using simultaneous multielement analysis

Abstract

Simultaneous multielement analysis in acacia honey samples was performed in order to determine their geographic origin. We investigated a total of 165 samples collected from two countries; 100 samples from Japan and 65 samples from China. Thirteen elements (Li, Mg, P, K, Ca, Mn, Fe, Rb, Sr, Y, Cd, Ba and La) were measured by inductively coupled plasma mass spectrometry (ICP-MS). Principal component analysis (PCA) of the obtained data revealed good separation between the Japanese and Chinese samples, and indicated the ability to distinguish between samples. Therefore, we performed two statistical analyses : support vector machine (SVM) and linear discriminant analysis (LDA), and compared their results. Consequently, the LDA model, using log 10 ([Ca]/[P]), log 10 ([Mn]/[P]), log 10 ([Rb]/[P]), and log 10 ([Sr]/[P]), where [E] is the concentration of element E, was selected as the most appropriate model. This model correctly predicted 97% of the Japanese samples and 100% of the Chinese samples. It was concluded that the simultaneous multielement analysis can be useful in discriminating the geographic origin of acacia honey.