ABSTRACT
Forty-nine green coffee (Coffea arabica L.) bean samples, representing the Harar, Jimma, Kaffa, Wollega, Sidama, and Yirgachefe varieties from east, west, and south Ethiopia were analyzed for Mg, P, S, Ca, Mn, Fe, Cu, Ba, Si, and K using inductively coupled plasma-optical emission spectroscopy. Principal component analysis was used to group the samples based on their elemental compositions. Application of linear discriminant analysis provided models with an accuracy of 92% for region of production and 79% for variety. The elements P, Mn, S, Cu, and Fe were the most discriminating elements. These results show that initial screening may be performed to characterize possible fraudulently labelled products with verification by the existing paper-based traceability system and sensory evaluation.
Acknowledgment
The authors thank the Ethiopia Commodity Exchange for providing many of the samples used in this study.