Origin determination of emeralds is a fascinating subject. In our efforts to gain greater insights into gemstones from different origins, SSEF has developed a new machine-learning-assisted chemical data visualization approach for emerald characterisation.

The case study compared 168 emeralds from different origins using a t-SNE (t-distributed stochastic neighbor embedding) algorithm. The image below shows how through compilation of56 elements in the t-SNE calculation, clustered groups of emeralds from different origins can be characterised.

This type of approach will never replace the need for human gemmologists, but provides a deeper level of understanding into how emeralds form and the chemical fingerprinting of emeralds.

This research has just been published in Journal of Analytical Atomic Spectrometry of the Royal Society of Chemistry in the UK.