A COMPOSITIONAL ANALYSIS OF ZONED WATERMELON TOURMALINES USING LASER-INDUCED BREAKDOWN SPECTROSCOPY (LIBS)
The pink and green zones of watermelon tourmalines from the Otjua Mine in Namibia and Brown Derby Mine in Colorado were analyzed using LIBS. Spectra were analyzed using the multivariate technique Principal Component Analysis (PCA). PCA discerns relationships within a complex data set and reduces variability by calculating linear regressions through the data set, principal components. A PCA score plot is a scatter plot in which data points are plotted against two principal components (PC); spectra that cluster together are compositionally similar and those that are dissimilar plot far apart. PCA loading plots show the relative influence of each wavelength on the direction of the PC through the dataset. Loading plots can be data-mined to learn which elements are more abundant in the respective zones.
Thirty spectra were acquired from each of the pink and green zones of the samples; these spectra were averaged by five to create six spectra for each zone for use in PCA. Pink zones from Otijua Mine have higher concentrations of Li, Si, Na, Al and Ca; green zones are enriched in Mg, Fe, Mn, and H. Pink zones from Brown Derby Mine have higher concentrations of Si, Al, Mg, Ti, Ca, Ba, H, Li, and K; green zones are enriched in B, Mn, Zn, and Fe. Mn is commonly purported to be the chromophore in the pink zone; however, it appears here to be concentrated in the green zones. Similar results were observed by Tollefson and Ihinger (2018 GSA Abstracts with Prog.) with the possible explanation that the oxidation state of Mn, rather than concentration, plays a large role in producing color. Further analysis is required to determine whether these compositional trends hold true for various localities.