Spectral library of gemological minerals from NE Brazil: A reference to aid the reflectance spectra-based discrimination of gem-quality materials

Abstract

We report a spectral library of gem-quality minerals from NE Brazil and imitation material as a strategic, rapid, nondestructive, and relatively cost-effective tool to be used as a reference to aid in discriminating gemological materials. Our study used spectra data collected by ASD-FieldSpec spectroradiometers (Standard Resolution and High-Resolution Next Generation) for 42 gem-quality mineral samples of beryl, emerald, elbaite tourmaline, euclase, vesuvianite, and apatite crystals from the Borborema Province. In addition, five emerald imitations and nine faceted emeralds from unknown provenances were also used to compose the spectral dataset. The main metals- and molecules-related absorption features were interpreted to show the similarities and differences of the spectral signatures from the minerals and imitation samples. In addition, the principal component analysis (PCA) was applied for data dimensionality reduction of the normalized reflectance spectra (n = 115), and the significance of PCs (99% of variance) was interpreted to highlight the main mineral groups. FreeViz algorithm was employed as an alternative for efficient visualization of PCs’ axes, showing a suitable discrimination of the samples. For example, the clustering of (a) beryl and emerald samples, (b) iron-rich elbaites (blue and green colors), and (c) Cu- and Mn-rich elbaites (Paraíba tourmaline and green varieties) was revealed, summarizing the similarities already expected from the qualitative spectral signature inspections and automating the discrimination of main mineral groups. Beyond introducing a spectral library of gem materials from NE Brazil, the methodology supports studies of gem classification and the authenticity assessment of diverse gem materials. The methodology can also be expanded to other reflectance datasets of gem-quality minerals worldwide.