gems, emerald, beryl, Brazil, mafic-ultramafic, phlogopite, schist, actinolite

This study unravels the spectral footprint of the Paraná deposit, northeastern Brazil, and provides strategies for emerald exploration in mafic- and/or ultramafic-hosted deposits using point spectral data and hyperspectral imaging. Potential host rocks (phlogopite and actinolite-phlogopite schists) were discriminated from other associated lithotypes after petrography and whole-rock geochemistry, using binary MgO versus Al₂O₃ and PC₁ versus PC₃ diagrams. Spectrally, phlogopite schist is marked by Fe–OH absorption at ~2,250 nm and Mg–OH absorptions at ~2,330 and ~2,388 nm. When actinolite is present, a shoulder at ~2,296 nm is recorded, and the first Mg–OH absorption feature is shifted to ~2,315 nm. Emerald crystals have their spectral signature attenuated in the presence of phlogopite. In a 5% emerald and 95% schist mixture simulation, two emerald absorption features are recognizable at ~1,150 (H₂O absorption) and ~2,155 nm. The emerald indices EI₁, EI₂, and EI₃ were chosen for the automated identification of emeralds in the Paraná samples based on the features at ~1,150, ~1,460, and ~2,155 nm. For the host schists, three indices were provided (MI_depth, MI_ratio, and ACI) considering the Mg–OH features of phlogopite and actinolite, allowing automated distinction between potential mineralized and other associated rocks of the Paraná deposit. Our results show that, operating in high-resolution mode (pixels of ~1 mm), imaging spectroscopy can detect the presence and characterize emerald in handpicked samples and drill cores based on chromium- and iron-derived absorption features in the visible to near infrared range. It thus represents a swift, reliable, nondestructive, and relatively cost-effective method for exploring for emerald and, potentially, other gems.