Abstract Raman spectroscopy can be a valuable tool for mineral identification; particularly when supported by a dependable library of reference spectra. Dependability, here, includes both that the library can be expected to cover all the minerals in the user’s sample, and that the library contains high signal-to-noise spectra, devoid of non-Raman artifacts, and taken from fully verified (structure and chemistry) samples. Additional ideals for a reference library include display of instrumental metadata, institutional drive for ongoing content growth, and the infrastructure to provide consistent data quality and data-resource longevity. The Raman Spectra Database (RSDB) is designed to have all of these qualities. In order to quickly provide a practical level of content dependability, a “likelihood of encountering” (LE) index was quantified for all known minerals by combining factors that reflect both how common a mineral is and the degree to which geoscientists (both academic and industry) actively seek out certain minerals. The minerals with the top 600 LE values were chosen as the initial content target. The methods for fulfilling this content target include curating reference spectra from the RRUFF spectra database (www.rruff.info) and collecting new Raman spectra on samples that were obtained from major museums and verified (via powder XRD and EDS/EPMA). Some analytical experience with mineral groups that present unique analytical challenges are shared and results achieved to date and plans for the future are detailed.
Bartholomew et al. (Tue,) studied this question.