IS RAMAN READY FOR PRIME TIME? COMPARING THE MINERAL IDENTIFICATION SUCCESS RATE OF RAMAN SPECTROSCOPY AND POWDER XRD

Raman Spectroscopy (RS) has been employed in the geosciences since the mid-1960’s, primarily for a few specialized research applications. Falling instrument prices and an extensive and growing database of reference spectra (http://rruff.info) give RS the potential to become a widely used tool for routine mineral identification – just as powder XRD has been for decades. Despite its potential advantages (e.g. little or no sample preparation, multi-micron-scale single grain analysis, in-situ analysis) RS has yet to be widely employed by the geoscience community – partly because this community does not know whether RS is dependable for routine mineral identification and partly because there are some remaining challenges to applying RS. In order to begin to address the dependability question a blind study was performed on a commercial set of 56 mineral separates using standardized methodologies for sample preparation, instrument operation, data processing and search/match identification. Results obtained using recent model Raman (780nm) and XRD (Cu Kα) instruments were quantified with a grading system that associated a success value depending on whether the correct mineral name came up first, in the top 5, or in the top 10 of the list of probable matches produced by search/match software. A few of the mineral separates were found to be impure, which is not a problem for RS which can gather spectra on individual grains, but which causes problems for search/match success of XRD. Since complete mineral separation is normally a basic objective in sample preparation for XRD the impure separates were excluded when calculating the average success grade for XRD which was 89 (a perfect grade being 100). The major remaining challenge in applying RS for mineral identification is specimen fluorescence interference, but progress is being made in developing solutions for this interference. In the present study 15% of the minerals produced fluorescence that prevented obtaining RS spectra resulting in an average success grade for RS of 77. However, if the fluorescing minerals are excluded the success grade for RS rises to 90. Therefore, if further development of RS technology does circumvent fluorescence interferences, the success rate of Raman Spectroscopy for routine mineral identification can be expected to be comparable to powder XRD.