Paper No. 1
Presentation Time: 9:00 AM


KOCHELEK, Kristen1, MCMILLAN, N.J.1, MCMANUS, Catherine E.2 and DANIEL, David3, (1)Geological Sciences, New Mexico State University, Box 30001, MSC 3AB, Las Cruces, NM 88003, (2)Materialytics, LLC, P.O. Box 10988, Killeen, TX 76547, (3)Economics and International Buisiness, New Mexico State University, Box 30001, MSC 3CQ, Las Cruces, NM 88003,

Laser-induced breakdown spectroscopy (LIBS) analyses processed using multivariate techniques provide a quantitative method to determine gem provenance. The ability of the gem industry to determine provenance has financial, security, and societal implications. Using LIBS, a laser is pulsed to ablate a small amount of sample, creating a plasma. As the plasma expands and cools, the energized particles emit photons that are collected through optical fibers, diffracted, and recorded by a CCD camera. The resulting LIBS spectra are characteristic of the chemical compositions of the samples. Multivariate techniques are ideal for analyzing the volume of data contained in each LIBS analysis. This study acquired LIBS spectra from 569 ruby and sapphire specimens from 11 countries and 21 different localities. The spectra were classified using the multivariate technique Partial Least Squares (PLS) regression in separate matching algorithms for sapphires and rubies. Each algorithm is a sequence of PLS models in which the individual models compare the spectra from a locality of interest to the remaining spectra in the database. Success rates, determined by the percent of correct provenance classifications, are 98.8% (ruby) and 99.3% (sapphire) for country of origin and 98.2% (ruby) and 98.5% (sapphire) for deposit of origin. This work demonstrates that country or deposit of origin can be quantitatively determined with confidence.