SS14: Tantalum, Tin and Tungsten at the Margins
Organizers / Organisateurs: Lee Groat & Bill Mercer
Room / Salle: A2402
Time: 4:40 PM
Presenter: Catherine McManus
Fingerprinting and provenance determination of gold, coltan, cassiterite, and wolframite: Multivariate analysis of Laser-Induced Breakdown (LIBS) spectra
McManus, C., C.McManus@materialytics.com, Dowe, J., Materialytics, PO Box 10988, Killeen, TX 76547-0988, USA, and McMillan, N.J., Dept. Geol. Sci., New Mexico State University, Box 30001, MSC 3AB, Las Cruces, NM 88003, USA
The strategic elements gold, tantalum, tungsten, and tin are used in the manufacture of many high-tech items, making them sufficiently valuable to be treated as conflict minerals in some areas. Thus, it is important to be able to identify the provenance, either country of origin or mine of origin, of the ore minerals gold, coltan (columbite-tantalite), wolframite, and cassiterite. In this study, we determined the mine of origin for 162 samples of gold from six mines in five countries (Brazil, Italy, North Korea, United States, Unknown), 264 samples of coltan from 12 mines in seven countries (Brazil, Canada, Democratic Republic of Congo, Mozambique, Norway, Sweden, United States), 90 samples of wolframite from three mines in two countries (Great Britain, United States), and 150 samples of cassiterite from five mines in five countries (Bolivia, Great Britain, Namibia, North Korea, and United States). This study was conducted using the premise of Quantagenetics®, that all materials, natural or synthetic, contain a record of their origin and history; this unique fingerprint can be acquired by using Laser-Induced Breakdown Spectroscopy (LIBS). LIBS spectra were acquired by collecting photon emission from a cooling laser-induced plasma. These spectra contain an immense amount of information, reflecting the compositional, isotopic, and structural properties of the material analyzed. A multivariate statistical method was applied to single-shot spectra which compares a spectrum, pixel by pixel, to spectra of the reference sets of each location using standard full cross-validation. Success rates are calculated as the percentage of the spectra comprised of the true positive and the true negative results. For gold, every spectrum was correctly classified, with a 100% success rate. Success rates for coltan varied between 96.6% and 100.0%, averaging 98.7%. Three coltan mines in the study are from the Democratic Republic of Congo; these mines have success rates of 99.6%, 100.0%, and 97.0%. Success rates for cassiterite range between 91.3% and 95.3%, averaging 93.6%. Success rates for wolframite range between 97.8% and 98.9%, averaging 98.5%. These highly accurate results, coupled with the ease and speed of LIBS analysis, demonstrate the usefulness of this technique for fingerprinting strategic minerals.