Complications with Using Natural Minerals as Microbeam Standards: Pyroxenes
Abstract
In the 1970s, Gene Jarosewich and colleagues at the Smithsonian National Museum of Natural History (NMNH) Department of Mineral Sciences performed the invaluable service to the microanalytical community by culling their collections for minerals and glasses to be used as electron probe microanalytical standards. These materials have been widely distributed worldwide and are widely cited. Cautions about these materials, given by Jarosewich, however, many times have been overlooked by users eager to acquire data. In evaluating some experimental orthopyroxene (opx) crystals in Nov 2011 using the NMNH pyroxene stds at UW, it became obvious that the grains in our standard mount differed in composition from the published (wet chemical) analyses. It was impossible to achieve both good totals and good stoichiometries. This project resulted. Vials with multiple grains of the 4 NMNH px stds were received, mounted and probed using typical EPMA conditions (15 kV, 20 nA), to determine (1) inter and intra grain heterogeneity and (2) possible errors in stated compositons. In the process, it was "discovered" that the NBS/NIST K411 and K412 glasses are extremely homogeneous, accurately characterized, and are optimal pyroxene standards. Data were acquired on Kakanui Augite (54 grains, 270 pts); Natural Bridge Diopside (49 grains, 245 pts); Johnstown Hypersthene (17 grains, 85 pts); and NMNH Cr-Augite (24 grains, 120 pts). Some results: Johnstown Hy: 10% of Si values and 20% of Mg, are outside of 3 sigma BUT if all 84 points from 17 grains utilized, the average (=wet chem analysis) is verified. So it is a poor standard potentially if only 1-3 grains used. On other hand, Kakanui Augite showed tighter spread (6% of Si, 4% of Mg, 2% of Ca outside 3 sigma), but the experimental compositions (using K411/412) differed from the published, i.e. 3% (rel) difference for MgO (16.15 vs 16.65 wt%), 2% (rel) difference for CaO (16.14 vs 15.82 wt%), 1.3 % (rel) difference for SiO2 (50.08 vs 50.73 wt%). The NMHNH Cr-augite has compositions that match very closely with the experimentally determined ones, except the Al differs by 5% (relative) to the one using K412. The Natural Bridge Diopside appears to have a SiO2 published composition 1 wt% lower than that determined experimentally, and CaO slightly lower (0.3 wt%), whereas the MgO is dead on. In conclusion, uncritically trusting a handful of analyses from 1-3 grains of these natural minerals can result in erroneous EPMA results. Jarosewich explicitly warned against this in his 1980 publication attached to the composition of these natural mineral and glass standards: "all the reference samples are very homogeneous provided a reasonably large number of counts are taken on a reasonably large number of grains. In practice, however, fewer counts and grains are normally used for standardization, and under these circumstances a grain having a slightly different composition may influence the microprobe results adversely."
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2012
- Bibcode:
- 2012AGUFM.V23C2827F
- Keywords:
-
- 1094 GEOCHEMISTRY / Instruments and techniques;
- 3694 MINERALOGY AND PETROLOGY / Instruments and techniques;
- 8194 TECTONOPHYSICS / Instruments and techniques;
- 8494 VOLCANOLOGY / Instruments and techniques