Development of Preliminary New Reference Concentrations for ATHO-G Major Elements and Lipari Obsidian ID-3506 Trace Elements

Homogeneous glass reference materials are widely used for instrument standardization and data quality control for EPMA, LA-ICP-MS, and other analytical techniques.For these purposes, accurate multi-element characterization is important, ideally including major, minor, and trace elements and for some materials isotopic ratios.Such robust reference concentrations are typically determined by compiling data from multiple methods and laboratories.Major reference glasses include those prepared by NIST, USGS, Smithsonian, MPI-DING, and others.Among these, the most widely-used materials with natural rhyolitic compositions include ATHO-G [1], Lipari obsidian .As these glasses may subject to alkali element migration during EPMA, depending on glass composition and analytical conditions, it is important that any data utilized in the development of reference concentrations be carefully evaluated for effects like sodium-loss as this can compromise accuracy.The primary objectives of this work are to re-examine the major element composition of ATHO-G and to extend the characterization of Lipari obsidian ID-3506 to trace elements.For several other reference glasses, selected published and new data are also compiled.ATHO-G is a re-melted rhyolite glass derived from an Icelandic obsidian and is part of a set of eight glasses of geological composition referred to as the MPI-DING reference glasses [1].Between 50 and 100 g of material was prepared for each MPI-DING glass.ATHO-G has been characterized for most major to trace element abundances and select isotope ratios to produce compositional reference values, information values, and uncertainties at 95% confidence [1].At least three different samples of natural obsidian from the island of Lipari, Italy have been used as reference materials for microanalysis beginning with the pioneering work of [4] who utilized EPMA to characterize volcanic tephra glasses in 1968.Numerous other samples have also been analyzed for archaeological obsidian sourcing.Lipari obsidian ID-3506 is the most recently established of these materials for microanalysis [2].Originally a 250 g sample (∼1/2 of the original) was supplied by the Harvard Mineralogical & Geological Museum and chipped for initial distribution and analysis.Most chips were returned to Harvard, subsequently transferred to the International Association of Geoanalysts, and now are distributed through IAGeo Limited.Lipari obsidian ID-3506 has been well-evaluated for major and minor element concentrations [2], but uncertainty estimates and recommended trace-element concentrations were not prepared.For ATHO-G, the complete multi-method major element data of [1], data from the GeoREM database, and other published data have been compiled.The added data consists primarily of EPMA, one XRF, and several LA-ICP-MS analyses.For the Lipari glass, all identified trace-element data has been compiled and added to the previous major-element compilation of [2].In addition, ATHO-G, Lipari ID-3506, and more than 30 other reference glasses have been repeatedly analyzed by EPMA at Concord University using a family of relatively high-precision glass analytical methods for 15 to 20 major, minor, and trace elements.Importantly, these methods incorporate the time-dependent intensity (TDI) correction and spectral interference correction procedures in Probe for EPMA [5].Full EPMA method details [e.g.6,7] are reported using the method template available from the Tephra Community portal at the EarthChem data repository, and all results will be archived there.The several multistandard analytical sessions also make it possible to compare analyzed and reference compositions across a suite of glasses which have been harmonized to a multi-standard consensus on a per-session basis.Examination of time-dependent intensity effects reveals that ATHO-G is far more e-beam sensitive than most other widely used reference glasses (e.g. Figure 1).Not surprisingly, the compiled ATHO-G data of [1] shows a tail of lower sodium concentrations by EPMA methods, whereas most other methods and TDI-corrected EPMA show higher concentrations (Figure 2).To obtain accurate reference concentrations, it is necessary to filter out EPMA analyses compromised by sodium-loss.Doing so yields a higher consensus for Na2O (4.25 to 4.37 wt%) and a lower one for SiO2 (74.9 to 75.2 wt%).