WHAT IS THE “REAL” COMPOSITION OF THE LIQUID? COMPARING GLASS TRACE ELEMENT DATA FROM LASER ABLATION AND SOLUTION ANALYSES

Petrogenetic modeling of volcanic rocks often involves assessment of the trace element budgets of mineral and glass phases. Laser ablation inductively coupled plasma source mass spectrometry (LA-ICPMS) offers the ability to quantitatively analyze trace element concentrations in situ for samples of 10^-3 to 10^-5 milligrams. Spatial resolution of this magnitude allows for the characterization of fine scale heterogeneities in glasses, but does not necessarily illuminate the significance of such heterogeneities.

This study compares thin section LA-ICPMS data to whole rock and glass solution data from vitrophyre samples of high temperature rheomorphic rhyolites in the Central Snake River Plain volcanic province. 1.0 mg and 200 mg aliquots of glass fragments (< 1 mm diameter) were hand picked from apparently fresh rhyolite vitrophyre to exclude phenocryst fragments. These were digested using a conventional, open-vial, mixed-acid dissolution method in order isolate the glass composition and exclude less soluble mineral phases. Solutions were analyzed using an Agilent 7700x ICPMS. A Thermo-Finnigan Element 2 ICPMS coupled to a New Wave UP213 laser was used to collect LA-ICPMS data. Troughs of 12-20 microns wide by 500-800 microns long were ablated to an approximate depth of 10 microns on polished thin sections.

The data reveal significant concentration discrepancies between the different types of analyses (LA-ICPMS, glass solutions, and whole rock) for elements that are compatible in observed phenocrysts. For Ba, Sr, and Zr, concentrations in the glass aliquots are generally higher and more variable than in the laser data, indicating the presence of zircon and feldspar microcrysts in the glass separates. Microcrysts form prior to and during eruption, emplacement, and cooling of volcanic deposits. Therefore, it is presently unclear what type of data best represents the true compositions of igneous glasses. This represents a source of uncertainty in accurate petrogenetic modeling of volcanic rocks using trace element data.