MAJOR TO TRACE ELEMENT ANALYSIS OF MELT INCLUSIONS BY LASER-ABLATION ICP-MS: METHODS OF QUANTIFICATION

Abstract

Current techniques for the quantification of melt inclusion chemistry require that inclusions are compositionally homogeneous and that post-entrapment devitrification or crystallization onto the inclusion walls could be reversed by appropriate re-melting. Laser-ablation ICP-MS provides a technique by which single heterogeneous inclusions can be analysed, thus avoiding the above prerequisites. Because host mineral is ablated with the inclusion, quantification of the melt composition necessitates deconvolution of the mixed signal by an internal standard. This can be obtained in various ways, including: (1) a fixed, pre-determined, concentration of a given element in the melt; (2) whole rock differentiation trends in a given igneous suite; (3) a constant, measured, distribution coefficient between the host and the inclusion melt and (4) determination of the volume ratios between the inclusion and total ablated volume. These four approaches were tested on a large set of cogenetic inclusions from a single plagioclase crystal in a rhyodacitic intrusion. Results suggest that quantification through whole rock differentiation trends is the most widely applicable, the most accurate and the least time-consuming technique, provided that the resulting data are critically interpreted with regard to the underlying assumptions. Uncertainties on the calculated element concentrations in the inclusions depend on the mass ratio between the melt inclusion and the host for a given ablation. They are of the order of 10% if the melt inclusion contributes more than 20% to the bulk analytical signal of a particular element. Calculated limits of detection for spherical 10 μm melt inclusions are of the order of a few ppm for elements strongly enriched in the melt relative to the host crystal. Concentrations in the melt inclusions can be determined even for elements enriched in the host mineral, but in this case uncertainties and calculated limits of detection increase with the concentration in the host. The uncertainty on the melt composition from a set of cogenetic inclusions can be commonly decreased by calculating of an uncertainty-weighted average of the concentration and their uncertainty.