Elemental fractionation during LA-ICP-MS analysis of silicate glasses: implications for matrix-independent standardization

Abstract

To determine the degree to which matrix affects elemental abundance measurements in LA-ICP-MS analysis, and describe these effects in term of elemental properties, we measured elemental sensitivity ratios (ESRs) from a multi-element solution and from the laser ablation of a range of silicate reference materials (SRM 61X series, MPI-DING glasses, USGS basalt glasses). For all analyses, ESRs depend strongly on element mass and first ionization potential (FIP), as expected. For laser ablation of the nearly transparent SRM 612, ESRs show an additional dependence on elemental temperature of condensation (T_c), such that ESRs of refractory elements are lower than predicted by their mass and FIP dependences. To explore the relationship between this volatility dependence and matrix transparency, we exploited the range of transparency available in the SRM 61X series. With increasing transparency of glasses, LA-ICP-MS analyses showed progressive volatility dependence, with measured concentrations of refractory elements being lower, and measured concentrations of volatile elements being higher, than reported concentrations. For the most transparent glass (SRM 614), concentration offsets from the reported concentrations of the refractory and volatile elements were ∼20%; for SRM 612 offsets were ∼10–15%. We show that this effect is minimized, and possibly negated, by increasing laser energy output. As the offsets for ESRs of refractory elements are uniformly lower in more transparent SRM 612 relative to SRM 610 (11% ± 4%; 1.1 mJ laser energy), successful matrix-independent standardization of refractory elements is possible if the internal normalizing element is approximately the same T_c as the elements to be reported (e.g., Ca and REEs). Volatile elements, however, do not produce uniform ESRs in transparent samples and matrix-independent standardization of volatiles in transparent samples with a 213 nm wavelength laser ablation system will include element-specific systematic errors of the order of tens of percent. All non-transparent glasses analyzed (MPI-DING glasses and USGS BHVO-2G and BCR-2G) show no resolvable matrix-dependent elemental fractionation and may be used interchangeably with an accuracy of better than 10%. In non-transparent samples, the greatest limitation to accurate LA-ICP-MS standardization is the uncertainty in reported reference material concentrations. Matrix-independent standardization of LA-ICP-MS analysis of non-transparent silicates and ceramics, then, is an accurate and viable tool that can best be employed by avoiding calibration with transparent standards, unless using an internal normalizing element of similar T_c to that of the elements to be reported. LA-ICP-MS analysis of transparent silicates and ceramics can be performed at the highest laser energy possible to minimize volatility effects, and calibrated with an internal normalizing element of similar T_c to that of the elements to be reported, using both transparent and non-transparent standards.