Laser-ablation ICP-MS analysis of silicate and sulfide melt inclusions in an andesitic complex I: analytical approach and data evaluation

Quantitative microanalysis of entire silicate and sulfide melt inclusions by Excimer Laser-ablation inductively-coupled-plasma mass-spectrometry (LA-ICPMS) has been applied to extrusive and shallow intrusive rocks from the andesitic Farallon Negro Volcanic Complex (northwestern Argentina). Silicate melts are trapped in pyroxene, amphibole, plagioclase and quartz, and sulfide melts are trapped in amphibole. Details of the analytical approach and the quantification procedure are given and the results are evaluated to test the accuracy of the technique and the validity of the interpretation of analytical signals. Similar compositions of silicate melt inclusions trapped in truly co-precipitating minerals show that the quantification approach of melt inclusion compositions from LA-ICPMS signals through an internal standard is valid. This correspondence also shows that melt inclusions investigated in this study are not significantly influenced by the boundary layer around a growing crystal or by post-entrapment modifications. Post-entrapment diffusive re-equilibration only affected the Fe and Mg content of melt inclusions in mafic phases. Thus, melt inclusions are representative samples of the melt from which the host mineral crystallized, with regard to most major and trace elements. Sulfide melt inclusions (present as pyrrhotite with exsolution of Au and Cu in phases separated during cooling) were analyzed for their bulk Fe, Cu and Au content, and the abundance of these elements was quantified using a silicate glass as external standard. The validity of this calibration was tested by comparing electron microprobe analyses of Fe, Cu, Ni and Co in homogeneous sulfide minerals with LA-ICPMS results. Identical results within calculated uncertainty (one standard deviation of five to nine analyses, mostly between 1 and 5 wt% RSD) demonstrate that for these elements, measured element ratios are independent of the matrix using our analytical setup.