Element partitioning in fluid/melt systems has been addressed in a large number of studies. Most of these have employed bulk analytical techniques to determine of the compositions of the coexisting phases. In such studies the challenge of demonstrating good yields and excluding or correcting for quenching effects can be great. A complementary approach is to trap the fluid phase within the quenched glass as fluid inclusions and directly microanalyse the contents. We are concentrating on the area of fluid-melt partitioning during the degassing of silicic magmas. Using methods for the synthesis and physical characterisation of fluid inclusions in glass developed recently, synthetic aqueous halide-bearing fluid inclusions in a haplogranitic glass matrix (HPG8) are produced by rapidly quenching a fluid-saturated melt. These samples are then micro-analyzed by laser ablation ICP-MS. LA-ICP-MS serves as an effective and simple tool for the analysis of single fluid inclusions, and much work has been done on fluid inclusions in natural quartz samples using this technique.
Initial experiments have been performed with "doped" HPG8 (doped with 20 elements ranging in atomic number from Li to W, each at ca. 700 ppm) and NaCl-solutions (up to 10 wt% NaCl) or with undoped HPG8 together with doped NaCl-solutions (up to 4 mol% NaCl, up to 5 elements, each at 1000 ppm). The syntheses have been performed in hydrothermal cold-seal vessels at 2 kbars and 850°C. Quenched samples are polished to a thickness of up to 250 µm, such that the fluid inclusions are clearly visible in transmitted light microscopy.
The first investigations using LA-ICP-MS have been performed with a PQ2+"S" (Fisons Instruments) coupled with a quadroupled Nd:YAG laser at Memorial University, St. John's and with the ELAN 6000 (Perkin Elmer) in combination with an excimer laser at the ETH, Zürich. A series of analyses was performed to evaluate the effects of varying inclusion size (20-150 µm in diameter) and depth as well as analytical parameters (laser energy, repetition rate, diameter of the laser) in order to optimize the analytical protocol. For internal standardization we have chosen Cl because the bulk salinity can be determined independently by heating/cooling stage fluid inclusion microthermometry. For glass matrix calibration we use NBS-standards (612 and 610) and for the fluid inclusions, element aqueous standard solutions of 50 ppm. Preliminary investigations yield time-resolved mass spectra allowing a clear separation between the matrix and the fluid inclusion signals. We are presently refining the internal standardization and the calibration procedures.