New method for initial composition determination of crystallized silicate melt inclusions

Silicate melt inclusions (SMI) trapped in minerals and carried up to the surface by volcanism are routinely studied in order to determine the pre-eruptive volatile budgets of volcanic systems. The volatile contents of SMI that are affected by post-entrapment processes, such as crystallization (PEC) and/or bubble nucleation during cooling, are generally difficult to interpret. Therefore, there is a general preference to select the melt inclusions that experienced minimal post-entrapment effects. Conversely, SMI can be homogenized at high temperature and quickly quenched. This method is controversial because heating may induce leakage of water from the silicate melt during experiments in addition to loss by diffusion. Here, we propose a new multi-disciplinary method to study crystallized SMI. This method involves the following processes: (1) X-ray microtomography for measuring phase proportions; (2) NanoSIMS for measuring C and H concentrations in glass as a proxy for CO2 and H2O; (3) EPMA (electron probe microanalyzer) for major and volatile element concentrations in all phases; and (4) thermodynamical modeling using Rhyolite-MELTS for volatile saturation determination. To test this approach, we studied the Los Humeros (East Trans Mexican Volcanic Belt) volcanic complex SMI. Three-dimensional (3D) petrographic observations of the studied SMI demonstrate that minerals inside the SMI were crystallized from the trapped silicate melt and were not trapped with the silicate melt (protogenetic). Accordingly, the mathematical reconstruction of the SMI is performed with the following equation: XSMI = Σ[VMineral(i) ∗ XMineral(i)] + [VGlass ∗ XGlass] + [VBubble ∗ XBubble]. Bubbles are associated with a heterogeneous trapping of a melt + fluid mixed phase. The PEC of large minerals inside SMI is consistent with slow cooling rates of the SMI, meaning that exsolved volatiles diffused toward the bubble in equilibrium with the melt before the glass transition. The volatile pressure determined in XSMI and using Rhyolite-MELTS is therefore equal to the pressure of the bubble. The volatile mass in the bubble is determined and added to the bulk volatiles of the silicate melt. This method can be used for a wide range of samples and may significantly improve our knowledge of volatile degassing processes.