Conference Agenda

The sulfur (S) budget in silicate melts is a crucial factor in crustal processes from ore formation to volcanic outgassing, yet its behavior remains incompletely understood. To investigate how S oxidation states depend on melt composition and pressure in lower-crustal arc magmas, we performed piston-cylinder experiments on sulfur-saturated, hydrous dacitic melt (corresponding to Pinatubo dacite) at 700 MPa and 950 ^°C. Using various solid-state buffers, we explored an fO₂ range from ΔFMQ -2 to +4.5. To address the compositional effect, we compared Fe-bearing and Fe-free synthetic glasses. Results reveal that S solubility in the dacitic melt range from 68 to 1328 ppm with increasing fO₂ and there is no significant difference between the Fe-bearing and Fe-free melts. S K-edge XANES results indicate the sulfide-sulfate transition occurs between ΔFMQ+1 and +2 (at 950-1000 °C, 300-1000 MPa) with a negligible pressure effect. Importantly, compared to basaltic melts, the sulfide-sulfate transition in dacitic melt is shifted to higher fO₂ by +0.4 log unit. Moreover, dissolved water content significantly influences both sulfur content at sulfide saturation (SCSS) and sulfate saturation (SCAS). Applying these findings to volatile degassing models for the 1991 Mt. Pinatubo eruption, the observed decrease in the S⁶⁺/ΣS ratio in the melt indicates the reduction of S⁶⁺ to S²⁻ during decompression and degassing as fO₂ decreases. Our findings imply that dacitic melts in the lower crust could dissolve ~1300 ppm S at fO₂ > ΔFMQ +1, even if fO₂ decreases during degassing.

Lithium and boron are light metals with contrasting properties in the silicate melt network and their diffusion rates differ by several orders of magnitude in melts. Nevertheless, the complex processes during pegmatite crystallization can lead to an enrichment of both, Li and B.

To investigate the role of a fluid phase in the transport of Li and B in pegmatites, we developed a novel experimental setup simulating their transport from a Li- and B-bearing source to a Li- and B-free sink melt. These melts were physically separated by a porous filling material (quartz or zircon powder), so that the transport of Li and B between the melts only occurred via a hydrous fluid phase in the pore space. Experimental runs were carried out in a rapid-heat/rapid-quench cold seal pressure vessel at 100 MPa. Among other parameters, the experimental duration (1–96 h) and the temperature gradient between the two melts (850–850 °C or 830–770 °C from capsule top to bottom) were varied.

The different Li or B concentrations in the sink melt after the experiments indicate a significant influence of the experimental duration and the temperature distribution on the effectivity of transport. The sink melt shows an enrichment of the heavier ⁷Li, which can be best explained by equilibrium isotope fractionation between silicate melt and fluid. For B, the determined melt–fluid isotopic fractionation is smaller than our analytical uncertainty (± 4.7 ‰), which is likely related to the similar bonding environments of B in melts and fluids.

Thermodynamics is written in scalars [P, V, T]; it applies only to isotropic loading. Anisotropic deformation has so far been a part of continuum mechanics (Euler-Cauchy). However, elastic deformation is by nature a change of state in the sense of the First Law of thermodynamics; thus a theory of elasticity must start with the First Law and the EOS. Euler-Cauchy is many decades older, it is therefore obsolete since 1847. Its relative success is restricted to orthogonal boundary conditions where the error does not show. For simple shear Euler-Cauchy fails systematically and completely. A new approach (1), thermodynamics in vector field form [f, r, T], duplicates the results of Euler-Cauchy where they look useful; but it also delivers reliable explanations for the simple shear phenomena which were so far enigmatic. It is found that all anisotropically loaded solids are constitutionally expanded (Poynting effect), relative to the ideal isotropic loading state. The cause is the work done by shear forces, which the Euler-Cauchy theory is unable to consider. This effect, called shear dilation, amounts to a new thermodynamic state function that shifts the loading path in PV-space towards higher-than-ideal volumes. Consequently, the phase boundaries must shift towards lower PT conditions. The effect is known experimentally: coesite is observed deep in the qz field in simple shear; sheared qz melts at 1350°C instead of 1720°C; sheared olivine melts up to 400° below equilibrium T_M.

(1) Int J modern Physics B 22, 2617, 2008

Pre-eruptive H₂O of rhyolites are difficult to determine because of loss during ascent and degassing. Quartz-hosted melt inclusions (MIs) are potential recorders of pre-eruptive H₂O contents. The main uncertainty is whether the MIs if some is lost via decrepitation or diffusion through the host crystal. We investigated this by performing dehydration experiments and measuring both H₂O concentration and the deuterium-to-hydrogen ratio (D/H) by SIMS. Fractionation of D/H with increasing dehydration is an unambiguous signal of diffusive loss through the host crystal. Experiments were conducted by heating single crystals of Bishop Tuff quartz in a Vernadsky stage at 750 °C for up to 336 hours. Within the first ~12 hours, some of the inclusions formed vapor bubbles while others remained bubble free. The concentration of H₂O in MIs containing vapor bubbles decreased by ~50 %, but with no corresponding increase of D/H. This loss is due to decrepitation. A few of the bubble-free MIs, however, show decreasing H₂O and increasing D/H after 120 hours, indicating diffusive loss. Melt invlusions in an experiment conducted 336 hours, however, show little-to-no evidence for diffusive loss. We interpret this as attributable to interactions among H₂O being diffusively lost from multiple MIs within a single grain. When a quartz grain contains a small number of inclusions diffusive loss can be detected, whereas in a grain containing a large number of inclusions it cannot. Our results demonstrate that quartz grains containing multiple, bubble-free inclusions are likely to provide the most accurate record of pre-eruptive H₂O.