Determination of the Entropy of Sapphirine at High Temperatures

Julie Hollis (jhollis@glg.ed.ac.uk) & Simon L. Harley (sharley@glg.ed.ac.uk)

Department of Geology and Geophysics, University of Edinburgh, Grant Institute, West Mains Rd, Edinbugh EH9 3JW, United Kingdom

The determination of sapphirine at high temperatures has implications for the positioning of sapphirine-bearing reactions and as a consequence to the interpretation of, in particular, the metamorphic evolution of granulite-facies terranes. In this study, experiments on the reaction 4 Mg-Tschermak's orthopyroxene = sapphirine + 2 quartz were used to extract the entropy of sapphirine at high temperatures in the MgO-Al₂O₃-SiO₂ system. Five reversed experiments at 1.2 GPa and 1250°C to 1350°C were achieved, along with reversed experiments at 1325°C / 1.4 GPa and 1325°C / 1.6 GPa and an unreversed experiment at 1325°C and 1.0 GPa. Starting materials were quartz, natural sapphirine (Mg_3.60Fe_0.03Al_8.73Si_1.62O₂₀) and natural and synthetic orthopyroxenes (Mg_1.88Fe_0.02Al_0.20Si_1.89O₆ and Mg_1.70Al_0.59Si_1.71O₆ respectively). The compositions of sapphirine and orthopyroxene in isobaric experiments were used to determine the entropy change of reaction over the temperature interval 1250°C to 1350°C, assuming that the heat capacity change of reaction over this temperature interval was negligible. The best fit linear regression of RTlnK vs T using ideal activity models for sapphirine and orthopyroxene yielded an entropy change of reaction of -108 ± 25 JK^-1mol^-1. Using the non-ideal orthopyroxene activity model of Berman and Aranovich (1996) resulted in a calculated entropy change of reaction of -182 ± 39 JK^-1mol^-1. Different activity models for sapphirine had negligible effect on the entropy change of the reaction, as there was no discernible change in the composition of sapphirine with changing temperature. Using the thermodynamic data set of Holland and Powell (1998) and thermodynamic data from this study, the average entropy of sapphirine was calculated as 1460 ± 25 JK^-1mol^-1 using an ideal mixing model for orthopyroxene, or 1386 ± 39 JK^-1mol^-1 using the non-ideal mixing model of Berman and Aranovich (1996). These values differ by -5% and -10% from those proposed by Holland and Powell (1998).

Berman RG & Aranovich LY, Contrib. to Mineral. & Petrol., 126, 1-24, (1996).

Holland TJB & Powell R, J. Metamorphic Petrol., 16, 309-343, (1998).

Solubility of Sulfur in Silicate Melts Saturated with Metal Sulfide: The Influence of Temperature, Pressure and Silicate Composition

Astrid Holzheid (holzheid@uni-muenster.de)¹ & Timothy Grove (tlgrove@mit.edu)²

¹ Univeristät Münster, Institut für Mineralogie, 48149 Münster, F. R. of Germany

² Mass. Inst. of Tech., Dept. of Earth, Atmos. and Planet. Sci., Cambridge, MA 02139-4307, USA

This study explores the controls of temperature, pressure, and silicate melt composition on sulfide solubility in silicate liquids. Sulfide saturation experiments were carried out in piston cylinder apparati within a temperature range of 1300 to 1600°C and a pressure range from 0.9 to 2.7 GPa. Starting materials consisted of powder mixtures of metal sulfide (Fe_1.45S), elemental sulfur, and oxide mixes and/or silicate glasses. Basaltic to ultramafic silicate starting materials were chosen to explore the controls of silicate melt composition on the S-solubility in the silicate liquid. Liquid metal sulfide, silicate melt, high-Ca pyroxene, and/or olivine crystals coexisted during the experiments. The solubility of S in silicate melts in equilibrium with metal sulfide increases with increasing temperature at isobaric conditions and decreases with increasing pressure at isothermal conditions. The silicate melt structure also exercises a control of sulfide solubility. Increasing the degree of polymerization of the silicate melt structure lowers the sulfide solubility in the silicate liquid. These experimentally based observations can be applied to models of ore formation that rely on injections of primitive magma into evolving magma chambers and to metal-silicate separation scenarios. In order to precipitate magmatic sulfides, a magma must first cool and undergo fractional crystallization to reach S-saturation. Resupply of primitive magma into an open system magma chamber is therefore not likely to trigger S-saturation. The Earth's bulk mantle sulfur content is lower than the sulfide solubility limit of a S-saturated silicate liquid with an nbo/t-value similar to that of the Earth's mantle. Equilibrium core formation scenarios that involve separation of an S-bearing metal phase from a completely molten proto-mantle in a S-saturated system can be therefore excluded.