Structural Implications of Water Dissolution in Haplogranitic Glasses from nmr Spectroscopy: Influence of Total Water Content and Mixed Alkali Effect

Burkhard C. Schmidt (burkhard.schmidt@uni-bayreuth.de)¹, Thomas Riemer², Simon C. Kohn³, Francois Holtz⁴ & Raymond Dupree⁵

¹ Department of Physics, University of Warwick, Coventry, CV4 7AL, United Kingdom, present address: Bayerisches Geoinstitut, Universität Bayreuth, D-95440 Bayreuth, Germany

² Department of Physics, University of Warwick, Coventry, CV4 7AL, United Kingdom, present address: IZKF/Klinische NMR-Spektroskopie, Universität Leipzig, 04103 Leipzig, Germany

³ Department of Earth Sciences, University of Bristol, Bristol, BS8 1RJ, United Kingdom

⁴ Institut für Mineralogie, Universität Hannover, 30167 Hannover, Germany

⁵ Department of Physics, University of Warwick, Coventry, CV4 7AL, United Kingdom

In order to study the effects of total water content and alkali variation on the structure of aluminosilicate glasses two series of glasses belonging to the ternary system Quartz (Qz) - Albite (Ab) - Orthoclase (Or) were synthesized and investigated with NMR spectroscopy. Series I consists of 7 glasses with normative composition Ab₃₉Or₃₂Qz₂₉ (AOQ) with water contents ranging from 0 to 6 wt%. Series II consists of dry and hydrous glasses (~2.0 wt% water) with 5 different compositions along the join Qz₃₇Ab₆₃ - Qz₃₄Or₆₆ (AQ-OQ) varying the alkali content (Na/K) at constant Si/Al ratio. All glasses were investigated with ¹H, ²³Na, ²⁷Al and ²⁹Si magic angle spinning (MAS) NMR.

¹H NMR results of AOQ glasses show the presence of molecular water and at least two types of OH groups of which one may be related to Al-OH. However, the small intensity of this signal indicates that only a small fraction of OH groups is present in this species. ²⁹Si MAS spectra show no change upon hydration, suggesting little variation of the Si environments although the large linewidth of the ²⁹Si signal may hide the presence of some Si Q³-OH. ²⁷Al NMR results show the decrease of the mean quadrupolar coupling constant (C_q) indicating a symmetry increase of the Al sites and thus the absence of significant amounts of Al-OH. Nonetheless, the evolution of C_q upon hydration suggests a correlation with OH concentration. The evolution of ²³Na isotropic chemical shifts upon hydration appears to be correlated with dissolved H₂O molecules.

NMR data for Series II show only a significant mixed alkali effect (non linear behavior) on NMR parameters for ²³Na but not for ²⁹Si or ²⁷Al. Therefore, these data suggest that the mixed alkali effect is rather related to the charge balancing cation than a modified aluminosilicate network.

Comparison of Spectroscopic Pressure Sensors for Hydrothermal Diamond-Anvil Cell Experiments

Christian Schmidt (hokie@gfz-potsdam.de)¹ & Martin Ziemann (marti@gfz-potsdam.de)²

¹ GeoForschungsZentrum Potsdam, Telegrafenberg D329, Potsdam 14473, Germany

² GeoForschungsZentrum, Telegrafenberg C7, Potsdam 14473, Germany

Difficulties in accurate pressure determination are perhaps the main limitation in the application of diamond-anvil cells to hydrothermal studies. A number of techniques for determining the pressure in these cells have been developed in the past. In this study, some spectroscopic pressure sensors were tested against each other for the same elevated pressures and temperatures. Water (which served as the pressure medium), quartz, ruby, and Sm-doped YAG were placed in a hydrothermal diamond-anvil cell. Spectra for each solid were recorded at pressures less than 20 kbar and temperatures between 23°C and 405°C using a Dilor XY Raman microprobe. The pressures were calculated from the frequency shifts (relative to the frequencies at ambient conditions) of the ruby and the Sm:YAG fluorescence doublets as well as from the frequency shifts of the 206 and 464 cm^-1 Raman lines of quartz. The data obtained so far show good agreement between these techniques at low temperatures (to 75°C). At elevated temperatures (above about 200°C), use of the 464 cm^-1 Raman line of quartz as a pressure sensor appears to be more accurate than application of the other techniques. This is caused by the smaller increase in linewidth of the 464 cm^-1 mode with temperature compared to that of the fluorescence doublets. The more pressure sensitive 206 cm^-1 Raman line of quartz is only calibrated near room temperature. At higher temperatures, application of the 206 cm^-1 mode as pressure sensor is hampered by the insufficiently known nonlinear temperature dependence of the frequency shift as a function of pressure.

Differential Melt Productivity of MORB, Greywacke and Pelite During Subduction

Max W. Schmidt (max@opgc.univ-bpclermont.fr) & Daniel Vielzeuf

CNRS - UMR 6524, 5, rue Kessler, 63038 Clermont-Ferrand, France

Major lithologies in subducting oceanic crust are MOR basalts, greywackes, and pelites, with different phase assemblages at crustal pressures. However, at 3-3.5 GPa zoisite and biotite decompose yielding almost identical assemblages of garnet-cpx-phengite-coesite±kyanite±rutile (at the wet solidus) and thus appearance and compositions of first melts in these lithologies are controlled by a single reaction (despite different phase proportions and compositions).

We performed experiments on average MORB, greywacke, and pelite (approx. 2 wt% H₂O) using octahedra of 25 mm edge length in multi-anvil apparatus. This assemblage allows simultaneous loading of three welded Au-capsules, thermal gradients are relatively low (<30°C).

As our charges are slightly water-saturated at the investigated pressures, small amounts of melt are present at the wet solidus (760-780°C at 4 GPa) and can be ascribed to a wet melting reaction. The first larger amounts of melt (10-15%) appear through a reaction phengite+quartz+cpx=garnet+liquid which occurs in all three lithologies around 950°C (at 4 GPa). At 950-1000°C, phengite+coesite are entirely consumed in basalt, cpx+coesite in greywacke, and cpx in pelite. At higher temperatures MORB proceeds melting through cpx=garnet+melt, greywackes through phengite=garnet+liquid, and pelites through phengite+coesite = garnet±kyanite+liquid. With temperature, melt percentages in basalt only slightly increase whereas in metasediments 30-40% of melt is formed at 1000 oC. Phase relations stay identical with pressure until phengite decomposes to K-hollandite (8-9 GPa in metasediments, 9.5-10 GPa in basalts). Preliminary experiments locate the major increase of melt percentages at 8 GPa near 1200°C.

Two scenarios involving melting of sediments but not of basalts can be imagined: Basalts (and peridotite underlying the oceanic crust) dehydrate and cause fluid-saturated 'flush'-melting in the greywackes/pelites (which are 50-150°C warmer than the basalts) or basalts are dry (apart from minor phengite) and fluid-absent melting could take place in the metasediments without significant melt extraction from the basalts.