V41C-1398 0800h
Partitioning of Trace Elements Between Clinopyroxene and Peridotitic Partial Melts: Implications for the Mechanism(s) by Which Piston-Cylinder Partial Melting Experiments Equilibrate
Earlier piston-cylinder experiments in our laboratory produced a collection of mantle melting run products at 1.0 GPa that have now been analyzed by ion probe for selected REE, Ti, Cr, Rb, Sr, Y, Zr, and Nb. The glass phase in 18 run products, representing melt percentages of ~2-20 wt.%, yielded excellent data that were inverted to yield the first estimates ever of clinopyroxene/melt distribution coefficients, Ds, derived from direct peridotite partial melting experiments. Uncertainties were estimated with a Monte Carlo method. For the REE and Y, these Ds were then compared to Ds calculated with the widely-used model of Wood and Blundy (1997) and the two sets overlap at the 2sigma level in 123 of 128 cases (96%). This indicates to us that the experiments analyzed here are well-equilibrated with respect to major and trace element distributions and it suggests that the cpx/melt Ds derived here for other elements, not modeled by the Wood and Blundy formulation, are probably also correct for peridotite melting to within their 2sigma uncertainties. The degree to which our experiments appear to have equilibrated seems at odds with recent measurements of the diffusivities of REE in diopside which suggest that relatively small percentages of our starting mineral grains should have equilibrated diffusively. Instead, we suggest that equilibration occurs much more rapidly through the processes of recrystallization and grain coarsening, accomplished through dissolution and reprecipitation. This suggestion is supported by the observation that our final grain sizes are typically 5-10 times larger than the ~10-20 mm starting sizes, indicating that substantial mass transfer occurred in our experiments, probably mediated by the melt phase in which diffusion is faster. Preliminary results from a time series of experiments suggests that disequilibrium melting during the first day of long-duration (3-5 days) experiments results in excess melt which is subsequently reprecipitated as overgrowths on existing crystal substrates.
V41C-1399 0800h
An Empirical Calibration of an Al-in-olivine Geothermometer for Mantle-derived Materials
The concentration of aluminum in olivine from mantle peridotites is variable and strongly temperature dependent, and can therefore be used as a geothermometer. A suite of fresh mantle peridotite xenoliths from the Kaalvallei kimberlite (South-Africa) was used for calibration of the thermometer. The samples contain olivine, opx, cpx, garnet $\pm$ spinel. Aluminum contents of the olivines, determined by LA-ICP-MS, range from 8-140 ppm, with a few anomalous values up to 310 ppm. Mg\# lie between 91.5 and 93.3. {\it P-T} conditions for the peridotites were estimated with the Al-in-opx geobarometer and cpx-opx geothermometer [1], and plot on a conductive geotherm of ca. 40 mW/m$^{2}$. The temperature range is 900-1380$\deg$C. Only samples plotting on the geotherm were considered, as the others showed disequilibrium features. The expression for the thermometer is: {\it T}$_{Al-in-ol }$($\deg$C) = 11390 / [ 11.88 - ln(ppm Al) ] - 273 with an average residual of 12$\deg$C. As the compositional range of mantle olivine is very small, no correction for major chemical components is necessary. In addition, no correction for Al activity of the system is necessary, as long as an Al-saturated phase such as garnet is present. Combined with the Ca-in-olivine barometer [2], the new thermometer has the potential to determine {\it P-T} conditions of single olivines. As olivine is an abundant component of heavy mineral separates from kimberlites, it could serve as a new tool for diamond exploration. Vanadium and Cr show similar temperature-dependent variations to Al, but to a lesser degree, and would therefore yield less accurate geothermometers. In addition, partitioning of these elements is sensitive to variations in oxidation state. The pressure dependence of the thermometer is the subject of future research. Considering the significant pressure effect on Ca and Ti partitioning into olivine, it is recommended that the Al-in-olivine thermometer in its current form is applied to rocks derived from comparable (cratonic) geotherms only. [1] Brey and Kohler, 1990, J Petrol, 31, 1353-1378 [2] Kohler and Brey, 1990, GCA, 54:2375-2388
V41C-1400 0800h
Crystal chemistry of NaMgF3 perovskite at high pressure and temperature
The crystal structure of NaMgF3 perovskite (neighborite) is studied at 4GPa and temperatures up to $1000\deg$C using Rietveld structure refinement method and In situ synchrotron X-ray powder diffraction collected using monochromatic radiation. The orthorhombic (Pbnm) to cubic (Pm3m) transition is observed when temperature increases from $900\deg$C to $1000\deg$C. Structure refinements show that the ratio of polyhedral volumes of the A- and B-sites (VA/VB) of the orthorhombic phase increases with temperature, approaching the ideal value (5) for the cubic structure. However, this ratio becomes smaller at 4GPa compared to the result from previous studies at the same temperature but ambient pressure, indicating that pressure makes it more difficult to transform from the orthorhombic phase to the cubic phase in this kind of perovskite.
V41C-1401 0800h
Relationship between grain/interphase boundary energies and phase diagrams
We have examined the relationship between the type of phase diagram and relative grain boundary to interphase boundary energy in order to answer a long-term unsolved question: Why do crystals prefer different neighbors? For minerals as well as metals and ceramics, dihedral angles formed at triple grain junctions involving two different crystalline phases (A and B) are commonly $<$$120\deg$. Based on the interface tension balance equation 2cos($\theta$/2) = $\gamma$$_{gb}$/$\gamma$$_{int}$ (Eq. 1) -- where \theta is dihedral angle and $\gamma$ is either grain or interphase boundary energy -- a value of $\theta$ $<$$120\deg$ indicates that the interphase boundary energy is lower than of the grain boundary energy. Grain boundary energies of metals correlate linearly with latent heat of fusion and/or melting temperature. Systems composed of two crystalline phases often have a eutectic point at a lower temperature than the melting points of pure phases; accordingly, the latent heat of fusion is lower for the two-phase system than for the individual component phases. Therefore, we predict that the interphase boundary energy is lower than that of the grain boundaries in the case of a eutectic system; however, the opposite holds for monotectic systems. We examined grain boundary versus phase boundary energies for binary systems in which one phase is Ag and the other is Fe, Co, Ni, Cu, Ge, or Si. The systems formed from Ag and any of the former three elements are monotectics, while the systems composed of Ag and any of the latter three elements are eutectics. To obtain binary polycrystalline materials, we sintered the powders of 5-10 $\mu$m of Ag plus either Fe, Co, Ni or Cu at vacuum conditions. We also made amorphous ribbons of Ag-Si and Ag-Ge by the rapid rolling technique, which we then annealed for 10-15 h to cause crystallization and grain growth. After surface etching, we measured dihedral angles with a field emission SEM. Also, we used dihedral angle data for synthetic and natural mineral assemblages of quartz-feldspar and olivine-orthopyroxene. In terms of Eq. (1), the interphase boundary energy is lower than grain boundary energies in all of the eutectic systems. In contrast, the interphase boundary energy ($\gamma$$_{AB}$) is higher than at least one grain boundary energies ($\gamma$$_{AA}$ and/or $\gamma$$_{BB}$) in all of the monotectic systems. Based on published molecular dynamic simulations, high-resolution microscopy observations, and grain boundary segregation analyses, a general high-angle grain boundary assumes a melt-like structure with a width of 1-3 atomic layers; this region cannot be defined as a thin boundary phase. Extending this concept, we can predict the type of phase diagram for any multi-phase system simply by measuring the dihedral angle. If a system changes from eutectic to monotectic or vice versa by adding trace elements such as water, this transition will be evident from the dihedral angle.
V41C-1402 0800h
Effect of Titanium on REE and HFSE Partitioning Between Garnet and Melt
Garnet is a strong fractionator of trace elements and plays an important role in the petrogenetic history of planetary interiors at high pressure. In order to model petrogenetic processes that operate within terrestrial planets accurately, it is important to understand how garnet partitions rare earth and high field strength elements. Here we assess the influence of Ti on garnet-melt trace element partitioning with a view both to constrain important crystal-chemical effects and to evaluate possible roles for garnet in lunar petrogenesis. Experiments were performed at $\sim$5 GPa and $1650-1675\deg$C in a Walker-style multi-anvil high pressure apparatus using an Apollo 14 black picritic glass composition ($\sim$17 wt% TiO$_{2}$) to assess the effect of Ti on garnet partitioning. These experiments were also designed to examine the possible presence of garnet in mare source regions. Experimental charges were analyzed for major and trace elements by EPMA and SIMS, respectively. D-values measured in this study using the Apollo 14 black Ti-rich composition are consistently higher than those measured by Draper et al. (2004, LPSC XXXV:1297), who used Apollo 15 green C glass ($<$0.5 wt% TiO$_{2}$). D vs. ionic radii are well-described for the trivalent cations by the lattice-strain partitioning model of Blundy and Wood (1994, Nature 372:452), with D$_{0}$ = 2.27 $\pm$ 0.40, E = 159 $\pm$ 58 GPa, and r$_{0}$ = 0.879 $\pm$ 0.044 $\AA$ (r$_{2}$ = 0.957). For comparison, this model applied to the low-Ti experiments of Draper et al. (2004) yields D$_{0}$ = 2.93 $\pm$ 0.25, E = 572 $\pm$ 40 GPa, and r$_{0}$ = 0.926 $\pm$ 0.005 $\AA$ (r$_{2}$ = 0.996) at $\sim$3.5 GPa. Both these fits show significant mismatch to the partitioning predicted by the formulations of van Westrenen et al. (2001, CMP 142:219), as previously shown for Fe-rich systems by Draper et al. (2003, PEPI 139:149). Use of our D-values (for rare earth and high field strength elements in batch-melting models) provisionally supports the hypotheses of Neal (2001, JGR 106:27865) and Neal and Shearer (2004, LPSC XXXV:2135), who proposed that garnet was present in the source regions of lunar mare basalts. The results of these calculations are similar to those using data from garnet-melt partitioning in very Ti-poor systems (Draper et al. 2004).
V41C-1403 0800h
Liquidus Phase Relations in Pyrolite at Pressures of the Transition Zone
According to some models of planetary accretion, the Earth may have been largely molten as a consequence of numerous impacts during its accretional history, leading to the formation of a global magma ocean. Subsequently, cooling of this putative magma ocean from the bottom up could have led to the stratification of the mantle by fractional crystallization. To constrain the geochemical evolution of mantle after crystallization of the magma ocean, knowledge of the pressure-temperature stability field of mantle minerals above solidus conditions is necessary. Recent experimental studies indicate that the liquidus and solidus of anhydrous fertile peridotite at pressures of the transition zone are at a considerably lower temperature than previously estimated. In view of these recent discrepancies, we reexamine the liquidus phase relations in pyrolite in the 10-14 GPa pressure range. Experiments were performed in a multi-anvil device, using short Re- and C-capsules and thick zirconia insulators to minimize thermal gradients. Preliminary results reveal that the liquidus is located between 2050-2100 C at 14 GPa and between 1950-2000 C at 10 GPa. These brackets are significantly lower than reported earlier at identical pressures. Furthermore, we find that Ca-poor pyroxene is stable in the melting interval at 10 GPa, i.e. about 2-3 GPa higher than previously observed. Experiments to clarify the phase relations between orthopyroxene, olivine and garnet at pressures of the transition zone are in progress.
V41C-1404 0800h
Preliminary Experiments and Determination of the Thermal Gradient in a 12.7 mm CaF$_{2}$ Furnace Assembly, Humboldt State University Piston-Cylinder Laboratory
A 12.7 mm piston-cylinder laboratory has been established at Humboldt State University. A series of double-thermocouple (DTC) experiments were performed to measure the thermal profile of the furnace assembly following the procedures of Pickering et al. (1998, Amer. Min.). Furnace assemblies consist of a 304 stainless base plug, CaF&_{2}$ sleeve, straight-walled graphite heater tube, crushable MgO inner parts, and lower graphite plug and ring which allow for extrusion of the graphite heater tube during shortening of the sleeve and MgO pieces. Careful measurement of pre- and post-run assembly parts indicate an average 30-$35%$ shortening of the assembly. DTC results show a thermal peak that is displaced $\sim$2.0 mm above the center of the effective furnace, defined as the length of inner MgO pieces post-run. This offset is in the same direction (upward, toward base plug), but slightly less than the offset described by Pickering et al. (1998). A secondary measure of the thermal profile using spinel growth via reaction between MgO and Al$_{2}$O$_{3}$ assembly parts (e.g., Watson et al., 2002, CMP) is underway. A single partial melting experiment was performed at 1.0 GPa and $1330\deg$C for 72 hours using intermediate peridotite starting material INT-A in a graphite-lined Pt capsule with vitreous carbon spheres as a melt sink. Phase compositions were determined by electron microprobe and mass balance calculations were made to determine melt fraction and mineral mode. Initial calculations yield glass:olivine:cpx:opx:spinel proportions of: 5.0:54.2:15.9:23.9:1.1. These preliminary results correspond well with previous work performed at the University of Oregon on the same starting material (Schwab and Johnston, 2001). The best match is with a $1315\deg$C experiment (INT-A13) in terms of mode (6.9:53.7:13.4:25.0:1.0) and glass composition, indicating that the temperature of this initial experiment may be slightly cooler than the target temperature, however the results of this interlaboratory comparison are still within the $\pm$10-$15\deg$C temperature uncertainty of the piston cylinder apparatus. Additional calibration experiments are ongoing.
V41C-1405 0800h
Examination of the Thermal Transformation of Chrysotile by Using Dispersion Staining and Conventional X-ray Diffraction Techniques
Chrysotile has been used industrially as a component of refractory products, and in friction products, such as brake linings. Examining the decomposition or transformation of chrysotile as a function of time and temperature will help clarify the characteristics of particulates released during processes such as automotive braking. Previous studies have reported that the thermal treatment of chrysotile alters both its surface and structure, resulting in deviations from its natural properties, possibly reducing its biological activity (Langer, 2003: Reg Tox Pharm, v38, p71). In past studies, the nonequilibrium thermal decomposition of chrysotile has been investigated by using static dehydration, X-ray diffraction, differential thermal analysis, and thermogravimetric analysis. These studies suggest that the thermal transformation of chrysotile follows a two step sequence of dehydroxylation/dehydration and recrystallization where (1) chrysotile yields forsterite + silica + water with (2) forsterite + silica later forming enstatite (Ball and Taylor, 1963: Mineral. Mag. v33, p467, Brindley and Hayami, 1965: Mineral. Mag. v35, p189). In this research the decomposition of chrysotile from Thetford Quebec was studied. Samples were heated isothermally in air at temperatures from $200\deg$C to $1000\deg$C. After heating for up to 24 hours the refractive indices of remaining chrysotile fibers were measured by using dispersion staining. In addition, reaction products were identified by optical methods, electron probe microanalysis, and X-ray diffraction performed after the termination of the experiment. Preliminary results show that there is no change in optical properties of chrysotile heated to $400\deg$C for 24 hours. From $400\deg$C to $575\deg$C for 24 hours, the index of refraction increases parallel to the length of the fiber from 1.552 to 1.560. From $400\deg$C to $575\deg$C for 24 hours, the index of refraction perpendicular to the length of the fiber varies irregularly from 1.538-1.548. The variability of the refractive indices, both parallel and perpendicular, observed in unheated chrysotile reduces from approximately $\pm$ 0.04 to $\pm$ 0.01 when heated to $575\deg$C for 24 hours. Chrysotile heated to $575\deg$C for 24 hours shows an overall loss of X-ray intensity for the main chrysotile peaks and possible growth of forsterite. At $800\deg$C, although the fibrous morphology remains, X-ray diffraction analysis of the run products shows no remaining chrysotile and the index of refraction parallel to the length of the fiber is approximately 1.594. Initial data show that the reaction rates associated with the dehydroxylation/dehydration and recrystallization of chrysotile are dependent upon the texture of individual fibers.
V41C-1406 0800h
The Behavior of Pyroxenes During Partial Melting of Pyroxenite and Lherzolite in the Mantle: An Experimental and Numerical Study
Pyroxenes are the primary phases involved in partial melting of peridotite and pyroxenite lithologies in the upper mantle. In order to better understand the grain-scale processes of pyroxene melting and their effects on major and trace element distributions during magma generation, we carried out a series of kinetic melting experiments using reaction couple method. Partial melting experiments were conducted at $1340\deg$C and 1.5 GPa for 72 hrs using reaction couples formed by juxtaposing pre-synthesized rods of orthopyroxenite (90% opx, 5% olivine, 5% melt) against clinopyroxenite (90% cpx + 10% melt). These laboratory experiments were supplemented by numerical simulations of partial melting in binary and ternary systems. Reaction between orthopyroxenite and clinopyroxenite at $1340\deg$C and 1.5GPa results in a reactive boundary layer (RBL, 240 $\mu$m after 72 hrs) that consists of euhedral olivine (35%), cpx (45%), and melt (20%). The RBL is located on the orthopyroxenite side of the original interface. The grain sizes of ol and cpx in the RBL are significantly larger than those in the orthopyroxenite and the clinopyroxenite. Clinopyroxene compositions vary systematically across the RBL, for example the Na$_{2}$O and TiO$_{2}$ content decrease from 1.05% and 0.35%, respectively, in the clinopyroxenite to 0.83% and 0.14% at the orthopyroxenite-RBL interface. The melt is concentrated in the RBL and appears to localize in those areas where olivine grains are more densely accumulated. The average melt in the RBL is a ne-normative alkali basalt (10.6% MgO, 49.2% SiO$_{2}$, 4.7% Na$_{2}$O, 0.28% K$_{2}$O) and has relatively high Al$_{2}$O$_{3}$ (15.5%) and CaO (10.7%). The characteristics of increasing grain size and varying cpx and melt compositions in the RBL are very similar to those observed in our clinopyroxenite-harzburgite partial melting experiments [1], consistent with the grain-scale melting processes that involve dissolution of opx at the orthopyroxenite-RBL interface, precipitation of ol and new cpx within the RBL, and production of Na$_{2}$O-Al$_{2}$O$_{3}$-SiO$_{2}$-rich melts in that region. This dissolution and reprecipitation process, due to two very different time-scales of crystal-melt interaction in partially molten systems, is reproduced in our numerical simulations of partial melting of bio-mineralic rocks in binary and ternary systems, and is likely to occur during lherzolite and pyroxenite partial melting in the laboratory and nature. One of the important consequences of cpx reprecipitation during peridotite and pyroxenite partial melting is the significant reduction in diffusive reequilibration time between the cpx and the surrounding melt, since diffusion coefficients of trace elements such as REE, U, and Th in cpx are much smaller than those of major elements in cpx and melt. The rate of cpx reprecipitation is dominated by the rate of chemical diffusion of the major components in cpx. This dissolution and reprecipitation process may help to explain the apparent equilibrium melting trends observed in some slab derived magmas that were produced at relatively low temperatures. The mechanisms outlined above can also be used to better understand the melting process of lherzolite and pyroxene-rich lithologies in the upper mantle and could partially explain the petrogenesis of high Na$_{2}$O-Al$_{2}$O$_{3}$-SiO$_{2}$ magmas, without the involvement of an eclogite component in the source region. [1] Lo Cascio et al., GRL, 31, L16605, doi: 10.1029/2004GL020602
V41C-1407 0800h
Experimental Determination of Allanite Stability in High-Silica Rhyolite
The rare earth elements (REE) are powerful tracers of the thermochemical evolution of rhyolitic/granitic magmas because their absolute and relative concentrations are primarily controlled by saturation of accessory minerals with discrete stability conditions. Allanite is the most important accessory mineral controlling the behavior of the light rare earth elements in metaluminous granites and rhyolites, yet the thermochemical conditions controlling its stability are poorly defined. Recent experiments (Hermann, 2002) have established allanite saturation in dacitic partial melts (63-65 wt.% SiO$_{2}$), but extrapolation of the results to the typical REE concentrations of high-silica rhyolite/leucogranite melts ($>$75 wt.% SiO$_{2}$) predicts temperatures that are at variance with those observed for natural allanite-bearing rhyolites (Bishop, Toba, & Bandelier Tuffs). In order to constrain allanite saturation in granitic magmas, we are performing experiments to determine the stability of allanite in high-silica rhyolite/leucogranite compositions at 1.6 to 7 kbar and 740 to 800$\deg$C. Experiments were conducted using piston-cylinder apparatus (7 kbar) and cold seal vessels (1.6 kbar), with Au capsules at H2O saturation and fO$_{2}$ at ~NNO. Chips of natural allanite were sealed with Late Bishop Tuff glass (77 wt.% SiO$_{2}$). Results at 7 kbar indicate that allanite saturation occurs between 760-770$\deg$C. At $\geq$780$\deg$C, anhydrous chevkinite (REE-Ti-rich silicate) replaced allanite, suggesting that allanite saturation may be associated with a discontinuous net transfer reaction involving chevkinite consumption. Preliminary results for experiments using cold seal vessels at 1.6 kbar pressure indicate that allanite is stable at temperatures from 750 to 770$\deg$C; higher temperature experiments are in progress. The stability of allanite in high-silica rhyolite melts containing REE concentrations greater than predicted by the experiments using dacitic melts suggests that the major element composition of magmas may affect allanite solubility. The apparent thermal stability of allanite at 7 kbar is similar to the 765$\pm$10$\deg$C inferred from Fe-Ti oxide thermometry of allanite-bearing Bishop Tuff that equilibrated at 1-2 kbar pressure (c.f., Hildreth, 1979), suggesting that allanite stability may be relatively insensitive to pressure for mid- to upper crustal granitic magmas. Ongoing experiments are examining the effects of water, fO$_{2}$, and melt REE concentration on allanite stability. References J. Hermann, Chem Geol 192: 289 (2002) W. Hildreth, GSA Special Paper 180: 43 (1979)
V41C-1408 0800h
A Solution Model for Silver Solubility in Galena in the Pb$_{2}$S$_{2}$-AgSbS$_{2}$-AgBiS$_{2}$ System
The extent of solubility of the Ag-bearing minerals miargyrite (AgSbS$_{2}$) and matildite (AgBiS$_{2}$) in galena has profound importance to models for ore genesis in Ag deposits. Sulfides are known to be reactive at temperatures of ore formation. Many of these sulfides likely do not preserve original compositions. Above $380\deg$C, all compositions in the Pb$_{2}$S$_{2}$ -AgSbS$_{2}$ -AgBiS$_{2}$ system are in the form of the $\alpha$-galena phase, which is the cubic (Fm-3m) structure of galena. X-ray data confirm that the high temperature structures of AgSbS$_{2}$, AgBiS$_{2}$, and ternary compositions are cubic, although the exact structures have not yet been determined and may not be quenchable. End-member compositions in this system are completely miscible at temperatures above $450\deg$C, even though the $\alpha$-galena structure cannot accommodate a high mole fraction of Ag at ambient temperatures. The solubility of Ag in galena is limited by a ternary miscibility gap, which is skewed toward the high Pb portion of the system. Reversal results from dry sinter experiments constrain the miscibility gap in the ternary system. A new asymmetric regular solution model for the ternary α-galena phase successfully reproduces observed coexisting compositions on the miscibility gap, which is predicted to close above about $450\deg$C. This solution model also satisfies data from previous workers on exchange equilibrium between Sb and Bi in the AgSbS$_{2}$-AgBiS$_{2}$ system. This thermodynamic model predicts that a low Ag/high Pb galena deposited at $400\deg$C may contain up to 8 mole$%$ Ag, and at $300\deg$C it may contain 3 mole$%$ Ag. For a deposit with molar Ag:Pb less than 1:30 and a temperature of formation at or above $300\deg$C, all of the Ag could have been accommodated originally in galena. This thermodynamic model can be used to evaluate original galena compositions in the context of ore grades and independent temperature of formation indicators. Results from this model may also have potential for improved ore genesis models and exploration tools.
V41C-1409 0800h
New Standard State Entropy for Sphene (Titanite)
Several recent papers have questioned the accepted standard state (STP) entropy of sphene (CaTiSiO$_{5}$), which had been considered to be in the range 129-132 J/mol.K (Berman, 1988: 129.3 Robie and Hemingway, 1995: 129.2 J/mol.K; Holland and Powell, 1995: 131.2 J/mol.K.). However, Xirouchakis and Lindsley (1998) recommended a much lower value of 106 J/mol.K for the STP entropy of sphene. Tangeman and Xirouchakis (2001) inferred a value less than 124 or 120 J/mol.K, based on based on enthalpy constraints combined with the tightly reversed reaction sphene+kyanite=rutile+anorthite by Bohlen and Manning (1991). Their recommendations are in conflict with the accepted values for STP entropy for sphene, including values calculated by direct measurement of Cp from 50 to 300 K by King (1954). In order to resolve this discrepancy, we have collected new data on the Cp of sphene between 5 and 300 K. Our measurements were made in the PPMS at Salzburg on a 21.4 g sample of sphene generously furnished by Tangeman and Xirouchakis (2001), the same sample as used in their experiments. The Cp data are slightly lower than those of King (1954) but merge smoothly with data of Tangeman and Xirouchakis (2001) from 330 to 483 K (or whatever) where a transition is recorded in the Cp data as a lambda anomaly. Tangeman and Xirouchakis also obtained data above the transition up to 950K. Integration of the new Cp data yields a STP entropy of 127.3 J/mol.K, lower than the generally accepted value by ca. 2 J/mol.K. A change in the STP entropy of sphene will have an effect on many Ti-bearing reactions which occur within the earth, although the magnitude of this change is not nearly as large as that suggested by Xirouchakis and Lindsley (1998). Above 700 K, the entropy calculated using the new STP entropy with the heat capacity equation of Tangeman and Xirouchakis (2001) is within 1 J/mol.K of the value tabulated in Robie and Hemingway (1995) and of that calculated from Berman (1988). The effect on most phase equilibrium calculations will not be large except for reactions with small $\Delta$S. The use of 127.2 J/mol.K as the standard entropy of sphene is recommended especially in calculations of geobarometers involving that phase.
V41C-1410 0800h
Extending Fe-Mg Olivine-liquid Partitioning to Picritic Compositions at One Atmosphere
Iron-magnesium partitioning between olivine and silicate melt exerts a fundamental control on the FeO and MgO contents of olivine-saturated liquids undergoing low-pressure fractionation as well as the compositions of liquids produced during partial melting of peridotite. Surprisingly, very little low-pressure olivine-liquid experimental data has been published on natural melt compositions with $>$ 16 wt% MgO. Results from high-pressure experiments [1] are scattered but they hint at an increasing value of the exchange coefficient, K$_{D}$=(FeO/MgO)$^{ol}$/(FeO/MgO)$^{liq}$, as melt MgO contents increase from 20 to 30 wt%; the K$_{D}$ values increase by as much as 30% relative over this range. Whether one uses K$_{D}$ = 0.30 or 0.38 has a dramatic effect on the compositions of reconstructed primary magmas, and it is therefore of interest to better determine the appropriate value(s). We conducted one-atmosphere experiments on a synthetic, alkali-free, Hawaiian picrite composition (25.7 wt% MgO, 45.7 wt% SiO$_{2}$) at QFM and temperatures ranging from $1300\deg$-$1500\deg$C. Samples were suspended on pre-saturated Pt-Fe wire loops and quenched after $\sim$12 to 72 hours. Time series experiments were conducted at $1300\deg$ and $1450\deg$C. Melt and olivine are present in all experiments; chrome spinel occurs in the $1300\deg$ and $1350\deg$C experiments. Over the temperature range investigated, wt% MgO in the glass varies from 13.4 to 23.8, while olivine forsterite contents increases from 88 to 92 (mole%). The weight fraction of glass shows a strong positive linear correlation with temperature, while the weight fraction of olivine is inversely correlated with temperature. The K$_{D}$ values calculated from the olivine-liquid pairs (after first calculating liquid FeO/Fe$_{2}$O$_{3}$ using [2]), are independent of temperature and melt composition (the mean value is 0.32$\pm$0.01, 1 sigma). Our mean value is the same as that calculated by [3] from a large body of lower temperature one-atmosphere experiments from the literature, and supports the results of the olivine-addition calculations on Mauna Kea glasses reported in [3]. [1] Kushiro & Walter (1998) Geophys. Res. Letters 25, 2337-2340. [2] Kress & Carmichael (1991) Contrib. Mineral. Petrol. 108, 82-92. [3] Stolper et al. (2004) GGG 5, doi 10.1029/2003GC000553.
V41C-1411 0800h
Effect of H2O on Ca-Na Partitioning Between Plagioclase and Melt: Implications for the Origin of An-rich Plagioclase in Low-Alkali Arc Tholeiite
We conducted high pressure (100-500 MPa) melting experiments on a low-alkali tholeiite (SiO2=53wt% MgO=6.5 wt%, CaO/Na2O=4.4, Al2O3/SiO2=0.33) at both H2O-undersaturated and H2O-saturated conditions to investigate the effect of H2O content on Ca-Na partitioning between plagioclase and melt. The experiments were carried out using IHPV at 100 and 200 MPa, and a solid media 0.5-inch piston cylinder apparatus at 500 MPa in the temperature ranges of 1000-1300Í‘, and varying H2O contents of 0-12wt%. Redox condition was about 0-2 log unit above NNO (nickel-nickel oxide) buffer. In this study, we focused on near liquidus plagioclases-melt partitioning, where melt composition remained constant and the effect of water/temperature on the partitioning was elucidated. Our experimental results show that, at each experimental pressure, An content of the near liquidus plagioclase and the Kd(Ca-Na) almost linearly increases as H2O content in melt increases. Each of the An content and the Kd(Ca-Na) variations in a low-alkali tholeiitic system (CaO/Na2O=4.0-4.5, Al2O3/SiO2=0.27-0.33) can be described by one equation using temperature, pressure(kb), and melt H2O content as parameters (lnAn=927.91/T-0.86298"C0.02693*P/T+0.01674*H2O, R=0.97; ln Kd(Ca-Na)=10695/T-6.7781-0.1009*P/T-0.00860*H2O, R=0.97). An content and the Kd(Ca-Na) of liquidus plagioclase increases with increasing melt H2O and decreasing pressure, indicating that the condition of nearly H2O-saturated at 200-300 MPa is preferable for the crystallization of most An-rich plagioclase ($>$An90) at nearly constant melt composition. We suggest this pressure condition of 200-300 MPa, corresponding to a depth of 7-10 km plays an important role for the crysatllization of An-rich plagioclase in H2O-rich arc tholeiite.
V41C-1412 0800h
Metal/Silicate Partitioning of W, Ge, Ga and Ni: Dependence on Silicate Melt Composition
Metal/silicate partition coefficients (Dm/s) for siderophile elements are essential to investigations of core formation when used in conjunction with the pattern of elemental abundances in the Earth's mantle (Drake and Righter, 2002; Jones and Drake, 1986; Righter et al. 1997). The partitioning of siderophile elements is controlled by temperature, pressure, oxygen fugacity, and by the compositions of the metal and silicate phases. In this work, we investigate the role of silicate melt composition on the partitioning of the siderophile elements W, Ge, Ga and Ni between metallic and silicate liquid. Experiments were performed in the Experimental Geochemistry Laboratory at the University of Arizona utilizing a non-end loaded piston cylinder apparatus with a barium carbonate pressure medium. Starting materials were created by combining the mafic and silicic compositions of Jaeger and Drake (2000) with Fe powder (~25 wt% of the total mixture) to achieve metal saturation. Small amounts of W, Ge, Ga2O3 and NiO powder (less than 2 wt% each) were also added to the starting compositions. The experiments were contained in a graphite capsule and performed with temperature and pressure fixed at 1400°C and 1.5 GPa. Experimental run products were analyzed with the University of Arizona Cameca SX50 electron microprobe with four wavelength dispersive spectrometers and a PAP ZAF correction program. All experiments in our set are saturated with metal and silicate liquid, indicating that oxygen fugacity is below IW. Several of the runs also contain a gallium-rich spinel as an additional saturating phase. Quench phases are also present in the silicate liquid in all runs. The experimentally produced liquids have nbo/t values (calculated using the method of Mills, 1993) that range from 1.10 to 2.97. These values are higher than those calculated for the liquids in the Jaeger and Drake (2000) study. The higher nbo/t values are due to uptake of Fe by the melt. The initial silicate composition contained no FeO, however the experimentally produced silicate liquids contained from 15 to 26 wt % FeO. We find that W is incompatible over the range of compositions used in this study. However, W compatibility increases as melts become more silicic, with D(W) = 0.0005 at nbo/t = 2.97 and D(W) = 0.09 at nbo/t = 1.1. The slope of the best fit line for the W data when plotted in nbo/t vs Log D space is -1.22 and close to the value of -1.34 found by Jaeger and Drake (2000). Ge is compatible at all compositions and follows a similar pattern to that of W becoming more compatible with decreasing nbo/t (D(Ge)= 14 at nbo/t = 2.97 and D(Ge) = 100 at nbo/t = 1.1). Ni and Ga display essentially flat slopes within the error of our analysis, with D(Ni) = 395 at nbo/t = 2.97 and D(Ni) = 870 at nbo/t 1.10 and D(Ga) = 0.08 at nbo/t = 2.97 and D(Ga) = 0.02 at nbo/t = 1.1. A second series of experiments is in progress to verify these data and extend the study to lower values of nbo/t. References: Drake, M.J. and Righter, K. (2002) Nature, v. 416, 39-44; Jones, J.H. and Drake, M.J. (1986) Nature, v. 323, 470-471; Righter, K., et al. (1997) Physics Earth and Planet. Int., v. 100, 115-134; Jaeger, W.L. and Drake, M.J. (2000) Geo. Cosmo. Acta, v. 64, 3887-3895; Mills, K.C. (1993) ISIJ International, v. 33, 148-155.
V41C-1413 0800h
Minor and Trace Element Partitioning Among $\alpha$-, $\beta$- and $\gamma$-(Mg,Fe)$_{2}$SiO$_{4}$ in Fertile Peridotite KLB-1: a High Pressure Experimental Study
Syngenetic mineral inclusions in diamond provide direct sampling of the mantle where diamond crystallized. Petrological models for the upper mantle suggest a pyrolitic composition where olivine, (Mg,Fe)$_{2}$SiO$_{4}$, and its high-pressure polymorphs comprise as much as 60 vol.%. Consistent with this, olivine is the most common inclusion in diamonds from the upper mantle, and rare diamonds from the uppermost lower mantle contain mineral inclusions with peridotitic compositions. However no peridotitic diamonds have been recovered from the transition zone. The most likely reason is that the crystal structure signature of the transition zone minerals may be erased upon decompression. However, the partitioning behavior of the (Mg,Fe)$_{2}$SiO$_{4}$ polymorphs may provide independent evidence for the origin of the diamond. Crystal chemistry and partitioning among $\alpha$-, $\beta$- and $\gamma$-(Mg,Fe)$_{2}$SiO$_{4}$ crystallized from fertile peridotite KLB-1 are being studied in multianvil experiments at 12 to 21 GPa and 1300 to 1600$\deg$C. At 12 GPa olivine is the dominant phase, together with garnet (Grt), diopsidic clinopyroxene (Cpx) and minor Al, Cr-spinel. At 13 GPa, Grt becomes strongly majoritic (jumping from 3.0 Si per 12 oxygens at 12 GPa to 3.7 Si at 13 GPa) with a high MgSiO3 component. Between 13 and 15 GPa olivine is replaced by wadsleyite. Wadsleyite contains higher Na, Ni, Ti, Cr and Al than olivine. Partition coefficients between wadsleyite and olivine (D$^{Wad/Ol}$) range from 2 (Na) to 7 (Al). Olivine contains higher Mg-numbers (av. 90.0 Ol, 89.5 Wad) and Mn than wadsleyite. Both olivine and wadsleyite show a strong negative correlation between Si and Mg. Ca, Na, Al, Cr and Ti, however, exhibit different behavior between the two phases. In olivine, Mg is negatively correlated with Ca, Na, Al, Cr and Ti, while in wadsleyite, Mg is positively correlated with Na, Al, Cr and Ti. The experiments demonstrate that in a fertile peridotite 1) the former presence of wadsleyite can be determined by analysis of trace elements in olivine when preserved in the confines of a diamond and 2) the majorite content of garnet increases sharply as pressure is increased from 12 to 13 GPa.
V41C-1414 0800h
HP and HT Polymorphism of AlPO$_{4}$ and its Solubility in SiO$_{2}$ -Stishovite : Implication for Phosphorus Geochemistry
The understanding of phosphorus behavior in the deep Earth requires to evaluate the ability of minor phosphorus to be incorporated in the octahedral site of high-pressure mineral structures. For that purpose, we studied the solubility of AlPO$_{4}$ in SiO$_{2}$-stishovite, the extend of which will indicate the phosphorus affinity for octahedral oxygen coordination. The choice of those compounds was also motivated by their importance in material sciences (e.g. pressure induced amorphization properties). Synthetic AlPO$_{4}$ (berlinite) as well as a 1 to 2 molar mixture of AlPO$_{4}$ (berlinite) : SiO$_{4}$ (quartz) where encapsulated separately in a Pt container and held together at $1600\deg$C, 18 GPa (4.5 h) in the 2000-ton split-sphere apparatus at GRC (Matsuyama, Japan). In the second capsule, micrometer-size stishovite crystals could be analyzed using the electron microprobe and revealed around 0.5 wt.% P$_{2}$O$_{5}$ (balanced by Al$_{2}$O$_{3}$). Berlinite, in the first capsule, totally reacted at HP and HT into a form which could not be quenched. This result was the incentive for a characterization of AlPO$_{4}$ polymorphism using in-situ HP-HT x-ray diffraction (Spring 8, BL04B1). Upon compression in the MgO-based assembly, the diffraction pattern of AlPO$_{4}$ berlinite showed indication of partial amorphization in keeping with the non-hydrostatic nature of the pressure cell at room temperature. Upon the first heating stage, crystallization of another AlPO$_{4}$ polymorph had occurred at $600\deg$C, 6.9 GPa (gold scale) in a first experiment and $900\deg$C, 17.3 GPa in a second one. The corresponding diffraction pattern could be indexed in the InPO$_{4}$ -structure (Cmcm) where Al is octahedrally coordinated whereas P is fourfold. No other transformation was identified up to 21 GPa, $1600\deg$C. Although, no rutile-type structure was observed for AlPO$_{4}$ up to 21 GPa, phosphorus, which is likely to be stored in garnet down to the transition zone, is expected to be further incorporated into silicate structures to greater depths.