Fe-FeO Buffer Research Articles

Experimental determination of tin partitioning between titanite, ilmenite and granitic melts using improved capsule designs

Abstract Investigating mineral/melt Sn partitioning at high temperatures and pressures is a difficult task because Sn is a redox-sensitive multivalent element and easily alloys with noble metal sample capsules. To obtain accurate Sn partition coefficients between titanite, ilmenite, and granitic melts, we developed single capsule Pt or Au and double capsule Pt95Rh5 (or Au)–Re designs to avoid significant Sn loss at a controlled oxygen fugacity (ƒO2). With these new capsule designs, we performed piston-cylinder experiments of Sn partitioning between titanite, ilmenite, and granitic melts. The experimental P–T–ƒO2 conditions were 0.5–1.0 GPa, 850–1000 °C and ~QFM+8 to ~QFM–4 (QFM: Quartz–Fayalite–Magnetite buffer), with ƒO2 controlled by the solid buffers of Ru–RuO2, Re–ReO2, Co–CoO, graphite, and Fe–FeO. The obtained mineral/melt Sn partition coefficients (DSnmin/melt) are 0.48–184.75 for titanite and 0.03–69.45 for ilmenite at the experimental conditions. The DSnmin/melt values are largely dependent on ƒO2 although the effects of temperature and melt composition are also observed. DSnTtn/melt strongly decreases with decreasing ƒO2, from ~46–185 at the most oxidizing conditions (Ru–RuO2 buffer), to ~2–16 at moderately oxidizing to moderately reducing conditions (Re–ReO2 to Co–CoO and graphite buffers), to < 1 at the most reducing conditions (Fe–FeO buffer). DSnIlm/melt exhibits a variation trend similar to DSnTtn/melt, but is always lower than DSnTtn/melt at a given ƒO2. These DSnmin/melt values can be applied to quantitatively assess the mineralization potential of granitic magmas. Using DSnTtn/melt, we estimate that Sn contents are ~150–400 ppm in the premineralization magmas of the tin-mineralized Qitianling plutons (South China).

Partitioning of nickel and cobalt between metal and silicate melts: Expanding the oxy-barometer to reducing conditions

Moderately siderophile elements (MSEs) are potential tracers of the thermodynamic conditions prevailing during planetary core formation because their metal–silicate partition coefficients (Dmet/sil) vary as a function of P, T, and oxygen fugacity (fO2). Those properties result in the production of planetary mantles with unique MSE depletion signatures. Among the MSEs, Ni and Co are reliable barometers in magma oceans because their Dmet/sil values are strongly correlated with pressure, decreasing by almost 3 orders of magnitude between 1 bar and 100 GPa. Current pressure-dependent expressions of Dmet/sil were calibrated based on experiments performed under relatively oxidizing conditions, mostly at fO2 slightly below the iron–wüstite Fe–FeO buffer (IW), which is relevant to the mantles of Earth and Mars. However, planets and asteroids formed under a wide range of redox conditions, from Mercury, the most reduced (∼IW − 5.5), to the most oxidized angrite parent body (IW − 1.5 to IW + 1). In this study, we performed and analyzed 38 metal–silicate partitioning experiments over a wide range of pressures (1 bar to 26 GPa) and oxygen fugacities (IW − 6.4 to IW − 1.9) to expand the available Ni and Co Dmet/sil values to reducing conditions. We then parameterized 255 Ni and 194 Co Dmet/sil values as a function of T (1573–5700 K), P (1 bar to 100 GPa), and fO2 (IW − 6.4 to IW + 0.2). We also modeled the evolution of Ni and Co Dmet/sil values along the liquidus of a chondritic mantle at various P and fO2 conditions to investigate the thermodynamic conditions of various planetary bodies’ magma oceans. The P and fO2 conditions we obtained for Earth, Mars, the Moon, and Vesta are consistent with previous studies using similar methods, and the pressure during core formation is strongly correlated to planetary size. Finally, we also applied our model to several achondrite parent bodies; our results indicate a wide variety of objects, from the asteroid-sized, oxidized angrite parent body to the planet-sized, highly reduced aubrite parent body.

Fe2+–Mg partitioning between olivine and liquid at low oxygen fugacity: an experimental and thermodynamic framework

A set of 1-atm gas-mixing experiments (n = 31) was run using Re wire loops at low-oxygen fugacity (ƒO2, within one half-log unit of the Fe–FeO buffer) and temperatures of 1175–1400 °C to explore the compositional dependence of the Fe2+–Mg olivine–liquid exchange coefficient, {mathrm{K}}_{mathrm{D},{mathrm{Fe}}^{2+}-mathrm{Mg}}^{mathrm{ol}/mathrm{liq}}, under conditions where corrections for liquid Fe3+ are small. The bulk compositions used for these experiments include a picrite, a high-alumina basalt, and a suite of three MORB compositions with variable Fe/Mg. The {mathrm{K}}_{mathrm{D},{mathrm{Fe}}^{2+}-mathrm{Mg}}^{mathrm{ol}/mathrm{liq}} values from the n = 14 experiments run on pre-saturated Re wire loops were fit to a regular solution model that relates {mathrm{K}}_{mathrm{D},{mathrm{Fe}}^{2+}-mathrm{Mg}}^{mathrm{ol}/mathrm{liq}} to temperature, melt composition, and olivine composition. Fe2+–Mg exchange in the experiments is well-described using two parameters: the SiO2 content of the liquid and the coexisting olivine composition. Combining our experiments with a literature compilation of low-pressure, low-ƒO2 experiments (largely on non-terrestrial bulk compositions) produced a combined data set spanning a broad region of composition space (e.g., liquid TiO2 and Na2O + K2O contents up to 18.4 wt% and 8.4 wt%, respectively). Fitting this expanded experimental database required two additional liquid compositional terms: Ti, and a Si-(Na + K) cross-term. Because the Fe3+ content of all of the experimental liquids is low, the compositional variation seen in {mathrm{K}}_{mathrm{D},{mathrm{Fe}}^{2+}-mathrm{Mg}}^{mathrm{ol}/mathrm{liq}} is effectively independent of the compositional effects on liquid Fe3+/Fe2+ ratios. Given olivine–liquid Mg or Fe2+ partition coefficient information, it is possible to eliminate the explicit dependence of {mathrm{K}}_{mathrm{D},{mathrm{Fe}}^{2+}-mathrm{Mg}}^{mathrm{ol}/mathrm{liq}} on olivine composition, enabling a simple iterative approach for calculating the composition of coexisting olivine given only the bulk composition (and Fe3+/Fe2+ ratio) of an olivine-saturated liquid.

An experimental investigation of F, Cl and H2O mineral-melt partitioning in a reduced, model lunar system

It is believed that the Moon formed following collision of a large planetesimal with the early Earth. Over the ∼4 Gyr since this event the Moon has been considerably less processed by geological activity than the Earth, and may provide a better record of processes and conditions in the early Earth-Moon system. There have been many studies of magmatic volatiles such as H, F, Cl, S and C in lunar materials. However, our ability to interpret variable volatile contents in the lunar sample suite is dependent on our understanding of volatile behaviour in lunar systems. This is currently constrained by limited experimental data. Here, we present the first experimental mineral-melt partitioning coefficients for F, Cl and H2O in a model lunar system under appropriately reduced conditions (log fO2 to IW-2.1, i.e. oxygen fugacity down to 2.1 log units below the Fe-FeO buffer). Data are consistent with structural incorporation of F, Cl and OH− in silicate melt, olivine and pyroxene under conditions of the lunar mantle. Oxygen fugacity has a limited effect on H2O speciation, and partitioning of H2O, F and Cl is instead largely dependent on mineral chemistry and melt structure. Partition coefficients are broadly consistent with a mantle source region for lunar volcanic products that is significantly depleted in F, Cl and H2O, and depleted in Cl relative to F and H2O, compared to the terrestrial mantle. Partitioning data are also used to model volatile redistribution during lunar magma ocean (LMO) crystallisation. The volatile content of lunar mantle cumulates is dependent upon proportion of trapped liquid during LMO solidification. However, differences in mineral-melt partitioning during LMO solidification can result in significant enrichment on F relative to Cl, and F relative to H2O, in cumulate phases relative to original LMO composition. As such, Cl depletion in lunar volcanic products may in part be a result of LMO solidification.

Methane-bearing fluids in the upper mantle: an experimental approach

The main obstacle to understanding of the geological role of reduced, CH4-bearing fluids is the absence of a reliable experimental technique applicable to solid-media high-pressure apparatuses, allowing their observation and direct characterisation under laboratory conditions. In this study, we describe the main pitfalls of earlier designs and technical aspects related to achievement of strongly reduced oxygen fugacity (fO2) conditions (i.e., Fe–FeO, IW) and maintenance of a constant fluid equilibrium during an experiment. We describe a new triple-capsule design made of an Au outer capsule with an Au-inner capsule containing a metal/metal oxide oxygen buffer and water, as well as an inner olivine container filled with a harzburgitic sample material and Ir powder that serves as a redox sensor. The bottom of the outer capsule is covered with a solid fluid source (e.g., stearic acid). The outer capsule is surrounded by a polycrystalline CaF2 pressure medium to minimise H2-loss from the assembly. Application of this design is limited to temperatures below the melting temperature of Au, which is pressure dependent. Metals other than Au can lead to fluid disequilibrium triggered by a dehydrogenation and carbonation of the methane. Test experiments were carried out at 5 GPa, temperatures < 1300 °C, at Mo–MoO2 and Fe–FeO buffer conditions. IrFe alloy sensors demonstrate successful achievement and maintenance of reduced fluid environment at ∆logfO2 ≈ IW + 0.5. The fluid phase was trapped in numerous inclusions within the olivine sample container. Raman spectra reveal that the fluid consists mainly of CH4, along with small amounts of higher hydrocarbons like C2H6. No water was detected, but H2 was found to be present in fluid and incorporated into the olivine structure. Our results are inconsistent with published fluid speciation models that predict significant H2O contents at these fO2 conditions. It is also apparent that fluids with significant CH4 contents are likely to be stable under the conditions recorded by some mantle samples.

Nitrogen solubility in the deep mantle and the origin of Earth's primordial nitrogen budget

The solubility of nitrogen in the major minerals of the Earth's transition zone and lower mantle (wadsleyite, ringwoodite, bridgmanite, and Ca-silicate perovskite) coexisting with a reduced, nitrogen-rich fluid phase was measured. Experiments were carried out in multi-anvil presses at 14 to 24 GPa and 1100 to 1800 °C close to the Fe–FeO buffer. Starting materials were enriched in 15N and the nitrogen concentrations in run products were measured by secondary ion mass spectrometry. Observed nitrogen (15N) solubilities in wadsleyite and ringwoodite typically range from 10 to 250 μg/g and strongly increase with temperature. Nitrogen solubility in bridgmanite is about 20 μg/g, while Ca-silicate perovskite incorporates about 30 μg/g under comparable conditions. Partition coefficients of nitrogen derived from coexisting phases are DNwadsleyite/olivine = 5.1 ± 2.1, DNringwoodite/wadsleyite = 0.49 ± 0.29, and DNbridgmanite/ringwoodite = 0.24 (+0.30/−0.19). Nitrogen solubility in the solid, iron-rich metal phase coexisting with the silicates was also measured and reached a maximum of nearly 1 wt.% 15N at 23 GPa and 1400 °C. These data yield a partition coefficient of nitrogen between iron metal and bridgmanite of DNmetal/bridgmanite ∼ 98, implying that in a lower mantle containing about 1% of iron metal, about half of the nitrogen still resides in the silicates. The high nitrogen solubility in wadsleyite and ringwoodite may be responsible for the low nitrogen concentrations often observed in ultradeep diamonds from the transition zone. Overall, the solubility data suggest that the transition zone and the lower mantle have the capacity to store at least 33 times the mass of nitrogen presently residing in the atmosphere. By combining the nitrogen solubility data in minerals with data on nitrogen solubility in silicate melts, mineral/melt partition coefficients of nitrogen can be estimated, from which the behavior of nitrogen during magma ocean crystallization can be modeled. Such models show that if the magma ocean coexisted with a primordial atmosphere having a nitrogen partial pressure of just a few bars, several times the current atmospheric mass of nitrogen must have been trapped in the deep mantle. It is therefore plausible that the apparent depletion of nitrogen relative to other volatiles in the near-surface reservoirs reflects the storage of a larger reservoir of nitrogen in the solid Earth. Dynamic exchange between these reservoirs may have induced major fluctuations of bulk atmospheric pressure over Earth's history.

Formation of N–С–О–Н molecules and complexes in the basalt–basaltic andesite melts at 1.5 Gpa and 1400°C in the presence of liquid iron alloys

The contents and speciation of nitrogen, carbon, and hydrogen were determined in basalt–basaltic andesite melts in equilibrium with liquid Fe alloys at 1.5 Gpa, 1400°C, and oxygen fugacity (fO2) 1.4–1.9 log units below that of the Fe–FeO buffer (ΔlogfO2(IW) =–1.4 …–1.9). Experiments were carried out on a piston- cylinder type apparatus using welded Pt capsules in the presence of excess С (graphite). Starting mixture consisted of natural ferrobasaltic glass and silicon nitride (Si3N4) as nitrogen source in the system. Experimental quench products representing glasses with spherical inclusions of iron alloy were analyzed using electron microprobe, Raman, and IR spectroscopy. With increase of Si3N4 in the starting mixture and, respectively, decrease of fO2, silicate melt forming during experiments became depleted in FeO and enriched in SiO2. It was established that the nitrogen content in the glasses increases from 0.13 to 0.44 wt % with decrease of ΔlogfO2(IW) from–1.4 to–1.9, whereas C content in the first approximation remains constant within 1.18–1.13 wt %, while the total water content (ОН– + Н2О) determined by IR spectroscopy decreases from 4.91 to 1.20 wt %. The N (0.13–0.48 wt %) and C (0.75–2.26 wt %) contents determined in the Fe alloy show no clear correlation with fO2. The IR and Raman spectroscopic study of the glasses indicates the formation of molecules and complexes with bonds N–H (NH3, NH2−, NH2+, NH4+), Н–О (Н2О, OH–), С–Н (СН4) as well as N2 and Н2 molecules in silicate melts. IR spectra also reveal the presence of complexes with С=О, С–N bonds and СО2 molecules. Obtained data are compared with results of previous studies on the solubility and speciation of N, С, and Н in the model FeO–Na2O–SiO2–Al2O3 melts in equilibrium with liquid iron alloys at 1.5 GPa (1400°C) and 4 GPa (1550°C) (Kadik et al., 2011, 2015).

Asteroid bombardment and the core of Theia as possible sources for the Earth's late veneer component.

The silicate Earth contains Pt-group elements in roughly chondritic relative ratios, but with absolute concentrations <1% chondrite. This veneer implies addition of chondrite-like material with 0.3-0.7% mass of the Earth's mantle or an equivalent planet-wide thickness of 5-20 km. The veneer thickness, 200-300 m, within the lunar crust and mantle is much less. One hypothesis is that the terrestrial veneer arrived after the moon-forming impact within a few large asteroids that happened to miss the smaller Moon. Alternatively, most of terrestrial veneer came from the core of the moon-forming impactor, Theia. The Moon then likely contains iron from Theia's core. Mass balances lend plausibility. The lunar core mass is ~1.6 × 1021 kg and the excess FeO component in the lunar mantle is 1.3-3.5 × 1021 kg as Fe, totaling 3-5 × 1021 kg or a few percent of Theia's core. This mass is comparable to the excess Fe of 2.3-10 × 1021 kg in the Earth's mantle inferred from the veneer component. Chemically in this hypothesis, Fe metal from Theia's core entered the Moon-forming disk. H2O and Fe2O3 in the disk oxidized part of the Fe, leaving the lunar mantle near a Fe-FeO buffer. The remaining iron metal condensed, gathered Pt-group elements eventually into the lunar core. The silicate Moon is strongly depleted in Pt-group elements. In contrast, the Earth's mantle contained excess oxidants, H2O and Fe2O3, which quantitatively oxidized the admixed Fe from Theia's core, retaining Pt-group elements. In this hypothesis, asteroid impacts were relatively benign with ~1 terrestrial event that left only thermophile survivors.

OH solubility in olivine in the peridotite–COH system under reducing conditions and implications for water storage and hydrous melting in the reducing upper mantle

Experimental studies of OH solubility in peridotite minerals are of crucial importance for understanding some key geochemical, geophysical and geodynamical properties of the upper mantle. In reducing depths of the upper mantle, C–O–H fluids are dominated by CH4 and H2O. However, available experimental H-annealing of olivine concerning water storage capacity in the reducing upper mantle has been exclusively carried out by equilibrating olivine with H2O only. In this study, OH solubility in olivine has been investigated by annealing natural olivine crystals under peridotite-bearing and CH4–H2O-present conditions with piston cylinder and multi-anvil apparatus. Experiments were performed at 1–7 GPa and 1100–1350 °C and with oxygen fugacity controlled by Fe–FeO buffer, and OH solubilities were measured from polarized infrared spectra. The olivines show no change in chemical composition during the experiments. The infrared spectra of all the annealed olivines show OH bands in the range 3650–3000 cm−1, at both high (>3450 cm−1) and low (<3450 cm−1) frequency, and the bands at ∼3400–3300 cm−1 are greatly enhanced above ∼3 GPa and 1300 °C. The determined H2O solubility is ∼90–385 ppm for the olivine coexisting with H2O (1–7 GPa and 1100 °C), and is ∼40–380 ppm for the olivine coexisting with CH4–H2O (1–7 GPa and 1100–1350 °C). When CH4 is present in the equilibrium fluid, the H2O solubility is reduced by a factor of ∼2.3 under otherwise identical conditions, indicating a strong effect of CH4 on the partitioning of water between olivine and coexisting fluid. The storage capacity of water in the reducing upper mantle is, modeled with the measured solubility of olivine and available partition coefficients of water between coexisting minerals, up to ∼2 orders of magnitude lower than some previous estimates. Considering the temperature along the geotherm in the reducing oceanic upper mantle, the required H2O concentration to trigger hydrous melting is 250 and 535 ppm at ∼100 and 210 km depth, respectively, and is even larger at greater depths. These values exceed the typical H2O abundance (∼100±50 ppm) in the upper mantle, suggesting that pervasive hydrous melting at reducing depths of the oceanic upper mantle is not likely. Similar arguments may also be casted for the reducing deep upper mantle in the continental regions.

Nitrogen distribution between aqueous fluids and silicate melts

The partitioning of nitrogen between hydrous fluids and haplogranitic, basaltic, or albitic melts was studied at 1–15 kbar, 800–1200 °C, and oxygen fugacities (fO2) ranging from the Fe–FeO buffer to 3log units above the Ni–NiO buffer. The nitrogen contents in quenched glasses were analyzed either by electron microprobe or by secondary ion mass spectrometry (SIMS), whereas the nitrogen contents in fluids were determined by mass balance. The results show that the nitrogen content in silicate melt increases with increasing nitrogen content in the coexisting fluid at given temperature, pressure, and fO2. Raman spectra of the silicate glasses suggest that nitrogen species change from molecular N2 in oxidized silicate melt to molecular ammonia (NH3) or the ammonium ion (NH4+) in reduced silicate melt, and the normalized Raman band intensities of the nitrogen species linearly correlate with the measured nitrogen content in silicate melt. Elevated nitrogen contents in silicate melts are observed at reduced conditions and are attributed to the dissolution of NH3/NH4+. Measured fluid/melt partition coefficients for nitrogen (DNfluid/melt) range from 60 for reduced haplogranitic melts to about 10 000 for oxidized basaltic melts, with fO2 and to a lesser extent melt composition being the most important parameters controlling the partitioning of nitrogen. Pressure appears to have only a minor effect on DNfluid/melt in the range of conditions studied. Our data imply that degassing of nitrogen from both mid-ocean ridge basalts and arc magmas is very efficient, and predicted nitrogen abundances in volcanic gases match well with observations. Our data also confirm that nitrogen degassing at present magma production rates is insufficient to accumulate the atmosphere. Most of the nitrogen in the atmosphere must have degassed very early in Earth's history and degassing was probably enhanced by the oxidation of the mantle.

Melting and subsolidus phase relations in peridotite and eclogite systems with reduced C[sbnd]O[sbnd]H fluid at 3–16 GPa

Melting phase relations of peridotite and eclogite coexisting with reduced C–O–H fluid have been studied at 3–16 GPa and 1200–1600 °C. In order to perform these experiments the double-capsule technique with fO2 control by outer Mo–MoO2 or Fe–FeO buffer capsule was designed and developed for multianvil experiments at pressures 3–21 GPa. Silicate phase assemblages resemble those in volatile-free lithologies, i.e. olivine/wadsleyite–orthopyroxene–clinopyroxene–garnet in peridotite and garnet–omphacite in eclogite. Melting was detected by the appearance of quenched crystals of pyroxene, feldspar and glassy silica. Estimated solidus temperatures for peridotite + C–O–H fluid with fO2=Fe–FeO are 1200 °C at 3 GPa and 1700 °C at 16 GPa. The solidus of the system with fO2=Mo–MoO2 was about 100 °C lower. Estimated solidus temperatures for eclogite + C–O–H fluid with fO2=Fe–FeO are 1100 °C at 3 GPa and 1600 °C at 16 GPa, and for eclogite at fO2=Mo–MoO2 solidus temperatures were 20–50 °C lower. These solidus temperatures are much higher (300–500 °C) than those for peridotite and eclogite systems with H2O and/or CO2, but are still 300–400 °C lower than the solidi of volatile-free peridotite and eclogite at studied pressures. The compositions of partial melt were estimated from mass-balance calculations: partial melts of peridotite have CaO-poor (6–9 wt.%) basaltic compositions with 44–47 wt.% SiO2 and 1.1–1.6 wt.% Na2O. Melts of eclogite contain more SiO2 (47–49 wt.%) and are enriched in CaO (9–15 wt.%), Na2O (9–14 wt.%), and K2O (1.3–2.2 wt.%). All runs contained graphite or diamond crystals along with porous carbon aggregate with micro-inclusions of silicates indicating that reduced fluid may dissolve significant amounts of silicate components. Analyses of carbon aggregates using a defocused electron microprobe beam reveal compositions similar to estimated partial melts. The diamonds formed from reduced C–O–H fluid may have natural analogues as polycrystalline diamonds. The oxygen fugacity in the Earth's mantle decreases with pressure from about fayalite–magnetite–quartz at shallow depths of 20–50 km to about iron–wustite at 250–300 km according to fO2 estimations from cratonic peridotite. We show significant increase of solidus temperatures in peridotite and eclogite coexisting with reduced CH4–H2O fluid relative to the systems with oxidized H2O–CO2 fluid. We emphasize that redox melting by change of oxidation state across a mantle section, a phase transition, or the lithosphere–asthenosphere boundary can be the dominant melting process in the deep Earth's interior.

Nitrogen solubility in upper mantle minerals

Nitrogen solubility in the upper mantle minerals forsterite, diopside, enstatite and pyrope has been quantified by SIMS measurements of nitrogen-saturated, synthetic samples. The crystals were grown in a 15N–H–O fluid buffered by Ni–NiO, Co–CoO, and Fe–FeO, at 1000–1300 °C and 15–35 kbar in a piston cylinder apparatus. Nitrogen solubility in minerals is significantly affected by temperature, pressure, mineral composition and, in particular, by oxygen fugacity. Nitrogen in all crystals buffered by Ni–NiO or Co–CoO is below detection limit or at most a few μg/g at very high pressures. Concentrations of 5–24 μg/g nitrogen have been quantified in diopside, enstatite and pyrope buffered by Fe–FeO at 1100 °C/15 kbar. Very high nitrogen solubility up to 100 μg/g is observed at the Fe–FeO buffer in enstatite at high-temperature or in Al-bearing enstatite and diopside. The nitrogen solubility in forsterite at the Fe–FeO buffer also clearly increases with temperature and pressure; a maximum solubility of 10 ppm is obtained at 1300 °C/35 kbar. The strong enhancement of nitrogen solubility under reducing conditions may be related to nitrogen dissolution as either NH+4 or as N3− directly replacing O2−. Both mechanisms require some charge compensation, consistent with the enhancement of nitrogen solubility with Al content in enstatite. Our results demonstrate that the reduced lower part of the upper mantle has a large nitrogen storage capacity, and may store ∼20–50 times more nitrogen than the present atmosphere. Therefore, some ‘missing’ nitrogen may still be retained in the Earthʼs deep, reduced mantle. The calculated nitrogen partition coefficients between upper mantle minerals and silicate melt reveal that an oxidized mantle source would lose almost its entire nitrogen during partial melting, whereas under reducing conditions a considerable fraction of nitrogen could be retained in the residual solids. The high nitrogen solubility in upper mantle minerals at reducing conditions also suggests that solidification of the magma ocean on the early Earth should have retained significant nitrogen, yielding higher N/Ar and N/C ratios in the young upper mantle as compared to the young atmosphere.

The oxidation state of tungsten in silicate melt at high pressures and temperatures

Although at 1atm pressure the oxidation state of W in silicate melts is 6+at oxygen fugacities from air down to several logfO2 units below Fe-FeO (IW) equilibrium, Cottrell et al. (2009) suggested that, at pressures above 6GPa, W becomes predominantly 4+in this oxygen fugacity range. Wade and Wood (2005), using a similar, but expanded metal-silicate partitioning dataset found, however, no evidence for an oxidation state change. In order to resolve the issue we collected tungsten L3 edge XANES spectra of a series of synthetic tungsten- bearing glass standards and of silicates from a range of high-pressure (1.5 to 25GPa) metal/silicate partitioning experiments. Glass standards were made at 1atm pressure and equilibrated at oxygen fugacities spanning a range from approximately 5.5 log units below the Fe-FeO buffer (IW-5.5) to Air. Metal-silicate partitioning experiments were performed at oxygen fugacities between IW-6.2 and IW-1 and at pressures between 1.5 and 25GPa. At low pressures and oxygen fugacities above IW-3.5, W exists in the silicate melt almost exclusively as W6+ (identical L3-edge energy to WO3) with the progressive reduction to W4+ completed by about IW-6. The XANES spectra of W from experiments at 6, 7 and 25GPa are completely consistent with those from 1atm and 1.5GPa experiments and with that of WO3. We conclude that there is no change of oxidation state with increasing pressure to 25GPa and that modelling of the oxygen-fugacity dependence of core formation requires use of a+6 oxidation state of W.

Determination of the formal Ge-oxide species in silicate melts at oxygen fugacities applicable to terrestrial core formation scenarios

High-temperature, high-pressure experiments were performed at 1350 °C and 0.5 GPa to determine the dependence of germanium partitioning behavior between coexisting solid metal and liquid silicate phases on oxygen fugacity and melt composition. A piston cylinder apparatus was used to perform experiments under effectively closed-system conditions to avoid possible loss of volatile Ge species. The oxygen fugacity of the experiments was adjusted by using silicate melts with different amounts of FeO. This causes a change of the melt composition. Additional experiments with different melt compositions but identical FeO contents were performed to determine the effect of melt composition on the metal–silicate partitioning behavior. The partitioning behavior of Ge depends on both oxygen fugacity ( f O2) and silicate melt composition (NBO/T). The dependences can be described by the equation: \batchmode \documentclass[fleqn,10pt,legalpaper]{article} \usepackage{amssymb} \usepackage{amsfonts} \usepackage{amsmath} \pagestyle{empty} \begin{document} \[logD^{met{\mbox{--}}sil}Ge= 3.32({\pm}0.07) {-}0.50({\pm}0.03) {\cdot} log\mathit{f}O_{2} {-} 0.29 ({\pm}0.04) {\cdot} \frac{NBO}{T} (\mathit{R}^{2} = 0.99).\] \end{document} The slope of the correlation between Dmet–silGe and oxygen fugacity allows one to determine the valence state of Ge in the silicate melt. Within the f O2 range of the experiments (−0.9 to −2.7 log units relative to the Fe-FeO (IW) buffer), the valence state of Ge is 2.00 (±0.12), i.e . GeO is the stable species in the silicate melt. This is in good agreement with an earlier study performed in 1 atm furnaces at 1300 °C and a comparable f O2 range (−0.5 to −2.7 log units relative to the IW buffer), but at variance with another study that determined tetravalent Ge cations dissolved as oxides in silicate melts in 1 atm experiments performed at more oxidizing conditions (−1.3 to +3.8 log units relative to IW buffer). The data of the present study in combination with these earlier data suggest a change in the dominant formal Ge oxide species in silicate melts at oxygen fugacities slightly below the IW buffer.

Solid solutions of (Mg, Fe2+)-cordierite: Synthesis, water content, and magnetic properties

The dependence of water concentration in synthetic (Mg, Fe2+)-cordierite on the composition of the solid solution was examined in experiments that lasted for 10 days at = 200–230 MPa, t = 600–700°C, and oxygen fugacity corresponding to the Fe-FeO buffer. Mass spectrometric data indicate that the dependence of water concentration in cordierite on its Fe mole fraction Fe2+/(Fe2+ + Mg) has maxima at compositions with F = 0.2–0.3. IR diffuse reflectance spectroscopic data and data on the structural setting of H2O molecules in the structural channels of alkali-free (Mg, Fe2+)-cordierite indicate that the H-H vector of some H2O molecules (H2O-II) is perpendicular to [001] of the crystal. The dependence of the magnetic properties of synthetic (Mg, Fe2+)-cordierite was studied by static magnetization technique at 5–300 K in an external magnetic field up to 20 kOe in strength.

The mobility of U and Th in subduction zone fluids: an indicator of oxygen fugacity and fluid salinity

The solubility of U and Th in aqueous solutions at P-T-conditions relevant for subduction zones was studied by trapping uraninite or thorite saturated fluids as synthetic fluid inclusions in quartz and analyzing their composition by Laser Ablation-ICPMS. Uranium is virtually insoluble in aqueous fluids at Fe-FeO buffer conditions, whereas its solubility increases both with oxygen fugacity and with salinity to 960 ppm at 26.1 kbar, Re-ReO2 buffer conditions and 14.1 wt% NaCl in the fluid. At 26.1 kbar and 800°C, uranium solubility can be reproduced by the equation: \( \log {\text{U}} = 2.681 + 0.1433\log f{\text{O}}_{2} + 0.594{\text{Cl,}} \) where fO2 is the oxygen fugacity, and Cl is the chlorine content of the fluid in molality. In contrast, Th solubility is generally low (<10 ppm) and independent of oxygen fugacity or fluid salinity. The solubility of U and Th in clinopyroxene in equilibrium with uraninite and thorite was found to be in the order of 10 ppm. Calculated fluid/cpx partition coefficients of Th are close to unity for all conditions. In contrast, Dfluid/cpx for uranium increases strongly both with oxygen fugacity and with salinity. We show that reducing or NaCl-free fluids cannot produce primitive arc magmas with U/Th ratio higher than MORB. However, the dissolution of several wt% of oxidized, saline fluids in arc melts can produce U/Th ratios several times higher than in MORB. We suggest that observed U/Th ratios in arc magmas provide tight constraints on both the salinity and the oxidation state of subduction zone fluids.

Mantle melting in equilibrium with an Iron–Wüstite–Graphite buffered COH-fluid

Partial melting experiments on a San Carlos peridotite were done in a Walker type multi-anvil press at pressures from 5 to 12.5 GPa. Experiments were done in the presence of a COH-fluid and at oxygen fugacity controlled by the Fe–FeO buffer. Olivine, clinopyroxene, garnet and orthopyroxene are stable in all but the highest temperature 10 GPa experiments where olivine and garnet coexist, and the highest temperature 5 GPa experiments where olivine is the single crystalline phase. The solidus at 5 GPa was found to be at approximately 1,200°C and the liquidus was estimated to be at 1,325°C, which is ∼500°C lower than has been reported for dry melting of peridotite. The aluminum concentration of the melts decreases with increasing melt fraction and decreases also with increasing pressure. At 5 GPa the melts have a CaO/Al2O3-ratio of 0.85–1.0, which is similar to that of undepleted komatiites; major element concentrations are also identical to those of undepleted komatiites such as the Munro komatiites. At 10 and 12.5 GPa the partial melts have CaO/Al2O3-ratios above 1.5 and major element composition almost identical to aluminum depleted komatiites such as the Barberton komatiites. We therefore conclude that in the presence of a reducing COH-fluid both aluminum-depleted and -undepleted komatiites could have formed at temperatures much lower than generally accepted.

Thermodynamic constraints on the formation conditions of winonaites and silicate-bearing IAB irons

Silicate inclusions in IAB irons and related winonaite meteorites have textures, mineralogies and mineral chemistries that indicate a complex formation history of heating, followed by brecciation and metamorphism. Using olivine-orthopyroxene-chromite assemblages in five IAB iron silicate inclusions (Caddo County, Campo del Cielo, Copiapo, Lueders, and Udei Station) and one winonaite (Winona), we calculated closure temperatures and oxygen fugacities for these meteorites. Calculated olivine-chromite Fe-Mg exchange temperatures are compared to two-pyroxene temperatures. Olivine-chromite closure temperatures range from ∼590°C to ∼700°C, while two-pyroxene temperatures range from ∼900°C to ∼1200°C. Oxygen fugacities of these meteorites, determined for the first time in this study, range from 2.3 to 3.2 log units below the Fe-FeO buffer and define a line between the Fe-FeO and Cr-Cr 2O 3 buffers. Highly variable temperatures were experienced by these rocks on the hand sample, and sometimes even the thin section, scale consistent with the idea that the winonaite-IAB iron parent body experienced collisional fragmentation and reassembly after peak temperatures were reached. Although modest reduction likely occurred during cooling, the oxygen fugacities and mineral compositions recorded at peak metamorphic temperatures suggest that the chondritic precursor for this parent body was initially more reduced than ordinary chondrites.

On type B CAI formation: experimental constraints on fO2 variations in spinel minor element partitioning and reequilibration effects

We report data from a series of dynamic crystallization experiments that focus on determining the partition coefficients (D’s) for V and Ti in the spinel + liquid system of an average type B1 CAI bulk composition for three different fO2 conditions. Partitioning data for Ca and Si are also obtained. We show that the D’s for V and Ti are fO2 dependent with DTi decreasing at low oxygen fugacity due to the presence of Ti3+. DV is essentially 0 in air, rises to 2.2 at the Fe-FeO buffer and drops to 1.4 at the C-CO buffer. This indicates that V3+ is highly compatible in spinel and that higher and lower valence states are much less compatible. We also report data from isothermal experiments that determine diffusion times for V and Ti in same system at a temperature close to the Tmax for type B1 CAIs. Diffusion of these elements between spinel and liquid is surprisingly rapid, with essentially total equilibration of Ti and V between spinel and liquid in 90 h run duration. Lack of equilibration of Cr, Si, and Ca shows that the Ti and V equilibration mechanism was diffusion and not dissolution and reprecipitation. Our experimental run durations set an upper limit of a few tens of hours on the time that type B1 CAIs were at their maximum temperature. Based on our data we argue that subsolidus reequilibration between spinel inclusion and host-silicate phases within type B CAIs likely explains the observed range of V and Ti concentrations in spinels which are inclusions in clinopyroxene.

Cordierite III: the site occupation and concentration of Fe3+

Cordierite has the ideal formula (Mg,Fe)2Al4Si5O18.x(H2O,CO2), but it must contain some Fe3+ to account for its blue color and strong pleochroism. The site occupation and concentration of Fe3+ in two Mg-rich natural cordierites have been investigated by EPR and 57Fe Mossbauer spectroscopy. In addition, powder IR spectroscopy, X-ray diffraction, and TEM examination were used to characterize the samples. Single-crystal and powder EPR spectra indicate that Fe3+ is located on T11 in natural cordierites and not in the channels. The amount in Mg-rich cordierites is very small with an upper limit set by Mossbauer spectroscopy giving less than 0.004 cations per formula unit (pfu). Fe3+ in cordierite can, therefore, be considered insignificant for most petrologic calculations. Heat-treating cordierite in air at 1,000 °C for 2 days causes an oxidation and/or loss of Fe2+ on T11, together with an expulsion of Na+ from the channels, whereas heating at the Fe–FeO buffer produces little Fe3+ in cordierite. Heating at 1,000 °C removes all class I H2O, but small amounts of class II H2O remain as shown by the IR measurements. No evidence for channel Fe2+ or Fe3+ in the heat-treated samples was found. The blue color in cordierite arises from a broad absorption band (E//b and weaker with E//a) around 18,000 cm−1 originating from charge-transfer between Fe2+ in the octahedron and Fe3+ in the edge-shared T11 tetrahedron. It therefore appears that all natural cordierites contain some tetrahedral Fe3+. The brown color of samples heated in air may be due to the formation of very small amounts of submicroscopic magnetite and possibly hematite. These inclusions in cordierite can only be identified through TEM study.