MR41A-1775
Hydrogen partitioning between FeNi and δ -AlOOH up to core-mantle boundary condition: Fate of water in the Earth's deep interior
The Earth's core consists of iron-nickel alloy containing of 10 % of light elements such as S, Si, O, C and H [Birch, 1952]. If hydrogen exists in the Earthfs core, the core density deficit can be explained by only 1 wt% addition of hydrogen due to its small atomic weight [Poirier, 1994]. δ-AlOOH is recently focused as a water carrier in the Earthfs deep mantle. Since δ-AlOOH is stable at least up to 134 GPa [Sano et al. 2008], this phase can bring water to the core-mantle boundary. If iron is in contact with hydrogen (H2) or water (H2O) under pressure, hydrogen can be incorporated into iron and form iron hydride (FeHx) [e.g., Fukai, 1984; Suzuki et al. 1984]. However, water behavior is unknown when iron alloy coexists with hydrous minerals at high pressure and high temperature. In this study, we have conducted in situ X-ray diffraction study of reaction between iron-nickel alloy and δ-AlOOH phase up to 126 GPa using diamond anvil cell and multi-anvil press and investigated hydrogen behaviour up to the core-mantle boundary. The results show that hydrogen dissolves into FeNi phase from low temperature conditions (900 K~) and to 1573-1673 K at 20 GPa. Although the temperature of iron-hydride (FeNiH) formation tends to increase (to around 2000 K) with increasing pressure (up to 126 GPa), this reaction temperature is well below the mantle geotherm. Therefore, hydrogen is likely to partition into the Earthfs core when δ-AlOOH brings the hydrogen to the core-mantle boundary. Dissolution of hydrogen into the core also affects significantly to the temperature condition of the Earth's core.
MR41A-1776
Partitioning of hydrogen between iron and ringwoodite and a potential water storages in the Martian core
Hydrogen is the most abundant element in the solar system. It has been proposed that the primordial Earth was covered by a hydrous magma ocean (Abe et al., 2000). In the hydrous magma ocean, metallic iron reacted with H2O and iron hydride was formed. Therefore, a large amount of hydrogen may have been transported into the Earthfs core during core formation. In Mars, the Martian transition zone composed mainly of wadsleyite or ringwoodite, which can contain large amount of water, may directly contacts with the Martian core. The water in the transition zone minerals (wadsleyite or ringwoodite) possibly reacts with the Martian core (metallic iron). In addition, there is also a possibility that hydrous silicate mantle reacts with iron- rich core of the icy satellites, such as Ganymede. In this study, we determined the partitioning behavior of water between iron and ringwoodite, which is the major constituent hydrous mineral in the deep mantle of Earth and Mars, at 16-21 GPa and 1273 K. Experiments were carried out using a Kawai-type multianvil apparatus (SPEED 1500) together with a synchrotron X-ray radiation at BL04B1 beamline of SPring-8, Japan. The amount of hydrogen dissolved in iron hydride was estimated based on the volume expansion of iron by interstitial hydrogen atoms. The result revealed that water (hydrogen) is partitioned strongly into iron compared to ringwoodite. Therefore, the planetary core possibility contains significant amount of hydrogen and hydrogen can be classified as a gsiderophileh element at high pressure. Reference: Y. Abe, E. Ohtani, T. Okuchi, K, Righter, M. Drake: In Origin of of the Earth and Moon, edited by R.M. Canup, K. Righter, Arizona Univ. Press., pp 413 (2000).
MR41A-1777
Partitioning Behavior of Oxygen Between Fe-S Alloy and Silicate
As known, the core is mainly composed of a Fe-Ni alloy. The core density deficit, compared to the density of pure iron, requires the presence of about 10 % of light elements. Among the likely candidates for light elements in the core are S, Si, and O. Previous studies have focused on the possibility of having S-Si and Si- O as the possible light elements in the core but few have focused on the possibility of an S-O combination as the light elements in the core. In this study, we report partitioning behavior of sulfur and oxygen between metal and silicate melts. The experiments were carried out using a piston cylinder apparatus and multi-anvil press at 2 GPa and 8 GPa for temperatures of 2273 K. We used 3 starting materials with a CI meteorite composition, prepared by varying the proportion of Fe-FeO and the sulfur content. Both MgO and graphite capsules were used in the experiments. Recovered samples were analyzed with an electron microprobe. The partitioning coefficient of sulfur and the weight percent of O in liquid metal alloy were studied as a function of pressure and oxygen fugacity. By using graphite capsule, our results demonstrate that as the system is more reduced, the O content in the metal decreases. In the same system, the oxygen content in the metal decreases with increasing pressure in our pressure range. By using MgO capsule, we also investigated the effect of the carbon on the Fe-S-O metallic alloy.
MR41A-1778
Chlorine Partitioning Between Mantle and Core: Implications for Early Earth Processes
The ratio of chlorine in the Bulk Silicate Earth (BSE) to primitive carbonaceous chondrites (or solar abundance) is an order of magnitude below that expected for a moderately volatile element. Given that crustal and mantle concentrations of Cl are fairly well constrained, the enormous Cl depletion can be attributed to any of three possible explanations: 1) a missing sink on Earth, 2) a ~500K overestimate of the 50% condensation temperature of Cl from the solar nebula or 3) an early volatile loss of Cl during Earth formation. Accepting the Cl concentrations of mantle and crust, McDonough (Treatise Geochem. 2003) proposed a core sink, with a Cl concentration of 200 ppm. We tested this hypothesis by conducting high pressure Cl partitioning experiments. Two experiments, one (A503) with equal proportions of primitive basalt (10.4 wt% MgO, Mg# 67) and pure Fe with 0.4 wt% FeCl2, and the second (A505) with equal proportions of basalt and Fe95.5S4.5 with 1.56 wt% FeCl2, were run in graphite capsules at 5 GPa and 1800° C in a Walker-style multianvil press. Both experiments produced homogeneous quenched silicate and metallic liquid. Average Cl contents in the quenched silicate melts were 5300±700 ppm for A503 and 10500±1000 ppm for A505 (2σ, n=6). Cl contents in metal were below detection (60 ppm) in A503 and in three of the six analyses of metal in A505; the average value for the above-detection points in that run was 140±50 ppm. The maximum metal-silicate DCl values are 0.011 and 0.018 for A503 and A505, respectively, three orders of magnitude lower than required for significant sequestration of Cl in the core. Unless there are strong effects of pressure and/or silicate liquid composition on DCl, our experimental results imply that Cl does not reside in the core. Explaining the low Cl concentrations on Earth via a lower 50% condensation temperature requires a decrease of >500K from previous estimates. This is inconsistent with all thermodynamic data for sodalite formation or the crystallization of NaCl(s) from vapor. The third possibility is that Cl was lost during expulsion of a primary atmosphere. Cl is classified as lithophile, but it is uniquely partitioned into water, making it 'hydrophile'. If Cl was strongly concentrated in an early ocean or atmosphere as HCl(g) or dissolved chloride, a large impact could volatilize the early ocean, causing Cl loss to space (e.g., Genda and Abe, Nature 2005), explaining the low Cl content on Earth.
MR41A-1779
Can Nitrogen be a Candidate for the Fe-Core Formation?
Among the light elements that have been added to mineral physics experiments concerning the Fe-rich core of the Earth, nitrogen is less favorable. In general, this is because metal-nitrides are thought to be rare within Earth. However this may not be because they are rare, but because nitrogen is difficult to detect by conventional electron microprobe analysis unless one is specifically looking for it. Theoretically, metal-nitrides could be equally considered as potential candidates for the light element in the core, not only because nitrogen forms strong metallic bonds, but also because metal-nitrides are common constituents of many iron meteorites. Some Fe-nitrides are found to be stable at extreme pressures and temperatures corresponding to Earth¡¯s core in both diamond anvil cell and shock experiments (Adler and Williams, 2005; Sekine et al., 2007). We have discovered a metal-nitride phase, TiN (osbornite) within a mantle mineralogical assemblage, opening a new opportunity to understand the history of Earth¡¯s core formation. The TiN was found in the mantle section of an unmetamorphosed Tibetan ophiolite, a fragment of former mid-ocean spreading center, which now marks the tectonic boundary between Asia and India. The osbornite occurs as inclusions in coesite pseudomorphic after stishovite, in association with FeTi alloy, native Fe, TiO2 II, cubic BN and diamond included in Os-Ir alloy, all from a massive chromitite ore body enclosed within harzburgite (Yang et al., 2007; Dobrzhinetskaya et al., 2007). The chromite also exhibits coesite and diopside exsolution lamellae (Yamamoto et al., 2007) that might suggest the calcium-ferrite polymorph of chromite as a precursor decompressed during upwelling. Measurements of δ15N with a Cameca 50 NanoSIMS using the same Focused Ion Beam foils prepared and used for earlier TEM studies suggest that the Tibetan osbornite is characterized by negative δ15N (-10 ‰). The δ15N results from the Tibetan osbornite are somewhat more negative than the most commonly measured value for Earth¡¯s uppermost mantle (δ15N = -3 to -5 ‰), and they are clearly different from the δ15N of shallow reservoirs. The latter include atmosphere, ocean, and crust having values of delta δ15N -- 0 - +5 ‰ for the atmosphere and ocean and +5 - +12 ‰.) for the crustal rocks and sediments. We conclude that the Tibetan osbornite contains mantle N, perhaps from an old and/or deep mantle reservoir. Apropos of the suggestion of N in the core, we point out that most iron meteorites have extremely negative δ15N values of -60 ‰ or more, hence it is conceivable that part of the N signal in our materials comes from a leaky core.
MR41A-1780
HCP vs. FCC vs. BCC iron phases in the Earth' s inner core
In order to understand the phase relations of the Earth' s core, we have established an internally consistent thermodynamic database for pure iron from the existing experimental data. Included phases are body- centered cubic (BCC) phases (α and δ phases), face-centered cubic (FCC) phase (γ phase), hexagonal close-packed (HCP) phase (ε phase), and liquid phase. We described fundamental thermodynamic relations as the Gibbs free energy divided into a thermochemical and a thermophyisical terms. The thermochemical data for the phases were evaluated from existing metallurgy database together with experimentally determined phase relations. The thermophysical term is obtained from the P-V-T equation of state (EoS) for the phases. To estimate the volume of the phases at high-P-T conditions beyond experimental conditions, we formulated the EoS in the framework of Mie-Grüneisen- Debye relation. For the FCC phase, we newly constructed an EoS from our internally-heated diamond cell experimental data. Using newly developed thermodynamic database, we calculated phase relations for pure iron to 360 GPa and 7000 K. The calculations showed that HCP-FCC-liquid triple point is located at 90 GPa and 2800 K. The FCC-HCP boundary has a curvature above 100 GPa and its metastable extension will be close and parallel to the HCP- liquid curve. The pressure-temperature dependence of the Gibbs energy is similar between the FCC and HCP phases. Therefore, there is a chance for the FCC phase to be stabilized at higher pressures. We have also explored the possibility of the BCC phase to be stable at high pressure. We investigated the effect of elastic parameter of the BCC phase on the iron phase relations. Depending on the elastic parameters within the reasonable range, the high-pressure BCC phase will be stabilized above 200 GPa. This hypothetical BCC phase will have a high melting temperature. In addition, a high-P BCC-HCP transition occurs at 200 GPa and 4000 K. These hypothetical phase relations may be the solutions for the existing discrepancies between the static and shock wave experimental studies. We will discuss the possible phase at the Earthfs inner core from our thermodynamic calculations.
MR41A-1781
Thermal expansion of Fe3C at high pressure and carbon in the Earth's inner core
Carbon is one of the major candidates for the principal light element in the Earth's core. Wood [1993] proposed that Fe3C, rather than iron-nickel alloy, is the dominant phase in the Earth's solid inner core. Testing the model of Fe3C-rich inner core requires knowledge on the thermal equation-of-state (EoS) of Fe3C under core conditions. To date, EoS data of Fe3C are only available at high pressure and 0 or 300 K [Scott et al., 2001, Li et al., 2002, Vocadlo et al., 2002] or at high temperature and 1 bar [Wood et al., 2004]. Wood et al. [2004] found that the thermal expansion coefficient is significantly affected by the ferromagnetic to paramagnetic transition above the Curie temperature. In this study, we have determined the thermal expansion coefficient of Fe3C up to 20 GPa and 1273 K, using a T-cup device and synchrotron x-ray diffraction techniques at beamline 13-ID of the Advanced Photon Source. Our results place constraints on the abundance of carbon the Earth's inner core. This work is supported by NSF EAR 06-09639. References: Gao et al. (2008), Geophys. Res. Lett., doi:10.1029/2008GL034817. Li, J. et al. (2002), Phys. Chem. Miner., 29(3), 166-169. Scott, H. P. et al. (2001), Geophys. Res. Lett., 28, 1875-1878 Vocadlo, L., et al. (2002), Earth Planet. Sci. Lett., 203(1), 567-575. 347. Wood, B. J. (1993), Earth Planet. Sci. Lett., 117(3-4), 593-607. Wood, I. G. et al. (2004), J. Appl. Crystallogr., 37, 82-90.
MR41A-1782
Phase stability of (Fe,Ni)3S at the core pressure conditions
It is considered that the Earth's core consists of iron and a few amount of light elements from the comparison between the density of pure iron and that of the Earth's core. Sulfur is one of the major candidates of the light element due to its highly solar abundance, easily making alloy with iron, existence in the iron meteorite. In the Fe-S system, Fei et al.(2000) reported that Fe3S phase is stable as subsolidus phase over 21 GPa. On the other hand, the Earth's core contains 5-10 % of nickel. Nickel affects to the stability of the crystal structure of iron-nickel alloy at ultra high pressure condition (e.g., Dubrovinsky et al., 2007). Although Fe3S with tetragonal structure is stable up to 80 GPa and 2500 K (Seagle et al., 2006), the stability of (Fe,Ni)3S is still uncertain. We report the stability of the (Fe, Ni)3S up to 111 GPa and 1800 K. (Fe,Ni)3S was synthesized at 21 GPa and 1173 K using a multi-anvil apparatus. Diamond anvil cell was used for high pressure experiments. Samples were placed in the sample chamber with sodium chloride pressure medium. We performed in situ powder diffraction experiments at BL10XU in SPring-8. (Fe,Ni)3S with tetragonal structure was observed up to 117 GPa which has lattice parameters of a = 8.131(9) Å, c = 3.940(5) Å. The unit cell volume of (Fe,Ni)3S was slightly larger than that of Fe3S calculated from the equation of state by Seagle et al. (2006) at 117 GPa. X-ray diffraction measurements show that tetragonal structure was stable up to 111GPa and 1800K and 117 GPa and 300K in the nickel bearing system.
MR41A-1783
Phase and Melting Relations of Fe-S System up to the Outer Core Conditions
The Earth's core may contain at least a few weight percent (wt%) sulfur based on the solar abundances [e.g., McDonough 2003] and high pressure partitioning experiments [Hillgren et al., 2000]. Phase relations of Fe-FeS system have been reported recently up to 80 GPa using a diamond anvil cell (DAC) [e.g., Chudinovskikh and Boehler, 2007; Campbell et al., 2007; Morard et al., 2008]. They reported the existence of Fe3S and Fe, which correspond to subsolidus phases in the Fe-FeS system. However, no experiments have been made under the core conditions. In this study, in situ X-ray experiments were performed up to 147 GPa and up to 2600 K using a laser-heated DAC at Photon Factory and SPring-8 in order to clarify the subsolidus phases of the Fe-FeS system under the core conditions. The starting material was a foil of Fe-13 atomic percent (at%) S (Fe-8wt%S) for the subsolidus experiments. For melting experiments, the starting sample which was a mixture of Fe3S and Fe was synthesized using multi anvil press from a powder mixture of Fe and FeS with a bulk composition of Fe-23.5at%S (Fe-15wt%S) at 21 GPa and 1113 K. The samples were sandwiched by powdered NaCl. Chemical compositions of the recovered samples were analyzed using EPMA. The present results show that Fe3S coexists with ε Fe in all of the PT conditions of the present experiments. Fe3S is stable up to 147 GPa and 2600 K. Therefore, Fe3S is likely to be the S- bearing iron-alloy phase under the core conditions. We did not observe any evidence for melting under the present experimental conditions. The present results also indicate that the eutectic temperatures of Fe-FeS system reported by Campbell et al. [2007] and Morard et al. [2008] are consistent with this study, whereas those reported by Chudinovskikh and Boehler [2007] are too low compared to our results. The sulfur content of Fe phase increases slightly with pressure, and the extrapolation of the pressure dependence of the sulfur content in iron indicates that the inner core can contain at least 3.7 at% (2.1 wt%) of sulfur. This value may explain the core density deficit (cdd) of the inner core, if we assume that the cdd of the inner core is 1% [Chen et al. 2007]. On the other hand, the cdd of the inner core might be larger up to 5% [e.g., Birch, 1964; Alfè et al., 2002]. If we adopt the inner core density deficit of 5%, 10.8 wt% (17.5 at%) of sulfur is required in the inner core. Since the sulfur content of the inner core is much higher than the amount of sulfur in Fe, coexistence of Fe and Fe3S are needed to account for the cdd of the inner core. The coexistence of the phases requires that the temperature of the inner core is lower than eutectic temperatures of Fe-FeS system, i.e., 5200K [Campbell et al., 2007] or 5950K [Morard et al., 2008] at the inner core boundary estimated by extrapolation of the Clausius-Clapeyron slopes of the eutectic temperature to higher pressure.
MR41A-1784
Phase relations of Fe-Si alloy up to core conditions: Implications for the Earth inner core.
Geophysical evidence indicates that the density of the inner core is about 2-5 wt% lower than the density of pure iron (Fe) under the conditions of the inner core, indicating the presence of light elements such as H, C, S, O, Si in the core. Recent studies report that a small amount of Si as an additional component can substantially affect the phase relations and thermodynamic properties of iron alloys. Therefore, study of the Fe-rich portion of the Fe-Si system and a knowledge of the effect of Si on the crystal structure and the equation of state of metallic Fe under the core conditions are essential for understanding the nature of the inner core. In this study, we determined the phase relation of Fe-3.4 wt% Si alloy up to 257 GPa and 2400 K using the double sided laser heated diamond anvil cell. X-ray diffraction experiments were conducted to 257 GPa and high temperature in situ on an iron-silicon alloy containing 3.4 wt% silicon, a candidate for the Earthfs inner core forming material. The results revealed that fcc and hcp phases coexist up to 104 GPa. A single hcp phase is stable at higher pressures at least up to 3600 K at 242 GPa and to 2400 K at 257 GPa. The cause of the seismic anisotropy has been a matter of great debates. The inner core phase responsible for the seismic anisotropy might be different depending on the Si content of the inner core. The present results indicate that it is likely that the inner core containing 3.4 wt% Si has an hcp structure, which can account for its seismic wave anisotropy.
MR41A-1785
Phase relations of iron-silicon alloys at high pressure and high temperature
The Earth's core is believed to be composed of iron-nickel alloy with one or more light elements. Although a number of elements lighter than iron, such as hydrogen, carbon, oxygen, silicon and sulfur, has been considered by various works, silicon is one of the most attracted candidates for the light elements in the core. If silicon is actually the major light element in the core, the phase relations of iron-silicon alloys, especially melting relations and solubility of silicon in solid iron at high pressure and high temperature, are keys to understand the structure, composition and crystallization of the inner core. The phase relations in the iron- silicon system at high pressure and high temperature, therefore, have been extensively studied. However, recent experimental studies with in-situ x-ray diffraction measurements have not given consistent results on the subsolidus phase relations in the iron-rich portion of iron-silicon alloys (Lin et al. 2002, Dubrovinsky et al. 2003, Asanuma et al. 2008). In this study, we have investigated the phase relations of iron-silicon alloys and solubility of silicon in hcp-iron based on in-situ synchrotron x-ray diffraction measurements in a laser-heated diamond-anvil cell (LH-DAC) along with chemical analysis of the quenched samples using a field-emission electron microprobe analyzer (JEOL JXA-8500F). For Fe-9.9 wt.% Si, high pressure experiments were conducted up to 130 GPa and 2600 K. Before heating, the sample has hcp structure. Upon heating to 1600 K at 45 GPa, the hcp phase transformed to a mixture of hcp and cubic phases, in agreement with Lin et al. (2008). In contrast, the sample remained in the hcp structure up to 2000 K at 130 GPa. These results indicate that the maximum solubility of silicon in solid hcp-iron increases with increasing pressure. Extrapolation of the present results suggested that solid hcp-iron at the inner core conditions could include up to about 10 wt.% Si. If silicon is one of the major light elements in the Earth's outer core, several percent of silicon may also dissolve into the inner core by the inner core solidification. We also studied more iron-rich compositions of iron-silicon alloys. Asanuma et al. (2008) previously conducted in-situ x-ray diffraction studies on Fe-3.4 wt.% Si and suggested that silicon might be a strong fcc stabilizer. According to this suggestion, the incorporation of larger amount of silicon should expand the stability field of the fcc phase. However, we did not observe the fcc phase for Fe-6.4 wt.% Si up to 2150 K at 65 GPa, indicating that the effect should be not so large.
MR41A-1786
In situ determination of binary alloy melt compositions in the LHDAC by X- Radiography
Constraining the light element in Earth's molten outer core requires an understanding of the melting phase relations in iron-light element binary systems. For example, it is critical to determine the composition of liquids at binary eutectics. Typically such measurements are carried out after the sample has been quenched in temperature and pressure. Such 'cook and look' methods possibly suffer from systematic errors introduced by exsolution of the light element from the melt on quench and error in the reintegration of the liquid composition [1]. Here, we present a novel method for the determination of melt compositions in iron-light element binary systems in situ in the LHDAC at simultaneous high-pressure, high-temperature conditions. Samples consist of a light element bearing compound, such as FeO, surrounded by a pure iron ring, forming a donut ~100 μm in diameter and ~15 μm thick. The donuts are loaded into stainless steel gaskets in the DAC, sandwiched between discs fabricated from sol-gel deposited nanocrystalline Al2O3 with similar dimensions to the donut. Pressure is monitored by ruby fluorescence during compression. The sample is heated at the boundary between the iron and light element compound using two 100 W IR lasers in a double-sided configuration at beamline 12.2.2 at the Advanced Light Source. Temperature is measured by spectroradiometry. Before, during and after melting, X-radiographic images of the sample are taken by shining a defocused beam of synchrotron X-rays through the sample and onto a CdWO4 phosphor. The visible light from the phosphor is then focused onto a high resolution CCD, where absorption contrast images are recorded. The absorption of the molten region is then determined, and it's composition calculated by linear interpolation between the absorption of the two solid end members. As a test of the reliability of the method we measured the Fe-FeS eutectic to 20 GPa and our results are in good agreement with previous studies that are based on various ex situ techniques. We measured the eutectic composition between Fe and Fe3C up to 44 GPa, and found that the carbon content of the eutectic drops rapidly above about 10 GPa, dropping to less that 1 wt% by 44 GPa. This result is generally consistent with the thermodynamic calculations of Wood [2]. Experiments on the Fe-FeSi eutectic yielded an increase in the Si content of the eutectic to 35 GPa, consistent with data from large volume press experiments [3] Notably, melting experiments at 35-43 GPa and ~2500 K on a boundary between Fe and FeO failed to yield evidence of a melt with a composition distinguishable from pure iron. However, an experiment at 12 GPa and 2700 K between Fe and FeO(OH) did yield a melt with a composition intermediate between the two end members. This suggests that O solubility in the Fe-O eutectic melt is low at mid-mantle pressures, but that H may dissolve into the melt by itself or in combination with O. [1] Walker, D., 2005. Core-Mantle chemical issues. Canad. Min., 43, 1553-1564 [2] Wood, B. J., 1993. Carbon in the core. Earth Planet Sci. Lett., 117, 593-607 [3] Kuwayama, Y. & Hirose, K., 2004. Phase relations in the system Fe-FeSi at 21 GPa. Am. Min., 89, 273-276.
MR41A-1787
Melting of Fe-Ni Alloy at High Pressure and Temperature
The Earth's core is believed to be composed primarily of Fe with substantial amount of Ni, and it is widely accepted that the Ni content is different between inner core and outer core due to the crystallization of liquid alloy. Therefore, the melting phase relation of Fe-Ni system at the core condition is important to constrain the composition and the structure of inner and outer core. However, while many experimental studies have investigated the subsolidus phase relations of Fe-Ni system at high pressure and temperature, its melting phase relation has been poorly constrained. Here, we investigated the melting relations of Fe-rich Fe-Ni alloys up to 65 GPa and 2800 K. High P-T conditions were generated using a laser-heated diamond- anvil cell technique. Chemical analyses of recovered sample were performed using a field-emission type electron probe microanalyser. The results at 7 GPa showed that the solid phase was depleted in Ni than coexisting liquid, which is consistent with phase relation at ambient pressure. In contrast, at pressure above 25 GPa, we observed the solid phase was enriched in Ni than coexisting liquid, furthermore the difference of Ni content between solid and liquid phase is larger than that in lower pressure. These results suggest that the inner core may include more than ten percent of Ni.
MR41A-1788
Effects of pressure and composition on Pt-Re-Os partitioning behavior between solid and liquid metal in the Fe-Ni-S system: Implication for Os isotopic anomalies in plume-derived lavas
Coupled 186Os/188Os and 187Os/188Os enrichments of plume-derived lavas have been suggested to reflect material contribution from the outer core (e.g., Brandon, 1998). This geochemical hypothesis is based on an assumption that the outer core shows coupled enrichments in 186Os/ 188Os and 187Os/ 188Os ratio, reflecting the decay of 190Pt and 187Re to 186Os and 187Os, respectively. In order to examine this hypothesis, partitioning experiments of Pt-Re-Os between solid metal and liquid metal were performed using an MA-8 Kawai-type multi-anvil apparatus at 5-20 GPa and 1250-1400C. Starting materials of Fe metal, Ni (7 wt.%) metal and FeS (5 wt.% S in the bulk) were doped with 3 wt.% of Pt, Re and Os metals. Concentrations of all elements were determined using JXA-8800M electron probe microanalyzer with wave-dispersive spectrometry. Measured partition coefficients of Pt, Re and Os increase with increasing sulfur content and almost constant with increasing pressure. Therefore, the effect of liquid composition on the partitioning behavior of highly siderophile elements is much more significant compared to the effect of pressure and temperature. On the basis of the present experimental results, it is unlikely to generate the required Pt-Re-Os fractionation during inner core crystallization assuming that the light element in the Earthfs core is sulfur only.
MR41A-1789
Formation And Deformation Of The Inner Core
Crystallization of the inner core is driven by the outer core heat transfer. The columnar convection in the outer core enhances the growth of the inner core in its equatorial region, generating a slow motion from the equator to the poles within the inner core (Yoshida et al, JGR 1996). During the inner core formation, the crystallization process increases the concentration of light elements in the liquid outer core. Therefore, the concentration of light elements in the solid phase also increases with time, generating a more and more stratified inner core. A numerical model of this two processes show interesting behaviours for Earth like parameters. This model offers the possibility to form a layered inner core. Different layers may present different textures and possibly explain the observed depth dependance of seismic anisotropy, including the presence of an innermost inner core and shallow isotropic layers.
MR41A-1790
Constraints on core formation from systematic study of metal-silicate partitioning on a great number of siderophile elements
The abundances of siderophile elements in the Earth's mantle are the result of core formation in the early Earth. Many variables are involved in the prediction of metal/silicate siderophile partition coefficients during core segregation: pressure, temperature, oxygen fugacity, silicate and metal compositions. Despite publications of numerous results of metal-silicate experiments, the experimental database and predictive expressions for elements partitioning are hampered by a lack of systematic study to separate and evaluate the effects of each variable. Only a relatively complete experimental database that describes Ni and Co partitioning now exists but is not sufficient to unambiguously decide between the most popular model for core formation with a single stage core-mantle equilibration at the bottom of a deep magma ocean (e.g. Li and Agee, 2001) and more recent alternative models (e.g. Wade and Wood, 2005; Rubie et al., 2007). In this experimental work, systematic study of metal silicate partitioning is presented for elements normally regarded as moderately siderophile (Mo, As, Ge, W, P, Ni, Co), slightly siderophile (Zn, Ga, Mn, V, Cr) and refractory lithophile (Nb, Ta). Using a new piston-cylinder design assembly allows us to present a suite of isobaric partitioning experiments at 3 GPa within a temperature range from 1600 to 2600° C and over a range of relative oxygen fugacity from IW-1.5 to IW-3.5. Silicate melts range from basaltic to peridotite in composition. The individual effect of pressure is also investigated through a combination of piston cylinder and multi anvil isothermal experiments from 0.5 to 18 GPa at 1900° C. Absolute measurements of partitioning coefficients combining EMP and LA-ICPMS analytical methods are provided. New results are obtained for elements whose partitioning behavior is usually poorly constrained and not integrated into any accretion or core formation models. We find notably that Ge, As, Mo become less siderophile with increasing temperature. In contrast Zn, Nb and Ta become more siderophile while Ga, W and P show negligible dependence with increasing temperature. Moreover, As, Mo, W, Ga and Nb become less siderophile with increasing pressure while a small influence of pressure is observed for Ta, Ge and Zn. At 3 GPa, regressions of the partitioning data show a 5+ valence state for As, Mo, W and Ta, 3+ for Ga, and 2+ for Ge and Zn. Finally regressions show that highly charged cations (Nb, P, W, Mo and As) are, as expected, the most sensitive to variations in silicate melt composition with the exception of Ta that shows a surprisingly small dependence. Generally, models of partitioning behaviors during core segregation are obtained for each element and seem to exclude the possibility of a single stage equilibrium scenario for the earth's core formation.
MR41A-1791
Fracture-induced flow and liquid metal transport during core formation
The most important event in the early history of the earth was the separation of its iron-rich core. Core formation induced profound chemical fractionations and extracted into the core most of Earth's iron and siderophile elements (Ni, Co, Au, Pt, W, Re), leaving the silicate crust and mantle with strong depletions of these elements relative to primitive planetary material. Recent measurements of radiogenic 182W anomalies in the silicate Earth, Mars and differentiated meteorites imply that planetesimals segregated metallic cores within a few Myr of the origin of the solar system. Various models have been put forward to explain the physical nature of the segregation mechanism (Fe-diapirs, 'raining' through a magma ocean), and more recently melt flow via fractures. In this contribution we present the initial results of a numerical study into Fe segregation in a deforming silicate matrix that captures the temperature-dependent effect of liquid metal viscosity on the transport rate. Flow is driven by pressure gradients associated with impact deformation in a growing planetesimal and the fracture geometry is constrained by experimental data on naturally deformed H6 chondrite. Early results suggest that under dynamic conditions, fracture-driven melt flow can in principle be extremely rapid, leading to a significant draining of the Fe-liquid metal and siderophile trace element component on a timescale of hours to days. Fluid transport in planetesimals where deformation is the driving force provides an attractive and simple way of segregating Fe from host silicate as both precursor and primary agent of core formation
MR41A-1792
Viscosity of Liquid Fe-17wt% Si at High Pressure and Temperature
In situ X-ray radiography falling-sphere experiments on liquid Fe-17wt% Si viscosity were carried out from 2 GPa to 7 GPa at APS and Spring-8 in multi-anvil apparati. Video images were recorded at speeds of up to 62 frames/sec. Both Re spheres coated with alumina and composite spheres of Pt or Re core and a mechanically prepared ruby mantle were used in the high pressure melts to avoid chemical reaction between the sample and the probing metallic spheres. The viscosity at the melting temperature was calculated from activation energy, which was determined from a combination of theoretical and experimental values of viscosity at ambient pressure. At the early stages of the compression (up to ~ 5.4 GPa), the viscosity increases but later appears to approach a constant value of 69 mPa.s in the higher pressure range. The constant relating activation energy to melting temperature, g, is 6.8 from this study. Assuming that temperature varies adiabatically in the core and melting temperature Tm at the inner core boundary is 4766 K and dTm dP = 10 K/GPa, the viscosity at the core-mantle boundary, inferred from this study, decreases to a value very close to the ambient pressure viscosity of 6 mPa.s for liquid metal.
MR41A-1793
Core Formation in Giant Gaseous Protoplanets
Sedimentation rates of silicate grains in gas giant protoplanets formed by disk instability are calculated for
protoplanetary masses between 1 MSaturn to 10 MJupiter. Giant protoplanets with masses of 5
MJupiter or larger are found to be too hot for grain sedimentation to form a silicate core. Smaller
protoplanets are cold enough to allow grain settling and core formation. Grain sedimentation and core
formation occur in the low mass protoplanets because of their slow contraction rate and low internal
temperature. It is predicted that massive giant planets will not have cores, while smaller planets will have
small rocky cores whose masses depend on the planetary mass, the amount of solids within the body, and
the disk environment. The protoplanets are found to be too hot to allow the existence of icy grains, and
therefore the cores are predicted not to contain any ices. It is suggested that the atmospheres of low mass
giant planets are depleted in refractory elements compared with the atmospheres of more massive planets.
These predictions provide a test of the disk instability model of gas giant planet formation.
The core masses of Jupiter and Saturn were found to be ~ 0.25 M⊕ and ~ 0.5
M⊕, respectively. The core masses of Jupiter and Saturn can be substantially larger if planetesimal
accretion is included. The final core mass will depend on planetesimal size, the time at which planetesimals
are formed, and the size distribution of the material added to the protoplanet. Jupiter's core mass can vary
from 2 to 12 M⊕. Saturn's core mass is found to be ~ 8 M⊕.
MR41A-1794
Investigating the Thermal Effects of Water on the Differentiation of Large Proto-planetary Bodies in the Early Solar System
Arguments regarding the origin of iron meteorite parent bodies suggest that they accreted and differentiated in the inner Solar System amongst the material that would ultimately become the terrestrial planets (Bottke et al. 2006). Results from high-precision W isotopic measurements (Markowski et al. 2006) suggest that these parent bodies experienced metal-silicate segregation no later that 1 Myr after CAI formation, making them some of the earliest solid bodies to have formed. However, there is considerable uncertainty regarding the water content of proto-planetary material that originated in the terrestrial planet region of the Solar System. Planetesimal formation in an optically thick dust disk may have resulted in bodies with water contents as high as 50% by mass (Machida & Abe 2006), whereas trends in the water contents of carbonaceous, ordinary and enstatite chondrites as a function of heliocentric distance suggest that the terrestrial planets formed from anhydrous material with water mass fractions of ~ 0.001% (Raymond et al. 2004). A possible range of four orders of magnitude in the initial water content of planetesimals in the inner Solar System would clearly influence the thermal evolution of these bodies: water acts as a thermal buffer due to its high heat capacity and could cause significant exo- and endothermic water-rock reactions (Cohen & Coker 2000). We will present thermal evolution calculations to investigate the effect of water on the timescales of differentiation for large (~100-1000 km size) proto-planetary bodies. The initial temperatures of these bodies will be set to that of the ambient solar nebula (~180 K) and their initial compositions will be assumed to be a mixture of water ice and silicates. The temperature evolution will primarily be dictated by the decay of short-lived radioactive isotopes like 26Al (Grimm & McSween 1993), which will be the dominant heat source for time scales of several half-lives (~3 Myr). Particular attention will be paid to the thermal effects that hydration and dehydration reactions will have on this evolution. We will consider specific serpentinization reactions and differences in the enthalpies of reaction for the forward and reverse reactions as a function of temperature. We will also explore the thermal buffering effect of water on the overall thermal evolution of these bodies. The ultimate goal of this study is to constrain the conditions in a thermal evolution parameter space (e.g. time of accretion, water abundance, heats of reaction) that will result in thermal histories that are consistent with the differentiation timescale constraints provided by the iron meteorites.