MR41A-01 INVITED 08:00h
New Constraints on the Earth's Core Chemical Composition
In the last forty years, there has been a considerable debate about which light element among sulfur, silicon, oxygen, carbon or hydrogen should be in the core [Poirier, {\it Phys. Earth Planet. Int.}, {\bf 85}, 319, 1994]. Not only the nature of these elements is a standing problem of prime importance, since it conditions the existence of a freezing point depression at the inner core boundary, but also their distribution within the core is unknown. It is indeed crucial to determine to what extent light elements are released in the liquid outer core, thus inducing solutal convection which in turn contributes to power the geodynamo [Loper, {\it J. R. Astron. Soc.}, {\bf 54}, 389, 1978]. In this respect, density and sound velocity measurements at high pressure in solid iron alloyed with different light elements are important to constrain core dynamics and coupling between liquid outer core and solid inner core. However, sound velocity data were not available until very recently, with sound velocity measurement in iron at high pressure [Fiquet et al.; {\it Science}, {\bf 291}, 468, 2001; Mao et al., {\it Science}, {\bf 292}, 914, 2001; Antonangeli et al., {\it Earth Planet. Sci. Lett.}, {\bf 225}, 243, 2004]. The question now is how to constrain the relative abundance of these light elements, and eventually rule out some of them based on a confrontation of seismic and mineralogical data. Here, we report direct measurements of acoustic sound velocity in iron alloyed with light elements supposedly entering in the composition of the Earth's core, {\it i.e.} oxygen, sulfur and silicon, and address the question of the composition of the core. In this work, we measured longitudinal sound velocities in light-element alloys of iron (FeS, FeO, FeS$_{2}$, and FeSi) at high pressure by inelastic X-ray scattering. This data set provides a new mineralogical constraint on the composition of the Earth's core, and completes the previous set formed by compressibility and density measurements for these compounds. The combination of these data sets and their comparison with the reference Earth models derived from seismology allows us to determine an average composition of the Earth's core.
MR41A-02 INVITED 08:15h
Study of sound velocities of hcp-Fe with nuclear resonant inelastic x-ray scattering in a laser-heated diamond cell
Iron is the most abundant component in the Earth's core. Understanding the physical properties of Fe under core conditions is essential for interpreting the seismological and geomagnetic observations deep in the Earth's interior. The physical properties of Fe have been extensively studied by theoretical calculations and dynamic and static high-pressure experiments, but direct static measurements of the sound velocities of iron under high pressures and temperatures are still lacking. We have studied hcp-Fe with nuclear resonant inelastic x-ray scattering technique in a laser-heated diamond anvil cell. Phonon density of states and sound velocities of hcp-Fe have been measured up to 70 GPa and 1700 K. We found that temperature has a strong effect on the sound velocities; the compressional (V$_{P}$) and shear wave velocities (V$_{S}$) of hcp-Fe decrease significantly with increasing temperature under high pressures. We note that our measurements in VS are relatively insensitive to the differences in the EOS data used in deriving the sound velocities. Therefore, the NRIXS technique is particular good at constraining V$_{S}$ with a precise measurement of V$_{D}$. Thermodynamic, elastic, and vibratonal properties of hot dense hcp-Fe have also been obtained from the measured phonon density of states. Our results have important implications for understanding the sound velocities of the Earth's inner core as well as the fundamental physical properties of iron under extreme pressures and temperatures.
MR41A-03 08:30h
Seismological Constraints on Core Composition From Core Liquid Immiscibility
Liquid immiscibility in iron-rich liquids at atmospheric pressure forms the basis of smelting, which separates an iron-rich liquid from a second liquid in which contaminating elements are concentrated. The earth's core is a similar iron-rich system that contains light elements that progressively enrich the outer core liquid as the core cools and the inner core crystallizes. Thermodynamic modelling of Fe-O-S liquids shows that immiscible liquids can exist at outer core pressures at temperatures less than 5000$^\circ$K. However, there is no evidence for liquid layers in the outer core in the travel times and waveforms of seismic waves that reflect internally in the core, P4KP. The absence of layers therefore constrains core compositions in the Fe-O-S system to be no richer than 4 wt.% oxygen and 2-11 wt.% sulfur, the remainder being iron.
MR41A-04 INVITED 08:42h
At the frontiers of the Earth's fluid core
The fluid core exchanges energy and momentum with the mantle at the Core-Mantle Boundary (CMB) and crystallizes at the Inner core Boundary (ICB). The dynamics of the core depends on the structure of both boundaries. We investigate here the roughness of the CMB, possibly conducive to dissipation, as well as the existence of a mush at the ICB. The roughness of the CMB, on the scale of centimeters to tens of meters, was modelled using a cellular automata method. Cubic cells, of the size of grains of the mantle material, on a 3-D grid, can be in one of three states, corresponding to mantle silicate or oxide, core fluid saturated in light element and unsaturated core fluid. We assumed a stationary stochastic process of evolution, without memory. Transitions of doublets of cells from one state to another are governed by parameters representing the rates of physical processes : dissolution and crystallization of mantle material at the CMB and diffusion of the light element in the core fluid. With reasonable values of the parameters, it is found that the boundary roughens on the scale of grains, and a boundary layer of saturated fluid, a few tens of centimeters thick, soon appears at the interface. An undulation with dominant wavelength of the order of a few tens of meters eventually appears. An order of magnitude of the resulting dissipation in the fluid flow, due to the roughness, is given. At the ICB, it is currently believed that a mushy layer develops, while the outer core material freezes onto the inner core. However, this view is derived from metallurgical or analog experiments on solidification carried out in the laboratory, in conditions widely different from those obtaining in the core. At the ICB, the temperature gradients are negative and many orders of magnitude smaller than the positive gradients in the laboratory, while the velocity of the solid-liquid interface is about 0.3 mm/year instead of, typically, 0.01 to 0.1 mm/s. Using a modified Mullins-Sekerka linear stability analysis, we investigated the conditions in which constitutional supercooling and instability of the interface (with dendrite and mush formation) coexist. We found that a mushy layer at the ICB is likely to exist if the liquidus slope (melting T vs concentration of light element) is of the order of 100 K or higher (as is currently assumed). However, if it is of the order of 0.1 K, which might occur if the major light element is oxygen, the interface might be stable and the mush would be replaced by a slurry, with a drizzle of solid particles onto the inner core.
MR41A-05 08:57h
The First Terrestrial Pb-isotope Paradox and the Core Formation
The first terrestrial Pb-isotope paradox refers to the fact that the Pb isotope compositions of most marine sediments and MORB plot well to the right of the area defined by meteorites in 207Pb/204Pb vs. 206Pb/204Pb space. This paradox requires at least one reservoir formed in the early Earth's history that plots to the left of the meteorite isochron. The core is one candidate that have been widely discussed in the past. In order to quantify possible fractionation into the metallic core, we have done different set of multi-anvil experiments at high pressure (up to 20 GPa) and high temperature (1900°C). The partitioning of Pb between liquid metal and silicates is investigated at different oxygen fugacity by using different metallic starting compositions. The oxygen fugacity in most experiments has been calculated by using two independent buffers : iron/wüstite (IW) and SiO2/Si, allowing to link consistently the Fe contents in silicates, the Si contents in metal and the temperatures of the experiments. The concentration of Pb in metal and in silicates has been measured by electron microprobe and ?PIXE. At oxygen fugacities 3 log units below IW, we have found partition coefficients between metal and silicate melt ranging from 6 to 9 depending on the starting composition of the metal. A first conclusion for this study is that Pb would remain siderophile even with a metal without sulphur. These first results heighten the importance of :(1) the hypothesis that the fisrt Pb-paradox could be the result of the Pb incorporation into a metallic phase at the time of core formation, (2) doing more experimental work at different pressures, temperatures and fO2 relevant to the early Earth.
MR41A-06 09:09h
Experimental constraints on silicon in the Earth's core
The light element composition of the core is fundamental to our understanding of the nature and dynamics of the Earth but is poorly constrained. One of the criteria for the principal light element in the core is that its partitioning between the inner core and outer core must reproduce the observed density contrast at the inner core boundary (ICB). To apply this criterion, we need to know the melting relations in the candidate Fe-light-element systems. Presently, experimental data on high pressure melting behavior of Fe-alloy systems are scarce. We have determined the melting relations in the Fe-Si binary system to 27 GPa, using in situ x-ray diffraction and x-ray radiograph techniques at BL04B1 beamline, SPring-8. Recovered samples have been analyzed using an electron probe microanalyzer at the Geophysical Laboratory, Carnegie Institution of Washington. Combining the structural data from x-ray diffraction and compositional data from electron probe analysis allows us to reevaluate the proposal of silicon as the major light element in the Earth's core.
MR41A-07 INVITED 09:21h
Numerical Models of the Geodynamo - Can They be Scaled to the Core?
Numerical dynamo models have been successful in the sense that they reproduce many properties of the geomagnetic field, such as its strength, dipole dominance, secular variation and stochastic reversals. They show that core-mantle coupling may explain non-axisymmetric features in the long-term magnetic field, for example preferred paths of the virtual geomagnetic pole during reversals. However, doubts remain on how realistic these models are, because several key parameters are far from Earth-like. Systematic parameter studies have become affordable and their goal is to establish scaling laws for the dependence of dynamo properties on the control parameters that can be extrapolated to Earth values. In a first step we recently derived a relation between the ohmic dissipation time and the magnetic Reynolds number. From this we estimate a rather moderate power requirement of the geodynamo, which implies that the inner core may be older than 2.5 Gyr. The amplitude of the zonal flow, which can be inferred from secular variation, scales with the square root of the buoyancy flux in the core. Assuming a dominant compositional source of buoyancy, this provides a maximum estimate of 0.1 mm/yr for the growth of the inner core, in agreement with an early formation of the inner core. An ambitious goal for the future is to determine what parameters control the strength of the magnetic field.
MR41A-08 09:36h
High-pressure alloying of xenon and iron: 'Missing' Xe in the Earth's core?
Noble gas xenon (Xe), with its tendency to not react with other elements, is ideal for studying the evolution of planets through investigation of its daughter products from radioactive isotopes (e.g., $^{244}$Pu (and $^{238}$U) --$>$ $^{136}$Xe, $^{129}$I --$>$ $^{129}$Xe systematics). However, the Earth is 'missing' some xenon in its atmosphere: Xe is much more depleted than expected from chondritic abundances. Mere hydrodynamic escape or impact vaporization, is not enough to account for the missing Xe since even lighter rare gases like argon and neon are less depleted than heavier xenon. Instead, the missing Xe may be hidden deep in the Earth's interior where there is little communication with the surface. To test the possibility of Xe allyoing with iron (Fe) during core formation and its thermodynamic stability as a solvent in the core, we perform density-functional based {\it ab-initio} calculations of Xe incorporation into the hexagonal close packed (hcp) high-pressure phase of iron, $\epsilon$-Fe. To that extent we set up supercells of hcp Fe of various sizes in which we incorporate Xe. Using the projector augmented wave method as implemented in the Vienna {\it Ab-initio} Simulation Package (VASP) we evaluate their energetics and stability relative to the elemental solids. We find that under static conditions up to 1 mol$%$ Xe can be alloyed into Fe at high pressure, suggesting that Xe may have been incorporated into the iron-rich core during core segregation. Substitutional incorporation of Xe into $\epsilon$-Fe causes the hcp structure to expand, depending on pressure and concentration, with 1 mol$%$ (~10$^{4}$ ppm by weight) substitution causing about 1$%$ volume expansion at core pressures. This potential alloying behavior could be crucial for understanding the accretion and evolution of the solid Earth and its atmosphere and could possibly provide an explanation to the Earth's 'missing' Xe.
MR41A-09 09:48h
Timing And Processes Of Earth's Core Differentiation.
Small $^{182}$W abundance excess of terrestrial W relative to W in bulk chondrites has been recently established (Yin et al. 2002, Kleine et al. 2002, Schoenberg et al. 2002). Rapid terrestrial accretion and early core formation, with completion of the bulk metal-silicate separation within less than 30 Myr have been proposed on this basis. These studies underline how much this $^{182}$W/$^{182}$Hf time scale agrees with dynamic accretion models (Wetherill, 1986) that predict a $\sim$10 Myr interval for the main growth stage of Earth's formation. This W model time scale for terrestrial accretion is shorter than current estimates based on Pb isotope systematics of mantle-derived basalts and terrestrial Xe isotope systematics. The end of metal-silicate differentiation and large scale mantle degassing has been defined $\sim$100 My after beginning of the accretion. These studies also indicate agreement of this time scale with dynamic accretion models that predict 100 My for the end of Earth's accretion. The Hf-W time scale for accretion and core formation assumes total equilibration of incoming metal and silicate of impactors with the bulk silicate Earth (BSE) during planet's growth. Recently, the assumption of incomplete equilibration of metal and silicate components with BSE has been investigated (Halliday, 2004). It is proposed that impacting core material has not always efficiently mixed with the silicate portions of the Earth before being added to the Earth's core Our approach also considers that equilibration between metal and silicate has not been complete in BSE during Earth's growth, and we argue that early part of the Earth's core has segregated through unmelted silicate material. When the baby Earth was large enough, the increase of the temperature induced Fe-FeS eutectic melting. The liquid metal segregated through the crystalline silicate matrix and formed the early part of the Earth's core. Experimental study (Yoshino et al. 2003) indicates the percolation threshold for molten iron-sulphur compounds of 5 vol% solid olivine, through channel on triple junction between minerals. This study allows us to reconsider the precedent proposition (Stevenson, 1990) based on experimental and theoretical considerations suggesting that percolation of metallic iron rich liquid through a mostly solid silicate matrix is largely prevented because of the high surface tension of iron. During formation and segregation of the Fe-FeS eutectic, W isotopic equilibration is limited by the diffusion through the solid silicate matrix. During the further Earth's growth, impact melting increased and has induced a progressive melting of BSE up to the formation of magma ocean at the end of the planet's accretion. Before the occurrence of the magma ocean, W equilibration between impactors and BSE has not been complete This incomplete isotopic exchange between terrestrial metal and metal originating from impactors with solid part of BSE during early accretion of the Earth leads to the observed excess of $^{182}$W of present BSE. It occurs when the $^{182}$W production in BSE is most significant, due to the short half-life of $^{182}$Hf. The change of segregation mechanisms of Earth's core during planet's growth and short-sightedness of Hf-W chronometer focused to the early segregation of Earth's core make the divergence with the U-Pb and I-Xe terrestrial records. Yin et al. 2002, Nature 418, 949-952. Kleine et al. 2002, Nature 418, 952-955. Schoenberg et al. 2002, Geochim. Cosmochim. Acta 66, 3151-3160. Wetherill 1986, in Origin of the Moon, eds Hartmann et al., LPI, 519-550. Yoshino et al. 2003, Nature 422, 154-157. Stevenson 1990, in Origin of the Earth, eds Newson et al., LPI, 231-249.