MR43A-0865 1340h
Observation of the Fluid Core Resonance With two Laser Extensometers at Gran Sasso, Italy
Since several years two laser extensometers are operating at a depth of 1400 m in the Gran Sasso underground observatory (Central Italy). The two orthogonal baselines, striking N66E and N24W, are 90-m long. Nominal sensitivity is about $3 \times 10^{-12}$ and recording rate is 5 Hz. Until 1999 only one Michelson interferometer measured difference in extension between the two baselines in an equal-arm configuration, after this year two independent unequal-arm interferometers are working, monitoring extension along two orthogonal directions. Tidal analysis of strain data has been performed using different codes in order to evidence the fluid core resonance effect in the diurnal tidal band and study the resonance function. The unusual depth of the station largely reduces contamination caused by atmospheric effects in the diurnal band of recorded tides, as indicated by the not so large amplitude of the S1 wave. Predicted ocean loading effects are small, mainly for one of the two monitored directions. In this condition uncomplete removal of oceanic tidal loadings and atmospheric effects can be expected to to affect results to a minor extent. Our results will be compared with predictions and previous works.
MR43A-0866 1340h
Interdiffusion of Iron and Nickel at High Pressure
To better understand diffusion-controlled properties of Earth's inner core, such as viscosity, we performed a series of multi-anvil experiments to measure Fe-Ni interdiffusion coefficients at pressures up to 23 GPa and temperatures up to 1800 $\deg$C. Diffusion couples consisting of a pure Fe and a pure Ni rod were compressed, heated and held at constant temperature for 0.5 to 6.5 hours, then quenched. An electron microprobe was used to analyze Fe and Ni concentrations perpendicular to the diffusion interface, and diffusion coefficients were determined as a function of alloy composition using the Boltzmann-Matano method. In all experiments, Fe-Ni interdiffusion coefficients were found to increase smoothly across the diffusion couple from pure Fe to pure Ni, with diffusion in almost pure nickel about two orders of magnitude faster than in iron. The diffusion coefficient at a given alloy composition decreases with increasing pressure, but the data are not consistent with a constant activation volume; instead the activation volume decreases with pressure. The entire data set for Fe-Ni interdiffusion in iron-rich alloys, including 1 atm and 4 GPa data of Goldstein et al. (1965; Trans AIME 233:812-820), is reproduced with reasonable accuracy by a simple exponential function of the reduced melting temperature, D=D$_{0}$exp(-bT$_{m}$/T), where b is a constant. This homologous temperature relation appears to be the most accurate method for extrapolating diffusion coefficients to high pressure. At inner core conditions (T/T$_{m}$ $\sim$ 0.9), the Fe-Ni interdiffusion coefficient for an alloy with the fcc structure is predicted to be $\sim$4$\times$10$^{-14}$ m$^{2}$/s. Diffusion coefficients in the inner core, which is thought to be composed of an iron-nickel alloy with the hcp or bcc structure, are likely to be higher since diffusion in hcp and bcc metals is faster than in fcc metals at the same homologous temperature.
MR43A-0867 1340h
Potassium as a Radioactive Heat Source in the Core? A High Pressure Experimental Study
The presence of potassium (K) in the core as a significant heat source was suggested over three decades ago. Experimental studies on K partitioning between metal and silicate have provided ambiguous results, because of experimental and analytical artefacts. It has been recognized that polishing of a run product for chemical analysis using water or oil lubricants results in substantial K loss from the metallic phase [Murthy et al., 2003, Nature 423]. Using a dry polishing technique, Murthy et al showed that K enters sulfide-rich metallic liquids with a strong dependence on temperature and silicate melt composition, but without a significant dependence on pressure over the range of their study (1-3 GPa). Extrapolating their data to conditions of pressure, temperature and melt structure, appropriate to core segregation, Murthy et al concluded that K is a substantial radioactive heat source in planetary cores. Their extrapolation technique is debatable, however, notably concerning the effects of composition and pressure on the partitioning. The aim of our study is therefore to reexamine the factors that can affect K partitioning between metallic liquid and silicate melt. We have performed multi-anvil experiments on a relatively wide pressure range, between 3 and 8 GPa, using graphite capsule. In contrast to Murthy et al who used compositions with high S and K contents, we used a CI-chondrite model composition (containing about 4000 ppm K) as starting material in order to obtain partitioning data directly applicable to planetary differentiation processes. Run products were analyzed by electron microprobe. Time-series experiments at 8 GPa reveal that equilibrium conditions are reached rapidly, within 10 s. The effect of temperature was investigated at 8 GPa on the 2000-2200 C temperature range. Results shows that over this temperature range, partition coefficients for K (DK) remain almost identical. The influence of pressure was investigated at 2000 C (3-8 GPa range). Results reveal that DK slightly increase from 3 to 8 GPa, with values of 0.074 and 0.082 respectively. However, this trend could be due to the increase of the S content of the metallic phase with pressure. Further experiments are currently underway to separate the effect of pressure from the effect of composition, and to extend the pressure range of the study.
MR43A-0868 1340h
Diffusion in Zinc at High Pressure and Rheology of the Earth's Inner Core
An attempt has been made here to estimate the viscosity of the Earth's inner core, and also to identify the primary mechanism by which the inner core deforms. Estimation of the viscosity and identification of the deformation mechanism(s) of the inner core require measurements of diffusion in the hcp (hexagonal close-packed, or epsilon) phase of iron, suggested to be stable at the Earth's inner-core pressure conditions. However, owing largely to experimental and analytical challenges, actual measurements on the diffusivity in this phase are non-existent. To overcome this problem, we have focused on the divalent transition metal, zinc (Zn), which has the hcp structure over a wide range of pressures. Hcp metals are known to have similar diffusivities at the same homologous temperature. A primary goal of this work was to explore the effect of increasing pressure on diffusion in zinc. Zinc has high compressibility, allowing diffusion measurements to be made on normalized pressures (P/K, pressure/bulk modulus) approaching those of Earth's core. We focused on diffusion of gold (Au), which has been extensively studied at atmospheric pressure. We find that with increasing pressure from 10 to 25 GPa, the diffusion coefficient of Au in Zn decreases, and the data at high pressures are in good agreement with that at 1 atm. However, in a plot of log D versus homologous temperature (Tm/T, where Tm is the melting point), the slope besides being slightly shallower, the high-pressure diffusivity values retrieved are higher than predicted from extrapolation of the 1 atm data. This trend is more prominent in plot of log D versus pressure, where instead of being linear, the diffusion coefficient shows a slight parabolic dependence, indicating that the activation volume decreases with pressure. High-pressure diffusivity values are higher than predicted either from homologous temperature scaling, or those retrieved assuming constant activation volume from one atmosphere data. To a good approximation, the inner core is in hydrostatic equilibrium with the surrounding fluid. The shear stress on the inner core is thought to be low, and grain sizes are larger than in the lower mantle. Temperature and pressure both influence the effective viscosity, and their role is often included in rheological models by referring the temperature to the melting temperature, Tm; the pressure dependence enters implicitly through its influence on Tm. Calculated values of viscosity of the Earth's inner core using the experimental philosophy outlined above are at least 8-9 orders of magnitude lower than estimated so far. These low viscosities have important consequences for the origin of seismic anisotropy and diffusion in the inner core. On the basis of these measurements, it appears that the inner core can quickly erase memory of deformation that occurred early in its history. Thus, search for alternative explanations of the anisotropy is required. Unraveling the clues left behind by the growth of the inner core offers the hope of new insights into the evolution of the Earth's deep interior.
MR43A-0869 1340h
The Fe-Ni-(S) System at 23 GPa: The Possibility of Strong Chemical Fractionation Between Phases in the Cores of the Earth, Mars and Mercury
The presence of nickel in the Earths core is widely accepted based on cosmochemical and seismological arguments. However, experimental studies into core compositions rarely include nickel, thus adding a degree of simplicity to otherwise complex experiments. Diamond-anvil cell studies have discovered that Fe-Ni alloys appear to separate into two phases upon heating above 10 GPa: from a single hexagonally close-packed (hcp) phase to the presence of both hcp and face centered cubic (fcc) phases (Lin {\it et al.}, 2002). Unfortunately, due to the small size of diamond-anvil cell samples, meaningful quantitative analysis is commonly impossible. We have conducted multi-anvil experiments at 23 GPa into the Fe-Ni system and have confirmed the presence of two phases in the sub-solidus system. The starting material for these experiments contains 6 wt% nickel, approximating the amount expected to be found in the Earths core (McDonough, 2003). In experiments to $1500\deg$C (the highest temperature thus far examined), electron microprobe analyses show dramatic phase fractionation with charges separating into an iron-rich phase containing less than 1 wt% Ni and a nickel-rich phase containing as much as 98 wt% Ni. We have observed the effect over a range of more than $500\deg$°C; further experiments are underway to determine whether these phases both persist toward the melting point of the alloy. Multi-anvil experiments at 23 GPa have also been conducted to examine the effect of nickel on the Fe-S system. Sulphur is an element favoured by many researchers as the light element component in the core of the Earth as well as that of Mars. Previous research has suggested that the addition of nickel to the Fe-S system results in the lowering of eutectic temperatures by about $75\deg$C (Pike {\it et al.}, 1999). The starting material for these experiments is the same as that used for the pure Fe-Ni experiments discussed above, with the addition of sulphur. Our results indicate a pseudo-binary, (Fe, Ni)-S, eutectic point lying slightly below $1200\deg$C, roughly consistent with the results of Pike {\it et al.} (1999). The measured eutectic liquid composition contains 4.4 wt% Ni and 15.8 wt% S. This liquid composition fits closely to the ideal composition of a (Fe, Ni)$_{3}$S compound (16.0 wt% S with 4.4 wt% Ni in the alloy), suggesting the possible importance of this structure in Fe-Ni-S melts. At subsolidus temperatures in the Fe-Ni-S system, our results become very interesting with each charge showing at least 3 coexisting phases. Based on these results, solid cores of Mercury and Mars containing iron, nickel and sulphur will hold at least 3 phases. Extrapolating our results to the inner core of the Earth would suggest that multiple phases occur in our planet as well.
MR43A-0870 1340h
Equation of state of iron sulfide at the conditions of Galilean satellite cores
The Galileo mission has revealed that Jupiter's satellites Io, Europa, and Ganymede are differentiated and likely have dense, metallic cores. Current structural models of the Galilean moons have assumed that core compositions will be in the Fe-FeS system because of the cosmochemical abundance of sulfur and its low eutectic temperatures. However, these models have generally assumed densities for FeS that are not based on measurements directly under the pressure-temperature range expected for the cores of the Galilean moons. The probable temperatures and pressures of the cores of the Galilean moons make FeS in the NiAs structure (FeS V) the most likely phase at core conditions. We have obtained energy-dispersive x-ray diffraction data of FeS from 0-15 GPa and 300-1100 K using the 250-ton multi-anvil press at the GSECARS sector of the Advanced Photon Source. From this dataset, a thermal equation of state of FeS V using the third-order Birch-Murnaghan equation will be constructed. Preliminary low-pressure density results for FeS V are in the range of 4.7 to 4.9 g/cm$^{3}$ from 1 to 2 GPa and from 600 to 900 K. Using a linear mixing model for Fe-FeS, these results indicate that the eutectic density at core conditions of the Galilean moons is approximately 13% larger than values typically used in existing models. We will construct a range of new interior models that satisfy the new equation of state for FeS as well as additional considerations from thermodynamics, rheology, compositional inferences, and accretion models.
MR43A-0871 1340h
Torsional Alfven Waves Excited Inside the Earth's Core Because of Modulations by the Solar Cycle of the Electrical Currents Flowing in the Magnetosphere
Torsional Alfven waves consist of rigid rotations of geostrophic cylinders coupled together by the magnetic field permeating the Earth's core. Fluctuations in the rate of rotation of the solid mantle arise from the coupling between the torsional waves and the rotational motion of the mantle. Depending on the electrical conductivity of the Earth's mantle, more or less intense electrical currents are induced in a thin layer at the top of the Earth's core by variations of the magnetic field of external origin. We find that oscillations of the geostrophic cylinders near the equator of the Earth's core are readily excited by these time varying electrical currents. That sets off torsional waves, which propagate inward and dissipate through magnetic friction with the solid inner core. These waves take part to the fluid motions at the core surface and interact with the main magnetic field. As a result, there is an interior magnetic field associated to any time varying external magnetic field with a long enough period. We find that the ratio between the magnitudes of respectively these internal and external magnetic fields increases rapidly with the period. The exact values depend on the electrical conductivities of both the core and the mantle and on the intensity of the main magnetic field inside the core. With current estimates of these parameters, the ratio between the strength of the magnetic field generated by the torsional oscillations and of the strength of the inducing external magnetic field can be much larger than 1 at the period of the solar cycle. We suggest that part of the secular variation may have an external origin. Our mechanism may successfully explain the rapid succession, every ten years or so, of magnetic "jerks" from 1969 until now.
MR43A-0872 1340h
Trace Element Partitioning Between Earths Lower Mantle Minerals and Iron Alloy Melts
The physical conditions (pressure-temperature-oxygen fugacity) and iron alloy compositions involved in the formation of Earths core are relatively poorly constrained. Current geochemical core formation models rely heavily on arguments related to how trace elements are distributed between metal and silicate during metal segregation. Hypotheses concerning the timing of Earths accretion and core segregation also require knowledge of metal-silicate partition coefficients for parent and daughter isotopes in key isotopic systems (e.g., W-Hf, U-Pb). Many studies (e.g., Righter, Ann Rev Earth Planet Sci 2003) have focused on metal melt - silicate melt partitioning at upper mantle and transition zone pressures (P $<$ 25 GPa) and high temperatures. In contrast, little is known about the partitioning of trace elements between metals and {\it lower} mantle minerals and melts, even though the later stages of core formation, characterised by high-energy processes related to collisions of Mars-sized objects, likely involved liquid iron alloys percolating through a solid or molten lower mantle matrix. We will present results of a systematic study of the distribution of trace elements between lower mantle minerals (Ca and Mg perovskite, and ferropericlase) and a range of iron alloy melts, to assess the redistribution of trace elements as metallic melts percolate through the lower mantle. Experiments are performed using a 10/3.5 assembly in a conventional Walker-type multi-anvil press (P $<$ 26 GPa), and using a 7/2 assembly in a spherically-constrained multi-anvil press with sintered diamond pressure-transmitting cubes (P $>$ 30 GPa). Starting materials include end-member silicates (wollastonite, enstatite), MgO, and Fe-light element (S, Si) mixtures. Trace elements added include slightly siderophile (Mn, V, Cr), moderately siderophile (P, W, Co, Ni, Mo) and highly siderophile elements, as well as key parent-daughter pairs, and run products are analysed by laser ablation ICP-MS. Implications of our data for core formation models will be discussed.
MR43A-0873 1340h
Trace Element Partitioning Between Metal and Melt at High Pressure
Fractionations between siderophile trace elements are produced during crystallization of solid metal from a molten planetary core. It has been proposed (e.g., Brandon et al., 2003) that fractionations of Re/Os and Pt/Os produced during crystallization of Earth's inner core are recorded in rocks whose source regions lie at the core-mantle boundary. However, the possible effects of pressure, temperature, composition, and metal crystal structure on metal-melt partitioning have not been fully evaluated experimentally. Accordingly, we have begun to measure the partitioning of major and trace siderophile elements between Fe-rich metal and metal-sulfide melt at high pressures and temperatures using laser ablation ICP-MS of multi-anvil press samples. The starting materials included iron meteorite powder (having natural abundances of PGEs at the $/sim$10 ppm level) and troilite. In some cases Ru powder was also added to the starting material to promote transformation of the Fe-rich metal to the hcp structure, as described by Campbell et al. (2003). The powders were loaded into an MgO or BN sample capsule in a 10/5 multi-anvil press assembly, and pressurized to 14.5 GPa. The sample chamber thickness was held to $<$0.4 mm to minimize the temperature gradient experienced by the samples. Melting was achieved at temperatures of 975 C or above, and run durations were 6 to 24 hours. The recovered run products were polished and examined by SEM or electron microprobe before LA-ICP-MS analysis. Laser ablation spot sizes ranged from 15 to 50 microns, depending on the available grain size. Partition coefficients (D) were determined for Co, Ni, Ru, Re, Os, Ir, and Pt. Errors on the D values were based on the reproduceability of at least 3 measurements in each phase, and were $<$10% for major elements and $<$25% for trace elements. The effects of temperature and composition were evaluated and implications on the chemistry of the core will be discussed. Brandon A. D. et al. (2003) EPSL 206:411-426. Campbell A. J. et al. (2003) Fall AGU Meeting.
MR43A-0874 1340h
Crystal Structure and Magnetic Properties of Fe$_{3}$C at High Pressures and High Temperatures
Fe$_{3}$C has been suggested as a major phase of the Earth's inner core. A number of experimental studies of equation of state at room temperature led to a density of this material at inner core pressures in excellent agreement with the PREM. However, first-principles computer simulations and measurements of thermal expansion of Fe$_{3}$C below and above ferromagnetic phase transition by neutron powder diffraction at ambient pressure have suggested that the physical properties of Fe$_{3}$C are incomparable with the probable range of the inner core density determined from seismic data. Disparate conclusions might be a result of fact that these studies were conducted at either at room temperature or at ambient pressure. In this work, the crystal structure and physical properties of the cementite phase of Fe$_{3}$C have been studied in situ at high pressures and temperatures in diamond anvil cell (DAC) with x-ray diffraction techniques and nuclear resonant spectroscopy. The Fe$_{3}$C samples were synthesised in two ways: ex-situ in a large volume press at 4 GPa and 1600K with further grinding before loading in DAC, and in-situ in a diamond anvil cell at high temperature and different pressures from a mixture of Fe and C powders. The $^{57}$Fe enriched sample was used for nuclear resonant spectroscopy. On compression at room temperature up to $\sim$60 GPa, a second order ferromagnetic transformation was observed in the pressure range of 20-30 GPa. On decompression, the reverse transformation to the magnetic state appeared at much lower pressures: 10-6 GPa. The crystal structure of Fe$_{3}$C remained orthorhombic (space group {\it Pnma}, Z=4) in the entire pressure range studied. Heating samples at $\sim$1800 K significantly affected the magnetic behavior of Fe$_{3}$C at high pressures. For example, the ferromagnetic transition pressure was reduced by $\sim$10 GPa. In situ x-diffraction at high pressures and high temperatures allowed us to structurally map the phase diagram of Fe$_{3}$C and its melting curve. The implications of these results to the composition of the Earth's inner core will be discussed.
MR43A-0875 1340h
Ab Initio Molcular Dynamics Study of the High-Temperature Elastic Constants of hcp-Iron at Inner-Core pressures.
It is well established that the inner core exhibits significant anisotropy with compressional P-wave velocities which are \(\sim 3\)~% faster along the polar axis than in the equatorial plane. Interpretation of this seismic anisotropy is hampered by lack of clear data on the physical properties of core phases at simultaneously high pressures and high temperatures. Although it is well known that there are light elements alloyed to iron in the inner core, even definitive results for the properties of pure iron at core conditions are proving elusive. Experimental groups have put an enormous effort over the last 10--15 years into obtaining the properties of pure iron at elevated pressures and temperatures, but above relatively modest \(P\) & \(T\) there is much uncertainty in the results. Theoretical calculations at inner core conditions on the hexagonal-close-packed phase of iron are achievable, but even here there are major uncertainties on the physical properties and the nature of the elastic anisotropy. In particular, results from previous theoretical calculations (Steinle-Neumann et al., 2001), based on the approximate ``particle in a cell'' method, suggest that the elastic properties of hcp-Fe change dramatically as a function of temperature. However previously we have shown these ``particle in a cell'' method results do not accurately describe the high \(P\), \(T\) behaviour of Fe (Gannarelli et al, 2003). Here we present the first \textit{ab initio} molecular dynamic simulation of the finite temperature elasticity of hcp-iron at core conditions. We conclude that again the previously reported (Steinle-Neumann et al., 2001) simulations of the high T behaviour of hcp Fe are in error, and we find that the sense of the elastic anisotropy at high \(T\) is essentially similar to its athermal behaviour. This supports the view originally expressed by Stixrude and Cohen (1995) that if the inner core anisotropy is associated with hcp Fe, then it is consistent with preferential alignment of Fe crystals with their \(c\)-axes closely aligned parallel to the rotation axis of the Earth. Steinle-Neumann, G., Stixrude, L., Cohen, R.E. and G\"{u}lseren O. (2001). ``Elasticity of iron at the temperature of the Earth's inner core'', Nature, \textbf{413}, 57. Gannarelli, C.M.S., Alf\`{e}, D. and Gillan, M.J. (2003). ``The particle-in-cell model for \textit{ab initio} thermodynamics: implications for the elastic anistropy of the Earth's inner core'', Phys. Earth Planet. Inter., \textbf{139}, 243. Stixrude, L. and Cohen, R.E. (1995). ``High-pressure elasticity of iron and anisotropy of Earth's inner core'', Science, \textbf{267}, 1972.
http://www.cmmp.ucl.ac.uk/~mjg
MR43A-0876 1340h
A Dynamo Mechanism for Rapid Decrease of the Geomagnetic Dipole Moment
The dipole moment of the geomagnetic field is decreasing at 6% per century, nearly one order of magnitude faster than its theoretical ohmic decay rate in the core. This change is significant because a sustained decline in dipole moment may indicate instability of the geodynamo. For example, according to the paleomagnetic record, polarity reversals typically begin with a large dipole moment decrease. Maps of geomagnetic secular variation on the core-mantle boundary show that much of the recent decrease in the dipole moment is associated with two phenomena: (i) emergence, growth, and poleward drift of reversed magnetic flux, and (ii) weakening of some normal polarity, high-intensity magnetic flux patches at located high latitudes. We have used numerical dynamo models to identify mechanisms for rapid decrease in dipole moment. Dynamo models indicate that rapid fluctuations in the dipole moment occur during bifurcations in the pattern of convection in the core. Bifurcations in the convection pattern alter the number, distribution, and intensity of the normal polarity flux patches, and trigger outbursts of reversed flux on the core-mantle boundary, all of which affect the dipole moment. We have analyzed a chaotic dynamo model that exhibits sudden decrease in dipole moment, similar to the recent geomagnetic dipole decrease, during bifurcation from an m=4 to an m=3 convection pattern. In this model, most of the reversed flux on the core-mantle boundary consists of poloidal magnetic field expelled from the core by convective plumes. During the bifurcation, reversed magnetic flux is expelled at low latitudes by transient plumes concentrated near the equatorial plane, transported poleward by meridional flow, and mixed with normal polarity field at higher latitudes, reducing the dipole moment.
MR43A-0877 1340h
Time-Average Core Flow: Mantle vs. Core Origins
We derive a time-average surface core flow by inverting geomagnetic secular variation data over a century. We use a grid-based numerical method with a helical-geostrophic assumption for the tangential velocity divergence to invert the frozen flux induction equation for the fluid flow below the core-mantle boundary at specific epochs. Those flows are then combined to form time-average and time-dependent parts. We compare our time-average core flow with two models of thermal wind flow. The first model is thermal wind driven by density gradients associated with convection in the core and a homogeneous core-mantle boundary. This model is obtained from numerical dynamo simulations. The second model is thermal wind driven by density gradients associated with core-mantle thermal coupling, without core convection. This model is obtained using lower mantle seismic tomography for the density gradients. Our results suggest that, over century time scales, only the zonal part of the flow at the top of the core may be considered steady. We find equatorially asymmetric zonal core flow at low and mid latitudes, consistent with an origin by thermal coupling to lower mantle heterogeneity. In contrast, at high latitudes, the time average zonal core flow consists of westward polar vortices, which are more consistent with an origin by core convection.
MR43A-0878 1340h
Evidence for a Generalized Core Resonance Phenomena in Tidal Gravimetry
It is well established that a resonance effect involving the Earth's fluid outer core affects gravimetric responses of tidal modes in the diurnal band. In this Earth-fixed case, the gravimeter is indirectly sensing the presence of the Nearly Diurnal Free Wobble (NDFW). Also well established with observations from Very Long Baseline Interferometry (VLBI) is the Free Core Nutation (FCN) - the corresponding interaction viewed from a space-fixed frame. Because contained, rotating fluids like the Earth's outer core support an infinite set of oscillatory modes, resonance effects in addition to the NDFW/FCN are possible in the diurnal band. In principle, these same oscillatory modes exist at semi-diurnal periods, where additional possibilities for resonance effects exist. Although this generalized resonance phenomena was suggested over a decade ago (e.g., Lumb et al., AGU Monograph 72, 51-68, 1993), observational evidence remains absent. The situation is further complicated by the presence of the Earth's solid inner core which is also expected to participate in the resonance process. With data from the globally distributed network of superconducting gravimeters involved in The Global Geodynamics Project (GGP), observational prospects appear enhanced. In fact Merriam (Geophys. J. Int., 123, 529-540, 1995) reports four residuals of 4-10 ngals in the semi-diurnal band after the removal of all known effects - including loading effects of shallow-water non-linear tides. Given that the spectrum of oscillatory modes is dense throughout this tidal band, these residuals may already be an indication of resonance effects at semi-diurnal periods. Merriam also reports validation of his single-station results from Cantley, Quebec (Canada) in single-station records from both Europe and Japan, thus emphasizing the global presence of these residuals. After briefly reviewing the generalized core resonance phenomena, and the theoretical basis for its detection in gravimetric data, attention is focused on a modified approach for data analysis. In particular, this approach exploits the fact that the oscillatory modes are traveling disturbances in the Earth-fixed frame. Because use of novel approaches like this one could be greatly accelerated, there is ample motivation for opportunities to Grid-enable the GGP (Lumb & Aldridge, Eos. Trans. AGU, 85(17), Joint Assembly Suppl., Abstract G34A-03, 2004).
MR43A-0879 1340h
Experimental Partitioning of U Between Liquid Iron Sulfide and Liquid Silicate: Implications for Radioactivity in the Core
The prodigious affinity of U for oxygen suggests that it may be difficult for U to be incorporated in a core formed from silicate mantle with residual partially oxidized Fe. If U is reduced enough to join the metal of the core, then Fe should be effectively completely reduced. We explored the ability for initially native U metal to be retained within polymetallic sulfide liquid solutions \{Fe+FeS\} in equilibrium with plausible mantle silicate liquids \{peridotite KLB-1\} and excess carbon. We found no conditions from 2-10 GPa, 1750-$2100\deg$C, and 0-28 wt% S in the metallic liquid where sufficient U could be retained in the metallic liquid to be interesting in the context of U-derived heat sources in Earth's core. We do find interesting structure to the variations of DU\{sulfide/silicate\}, which was always a very small number. Typical DU ranged from a minimum of $<$1.3\times10$^{-5}$ to a maximum of 0.001. A possible weak increase in DU with temperature was observed. Increasing pressure also may cause a weak increase in DU. Sulfur content in the Fe sulfide was the largest influence on DU indicating some degree of U chalchophility. Typical DU values increased an order of magnitude when S content in the sulfide increased from 7 to 28 wt%. It is important to note that even wild extrapolation of the most favorable data did not yield significant U in the sulfide at core conditions. Therefore, it is our conclusion that if there is U in the core, it most likely did not get there during core formation in a partially oxidized magma ocean scenario.
MR43A-0880 1340h
A suction mechanism for iron entrainment from the outer core into the lower mantle.
Variations in the earth's rotation rate (nutation data) strongly favors the existence of a thin high electrical conductivity layer, having a conductance of ~ 10$^8$ S, near the Core Mantle Boundary (CMB)(Buffett, JGR 1992). However, it is not clear if this layer is in the outer core-side or mantle-side of the CMB. Seismic data does not provide a clear-cut answer, but indicates that any core-side conducting layer must be very thin (Rost & Revenaugh, Science 2001). An earlier model of sediments containing lighter core elements at the top of the inner core can explain the nutation data only if very low permeabilities (and hence sub micron grain sizes) are assumed (Buffett et al., Science 2000). Here, we have investigated a simple 1-D, suction driven mechanism that can entrain significant quantities of iron into the lower mantle on time scales of ~ 10s of Ma - small compared to that of mantle convection. Our mechanical model is driven by the pressure difference between the liquid iron and the silicate mantle near the CMB. This "suction" drives compaction in a zone of several hundred meters of the lowermost mantle. Analytical solution to the linear problem with constant liquid (Fe) fraction and effective compaction viscosity indicates that sufficient iron can be entrained to explain the conductance required by nutation data. We then explored the 1-D nonlinear case, where the effective compaction viscosity, as well as the mantle viscosity near the CMB (hence the suction boundary condition at CMB) are strongly liquid fraction dependent - both decrease as the fraction increases. Compared to the linear problem, entrainment is harder in the non-linear problem. As the liquid fraction increases, suction at the CMB starts decreasing as the effective mantle viscosity drops. In addition, the effective compaction viscosity increases, reducing the liquid fraction entrainment rate. We solve the resulting coupled set of compaction and fluid continuity equations using an iterative finite difference scheme. We find that when the effective compaction viscosity is much smaller than mantle viscosity (by at least an order of magnitude), the maximum of the Fe liquid distribution is pushed deeper into the mantle. Sufficient iron gets entrained within a few hundred meters thick zone to yield the required conductance. Our results, although preliminary at this stage, would support the plausibility of a conducting layer on the mantle-side of CMB, thus obviating the need for complex assumptions regarding the chemistry and sedimentation dynamics in the outer core. It must be noted that the presence of partial melt as inferred from ULVZs, as well as recent research indicating the presence of Fe-rich silicate near the bottom of the mantle, will only enhance the conductance of our layer. We next propose to explore the effect of including the full two-phase physics of the problem (silicate matrix and Fe liquid) on our results.
MR43A-0881 1340h
Rotating Parametric Instability in Earth's Core and the Geodynamo
Records of paleomagnetic intensity from ocean sediment cores reveal cycles of growth and decay. This transient nature of the geomagnetic field is consistent with sequential excitation and collapse of rotational parametric instability (RPI) in the fluid core. Linear stability theory predicts RPI will develop in rotating contained fluids provided dissipation does not outpace the pure or ideal growth. Small amplitude, periodic strains from tides and precession provide elliptical and shear deformation of streamlines that couple with the core's stable rotational modes to drive the development of RPI in the fluid. The continuous action of these small strains on the fluid result in the build up of RPI through the linear regime into finite amplitude and inevitable collapse. Growth and decay of intensity are a measure of the dynamo action in the core which would be driven by turbulent flow due to RPIs. Alternating growths and decays have been recovered from records including single site records from the West Caroline Basin, and the North Atlantic, as well as the NAPIS-75 and GLOPIS-75 composite records. Neighbouring growth-decay and decay-growth pairs are combined to extract an estimate of the ideal growth of RPIs in Earth's fluid core. This estimate is consistent in the records and compares favorably with the ideal growth of intensity expected from the linear theory of RPI in the fluid core. Decay portions of the records show that elliptical and shear RPIs are a geophysically plausible mechanism to control the geodynamo.
MR43A-0882 1340h
Reaction Between Liquid Iron and Mg-perovskite and Solubility of Silicon and Oxygen in Liquid Iron
A reaction between liquid iron and Mg-perovskite was investigated at 28 GPa and 2550-2940 K to discuss the core formation process and light elements in earth's core. The earth's core contains small amount of light elements. These light elements were dissolved into liquid iron to form the core during core formation process. Deep magma ocean is supposed to have extended to a depth of lower mantle in the core formation stage (Ohtani et al., 1997; Li and Agee, 2001). Because Mg-perovskite is the most dominant mineral in the lower mantle, reaction between the liquid iron and the Mg-perovskite must have occurred at base of the deep magma ocean and could have provided Si and O as the light elements into the liquid iron. Knittle and Jeanloz (1991) and Hillgren and Boehler (1999) have studied this reaction with a laser-heating diamond anvil cell. But their studies had a large temperature gradient in the sample room resulting in a possibility of disequilibrium. In this study, high pressure and temperature experiments were conducted with a Kawai-type multi-anvil apparatus and samples were reached to equilibrium with uniform heating by Re cylindrical heater. Pure iron rod was packed into MgSiO$_{3}$ or (Mg$_{0.9}$, Fe$_{0.1}$) SiO$_{3}$ powder capsule, which transformed to Mg-perovskite in experimental conditions. The sample was compressed to a desired load and then heated to a desired temperature and finally quenched. Run products were analyzed with an electron microprobe. Magnesiow\"{u}site was formed at the boundary between liquid iron and Mg-perovskite. Quenched liquid iron contained oxide blobs in all runs and stishovite grew in the quenched liquid in the run made at 2940 K and 1.72 log units below IW buffer. The liquid iron reacted with Mg-perovskie to form the magnesiow\"{u}stite and Si and O dissolved into the liquid iron at temperatures above 2550 K. Si and O solubility in the liquid iron decreased and increased with increasing oxygen fugacity, respectively, and both increased with increasing temperature. Si and O contents in liquid iron were 1.70 wt% and 2.33 wt%, respectively, in the run made at 2940 K and 1.72 log units below IW buffer. They were 0.18 wt% and 7.52 wt%, respectively, in the run made at 2830 K and 0.20 log units above IW buffer.
MR43A-0883 1340h
Fe/Mn Ratios in Ocean Island Basalts: a Tracer for Outer Core-Lower Mantle Interaction?
Fe/Mn ratio can be used as a tracer to test the Outer Core-Lower Mantle interaction hypothesis about the source characteristics of some ocean island basalts (e. g. Hawaiian plume). Literature data for Fe/Mn can not resolve a difference in Fe/Mn of less than 15%. We developed an ICP-MS method to precisely measure Fe/Mn to better than 0.5%. The intensities of $^{57}$Fe and $^{55}$Mn peaks were monitored by ICP-MS, and corrected for background and interferences ($<$2%). Fe/Mn ratios were obtained by standardization against gravimetrically prepared standards. For the standards, a 0.5% difference in Fe/Mn ratio is resolvable at 2$\sigma$ level. Analysis of individual olivine and orthopyroxene grains picked from a single Kilbourne Hole peridotite xenolith showed that all five olivine and five orthopyroxene grains had uniform Fe/Mn ratio of 69.83$\pm$0.40 (2$\sigma$) and 44.27$\pm$0.16, respectively. We have recently reported precise Fe/Mn ratios for Hawaiian lavas, which showed they had resolvably higher Fe/Mn ratios (66-71) than Pacific MORBs (55-58) and Icelandic picrites. Here we report new results for OIBs, MORBs and komatiites. To avoid ubiquitous secondary Fe-Mn oxides, Fe/Mn ratios in Pacific, Atlantic and Indian MORBs were determined by Laser Ablation ICP-MS and the result showed ratios of 53 to 56, consistent with SN-ICP-MS results for fresh Pacific MORB. Icelandic picrites (MgO: 10-29%) had Fe/Mn ratios from 54 to 64. Six relatively fresh komatiites from Belingwe (MgO: 20-29%) yielded a Fe/Mn of 58.33$\pm$0.44. Seven Tahiti basalts (MgO: 8-21%) had Fe/Mn mostly between 64-70, overlapping the range for Hawaiian picrites. At the same MgO content, Tahitian and Hawaiian lavas exhibited higher Fe/Mn than MORBs, Icelandic and Belingwe lavas. This implies that the source regions of Tahiti lavas were enriched in Fe relative to other mantle sources (e. g. MORBs, Iceland, Belingwe). Together with previous results for Hawaii, this implies that high Fe/Mn is characteristic of at least 2 plumes from the Pacific Superswell. It is conceivable that this is evidence for excess Fe due to Core-Mantle interaction at the base of the Pacific Superswell.
MR43A-0884 1340h
Waveform Search for the Innermost Inner Core
Waveforms of the PKIKP seismic phase in the distance range 150$^{o}$ to 180$^{o}$ are analyzed for evidence of an inner-most inner core of the type proposed by Ishii and Dziewonski having an abrupt change in elastic anisotropy near radius 300 km. Seismograms synthesized in models having a discontinuity at 300 km radius in the inner core exhibit focused diffractions around the innermost sphere at antipodal range that are inconsistent with observed PKIKP waveforms. Successful models have either a transition in elastic properties spread over a depth interval greater than 100 km or an innermost sphere that exceeds 450 km radius. Evidence of a sharp discontinuity in the lower to mid-inner core is sparse in existing global seismic data. Some examples, however, can be found of PKIKP complexity near 161$^{o}$ and 164-165$^{o}$, consistent with a triplication created by a 475 km radius discontinuity. An abrupt change in either viscoelastic or scattering attenuation at this radius is also observed in PKIKP waveforms, suggesting the existence of an inner-most sphere with low, regionally uniform, seismic attenuation. In contrast to the relatively uniform inner-most inner core, a 0 to 100 km thick region at the top of the inner core exhibits strong lateral variation in seismic attenuation, suggesting lateral variations in the processes of solidification, flow and recrystallzation at the inner core/outer core boundary. Analogous to the evidence for an abrupt fabric change in the upper-most inner core, the seismic evidence for an inner-most inner core may represent another fabric change. This change may simply signify the end stage of solidification, flow and recrystallization, resulting in the highest ordering and largest grain sizes of intrinsically anisotropic crystals.
http://www.phys.uconn.edu/~cormier/innermost.html
MR43A-0885 1340h
Interconnectivity of liquid Fe-alloy in planetary mantles
The most important and fundamental aspect of planetary core formation is whether liquid Fe-alloy can segregate through crystalline silicates. The interconnectivity between silicate minerals (olivine, ringwoodite, garnet, majorite, and Mg-Perovskite) and liquid Fe-S-O has been investigated experimentally. Experiments were carried out using 1200 and 5000 tonne multi-anvil presses. In all experiments, starting powders were mixtures of Fe$_{61}$S$_{39}$ and synthetic silicates that were sealed in graphite capsules. Experimental conditions were 1.5-24.5 GPa, 1650-2200 K, and 5-12 hours duration. Observations of the sample textures were performed using SEM and TEM. The interconnectivity changes significantly with the solubility of light elements, especially oxygen in the Fe-alloy. Dihedral angles are observed to increase with increasing pressure between 1.5-5.0 GPa. This is most likely a result of the drastic decrease in Fe-alloy oxygen solubility with pressure. As a result, interconnected networks can form only up to approximately 3.5 GPa. This indicates that Fe-alloy can percolate through the mantles of small sized bodies such as planetesimals and asteroids at oxidized conditions. Dihedral angles in olivine and ringwoodite assemblages are $103-69\deg$ which is slightly lower (approximately $10\deg$) than that in the garnet and majorite assemblages using the same Fe-alloy composition. For (MgFe)SiO$_{3}$ perovskite, however, variations in composition and dihedral angle are rather complicated. The Fe content in Fe-alloy apparently increases with the Fe content of perovskite most likely as a result of the disproportionation of Fe$^{2+}$ to Fe + Fe$^{3+}$. Fe$^{3+}$/$\Sigma$Fe ratio of up to 0.25 in perovskite coexisting with Fe-S alloy were measured by EELS. The Fe$^{3+}$ content of Pv increases with Fe content of Pv. The additional metal formed dissolves in the Fe-S alloy and increases with the Fe/S ratio. This increase is coupled with a decrease in the dihedral angle. These results show that interconnected networks are limited to relatively small bodies ($\sim$ 1000 km radius) and are not likely to form in the bulk of planetary mantles.