Search

• Liquidus phase relations in a hydrous silicate system were determined at 24 GPa. • Melts are neutrally buoyant at 24 GPa before ∼98% crystallization of a magma ocean. • A transient melt-rich layer may form atop the lower mantle in a mushy magma ocean. The solidification of a deep magma ocean occurred early in Earth's history. Although the initial amount of H 2 O in Earth's magma ocean is predicted to be low (e.g., <3000 ppm), as an incompatible element it becomes highly enriched (e.g. >10 wt%) in the final few percent of crystallization. In order to understand how a hydrous magma ocean would crystallize at the top of the lower mantle, we determined liquidus phase relations in the MgO-FeO-CaO-Al 2 O 3 -SiO 2 -H 2 O system at 24 GPa. We find that the bridgmanite (brg) + stishovite (st) + melt and bridgmanite (brg) + ferropericlase (fp) + melt cotectic boundary curves trend to Mg-rich melt compositions with decreasing temperature and extend to very high H 2 O contents (∼80 mol% H 2 O). The brg+st+melt curve is a subtraction curve at < ∼18 mol% H 2 O and a reaction curve at higher H 2 O contents, whereas the brg+fp+melt is a subtraction curve throughout its length. The density of melts along the two cotectics leads to neutral buoyancy with respect to shallow lower mantle and transition zone minerals at H 2 O contents up to ∼25 mol%. A transient melt-rich layer can form at the top of the lower mantle during late-stage crystallization in a mushy magma ocean when melt percolation dominates. When crystallization exceeds ∼98%, hydrous melts (>25 mol% H 2 O) become buoyant and can percolate into and hydrate the mantle transition zone. [ABSTRACT FROM AUTHOR]

• Acoustic and electrical measurements were performed on ilmenite-rutile aggregates. • Experiments were conducted at the pressure and temperature conditions of the lunar mantle. • Fe-Ti oxides increase the electrical conductivity and the seismic velocities of the lunar mantle. • Field datasets are reproduced with overturned ilmenite-bearing cumulates containing 20% melt. The overturn of titanium-rich mantle cumulates has been invoked to explain the structure and dynamics of the Moon. These dense cumulates are stable at the core-mantle boundary (CMB) and could explain field observations inferred from geophysical studies. We report acoustic and electrical experiments on natural ilmenite-rutile aggregates up to 4.5 GPa and 1920 K. Seismic velocities show a weak pressure and temperature dependence, with Vs ∼4.2 (+/-0.2) km/s and Vp ∼ 8.0 (+/-0.2) km/s at the CMB conditions. Conductivity increases by a factor of 104 from 373 to 1920 K and is >103 S/m above 1573 K. Seismic and electrical models for the lunar mantle based on our results, considering mixtures of Fe-Ti oxides and olivine, indicate that field velocity and conductivity estimates are reproduced satisfactorily with 3–16 vol.% Fe-Ti oxides and 20 vol.% melt. Interactions between a Ti-rich, melt-bearing layer and the adjacent core likely affect the cooling and magnetic history of the Moon. [ABSTRACT FROM AUTHOR]

Velocity and density jumps across the 410-km seismic discontinuity generally indicate olivine contents of ∼30 to 50 vol.% on the basis of the elastic properties of anhydrous olivine and wadsleyite, which is considerably less than the ∼60% olivine in the widely accepted pyrolite model for the upper mantle. A possible explanation for this discrepancy is that water dissolved in olivine and wadsleyite affects their elastic properties in ways that can reconcile the pyrolitic model with seismic observations. In order to more fully constrain the olivine content of the upper mantle near the 410-km discontinuity, and to place constraints on the mantle water content at this depth, we determined the full elasticity of hydrous wadsleyite at the P-T conditions of the discontinuity based on density functional theory calculations. Together with previous determinations for the effect of water on olivine elasticity, we simultaneously modeled the density and seismic velocity jumps (Δ ρ , Δ V P , Δ V S) across the olivine-wadsleyite transition. Our models allow for several scenarios that can well reproduce the density and seismic velocity jumps across the 410-km discontinuity when compared to globally averaged seismic models. When the water content of olivine and wadsleyite is assumed to be equal as in a simple binary system, our modeling indicates a best fit for low water contents (<0.1 wt.%) with an olivine proportion of ∼50%, suggesting a relatively dry, non-pyrolitic mantle at depths of the 410-km discontinuity. However, our modeling can be reconciled with a pyrolitic mantle if the water content in wadsleyite is ∼0.9 wt.% and that in olivine is at its storage capacity of ∼500-1500 ppm. The result would be consistent with a hydrous melt phase produced at depths just above the phase transition. • Elasticity of hydrous wadsleyite at high P-T were obtained from ab initio method. • High-precision data was used to model 410-km discontinuity without extrapolation. • Hydrous MTZ can reconcile the long-standing discrepancy on olivine content. • Pyrolitic MTZ is globally hydrous at least at its top. [ABSTRACT FROM AUTHOR]

Abstract We report on laser-heated diamond anvil cell (LHDAC) experiments in the FeO–MgO–SiO 2 –CO 2 (FMSC) and CaO–MgO–SiO 2 –CO 2 (CMSC) systems at lower mantle pressures designed to test for decarbonation and diamond forming reactions. Sub-solidus phase relations based on synthesis experiments are reported in the pressure range of ∼35 to 90 GPa at temperatures of ∼1600 to 2200 K. Ternary bulk compositions comprised of mixtures of carbonate and silica are constructed such that decarbonation reactions produce non-ternary phases (e.g. bridgmanite, Ca-perovskite, diamond, CO 2 –V), and synchrotron X-ray diffraction and micro-Raman spectroscopy are used to identify the appearance of reaction products. We find that carbonate phases in these two systems react with silica to form bridgmanite ±Ca-perovskite + CO 2 at pressures in the range of ∼40 to 70 GPa and 1600 to 1900 K in decarbonation reactions with negative Clapeyron slopes. Our results show that decarbonation reactions form an impenetrable barrier to subduction of carbonate in oceanic crust to depths in the mantle greater than ∼1500 km. We also identify carbonate and CO 2 –V dissociation reactions that form diamond plus oxygen. On the basis of the observed decarbonation reactions we predict that the ultimate fate of carbonate in oceanic crust subducted into the deep lower mantle is in the form of refractory diamond in the deepest lower mantle along a slab geotherm and throughout the lower mantle along a mantle geotherm. Diamond produced in oceanic crust by subsolidus decarbonation is refractory and immobile and can be stored at the base of the mantle over long timescales, potentially returning to the surface in OIB magmas associated with deep mantle plumes. Highlights • Experiments were made in the systems FMS-CO2 and CMS-CO2 at lower mantle pressures. • Carbonate in these systems react with silica in the lower mantle. • Decarbonation reactions with negative P–T slopes form CO 2 or diamond plus oxygen. • Carbonate in oceanic crust is limited to ∼1500 km depth along subduction geotherms. • Diamond is the stable form of carbon in oceanic crust at ambient mantle temperature. [ABSTRACT FROM AUTHOR]

• Hydrous melts of the mantle above and below the transition zone have 40-50 mol% H 2 O. • Structure, density, and viscosity of hydrous melts are determined using ab initio MD. • Hydrous melts with Mg#s > 25 will be less dense than PREM mantle and are inviscid. • Hydrous melts are highly mobile and move upward above and below the transition zone. • Modelling shows the upper mantle and transition zone have similar H 2 O contents. We report ab initio atomistic simulations of hydrous silicate melts under deep upper mantle to shallow lower mantle conditions and use them to parameterise density and viscosity across the ternary system MgO-SiO 2 -H 2 O (MSH). On the basis of phase relations in the MSH system, primary hydrous partial melts of the mantle have 40-50 mol% H 2 O. Our results show that these melts will be positively buoyant at the upper and lower boundaries of the mantle transition zone except in very iron-rich compositions, where ≳ 75% Mg is substituted by Fe. Hydrous partial melts will also be highly inviscid. Our results indicate that if melting occurs when wadsleyite transforms to olivine at 410 km, melts will be buoyant and ponding of melts is unexpected. Box models of mantle circulation incorporating the upward mobility of partial melts above and below the transition zone suggest that the upper mantle becomes efficiently hydrated at the expense of the transition zone such that large differences in H 2 O concentration between the upper mantle, transition zone and lower mantle are difficult to maintain on timescales of mantle recycling. The MORB source mantle with ∼0.02-0.04 wt% H 2 O may be indicative of the H 2 O content of the transition zone and lower mantle, resulting in a bulk mantle H 2 O content of the order 0.5 to 1 ocean mass, which is consistent with geochemical constraints and estimates of subduction ingassing. [ABSTRACT FROM AUTHOR]

Abstract: Unraveling the conditions at which Earth''s metallic iron core formed yields important information about Earth''s early accretion and differentiation history. Multi-variable statistical modeling of siderophile element partitioning between core-forming metallic liquids and silicate melts form the basis for physical models of core formation. While it seems clear that core segregation in a deep peridotitic magma ocean is generally consistent with many mantle siderophile element abundances, there is considerable disparity among extant physical models in terms of the key parameters of pressure, temperature and oxygen fugacity at which the core formed. Moreover, there is ongoing debate over whether simple single-stage equilibrium or more complex multi-stage accretion models are required by the partitioning data. Here we consider how variations in the statistical regression of partitioning data affect the outcomes of physical models for core formation. Taking extant experimental data sets for four well-studied siderophile elements (Ni, Co, W and V) as examples, we find that the regression model exerts a fundamental control on physical model outcomes. Further, the experimental data are currently too imprecise to discriminate among various single-stage and continuous core formation scenarios. Progress in the development of physical models requires better isolation of the independent variables that affect partition coefficients and verification of activity models at high pressure and temperature in order to reduce the global uncertainty in multi-variable statistical models. [Copyright &y& Elsevier]

Abstract: Tungsten partitioning between liquid metal and liquid silicate, D(W), from 1723–2673 K, 0.5–18 GPa, and over a wide range of metal and silicate compositions provides constraints on planetary core formation models. We find that W partitioning is extremely sensitive to the carbon content of the metal alloy, becoming about an order of magnitude more siderophile at carbon saturation. Activity-composition corrections based on interactions between W–C and Fe–C in metal solution shift calculated D(W) to more lithophile values, and calculated oxygen fugacities (fO2) to more oxidizing values, than uncorrected data. W exists as a highly charged cation in silicate solution, displaying a mixture of oxidation states from +4 to +6 in experiments at fO2 of ∼0.5–2 log units below the iron–wüstite buffer. At constant fO2, the average calculated valence decreases with pressure from ∼+5.5 at 0.5 GPa to ∼+4 at 11–18 GPa. As a result of its high oxidation state, W partitioning is strongly dependent on silicate melt polymerization and fO2. In contrast to some previous results, we find that D(W) may decrease slightly in response to increasing temperature in the pressure range of our experiments. Pressure exerts a non-monotonic effect: D(W) increases with pressure up to ∼3 to 4 GPa, but decreases at higher pressures. Previous models for the effects of pressure and temperature on W partitioning that conflict with our results appear to result from a conflation of the intensive parameters of pressure, temperature, and carbon content. The mantle abundance of W could have been set by single-stage metal–silicate equilibration along the liquidus in a deep peridotite magma ocean at pressures from 20–50 GPa, and at oxygen fugacities consistent with the mantle''s present iron budget (IW−2 to −2.5). Equilibration at higher pressure is viable if the core-forming metal contained a significant, but not unreasonable, abundance of carbon (∼2 wt.%). Recent continuous accretion models involving multi-stage metal-silicate equilibration in a deepening magma ocean with progressive oxidation of the silicate remain permissible given our new treatment of W partitioning data. [Copyright &y& Elsevier]

Abstract: Experimentally determined phase relations in the system MgO–AlO1.5–SiO2 at pressures relevant to the upper part of the lower mantle indicate that Mg–silicate perovskite incorporates aluminum into its structure almost exclusively by a charge-coupled reaction. MgSiO3-rich bulk compositions along the joins MgSiO3–MgAlO2.5 and MgSiO3–MgAl2O4 crystallize assemblages of perovskite coexisting with periclase. MgO-saturated perovskites along these joins have ambient unit cell volumes consistent with those measured and calculated for aluminous perovskite along the charge-coupled substitution join, MgSiO3–AlO1.5. The compressibility of aluminous perovskite along the MgO-saturated joins is not anomalously low as predicted for oxygen-defect perovskites. The bulk moduli, however, are consistent with previous measurements made for aluminous perovskites along the charge-coupled substitution join. These results agree with first-principles calculations showing very limited stability of O-defects in Mg-perovskite at pressures and temperatures corresponding to lower mantle conditions, but are inconsistent with earlier experimental results showing unusually compressive aluminous perovskite. The maximum solubility of alumina in perovskite is ∼25 mol% along the MgSiO3–AlO1.5 join within the ternary MAS-system (i.e. pyrope composition), and the join is apparently binary. Although primitive mantle peridotite compositions are MgO-saturated and fall nearly on the oxygen vacancy join, alumina substitution into perovskite is expected to occur primarily by charge-coupled substitution throughout the lower mantle. The compressibility of aluminous perovskite in primitive mantle is expected to be only a few percent lower than for end member MgSiO3 perovskite. [Copyright &y& Elsevier]

The birth and infancy of Earth was a time of profound differentiation involving massive internal reorganization into core, mantle and proto-crust, all within a few hundred million years of solar system formation (t0). Physical and isotopic evidence indicate that the formation of iron-rich cores generally occurred very early in planetesimals, the building blocks of proto-Earth, within about 3 million years of t0. The final stages of terrestrial planetary accretion involved violent and tremendously energetic giant impacts among core-segregated Mercury- to Mars-sized objects and planetary embryos. As a consequence of impact heating, the early Earth was at times partially or wholly molten, increasing the likelihood for high-pressure and high-temperature equilibration among core- and mantle-forming materials. The Earth''s silicate mantle harmoniously possesses abundance levels of the siderophile elements Ni and Co that can be reconciled by equilibration between iron alloy and silicate at conditions comparable to those expected for a deep magma ocean. Solidification of a deep magma ocean possibly involved crystal–melt segregation at high pressures, but subsequent convective stirring of the mantle could have largely erased nascent layering. However, primitive upper mantle rocks apparently have some nonchondritic major and trace element refractory lithophile element ratios that can be plausibly linked to early mantle differentiation of ultra-high-pressure mantle phases. The geochemical effects of crystal fractionation in a deep magma ocean are partly constrained by high-pressure experimentation. Comparison between compositional models for the primitive convecting mantle and bulk silicate Earth generally allows, and possibly favors, 10–15% total fractionation of a deep mantle assemblage comprised predominantly of Mg-perovskite and with minor but geochemically important amounts of Ca-perovskite and ferropericlase. Long-term isolation of such a crystal pile is generally consistent with isotopic constraints for time-integrated Sm/Nd and Lu/Hf ratios in the modern upper mantle and might account for the characteristics of some mantle isotope reservoirs. Although much remains to be learned about the earliest formative period in the Earth''s development, a convergence of theoretical, physical, isotopic and geochemical arguments is beginning to yield a self-consistent portrait of the infant Earth. [Copyright &y& Elsevier]

The connectivity of molten Fe–S in peridotite has been experimentally investigated by means of in situ electrical conductivity measurements at high temperatures and 1 GPa. Starting materials were powdered mixtures of peridotite KLB-1 with various amounts (0, 3, 6, 13, 19, 24 vol.%) of the 1 GPa eutectic composition in the Fe–FeS binary system. At temperatures above the eutectic point in the Fe–FeS system (∼980 °C) and below the solidus of KLB1 (∼1200 °C), molten Fe–S in a solid silicate matrix interconnects when the volume fraction is over ∼5%. Conductivity–temperature paths indicate that in the presence of partial silicate melting the connectivity of molten Fe–S in a peridotite matrix is inhibited. Based on observations of retrieved samples, the percolation threshold of Fe–S melts in the presence of low to moderate degrees of silicate melt is estimated at 13±2 vol.%. These results indicate that if the volume fraction of Fe-alloy in a planetesimal was initially greater than 5%, and if early heating by decay of radionuclides raised the temperature of the interior above the Fe-alloy melting point, initial metal segregation was controlled by permeable flow of molten iron alloy in a solid silicate matrix. These conditions were likely met by many terrestrial objects in the early solar nebula. Efficient removal of residual Fe-alloy (5 vol.%) from silicate requires high-degree melting of silicate so that metal can segregate as droplets. Giant impacts during the final stage of accretion of large planetary objects could supply the energy required for high-degrees of melting. Alternatively, if initial metal segregation were delayed until a planetary object grew to large size (∼1000 km in diameter), release of gravitational potential energy due to metal segregation could contribute enough heat to form a magma ocean. [Copyright &y& Elsevier]

Subduction of oceanic lithosphere transports surface H 2 O into the mantle. Recent studies show that dense SiO 2 in the form of stishovite, an abundant mineral in subducted oceanic crust at depths greater than ∼270 km, has the potential to host and transport a considerable amount of H 2 O into the lower mantle, but the H 2 O storage capacity of SiO 2 phases at high pressure and temperature remains uncertain. We investigate the hydration of stishovite and its higher-pressure polymorphs, β -stishovite and seifertite, with in situ X-ray diffraction experiments at high pressures and temperatures. The H 2 O contents in SiO 2 phases are quantified based on observed increases in unit cell volume relative to the anhydrous SiO 2 system. Density functional theory (DFT) computations permit calibration of water content as a function of volume change based on interstitial substitution of H 2 O. Regression of our experimental data indicates an H 2 O storage capacity in stishovite of ∼3.5 wt% in the transition zone and shallow lower mantle, decreasing to about 0.8 wt% at the base of the mantle. We find that SiO 2 -bearing subducted oceanic crust can accommodate all the H 2 O in slab lithosphere that survives sub-arc dehydration. Hydration of silica phases in subducted oceanic crust and their unparalleled capacity to host significant amounts of H 2 O even at high mantle temperatures provides a unique mechanism for transport and storage of water in the deepest mantle. • Quantification of the H 2 O storage capacity in hydrous silica in the deep mantle. • Parameterization of the H 2 O storage capacity in hydrous silica as a function of P-T. • Interstitial substitution of H 2 O dominates incorporation water of hydrous silica. • Hydrous silica plays an important role in transporting water to the deep Earth. [ABSTRACT FROM AUTHOR]

Geochemical studies of planetary accretion and evolution have invoked various degrees of collisional erosion to explain differences in bulk composition between planets and chondrites. Here we undertake a full, dynamical evaluation of ‘crustal stripping’ during accretion and its key geochemical consequences. Crusts are expected to contain a significant fraction of planetary budgets of incompatible elements, which include the major heat producing nuclides. We present smoothed particle hydrodynamics simulations of collisions between differentiated rocky planetesimals and planetary embryos. We find that the crust is preferentially lost relative to the mantle during impacts, and we have developed a scaling law based on these simulations that approximates the mass of crust that remains in the largest remnant. Using this scaling law and a recent set of N -body simulations of terrestrial planet formation, we have estimated the maximum effect of crustal stripping on incompatible element abundances during the accretion of planetary embryos. We find that on average approximately one third of the initial crust is stripped from embryos as they accrete, which leads to a reduction of ∼20% in the budgets of the heat producing elements if the stripped crust does not reaccrete. Erosion of crusts can lead to non-chondritic ratios of incompatible elements, but the magnitude of this effect depends sensitively on the details of the crust-forming melting process on the planetesimals. The Lu/Hf system is fractionated for a wide range of crustal formation scenarios. Using eucrites (the products of planetesimal silicate melting, thought to represent the crust of Vesta) as a guide to the Lu/Hf of planetesimal crust partially lost during accretion, we predict the Earth could evolve to a superchondritic 176 Hf/ 177 Hf (3–5 parts per ten thousand) at present day. Such values are in keeping with compositional estimates of the bulk Earth. Stripping of planetary crusts during accretion can lead to detectable changes in bulk composition of lithophile elements, but the fractionation is relatively subtle, and sensitive to the efficiency of reaccretion. [ABSTRACT FROM AUTHOR]

Eutectic melting curves in the system MgO–SiO 2 have been experimentally determined at lower mantle pressures using laser-heated diamond anvil cell (LH-DAC) techniques. We investigated eutectic melting of bridgmanite plus periclase in the MgO–MgSiO 3 binary, and melting of bridgmanite plus stishovite in the MgSiO 3 –SiO 2 binary, as analogues for natural peridotite and basalt, respectively. The melting curve of model basalt occurs at lower temperatures, has a shallower d T / d P slope and slightly less curvature than the model peridotitic melting curve. Overall, melting temperatures detected in this study are in good agreement with previous experiments and ab initio simulations at ∼25 GPa ( Liebske and Frost, 2012 ; de Koker et al., 2013 ). However, at higher pressures the measured eutectic melting curves are systematically lower in temperature than curves extrapolated on the basis of thermodynamic modelling of low-pressure experimental data, and those calculated from atomistic simulations. We find that our data are inconsistent with previously computed melting temperatures and melt thermodynamic properties of the SiO 2 endmember, and indicate a maximum in short-range ordering in MgO–SiO 2 melts close to Mg 2 SiO 4 composition. The curvature of the model peridotite eutectic relative to an MgSiO 3 melt adiabat indicates that crystallization in a global magma ocean would begin at ∼100 GPa rather than at the bottom of the mantle, allowing for an early basal melt layer. The model peridotite melting curve lies ∼ 500 K above the mantle geotherm at the core–mantle boundary, indicating that it will not be molten unless the addition of other components reduces the solidus sufficiently. The model basalt melting curve intersects the geotherm at the base of the mantle, and partial melting of subducted oceanic crust is expected. [ABSTRACT FROM AUTHOR]

The melting curve of Ni has been determined to 125 GPa using laser-heated diamond anvil cell (LH-DAC) experiments in which two melting criteria were used: firstly, the appearance of liquid diffuse scattering (LDS) during in situ X-ray diffraction (XRD) and secondly, plateaux in temperature vs. laser power functions in both in situ and off-line experiments. Our new melting curve, defined by a Simon–Glatzel fit to the data where T M ( K ) = [ ( P M 18.78 ± 10.20 + 1 ) ] 1 / 2.42 ± 0.66 × 1726 , is in good agreement with the majority of the theoretical studies on Ni melting and matches closely the available shock wave melting data. It is however dramatically steeper than the previous off-line LH-DAC studies in which determination of melting was based on the visual observation of motion aided by the laser speckle method. We estimate the melting point ( T M ) of Ni at the inner-core boundary (ICB) pressure of 330 GPa to be T M = 5800 ± 700 K ( 2 σ ) , within error of the value for Fe of T M = 6230 ± 500 K determined in a recent in situ LH-DAC study by similar methods to those employed here. This similarity suggests that the alloying of 5–10 wt.% Ni with the Fe-rich core alloy is unlikely to have any significant effect on the temperature of the ICB, though this is dependent on the details of the topology of the Fe–Ni binary phase diagram at core pressures. Our melting temperature for Ni at 330 GPa is ∼2500 K higher than that found in previous experimental studies employing the laser speckle method. We find that those earlier melting curves coincide with the onset of rapid sub-solidus recrystallization, suggesting that visual observations of motion may have misinterpreted dynamic recrystallization as convective motion of a melt. This finding has significant implications for our understanding of the high-pressure melting behaviour of a number of other transition metals. [ABSTRACT FROM AUTHOR]

• Elastic properties of H-bearing Fe alloys were calculated using ab initio molecular dynamic simulations. • Superionic H-bearing alloys were found under inner-core conditions. • The hcp -Fe–Si–H alloys can explain the density and seismic velocities of the inner core. Geophysical and geochemical evidence suggests that Earth's core is predominantly made of iron (or iron-nickel alloy) with several percent of light elements. However, Earth's solid inner core transmits shear waves at a much lower velocity than expected from mineralogical models that are consistent with geochemical constraints. Here we investigate the effect of hydrogen on the elastic properties of iron and iron-silicon alloys using ab initio molecular dynamic simulations. We find that these H-bearing alloys maintain a superionic state under inner-core conditions and that their shear moduli exhibit a strong shear softening due to the superionic effect, with a corresponding reduction in V S. Several hcp -iron-silicon-hydrogen compositions can explain the observed density, V P , V S , and Poisson's ratio of the inner core simultaneously. Our results indicate that hydrogen is a significant component of the Earth's core, and that it may contain at least four ocean masses of water. This indicates that the Earth may have accreted wet and obtained its water from chondritic and/or nebular materials before or during core formation. [ABSTRACT FROM AUTHOR]

• Slab-derived, carbonated melt can persist in the reduced mantle over geological time. • Melt migrates from the slab through channels in the mantles. • The seismically observed LVL beneath the western US can sequester up to 25 MT CO2/yr. We present a theoretical model of the stability and migration of carbonate-rich melts to test whether they can explain seismic low-velocity layers (LVLs) observed above stalled slabs in several convergent tectonic settings. The LVLs, located atop the mantle transition zone, contain small (∼1 vol%) amounts of partial melt, possibly derived from melting of subducted carbonate-bearing oceanic crust. Petrological and geochemical evidence from inclusions in superdeep diamonds supports the existence of slab-derived carbonate melt, which may potentially explain the origin of the observed melt in the LVL. However, the presumptive reducing nature of the ambient mantle can be an impediment to the stability of carbonated melt. To reconcile this apparent contradiction, we test the stability and migration rates of carbonate-rich melts atop a stalled slab as a function of melt percolation, redox freezing, amount of carbon supplied by subduction, and the metallic Fe concentration in the mantle. Our results demonstrate that carbonate-rich melts in the LVL can potentially survive redox freezing over long geological time scales. We also show that the amount of subducted carbon exerts a stronger influence on the stability of carbonate melt than does the mantle redox condition. Concentration dependent melt density leads to rapid melt propagation through channels while a constant melt density causes melt to migrate as a planar front. Our calculations suggest that the LVLs can sequester significant fractions of carbon transported to the mantle by subduction. [ABSTRACT FROM AUTHOR]