Core formation and chemical equilibrium in the Earth—II: Chemical consequences for the mantle and core

We present here a new model of core formation which is based on the current understanding of planetary accretion and discuss its implications for the chemistry of the Earth's mantle and core. Formation of the Earth by hierarchical accretion of progressively larger bodies on a time scale much longer than that of solid body differentiation in the nebula indicates that a significant fraction of metal in the core could be inherited from preterrestrially differentiated planetesimals. An analysis of the segregation of this iron to form the core suggests that most of the metal settles to the core without interaction with silicates; only a small fraction of the metal chemically equilibrates at high temperatures and pressures with the silicates. The siderophile element abundances in the mantle are considered to be a consequence of a two-step equilibration with iron, once preterrestrially in the planetesimals at low temperatures and pressures, and later in the Earth at high temperatures and pressures. The highly siderophile elements such as Re, Au and the platinum group elements in the mantle are essentially excluded from silicates from the preterrestrial equilibration. We attribute the abundances of these elements in the mantle to the later equilibration in the Earth at substantially reduced metal-silicate partition coefficients (D^met/sil), for which there is a considerable experimental evidence now. Mass balance considerations constrain the fraction of core metal involved in such an equilibration at approximately 0.3 – 0.5%. The model accounts for the levels and the near-chondritic ratios of the highly siderophile elements in the mantle. The mantle abundances of the less siderophile elements are largely determined by preterrestrial metal-silicate equilibrium and are not significantly affected by the second equilibration. The extreme depletion of sulfur and the lack of silicate melt-sulfide signature in the noble metal abundances in the mantle are natural consequences of this mode of core formation. Sulfur was added to the magma ocean during the high-T, high-P equilibration in the Earth, not extracted from it by sulfide segregation to the core. Except for Ni and Co, the overall siderophile abundances of the mantle can be well matched in this two-step equilibration model.

The mantle characteristics of Ni and Co are unique to the Earth and hence suggest a terrestrial process as the likely cause. One such process is the flotation and addition of olivine to the primitive upper mantle. In our model of core formation, neither the elemental and isotopic data of Re---Os, nor the low sulfur content of the mantle remains as an objection to the existence of a magma ocean and olivine flotation.

The small fraction of core metal that equilibrates with silicates at high T and P suggests that the light elements O, Si or H are unimportant in the core, leaving S (and possibly C) as prime candidates. Sulfur, as FeS associated with incoming iron metal, is directly sequestered to the core along with the bulk of the iron metal. It appears unlikely that other light elements can be added to the core after its formation. U and Th are excluded from the core but the model allows for entry of some K; however, the extent to which K serves as a heat source in the core remains uncertain.

The model is testable in two ways. One is by investigation of the metal-silicate partitioning at high temperatures and pressures under magma ocean conditions to determine if the (D^met/sil) values are lowered to the levels required in the model. The other is by experiments to determine if a solvus closure between metal and silicate liquids occurs at high temperatures relevant to a magma ocean.     

Trace elements in shergottite meteorites: Implications for the origins of planets

The average concentrations of 19 siderophile and volatile elements in shergottite meteorites differ from those in terrestrial basalts by less than a factor of ten. This observation undermines claims that the abundances of siderophile and volatile elements in the Earth's upper mantle are uniquely terrestrial. Claims that similarities in the Moon's siderophile element pattern imply a terrestrial origin for the Moon are also weakened. The implication that basalt source regions on the asteroidal parent body of the shergottites resembled the terrestrial upper mantle constrains models of planetary formation and evolution. Heterogeneous accretion models may explain many of the similarities between these planets. Alternatively, separation of sulfide from basaltic magmas or their source regions on the Earth and the shergottite parent body may explain some of these similarities.     

The influence of silicate melt composition on distribution of siderophile elements among metal and silicate liquids

Liquid metal-liquid silicate partitioning of Fe, Ni, Co, P, Ge, W and Mo among a carbon-saturated metal and a variety of silicate melts (magnesian-tholeiitic-siliceous-aluminous-aluminosiliceous basalts) depends modestly to strongly upon silicate melt structure and composition. Low valency siderophile elements, Fe, Ni and Co, show a modest influence of silicate melt composition on partitioning. Germanium shows a moderate but consistent preference for the depolymerized magnesian melt. High valency siderophile elements, P, Mo, and W, show more than an order of magnitude decrease in metal-silicate partition coefficients as the silicate melt becomes more depolymerized. Detailed inspection of our and other published W data shows that polymerization state, temperature and pressure are more important controls on W partitioning than oxidation state. For this to be true for a high and variable valence element implies a secondary role in general for oxidation state, even though some role must be present. Equilibrium core segregation through a magma ocean of ‘ultrabasic’ composition can provide a resolution to the ‘excess’ abundances of Ge, P, W and Mo in the mantle, but the mantle composition alone cannot explain the excess abundances of nickel and cobalt in chondritic proportions.     

Siderophile and chalcophile element abundances in oceanic basalts, Pb isotope evolution and growth of the Earth''s core

We have investigated the hypothesis that mantle Pb isotope ratios reflect continued extraction of Pb into the Earth's core over geologic time. The Pb, Sr and Nd isotopic compositions, and the abundance of siderophile and chalcophile elements (W, Mo and Pb) and incompatible lithophile elements have been determined for a suite of ocean island and mid-ocean ridge basalt samples. Over the observed range in Pb isotopic compositions for oceanic rocks, we found no systematic variation of siderophile or chalcophile element abundances relative to abundances of similarly incompatible, but lithophile, elements. The high sensitivity of theMo/Pr ratio to segregation of Fe-metal or S-rich metallic liquid (sulfide) and the observed constantMo/Pr ratio rules out the core formation model as an explanation for the Pb paradox. The mantle and crust have the sameMo/Pr and the sameW/Ba ratios, suggesting that these ratios reflect the ratio in the Earth's primitive mantle.

Our data also indicate that thePb/Ce ratio of the mantle is essentially constant, but the presentPb/Ce ratio in the mantle ( 0.036) is too low to represent the primitive value ( 0.1) derived from Pb isotope systematics. HigherPb/Ce ratios in the crust balance the lowPb/Ce of the mantle, and crust and mantle appear to sum to a reasonable terrestrialPb/Ce ratio. The constancy of thePb/Ce ratio in a wide variety of oceanic magma types from diverse mantle reservoirs indicates this ratio is not fractionated by magmatic processes. This suggests crust formation must have involved non-magmatic as well as magmatic processes. Hydrothermal activity at mid-ocean ridges may result in significant non-magmatic transport of Pb from mantle to crust and of U from crust to mantle, producing a higherU/Pb ratio in the mantle than in the total crust. We suggest that the lower crust is highly depleted in U and has unradiogenic Pb isotope ratios which balance the radiogenic Pb of upper crust and upper mantle. The differences between thePb/Ce ratio in sediments, this ratio in primitive mantle, and the observed ratio in oceanic basalts preclude both sediment recycling and mixing of primitive and depleted reservoirs from being important sources of chemical heterogeneities in the mantle.     

Mechanisms of metal–silicate equilibration in the terrestrial magma ocean

It has been proposed that the high concentrations of moderately siderophile elements (e.g. Ni and Co) in the Earth’s mantle are the result of metal–silicate equilibration at the base of a deep magma ocean that formed during Earth’s accretion. According to this model, liquid metal ponds at the base of the magma ocean and, after equilibrating chemically with the overlying silicate liquid at high pressure (e.g. 25–30 GPa), descends further as large diapirs to form the core. Here we investigate the kinetics of metal–silicate equilibration in order to test this model and place new constraints on processes of core formation. We investigate two models: (1) Reaction between a layer of segregated liquid metal and overlying silicate liquid at the base of a convecting magma ocean, as described above. (2) Reaction between dispersed metal droplets and silicate liquid in a magma ocean. In the liquid-metal layer model, the convection velocity of the magma ocean controls both the equilibration rate and the rate at which the magma ocean cools. Results indicate that time scales of chemical equilibration are two to three orders of magnitude longer than the time scales of cooling and crystallization of the magma ocean. In the falling metal droplet model, the droplet size and settling velocity are critical parameters that we determine from fluid dynamics. For likely silicate liquid viscosities, the stable droplet diameter is estimated to be ∼1 cm and the settling velocity ∼0.5 m/s. Using such parameters, liquid metal droplets are predicted to equilibrate chemically after falling a distance of <200 m in a magma ocean. The models indicate that the concentrations of moderately siderophile elements in the mantle could be the result of chemical interaction between settling metal droplets and silicate liquid in a magma ocean but not between a segregated layer of liquid metal and overlying silicate liquid at the base of the magma ocean. Finally, due to fractionation effects, the depth of the magma ocean could have been significantly different from the value suggested by the apparent equilibration pressure.     

Targeting the impactors: siderophile element signatures of lunar impact melts from Serenitatis

Highly siderophile element compositions of lunar impact melt breccias provide a unique record of the asteroid population responsible for large cratering events in the inner Solar System. Melt breccias associated with the 3.89 Ga Serenitatis impact basin resolve at least two separate impact events. KREEP-rich melt breccias representing the Apollo 17 poikilitic suite are enriched in highly siderophile elements (3.6-15.8 ppb Ir) with CI-normalized patterns that are elevated in Re, Ru and Pd relative to Ir and Pt. The restricted range of lithophile element compositions combined with the coherent siderophile element signatures indicate formation of these breccias in a single impact event involving an EH chondrite asteroid, probably as melt sheet deposits from the Serenitatis Basin. One exceptional sample, a split from melt breccia 77035, has a distinctive lithophile element composition and a siderophile element signature more like that of ordinary chondrites, indicating a discrete impact event. The recognition of multiple impact events, and the clear signatures of specific types of meteoritic impactors in the Apollo 17 melt breccias, shows that the lunar crust was not comprehensively reworked by prior impacts from 3.9 to 4.5 Ga, an observation more consistent with a late cataclysm than a smoothly declining accretionary flux. Late accretion of enstatite chondrites during a 3.8-4.0 Ga cataclysm may have contributed to siderophile element heterogeneity on the Earth, but would not have made a significant contribution to the volatile budget of the Earth or oxidation of the terrestrial mantle. Siderophile element patterns of Apollo 17 poikilitic breccias become more fractionated with decreasing concentrations, trending away from known meteorite compositions to higher Re/Ir and Pd/Pt ratios. The compositions of these breccias may be explained by a two-stage impact melting process involving: (1) deep penetration of the Serenitatis impactor into meteorite-free lower crust, followed by (2) incorporation of upper crustal lithologies moderately contaminated by prior meteoritic infall into the melt sheet. Trends to higher Re/Ir with decreasing siderophile element concentrations may indicate an endogenous lunar crustal component, or a non-chondritic late accretionary veneer in the pre-Serenitatis upper crust.     

A metal particle from a Ca,Al-rich inclusion from the meteorite Allende,and the condensation of refractory siderophile elements

A small particle (ca. 10^?6 g) was magnetically separated from a Ca,Al-rich inclusion of the Allende meteorite. By using instrumental neutron activation analysis it was found that the elements Os, W, Re, Ir, Mo, Ru and Pt were enriched by a mean factor of about 7000 relative to Cl chondrites.A polished section of the grain showed that it consisted mainly of silicates, with a rounded particle of metal and sulfide (20 μm across) attached to it.Concentrations of up to 11% Pt were determined with the microprobe in the Ni-Fe center of the particle. Furthermore, Rh was for the first time measured in an Allende inclusion. It is enriched in about the same degree as Pt, Ir and W. The Ni-Fe center was surrounded by troilite. Mo was concentrated in the sulfide, while Os and Ru were inhomogeneously distributed over the metal + sulfide phases. The particle is interpreted as direct product of metal condensation of the solar nebula. The sulfide phase formed at lower temperatures and caused redistribution of the refractory siderophile elements. Condensation calculations for a metal alloy show that Fe and Ni are expected to be already present at higher temperatures than the condensation temperatures of pure Fe. Pt and Rh, having lower condensation temperatures than pure Fe should also be completely condensed above the condensation temperature of pure Fe. Kinetic considerations show that minimum times to grow this kind of particle should be of the order of 500 years at 10^?3 atm.     

Heterogeneous accretion,composition and core–mantle differentiation of the Earth

A model of core formation is presented that involves the Earth accreting heterogeneously through a series of impacts with smaller differentiated bodies. Each collision results in the impactor's metallic core reacting with a magma ocean before merging with the Earth's proto-core. The bulk compositions of accreting planetesimals are represented by average solar system abundances of non-volatile elements (i.e. CI-chondritic), with 22% enhancement of refractory elements and oxygen contents that are defined mainly by the Fe metal/FeO silicate ratio. Based on an anhydrous bulk chemistry, the compositions of coexisting core-forming metallic liquid and peridotitic silicate liquid are calculated by mass balance using experimentally-determined metal/silicate partition coefficients for the elements Fe, Si, O, Ni, Co, W, Nb, V, Ta and Cr. Oxygen fugacity is fixed by the partitioning of Fe between metal and silicate and depends on temperature, pressure and the oxygen content of the starting composition. Model parameters are determined by fitting the calculated mantle composition to the primitive mantle composition using least squares minimization. Models that involve homogeneous accretion or single-stage core formation do not provide acceptable fits. In the most successful models, involving 24 impacting bodies, the initial 60–70% (by mass) of the Earth accretes from highly-reduced material with the final 30–40% of accreted mass being more oxidised, which is consistent with results of dynamical accretion simulations. In order to obtain satisfactory fits for Ni, Co and W, it is required that the larger (and later) impactor cores fail to equilibrate completely before merging with the Earth's proto-core, as proposed previously on the basis of Hf-W isotopic studies. Estimated equilibration conditions may be consistent with magma oceans extending to the core–mantle boundary, thus making core formation extremely efficient. The model enables the compositional evolution of the Earth's mantle and core to be predicted throughout the course of accretion. The results are consistent with the late accretion of the Earth's water inventory, possibly with a late veneer after core formation was complete. Finally, the core is predicted to contain ~ 5 wt.% Ni, ~ 8 wt.% Si, ~ 2 wt.% S and ~ 0.5 wt.% O.     

Siderophile element fractionation in enstatite chondrites

The concentration of 10 to 15 siderophile elements was determined in the magnetic and non-magnetic portions of Abee (E4) and Hvittis (E6). The results indicate that, with the exception of Cu, W and Fe, all elements are strongly concentrated in the metal phase. Unlike ordinary chondrites, the metal phase of Abee and Hvittis consists exclusively of kamacite, which is very homogeneous and shows no systematic variation in composition with grain size.Differences in siderophile element content between Abee and Hvittis can be accounted for exclusively by differences in metal content and composition. These differences reflect different degrees of refractory siderophile loss, metal-silicate fractionation and loss of moderately volatile elements. The Ir/Ni ratio is 25% lower in Abee than in Hvittis, indicating that more Ir (Os, Pt, etc.) was lost from Abee during the refractory element fractionation. Abee and the other E4–5 members have also lost no metal and are not depleted in moderately volatile elements. In Abee the non-refractory elements Fe to Ge are present in CI ratios, and this meteorite has also Ir/Re ratios ?CI.These differences, which are recorded in the composition of the metal phase, make a straightforward genetic relationship between the two enstatite chondrite groups difficult to accept. In particular, the different Ir/Ni ratios, which were established very early in the chemical history of these chondrites, at the time of the refractory element fractionation, force us to conclude that E4–5 and E6 chondrites evolved from two different reservoirs, and that exchange of material among them never occurred. However, members of both groups have similar cosmic ray exposure ages suggesting derivation from the same parent body, which poses some interesting problems.     

Element partitioning between metallic liquid, silicate liquid, and lower-mantle minerals: implications for core formation of the Earth

We determined the partition coefficients of 19 elements between metallic liquid and silicate liquid at 20 GPa and 2500°C, and between metallic liquid and silicate perovskite at 27 GPa and 2200°C. Remarkable differences were observed in the partitioning behaviors of Si, P, W, Re, and Pb among the silicate liquid, perovskite, and magnesiowüstite coexisting with metallic liquid, reflecting incompatibility of the elements in the silicate or oxide phase. We could not observe any significant difference in the partitioning behaviors of V, Cr, Mn, Co, Ni, and Cu among the phases coexisting with metallic liquid.

Comparison of the present partitioning data with those obtained previously at lower pressure and temperature suggests that the exchange partition coefficients, K_met/sil, of Co, Ni, Mo, and W decrease, whereas those of V, Cr, and Mn increase and tend to approach unity with increasing pressure and temperature. We also made preliminary experiments to clarify the effect of sulfur on the partitioning behaviors. Sulfur lowers the exchange partition coefficients, K_met/sil, of Mo and W between metallic liquid and silicate liquid significantly at 20 GPa and 2300°C.

The mantle abundances of Co, Ni, Cu, Mo, and W calculated for the metal-silicate equilibrium model are lower than those of the real mantle, whereas P, K, and Mn are overabundant in the calculated mantle. The discrepancies in the abundances of Co and Ni could be explained by the chemical equilibrium at higher pressure and temperature. Large discrepancies in Mo and W between the calculated and real mantles could be accounted for by the effect of sulfur combined with the effects of pressure and temperature on the chemical equilibrium. The mantle abundances of P, K, and Cu could be accounted for by volatile loss in the nebula, perhaps before accretion of the Earth, combined with the chemical equilibrium at higher pressure and temperature. Thus the observed mantle abundances of P, K, Co, Ni, Cu, Mo, and W may be consistent with a model of sulfur-bearing metal-silicate equilibrium in lower-mantle conditions.     

Chemical composition of Archaean komatiites: Implications for early history of the earth and mantle evolution

Estimates of the chemical composition of the Archaean mantle, derived from elemental abundance ratios in komatiites combined with ultramafic xenolith data, support a model of a multistage heterogeneous accretion history of the Earth and synchronous core formation, 4.6 Ga ago.Most refractory lithophile element abundance ratios in komatiites and xenoliths are close to chondritic except for V/Ti and Ca/Al. Depletion of vanadium is likely due to its partial incorporation into the iron core; whereas fractionation of Ca/Al observed in Archaean Al-undepleted komatiites (1.20 times chondrites) and in some modern fertile spinel lherzolite xenoliths (1.15 times chondrites) could be due to small amounts of garnet (rich in Al but poor in Ca) segregation into the lower mantle during partial or complete melting of the upper mantle in the very early history of the Earth. However, this process may have had only a small effect on the overall chemical composition of the upper mantle.Simultaneous occurrence of early Archaean Al-undepleted (Al/Ti chondrites) and Al-depleted (Al/Ti 0.5 chondrites, and depletion of Sc and heavy REE) peridotitic komatiites in the Barberton area, S. Africa, and late Archaean Newton Township, Canada, argue against derivation of peridotitic komatiites from a circum-global magma ocean. Garnet separation from a mantle diapir which intersects the solidus at great depth ( 200 km) in a hotter early Archaean mantle could explain the chemical characteristics of Al-depleted komatiites. Alternatively, these two types of komatiites could have been derived from different layers in a fractionated mantle. A limited amount of Hf isotope data for Archaean komatiites seems to suggest that both mechanisms are important. This chemically and minerallogically layered mantle, if it existed, was homogenized by mantle convection after early Archaean times.Constant P₂O₅/TiO₂, Ni/Mg, Co/Mg, Fe/Mg ratios (siderophile/lithophile) and PGE abundances, estimated for the mantle sources of komatiites from Archaean to modern times, strongly argue against continuous growth of the Earth's core since the early Archaean.Extensive crustal contamination might have been involved in the generation of Archaean-early Proterozoic siliceous high magnesian basalts with “boninite affinity”. However, involvement of chemically modified ancient continental lithosphere may also be important in the generation of these basalts.     

A multi-isotopic and trace element investigation of the Cretaceous-Tertiary boundary layer at Stevns Klint, Denmark - inferences for the origin and nature of siderophile and lithophile element geochemical anomalies

Os, Sr, Nd and Pb isotope data were collected from a profile across the Cretaceous-Tertiary (K-T) boundary layer at Stevns Klint, Denmark. ?_Nd [T=65 Ma] values from within the boundary layer (Fish Clay) are lower by ∼1 ? unit than those of the underlying Maastrichtian limestone and the overlying Danian chalk sequences. Systematic profile-upward changes of Pb, Sr and Os isotopic compositions and concentrations in the boundary layer cannot be accounted for by in situ growth of daughter products since the sedimentation of the Fish Clay. While Os, Nd and Pb isotopes indicate the admixing of less radiogenic components to the Fish Clay, Sr isotopes show elevated radiogenic values in the boundary layer, relative to the carbonate sequences beneath and above it. The sudden change in lithophile (e.g., Sr, Pb and Nd) isotope compositions at the base of the Fish Clay and profile-upward trends of ⁸⁷Sr/⁸⁶Sr and ²⁰⁶Pb/²⁰⁴Pb ratios towards those of the overlying Danian chalk are interpreted to reflect recovery from enhanced, acid rain-induced continental (local?) weathering input to the seawater. However, a continental crustal source is invalid for the siderophile element Os. In the light of evidence from chromium isotopes for a cosmic origin of the platinum group elements (PGEs) and certain moderately siderophile elements (Cr, Ni, Co, V) in K-T boundary sediments, including Stevns Klint [Shukolyukov and Lugmair, Science 282 (1998) 927-929], and supported by the finding of projectile debris [Bauluz et al., Earth Planet. Sci. Lett. 182 (2000) 127-136] and the occurrence of abundant Ni-rich spinel at many K-T sites [Robin et al., Nature 363 (1993) 615-617; Kyte, Nature 396 (1998) 237-239], we favor to explain the sudden drop of ¹⁸⁷Os/¹⁸⁸Os ratios from 0.210 to 0.160 at the K-T boundary to derive from global fall-out of extraterrestrial matter. The present ¹⁸⁶Os/¹⁸⁸Os ratio of 0.119836±0.000004 measured in the basal layer of the Fish Clay is within the uncertainty a chondritic value. We therefore exclude the possibility of a major contribution of PGEs to the sediment from iron meteorites. Chondrite-normalized (Ru/Ir)_N ratios of ∼0.95±0.14 and (Os/Ir)_N ratios of ∼0.93±0.14 in the Fish Clay cannot distinguish between abundance ratios of different types of chondrites, and strongly sub-chondritic (Pt/Ir)_N ratios of ∼0.62±0.09 (2σ) suggest differential PGE remobilization through the sedimentary column (and consequently the alteration of inter-element ratios). PGEs and the moderately siderophile elements Cr, Ni, V, and Co form an elemental association with systematically upward-decreasing concentrations in the Fish Clay. Low Co/Ni ratios of ∼0.12 in the Fish Clay relative to values of ∼0.35 in the over- and underlying carbonate sequences support mixing of meteorite-derived (Co/Ni ∼0.05) and terrestrial upper mantle/crustal (Co/Ni >∼0.3) sources. While lithophile element isotope data indicate an increased continental crustal input to the Fish Clay at the K-T transition, the uncertainty with respect to possible post-depositional alteration of abundance patterns of siderophile and moderately siderophile elements - though not affecting the chondritic isotopic composition of Os - does not allow confirmation of indications from chromium isotopes for a carbonaceous (CV-type) meteorite as the preferred K-T impactor type by Shukolyukov and Lugmair [Science 282 (1998) 927-929].     

Prediction of siderophile element metal-silicate partition coefficients to 20 GPa and 2800°C: the effects of pressure, temperature, oxygen fugacity, and silicate and metallic melt compositions

We report new metal-silicate partition coefficients for Ni, Co and P at 7.0 GPa (1650–1750°C), and Ni, Co, Mo, W and P at 0.8, 1.0 and 1.5 GPa (1300–1400°C). Guided by thermodynamics, all available metal-silicate partition coefficients, D(i), where i is Ni, Co, P, Mo and W, are regressed against 1/T, P/T, lnf(O₂), ln(1 − X_s) (X_S is mole fraction of S in metallic liquid) and nbo/t (non-bridging oxygen/tetrahedral cation ratio, a silicate melt compositional-structural parameter) to derive equations of the following form: ln D(i) = aln f(O₂) + (b/T) + (cP/T) + d(nbo/t) + eln(1 − X_S) + f. Expressions for solid metal-liquid silicate and liquid metal-liquid silicate partition coefficients are derived for S-free and S-bearing systems.

We investigate whether Earth's upper-mantle siderophile element abundances can be reconciled with simple metal-silicate equilibrium. Sulfur-free metallic compositions do not allow a good fit. However, Ni, Co, Mo, W and P abundances in the upper mantle are consistent with simple metal-silicate equilibrium at mantle pressures and temperatures (27 GPa, 2200 K, ΔIW(iron-wüstite) = −0.15, nbo/t = 2.7; X_S = 0.15). Although these conditions are near the anhydrous peridotite solidus, they are well above the hydrous solidus and probably closer to the liquidus. A hydrous magma ocean and early mantle are consistent with predicted planetary accretion models. These results suggest that siderophile element abundances in Earth's upper mantle were established by liquid metal-liquid silicate equilibrium near the upper-mantle-lower-mantle boundary.     

Trace elements in ocean ridge basalts

A correlary of sea floor spreading is that the production rate of ocean ridge basalts exceeds that of all other volcanic rocks on the earth combined. Basalts of the ocean ridges bring with them a continuous record in space and time of the chemical characteristics of the underlying mantle. The chemical record is once removed, due to chemical fractionation during partial melting. Chemical fractionations can be evaluated by assuming that peridotite melting has proceeded to an olivine-orthopyroxene stage, in which case the ratios of a number of magmaphile elements in the extracted melt closely match the ratios in the mantle. Comparison of ocean ridge basalts and chondritic meteorites reveals systematic patterns of element fractionation, and what is probably a double depletion in some elements. The first depletion is in volatile elements and is due to high accretion temperatures of a large percentage of the earth from the solar nebula. The second depletion is in the largest, most highly charged lithophile elements (“incompatible elements”), probably because the mantle source of the basalts was melted previously, and the melt, enriched in these elements, was removed. Migration of melt relative to solid under ocean ridges and oceanic plates, element fractionation at subduction zones, and fractional melting of amphibolite in the Precambrian are possible mechanisms for depleting the mantle in incompatible elements. Ratios of transition metals in the mantle source of ocean ridge basalts are close to chondritic, and contrast to the extreme depletion of refractory siderophile elements, the reason for which remains uncertain. Variation of ocean ridge basalt chemistry along the length of the ridge has been correlated with ridge elevation. Thus chemically anomalous ridge segments up to 1000 km long appear to broadly coincide with regions of high magma production (plumes, hot spots). Basalt heterogeneity at a single location indicates mantle heterogeneity on a smaller scale. Variation of ocean ridge basalt chemistry with time has not been established, in fact, criteria for recognizing old oceanic crust in ophiolite terrains are currently under debate. The similarity of rare earth element patterns in basalt from ocean ridges, back-arc basins, some young island arcs, and some continental flood basalts illustrates the dangers of tectonic labeling by rare earth element pattern.     

The cosmic abundance of molybdenum

Fifteen carbonaceous chondrites were analysed for Mo and Ir by neutron activation analysis combined with a metal extraction method. The results of two Orgueil analyses gave a mean concentration of 915 ppb Mo. This corresponds to 2.51 atoms Mo/10⁶ atoms Si, which is 50% lower than data reported by Case et al. [3]. The lower Mo concentration for Orgueil was predicted by Suess and Zeh [4] from semi-empirical abundance rules. A constant Mo/Ir ratio is found for C1, C2, and C3V chondrites; C3Os have variable Mo/Ir ratios. These variations are due to variable Ir concentrations. Micron-sized grains enriched in Ir but not in Mo are presumably responsible for these variations. The Mo content of Karoonda is nearly a factor of four lower than that of C3V chondrites.     

Early crust on top of the Earth's core

In an effort to resolve the current conflict between geochemical requirements for an apparently isolated mantle reservoir and recent geophysical evidence for whole-mantle convection, we investigate the possibility that the region above the core-mantle boundary, termed D″, serves as an early-isolated, rare-gas- and incompatible-element-bearing reservoir, and we propose a mechanism for its formation that is a likely outcome of Earth accretion models. In these models, the most cataclysmic event in Earth history, the moon-forming giant impact on the proto-Earth of a Mars-size planet (perhaps as early as 4540 Ma ago) was followed by accretion of smaller bodies long afterwards (until ∼3900 Ma). Some collisions probably triggered melting, metal segregation and degassing. However, the small bodies, fragments, particles, dust, including those of chondrite-like composition, existed on near-earth orbits, were irradiated by intense solar wind, and finally fell on an early-formed, incompatible-element-enriched basaltic crust without causing extensive melting. The respective regions of the crust, loaded with chondrite-like debris, were therefore significantly enriched in iron. When this mixed material was subducted, the bulk density of the subducted lithosphere exceeded that of the bulk silicate mantle, which had already lost its metallic iron to the core. Segregation of this denser material at the base of the mantle was facilitated by the high temperatures at the core-mantle boundary, which greatly reduce the viscosity, as was quantitatively modelled by Christensen and Hofmann (Christensen, U.R., Hofmann, A.W., 1994. Segregation of subducted oceanic-crust in the convecting mantle. J. Geophys. Res.-Solid Earth 99 (B10), 19867–19884). Assuming a basalt/chondrite mass ratio of about 4/1, we obtain a density contrast of ∼7%, which would stabilize the subducted material between the metal core and silicate mantle.Mass balance considerations and preliminary results of geochemical modelling of the above scenario (similar to that performed by Tolstikhin and Marty [Tolstikhin, I.N., Marty, B., 1998. The evolution of terrestrial volatiles, A view from helium, neon, argon and nitrogen isotope modeling. Chem. Geol. 147, 27–52]) show the potential geochemical importance of D″. (1) Modelling of Pu–U–I–Xe isotope systematics predicts formation of this reservoir early in Earth history, ∼100 Ma after formation of the Solar system. (2) The total amount of heat-generating U, Th, K (and other highly incompatible elements) in D″ exceeds 20% of the Earth inventory, and a similar portion of terrestrial heat is being transferred from the core + D″ into the base of the overlying convecting mantle. (3) D″ is enriched in solar implanted rare gases because the small (re)-accreting debris with high surface/mass ratios will have been subjected to intense radiation by the early sun. (4) Rare gases diffuse from D″ into the overlying mantle and are then transferred into upwelling plumes, which originate above D″. In addition, small amounts of D″ material may be entrained by the mantle convective flow as was recently discussed by Schott et al. [Schott, B., Yuen, D.A., Braun, A., 2002. The influences of composition and temperature-dependent rheology in thermal-chemical convection on entrainment of the D″ layer. Physics Earth Planet. Inter. 129, 43–65]. From the rare-gas modelling it follows that initially (∼4500 Ma ago) D″ could have been more massive by a factor of ∼1.2 than at present (about 2 × 10²⁶ g). The present-day mass flux from D″ into the convecting mantle is estimated to be ≤0.05 × 10¹⁶ g year⁻¹, a factor of ∼100 less than the rate of ridge magmatism. This small contribution of D″ material makes it difficult to trace fingerprints of D″ even using such sensitive tracers as Pb isotope ratios. (5) The density contrast that stabilizes D″ is maintained by its higher intrinsic density due to the iron-rich chondrite-like component. Subduction of this material, its entrainment by convective mantle flow and mixing could also account for the preservation of the chondritic relative abundances of siderophile elements in the mantle. If D″ is partially molten, the density contrast may be caused by a high-density melt fraction.     

FeO and H2O and the homogeneous accretion of the earth

We present new shock devolatilization recovery data for brucite (Mg(OH)₂) shocked to 13 and 23 GPa. These data combined with previous data for serpentine (Mg₃Si₂O₅(OH)₄) are used to constrain the minimum size terrestrial planet for which planetesimal infall will result in an impact-generated water atmosphere. Assuming a chondritic abundance of minerals including 3–6%, by mass water, in hydrous phyllosilicates, we carried out model calculations simulating the interaction of metallic iron with impact-released free water on the surface of the accreting Earth. We assume that the reaction of water with iron in the presence of enstatite is the prime source of the terrestrial FeO component of silicates and oxides. Lower and upper bounds on the terrestrial FeO budget are based on mantle FeO content and possible incorporation of FeO in the outer core. We demonstrate that the iron-water reaction would result in the absence of atmospheric/hydrospheric water, if homogeneous accretion is assumed. In order to obtain10²⁵g of atmospheric water by the end of accretion, slightly heterogeneous accretion with initially 36% by mass iron planetesimals, as compared to a homogeneous value of 34% is required. Such models yield final FeO budgets, which either require a higher FeO content of the mantle (17 wt.%) or oxygen as a light element in the outer core of the Earth.     

Mantle geochemistry: Insights from ocean island basalts

The geochemical study of the Earth's mantle provides important constraints on our understanding of the formation and evolution of Earth, its internal structure, and the mantle dynamics. The bulk Earth composition is inferred by comparing terrestrial mantle rocks with chondrites, which leads to the chondritic Earth model. That is, Earth has the same relative proportions of refractory elements as that in chondrites, but it is depleted in volatiles. Ocean island basalts(OIB) may be produced by mantle plumes with possible deep origins; consequently, they provide unique opportunity to study the deep Earth. Isotopic variations within OIB can be described using a limited number of mantle endmembers, such as EM1, EM2 and HIMU, and they have been used to decipher important mantle processes. Introduction of crustal material into the deep mantle via subduction and delamination is important in generating mantle heterogeneity; however, there is active debate on how they were sampled by mantle melting, i.e.,the role of olivine-poor lithologies in the OIB petrogenesis. The origin and location of high 3He/4He mantle remain controversial,ranging from unprocessed(or less processed) primitive material in the lower mantle to highly processed materials with shallow origins, including ancient melting residues, mafic cumulates under arcs, and recycled hydrous minerals. Possible core-mantle interaction was hypothesized to introduce distinctive geochemical signatures such as radiogenic 186 Os and Fe and Ni enrichment in the OIB. Small but important variations in some short-lived nuclides, including 142 Nd, 182 W and several Xe isotopes, have been reported in ancient and modern terrestrial rocks, implying that the Earth's mantle must have been differentiated within the first 100 Myr of its formation, and the mantle is not efficiently homogenized by mantle convection.     

Accretion of the ordinary chondrites

The narrow size distributions of silicate and metal particles in 19 unequilibrated ordinary chondrites and other textural properties of these meteorites strongly suggest that chondritic material was sorted before or during its accumulation in parent bodies. Gravitational sorting during accretion is possible, but the conditions which it requires are implausible. Aerodynamic sorting - exclusion of small and/or low-density particles from a planetesimal moving through a mixture of gas and dust - can account for the textures of ordinary chondrites. It may also explain observed variations of siderophile element contents among and within the three groups of ordinary chondrites.     

Equilibrium thallium isotope fractionation and its constraint on Earth’s late veneer

Equilibrium isotope fractionation of thallium(Tl) includes the traditional mass-dependent isotope fractionation effect and the nuclear volume effect(NVE). The NVE dominates the overall isotope fractionation, especially at high temperatures. Heavy Tl isotopes tend to be enriched in oxidized Tl^3+-bearing species. Our NVE fractionation results of oxidizing Tl^+ to Tl^3+ can explain the positive enrichments observed in ferromanganese sediments. Experimental results indicate that there could be0.2–0.3 e-unit fractionation between sulfides and silicates at 1650 ℃. It is consistent with our calculation results,which are in the range of 0.17–0.38 e-unit. Importantly,Tl’s concentration in the bulk silicate Earth(BSE) can be used to constrain the amount of materials delivered to Earth during the late veneer accretion stage. Because the Tl concentration in BSE is very low and its Tl isotope composition is similar with that of chondrites, suggesting either no Tl isotope fractionation occurred during numerous evaporation events, or the Tl in current BSE was totally delivered by late veneer. If it is the latter, the Tl-contentbased estimation could challenge the magnitude of late veneer which had been constrained by the amount of highly siderophile elements in BSE. Our results show that the lateaccreted mass is at least five-times larger than the previously suggested magnitude, i.e., 0.5 wt% of current Earth’s mass. The slightly lighter 205 Tl composition of BSE relative to chondrites is probable a sign of occurrence of Tlbearing sulfides, which probably were removed from the mantle in the last accretion stage of the Earth.