Metal-silicate Partition Coefficients Research Articles

Earth's “Missing” Chlorine May Be in the Core

AbstractThe budget of volatile halogens in the bulk silicate Earth, fluorine and chlorine, differs distinctly from that of chondritic meteorites. Arguably, the bulk silicate Earth shows a low abundance of Cl, while the F budget is in line with the expected volatility trend. One hypothesis for the missing Cl is its sequestration into Earth's core during planetary segregation, but experimental data on partitioning between silicate and metallic melts are limited in pressure and remain inconclusive. Here we use computational quantum mechanical methods to study F and Cl geochemistry during core–mantle differentiation over a wide pressure and temperature range. Our calculations reveal that with increasing pressure and temperature, chlorine shows an enhanced affinity for iron metal. The Cl metal–silicate partition coefficient increase from −1.89 ± 0.84 at 10 GPa and 3000 K to 1.62 ± 0.80 at 130 GPa and 5000 K, while corresponding calculations for F show minimal variations in metal–silicate partitioning across the pressure–temperature conditions investigated, yielding values between −3.61 ± 0.81 and −3.37 ± 0.80. The shift in shows a transition from lithophile to siderophile behavior. Further calculations on isotopic partitioning show negligible fractionation of Cl isotopes (37Cl/35Cl) between the mantle and core. Our results suggest that metallic iron within Earth's mantle and core may serve as an important Cl reservoir, potentially accounting for up to 40% of Earth's overall Cl inventory.

Partitioning of nickel and cobalt between metal and silicate melts: Expanding the oxy-barometer to reducing conditions

Moderately siderophile elements (MSEs) are potential tracers of the thermodynamic conditions prevailing during planetary core formation because their metal–silicate partition coefficients (Dmet/sil) vary as a function of P, T, and oxygen fugacity (fO2). Those properties result in the production of planetary mantles with unique MSE depletion signatures. Among the MSEs, Ni and Co are reliable barometers in magma oceans because their Dmet/sil values are strongly correlated with pressure, decreasing by almost 3 orders of magnitude between 1 bar and 100 GPa. Current pressure-dependent expressions of Dmet/sil were calibrated based on experiments performed under relatively oxidizing conditions, mostly at fO2 slightly below the iron–wüstite Fe–FeO buffer (IW), which is relevant to the mantles of Earth and Mars. However, planets and asteroids formed under a wide range of redox conditions, from Mercury, the most reduced (∼IW − 5.5), to the most oxidized angrite parent body (IW − 1.5 to IW + 1). In this study, we performed and analyzed 38 metal–silicate partitioning experiments over a wide range of pressures (1 bar to 26 GPa) and oxygen fugacities (IW − 6.4 to IW − 1.9) to expand the available Ni and Co Dmet/sil values to reducing conditions. We then parameterized 255 Ni and 194 Co Dmet/sil values as a function of T (1573–5700 K), P (1 bar to 100 GPa), and fO2 (IW − 6.4 to IW + 0.2). We also modeled the evolution of Ni and Co Dmet/sil values along the liquidus of a chondritic mantle at various P and fO2 conditions to investigate the thermodynamic conditions of various planetary bodies’ magma oceans. The P and fO2 conditions we obtained for Earth, Mars, the Moon, and Vesta are consistent with previous studies using similar methods, and the pressure during core formation is strongly correlated to planetary size. Finally, we also applied our model to several achondrite parent bodies; our results indicate a wide variety of objects, from the asteroid-sized, oxidized angrite parent body to the planet-sized, highly reduced aubrite parent body.

Hafnium-tungsten evolution with pebble accretion during Earth formation

We combine calculations of pebble accretion and accretion by large and giant impacts to quantify the effects of pebbles on the hafnium-tungsten system during Earth formation. Our models include an early pebble accretion phase lasting 4–6 Myr with a global magma ocean and core segregation, a 20–50 Myr phase of large impacts, and a late giant impact representing the Moon-forming event. We consider various mass additions during each accretion phase, vary the metal-silicate partition coefficient for tungsten over a wide range, and track Hf180, Hf182, W182 and W184 in proto-Earth and impactor models over time using standard chondritic values for these isotopes in the pebbles. We find that an early phase of pebble accretion is compatible with the tungsten anomaly of Earth's early mantle as well as the present-day Hf/W ratio, but under restricted conditions. In particular, the pebble mass of proto-Earth is limited to 0.7 Earth masses or less, the average metal-silicate partition coefficient for tungsten is 30–50, and because the metal-silicate equilibration efficiency for giant impacts is low, the equilibration efficiency must be high for the large impactors.

I/Pu reveals Earth mainly accreted from volatile-poor differentiated planetesimals.

The observation that mid-ocean ridge basalts had ~3× higher iodine/plutonium ratios (inferred from xenon isotopes) compared to ocean island basalts holds critical insights into Earth's accretion. Understanding whether this difference stems from core formation alone or heterogeneous accretion is, however, hindered by the unknown geochemical behavior of plutonium during core formation. Here, we use first-principles molecular dynamics to quantify the metal-silicate partition coefficients of iodine and plutonium during core formation and find that both iodine and plutonium partly partition into metal liquid. Using multistage core formation modeling, we show that core formation alone is unlikely to explain the iodine/plutonium difference between mantle reservoirs. Instead, our results reveal a heterogeneous accretion history, whereby predominant accretion of volatile-poor differentiated planetesimals was followed by a secondary phase of accretion of volatile-rich undifferentiated meteorites. This implies that Earth inherited part of its volatiles, including its water, from late accretion of chondrites, with a notable carbonaceous chondrite contribution.

Cadmium isotope fractionation during metal-silicate partitioning – Results and implications for Earth's volatile accretion

Metal-silicate partitioning experiments were carried out at 1.5 GPa and 1508 to 1843 K to constrain Cd partitioning and isotope fractionation during core formation. At the studied conditions, there was no significant stable isotope fractionation during Cd partitioning between metal and silicate phases with a mean ∆114Cdmet-sil = −0.02 ± 0.09‰ (2SD, n = 7). Two experiments that investigated sulphide-silicate partitioning of Cd yielded fractionation factors of −0.04 ± 0.06‰ and − 0.23 ± 0.07‰ (2SE), whereby the latter result was obtained for a short run that may not represent full equilibrium. In summary, the findings suggest that Cd isotope fractionation during segregation of Earth's core was either absent or very minor. The Cd partitioning data of this and previous investigations were combined in multiple linear regression analyses to better constrain Cd metal-silicate partitioning during core formation. In accord with earlier work, the analyses reveal that Cd metal-silicate partitioning is not significantly impacted by temperature and pressure but affected by the S content of the metal phase. In addition, it is shown that the presence of C and Si in the metal reduce the siderophile character of Cd. Based on estimates for the composition of Earth's core, the data suggest a metal-silicate partition coefficient DCd of about 0.4 for a single-stage core formation event. However, given uncertainties about the light element composition of Earth's core, DCd values larger than 1 cannot be ruled out at present for core formation. The results of this study and data on the composition of the bulk silicate Earth and chondritic meteorites were applied in mass balance calculations to constrain the Cd signature of Earth's main stage accretion material prior to delivery of the late veneer. The modelling indicates Earth's main stage of accretion involved material with an average Cd isotope composition that was lighter than that of known carbonaceous and enstatite chondrites. Most likely, this reflects either poor characterisation of these meteorites by the few precise data currently available or that a significant fraction of the terrestrial volatile inventory was acquired from material not directly related to carbonaceous and enstatite chondrites. Furthermore, terrestrial accretion most likely did not encompass the addition of a volatile-rich late veneer exceeding 2% of Earth's mass, in accord with accretion models, which invoke that volatile delivery occurred primarily during main stage accretion, alongside core formation.

Metal compositions of carbonaceous chondrites

FeNi metals represent an important fraction of chondritic components that remains relatively unexplored within most carbonaceous chondrite groups. The compositions of these metals can place constraints on the nature of their precursor materials as well as the physicochemical conditions of chondrule formation. In this study, we have analyzed the major, minor and trace element compositions of metal grains from relatively unaltered carbonaceous chondrites NWA 801 (CR), Leoville (CV3.1), Paris (CM2.9), Maribo (CM2.8) and Bells (CM-an). We observe a predominant and constant sub-solar Co/Ni ratio of CR, CM and CM-an metal grains. In Ni versus Co space, the metal grains fall below modelled curves for equilibrium condensation of metals from a solar gas. From Ni versus Cr plots, we infer that Paris (and possibly Leoville) metal grains could have maintained a primary condensation signature, although for most grains, condensation must have occurred under disequilibrium conditions. CR and isolated CM-an metals mostly fall outside of the predicted condensation fields. Based on metal-silicate partition coefficients of Ni and Co that vary with pressure, we interpret their Co/Ni signatures as having a planetary origin, with presumable extraction by impact jetting. Considering that almost all CM and CR metal grains have the same Co/Ni ratio, we cannot rule out a planetary origin for CM metal grains. We relate the highly siderophile element (HSE) patterns of carbonaceous chondrite metal to mixing and subsequent equilibration of refractory metal nuggets (RMN), FeNi alloys and silicate chondrule precursors. As with the Co/Ni ratios, the HSE patterns of CM, CM-an, CR and CV metal grains are nearly identical, suggesting that the abundance and nature of the metal precursor materials were similar for carbonaceous chondrites. The overall volatility patterns of CV, CM and CR chondrites, suggest that the latter form under more oxidizing conditions than CV chondrites. The volatility patterns of Paris metal grains overlap with CV and CR chondrule metals, implying variable P-T-fO2 conditions during CM chondrule formation. Finally, we comment on the origin of metal grains in various petrological settings. Chondrule rim and isolated metal grains are likely derived and expelled from the equilibrated core metal and were subsequently altered to include and re-equilibrate with materials from the disk. Trace element analyses of the anomalous CM chondrite Bells metal grains show potential relationships with CM chondrite and CH chondrite metal for the chondrule cores and isolated grains, respectively. Small metal grains from CM chondrite Maribo, which are located in the chondrite matrix, potentially have distinct volatility patterns from CR and Paris isolated grains, hinting at a distinct origin for small metal grains and large chondrule-derived metal. Future work on carbonaceous chondrite metal should include an investigation of small (micron scale) versus large isolated metal grains.

Determination of the refractory enrichment factor of the bulk silicate Earth from metal-silicate experiments on rare Earth elements

This study investigates the partitioning of rare earth elements (REE) from La to Gd between molten metal and silicate to evaluate potential fractionation occurring during core-mantle differentiation. We report molten metal-silicate liquid partition coefficients from 24 multi-anvil experiments, extending the range of pressure, previously ranging from 1 to 8 GPa, up to 14 GPa. Experiments were performed at temperatures of between 2300 and 2560 K, and for oxygen fugacities ranging from the IW (Iron-Wüstite buffer) to IW–4. Metal-silicate partition coefficients for the studied REE vary with the oxygen fugacity and S concentration in the metallic phase of the system. These elements were all lithophile during the Earth's accretion. By compiling all existing data on molten metal-silicate liquid partitioning, REE partitioning between the mantle and core during the Earth's accretion can be determined for a wide range of P, T and fo2 conditions representing the early evolution of planetary bodies from planetesimals to planets. REE concentrations of the bulk silicate Earth (BSE) are calculated from accretion scenarios using varying proportions and compositions of chondritic building blocks. The models selected are those that reproduce the Earth's nucleosynthetic isotope signature and the Ni/Co, Th/U and Nb/Ta ratios of the BSE. The BSE refractory element enrichment factor determined from REE data is equal to 2.88 (relative to CI chondrites). This calculation takes into account the depletion in volatile elements in the Earth compared to chondrites. This new estimate is in good agreement with previous determinations based on analysis of the upper mantle rocks, which supports the idea of a chemically homogeneous mantle. We also confirm that the formation of the core, with or without segregation of a sulfide phase, does not fractionate Sm/Nd and cannot be responsible for the 142Nd excess measured in modern terrestrial samples relative to chondrites.

Uranium and thorium partitioning in the bulk silicate Earth and the oxygen content of Earth’s core

This study investigates the partitioning of U and Th between molten metal and silicate liquid (DU and DTh) during Earth’s core-mantle differentiation. We report new Th partition coefficients between chondritic silicate melt and various Fe-rich alloys in the system Fe-Ni-C-S-Si as determined by experiments in a multi-anvil apparatus at 3–8 GPa, 2073–2373 K, and oxygen fugacities from 1.5 to 5 log units below the iron-wüstite (IW) buffer. By compiling all existing data on molten metal-silicate liquid partitioning of U and Th, we develop global models of U and Th partitioning between the mantle and core throughout Earth’s accretion. The calculated concentrations in the Bulk Silicate Earth (BSE) are in agreement with previous studies (UBSE = 11.42 ± 0.45 ppb and ThBSE = 43.20 ± 1.73 ppb), whereas the contents of these radioactive elements in the Earth’s core remain negligible. Compared to recent geochemical estimations, the calculated (Th/U)BSE supports EL rather than EH enstatite chondrites as the reduced building blocks of the Earth. Furthermore, we demonstrate that Th is much more sensitive than U to the oxygen content of the metallic phase. To reproduce the Th/U ratio of the BSE within its uncertainties, the oxygen content of the Earth’s core must be lower than 4.0 wt%. By combining other existing constraints, this suggests that the core contains 2.0–4.0 wt% O. The calculations of U and Th concentrations and Th/U in the BSE developed herein can be used as new constraints for determining the concentrations of other refractory lithophile elements in the BSE as soon as their metal-silicate partition coefficients are well constrained over the conditions of core segregation.

Geochemical constraints on core-mantle differentiation in Mercury and the aubrite parent body

Differentiation of Mercury and the aubrite parent body (AuPB) at extremely reducing conditions is implied from the low FeO and high S contents of their mantles. Here, the mantle abundances of these elements as derived from remote sensing (in the case of Mercury) and aubrite meteorite analysis are used in conjunction with new experimentally determined metal-silicate partition coefficients at reducing conditions presented in Steenstra et al. (2020a) to quantify the geochemical consequences of core formation in highly reduced planetary bodies. Plausible core compositions of Mercury and the AuPB are assessed, and the distribution of Si and other elements during core formation are quantified.Combining the new results with previous remote sensing observations of the surface of Mercury it is found that its core is likely to be Si-rich (>4.5–21 wt%). The estimated core Si contents depend mostly on the assumption of C saturation in Mercury's mantle and the type of bulk composition considered. Calculations for C-free conditions also yield significant quantities of Si in the AuPB core (>6–20 wt%, depending on bulk composition and redox state). The amount of Si in the AuPB core is greatly decreased if C-saturation is assumed.The new experimental data and related thermodynamic parameterization for predicting the C contents at graphite saturation of FeSi alloys (CCGS) shows that Mercury's mantle was likely C-saturated following formation of a Si-rich core, allowing for the separation of a graphitic flotation crust. Due to the lower Si contents calculated for the AuPB core under C-saturated conditions, a graphitic flotation crust is unlikely to form in smaller-sized reduced asteroids such as the AuPB.The Si-rich nature of the AuPB core for C-undersaturated scenarios is expected to have resulted in preferential partitioning of the majority of the volatile siderophile elements (VSE) into the AuPB core. However, depletions for the most volatile VSE cannot be reconciled with core formation depletion only. Their depletions in aubrites require additional depletion from (I) segregating sulfide liquids during AuPB differentiation and/or (II) compatible behavior during mineral-melt fractionation and/or (III) loss of these elements in a vapor phase during and/or after AuPB differentiation.

The effect of melt composition on metal-silicate partitioning of siderophile elements and constraints on core formation in the angrite parent body

We present 275 new metal-silicate partition coefficients for P, S, V, Cr, Mn, Co, Ni, Ge, Mo, and W obtained at moderate P (1.5GPa) and high T (1683–1883K). We investigate the effect of silicate melt composition using four end member silicate melt compositions. We identify possible silicate melt dependencies of the metal-silicate partitioning of lower valence elements Ni, Ge and V, elements that are usually assumed to remain unaffected by changes in silicate melt composition. Results for the other elements are consistent with the dependence of their metal-silicate partition coefficients on the individual major oxide components of the silicate melt composition suggested by recently reported parameterizations and theoretical considerations. Using multiple linear regression, we parameterize compiled metal-silicate partitioning results including our new data and report revised expressions that predict their metal-silicate partitioning behavior as a function of P-T-X-fO2. We apply these results to constrain the conditions that prevailed during core formation in the angrite parent body (APB). Our results suggest the siderophile element depletions in angrite meteorites are consistent with a CV bulk composition and constrain APB core formation to have occurred at mildly reducing conditions of 1.4±0.5 log units below the iron-wüstite buffer (ΔIW), corresponding to a APB core mass of 18±11%. The core mass range is constrained to 21±8mass% if light elements (S and/or C) are assumed to reside in the APB core. Incorporation of light elements in the APB core does not yield significantly different redox states for APB core-mantle differentiation. The inferred redox state is in excellent agreement with independent fO2 estimates recorded by pyroxene and olivine in angrites.

Chalcophile elements in Martian meteorites indicate low sulfur content in the Martian interior and a volatile element-depleted late veneer

It is generally believed that the Martian mantle and core are rich in sulfur and that shergottites originated from sulfide-saturated magma. However, recent work suggests that the high FeO contents would require very high S concentrations in shergottite parent magmas at sulfide saturation. Here we combine new and published data on chalcophile elements in shergottites, nakhlites and ALH84001 to constrain the sulfide saturation state of the parent magmas and the chalcophile element concentrations in their mantle sources.Regardless of the MgO content and the long-term depletion history of incompatible lithophile elements as indicated by initial ε143Nd, different groups of shergottites display limited variations in ratios of Pt, Pd, Re, Cu, S, Se and Te. The emplacement of most shergottites within the crust and limited variations of ratios of chalcophile elements with substantial differences in volatility during eruption (e.g., Cu/S, Cu/Se and Pt/Re) indicate little degassing losses of S, Se, Te and Re from shergottites. Limited variations in ratios of elements with very different sulfide–silicate melt partition coefficients and negative correlations of chalcophile elements with MgO require a sulfide-undersaturated evolution of the parent magmas from mantle source to emplacement in the crust, consistent with the FeO-based argument. Sulfide petrography and the komatiite-like fractionation of platinum group elements (PGE) in shergottites also support this conclusion. The absence of accumulated sulfides in the ancient Martian cumulate ALH84001 results in very low contents of PGE, Re, Cu, Se and Te in this meteorite, hinting that sulfide-undersaturated magmas may have occurred throughout the Martian geological history. The negative correlation of Cu and MgO contents in shergottites suggests approximately 2±0.4(1s) μg/g Cu in the Martian mantle. The ratios of Cu, S, Se and Te indicate 360±120 μg/g (1s) S, 100±27 ng/g (1s) Se and 0.50±0.25 ng/g (1s) Te in the Martian mantle. At such low S concentrations, all S in Martian mantle sources may dissolve in basaltic melts that form at >5 % partial melting.Assuming equilibrium metal–silicate partitioning, and provided that the compositional model of the Martian mantle based on SNC meteorites is correct, Martian mantle inventories of Cu, S and Se were mostly established by core formation and the Martian core should contain <5–10 wt.% S only (depending on the choice of metal–silicate partition coefficients). The low S content in the Martian interior is consistent with the low Zn content in the Martian mantle, which indicates about 5 wt.% S in the core. In contrast, the highly siderophile PGE, Re and Te were added to the mantle by late accreted material after the Martian core formed. The near chondritic PGE ratios and the very low ratio of volatile Te to refractory PGE reflect a strongly volatile element-depleted late veneer and imply that the delivery of Martian water, presumably from carbonaceous chondrite like materials, must have occurred before accretion of the late veneer, likely within 2–3 million years after formation of the solar system.

Distribution of Sb, As, Ge, and In between metal and silicate during accretion and core formation in the Earth

A large number of siderophile (iron-loving) elements are also volatile, thus offering constraints on the origin of volatile elements in differentiated bodies such as Earth, Moon, Mars and Vesta. Metal–silicate partitioning data for many of these elements is lacking, making their overall mantle concentrations in these bodies difficult to model and origin difficult to distinguish between core formation and volatile depletion. To address this gap in understanding, we have undertaken systematic studies of four volatile siderophile elements – Sb, As, Ge and In – at variable temperature and variable Si content of metal. Several series were carried out at 1GPa, and between 1500 and 1900°C, for both C saturated and C-free conditions. The results show that temperature causes a decrease in the metal/silicate partition coefficient for all four elements. In addition, activity coefficients for each element have been determined and show a very strong dependence on Si content of Fe alloy. Si dissolved in metal significantly decreases the metal/silicate partition coefficients, at both 1600 and 1800°C. The combination of temperature and Si content of the metal causes reduction of the metal–silicate partition coefficient to values that are close to those required for an origin of mantle As, Sb, Ge, and In concentrations by metal–silicate equilibrium processes. Combining these new results with previous studies on As, Sb, Ge, and In, allowed derivation of predictive expressions for metal/silicate partition coefficients for these elements which can then be applied to Earth. The expressions are applied to two scenarios for continuous accretion of Earth; specifically for constant and increasing fO2 during accretion. The results indicate that mantle concentrations of As, Sb, Ge, and In can be explained by metal–silicate equilibrium during an accretion scenario. The modeling is not especially sensitive to either scenario, although all element concentrations are explained better by a model with variable fO2. The specific effect of Si is important and calculations that include only S and C (and no Si) cannot reproduce the mantle As, Sb, Ge, and In concentrations. The final core composition in the variable fO2 model is 10.2% Si, 2% S, and 1.1% C (or XSi=0.18, XS=0.03, and XC=0.04. These results suggest that core formation (involving a Si, S, and C-bearing metallic liquid) and accretion were the most important processes establishing many of Earth’s mantle volatile elements (indigenous), while post-core formation addition or re-equilibration (exogenous) was of secondary or minor importance.

Effect of H2O on metal–silicate partitioning of Ni, Co, V, Cr, Mn and Fe: Implications for the oxidation state of the Earth and Mars

This study investigates the metal–silicate partitioning of Ni, Co, V, Cr, Mn and Fe during core mantle differentiation of terrestrial planets under hydrous conditions. For this, we equilibrated a molten hydrous CI chondrite model composition with various Fe-rich alloys in the system Fe–C–Ni–Co–Si–S in a multi-anvil over a range of P, T, fO2 and water content (5–20GPa, 2073–2500K, from 1 to 5 log units below the iron–wüstite (IW) buffer and for XH2O varying from 500ppm to 1.5wt%). By comparing the present experiments with the available data sets on dry systems, we observes that the effect of water on the partition coefficients of moderately siderophile elements is only moderate. For example, for iron we observed a decrease in the partition coefficient of Fe (Dmet/silFe) from 9.5 to 4.3, with increasing water content of the silicate melt, from 0 to 1.44wt%, respectively. The evolution of metal–silicate partition coefficients of Ni, Co, V, Cr, Mn and Fe are modelled based on sets of empirical parameters. These empirical models are then used to refine the process of core segregation during accretion of Mars and the Earth. It appears that the likely presence of 3.5wt% water on Mars during the core–mantle segregation could account for ∼74% of the FeO content of the Martian mantle. In contrast, water does not play such an important role for the Earth; only 4–6% of the FeO content of its mantle could be due to the water-induced Fe-oxidation, for a likely initial water concentration of 1.8wt%. Thus, in order to reproduce the present-day FeO content of 8wt% in the mantle, the Earth could initially have been accreted from a large fraction (between 85% and 90%) of reducing bodies (similar to EH chondrites), with 10–15% of the Earth’s mass likely made of more oxidized components that introduced the major part of water and FeO to the Earth. This high proportion of enstatite chondrites in the original constitution of the Earth is consistent with the 17O,48Ca,50Ti,62Ni and 90Mo isotopic study by Dauphas et al. (2014). If we assume that the CI-chondrite was oxidized during accretion, its intrinsically high water content suggests a maximum initial water concentration in the range of 1.2–1.8wt% for the Earth, and 2.5–3.5wt% for Mars.

New geochemical models of core formation in the Moon from metal–silicate partitioning of 15 siderophile elements

We re-examine the conditions at which core formation in the Moon may have occurred by linking the observed lunar mantle depletions of 15 siderophile elements, including volatile siderophile elements (VSE) to predictive equations derived from a database compilation of metal–silicate partition coefficients obtained at lunar-relevant pressure–temperature–oxygen fugacity (P–T–fO2P–T–fO2) conditions. Our results suggest that at mantle temperatures between the solidus and liquidus the depletions for all elements considered can be satisfied, but only if the Moon was essentially fully molten at the time of core formation while assuming a S-rich (>8 wt%) core comprising 2.5 wt% of the mass of the Moon. However, we observe that at temperatures exceeding the mantle liquidus, with increasing temperature the core S content required to satisfy the element depletions is reduced. As a S-poor core is likely from recent lunar mantle estimates of S abundance, this suggests much higher temperatures during lunar core formation than previously proposed. We conclude that the VSE depletions in the lunar mantle can be solely explained by core formation depletion, suggesting that no significant devolatilization has occurred in later periods of lunar evolution. This is in agreement with the discovery of significant amounts of other volatiles in the lunar interior, but hard to reconcile with current lunar formation models.

Highly Siderophile Elements in Earth, Mars, the Moon, and Asteroids.

The highly siderophile elements (HSE: Os, Ir, Ru, Rh, Pt, Pd, Re, Au) are key tracers of planetary accretion and differentiation processes due to their affinity for metal relative to silicate. Under low-pressure conditions the HSE are defined by having metal–silicate partition coefficients in excess of 104 (e.g., Kimura et al. 1974; Jones and Drake 1986; O’Neill et al. 1995; Borisov and Palme 1997; Mann et al. 2012). The HSE are geochemically distinct in that, with the exception of Au, they have elevated melting points relative to iron (1665 K), low vapour pressures, and are resistant to corrosion or oxidation. Under solar nebular conditions, Re, Os, Ir, Ru, Rh, and Pt, along with the moderately siderophile elements (MSE) Mo and W, condense as refractory-metal alloys. Palladium and Au are not as refractory and condense in solid solution with FeNi metal (Palme 2008). Assuming abundances of the HSE in materials that made up the bulk Earth were broadly similar to modern chondrite meteorites, mass balance calculations suggest that >98% of these elements reside in the metallic core (O’Neill and Palme 1998). In practical terms, the resultant low HSE abundance inventories in differentiated silicate crusts and mantles enables the use of these elements in order to effectively track metallic core formation and the subsequent additions of HSE-rich impactors to planets and asteroids (Fig. 1). In detail, the absolute and relative abundances of the HSE in planetary materials are also affected by mantle and crustal processes including melting, metasomatism, fractional crystallization, and crust-mantle remixing, as well as later impact processing, volatility of Re under oxidizing conditions, and low-temperature secondary alteration (cf., Day 2013; Gannoun et al. 2016, this volume). In the absence of metal, the HSE are chalcophile, so these elements are also affected by processes …

Accretion and core formation: The effects of sulfur on metal–silicate partition coefficients

The accretion of the Earth was marked by the high-pressure segregation of the core, accompanied by dissolution of about 10% of one or more “light” elements into the metal. Cosmochemical data suggest that, of these 10% “light” elements, the core contains ∼1.7% S (Dreibus and Palme, 1996) and there is evidence that volatile elements such as S accreted to the Earth late in planetary growth, plausibly as a sulfide “matte” (O’Neill, 1991). Given that metallurgical data indicate that dissolution of even small amounts of sulfur in liquid Fe can have profound effects on the activities of some trace components, we have undertaken a study of the affect of S on the metal–silicate partitioning of a number of the most important chalcophile and siderophile elements.We performed experiments at 1.5GPa and 1460–1650°C on metal–silicate partitioning of W, Mo, Ni, Co, Cu, Ag, Mn, Cd, Tl, Cr, Sb, In, Pb, Ga, Ge, V and Zn under conditions where the S content of the metal was varied from 0 to 37wt%. Mn and Ag were found to exhibit the highest ratio of chalcophile to siderophile behaviour while W, Ga and Sb are the most “chalcophobic” of the elements studied. In terms of the 1-parameter epsilon model (Wagner, 1962) we derived values for each element at 1600°C as follows (negative values indicate chalcophile behaviour): εCuS=-2.57;εMnS=-6.41;εAgS=-4.15;εSbS=4.36;εCdS=-3.78;εInS=-0.24;εTlS=-6.21;εGaS=6.54;εPbS=-3.73;εCrS=-3.70;εGeS=7.03;εVS=-3.14;εZnS=-1.68;εMoS=2.27;εWS=6.47;εNiS=2.17;εCoS=2.40.We use these new data in conjunction with published pressure–temperature dependences of metal–silicate partitioning to test the effects of accreted S on the calculated trace element concentrations in bulk silicate Earth. The approach employs a continuous accretion model in which the oxidation state of the Earth and pressure of core segregation both increase during accretion. We find that, without addition of S in the latter stages of accretion, the Mo/W ratio of silicate Earth would be several times larger than that observed. Addition of ∼2% S accompanied by small amounts of carbon in the last 15% of accretion, however, enables us to match the observed concentrations of these elements in silicate Earth. This confirms an earlier conclusion that the Mo/W ratio of silicate Earth requires late sulfide addition to the core (Wade et al., 2012).Further support for late sulfide addition to the core comes from the depletion factors of volatile chalcophile elements Cu, Ga, Sb, Ag, Zn, Pb, Cd, In and Tl in silicate Earth relative to lithophile elements of similar volatility. We find that depletions of these elements are well correlated with their partition coefficients into sulfide (FeS) liquids at 1.5GPa and temperatures of 1460–1650°C. In contrast there is essentially no correlation between their depletion factors and sulfur-free liquid Fe–silicate partition coefficients.

Unlocking the zinc isotope systematics of iron meteorites

Zinc isotope compositions (δ66Zn) and concentrations were determined for metal samples of 15 iron meteorites across groups IAB, IIAB, and IIIAB. Also analyzed were troilite and other inclusions from the IAB iron Toluca. Furthermore, the first Zn isotope data are presented for metal–silicate partitioning experiments that were conducted at 1.5 GPa and 1650 K. Three partitioning experiments with run durations of between 10 and 60 min provide consistent Zn metal–silicate partition coefficients of ∼0.7 and indicate that Zn isotope fractionation between molten metal and silicate is either small (at less than about ±0.2‰) or absent. Metals from the different iron meteorite groups display distinct ranges in Zn contents, with concentrations of 0.08–0.24 μg/g for IIABs, 0.8–2.5 μg/g for IIIABs, and 12–40 μg/g for IABs. In contrast, all three groups show a similar range of δ66Zn values (reported relative to ‘JMC Lyon Zn’) from +0.5‰ to +3.0‰, with no clear systematic differences between groups. However, distinct linear trends are defined by samples from each group in plots of δ66Zn vs. 1/Zn, and these correlations are supported by literature data. Based on the high Zn concentration and δ66Zn ≈ 0 determined for a chromite-rich inclusion of Toluca, modeling is employed to demonstrate that the Zn trends are best explained by segregation of chromite from the metal phase. This process can account for the observed Zn–δ66Zn–Cr systematics of iron meteorite metals, if Zn is highly compatible in chromite and Zn partitioning is accompanied by isotope fractionation with Δ66Znchr-met≈−1.5‰. Based on these findings, it is likely that the parent bodies of the IAB complex, IIAB and IIIAB iron meteorites featured δ66Zn values of about −1.0 to +0.5‰, similar to the Zn isotope composition inferred for the bulk silicate Earth and results obtained for chondritic meteorites. Together, this implies that most solar system bodies formed with similar bulk Zn isotope compositions despite large differences in Zn contents.

The solubility of platinum in silicate melt under reducing conditions: Results from experiments without metal inclusions

The solubility of Pt in silicate melt was investigated at conditions of 2073–2573K, 2GPa and ∼IW −1.5 to +3.5. These are the first measurements of Pt solubility under conditions more reducing than the iron-wüstite buffer (IW) which are demonstrably free from contamination by metal-inclusions. Pt solubility increases with increasing temperature and decreasing oxygen fugacity. The ability of carbon to enhance Pt solubility under reducing conditions (<IW +2) was tested by performing a set of isobaric, isothermal experiments which utilised graphite and PtIr alloy capsules. Analyses of the run-product glasses from these experiments show no correlation between dissolved C and Pt contents. In light of this result, the dissolution of Pt as a neutral species is suggested as being responsible for the observed relationship between solubility and fO2. A compilation of data from this study, Ertel et al. (2006) and Mann et al. (2012), suggests that pressure has no significant effect on solubility between 2 and 18GPa. Metal–silicate partition coefficients calculated from our Pt solubility data show that, for core–mantle equilibrium at IW −2, Pt concentrations in the primitive upper mantle (PUM) can be satisfied if the temperature of equilibration is >3500K. Under these conditions however, the estimated Pt/Os ratio is ∼40,000 times higher than that estimated for the PUM (Brandon et al., 2006). Instead, the PUM composition is generated most readily by metal–silicate equilibrium at more modest temperatures (∼3100K), followed by a late accretion of chondritic material subsequent to core formation.

Non-chondritic sulphur isotope composition of the terrestrial mantle

Core-mantle differentiation is the largest event experienced by a growing planet during its early history. Terrestrial core segregation imprinted the residual mantle composition by scavenging siderophile (iron-loving) elements such as tungsten, cobalt and sulphur. Cosmochemical constraints suggest that about 97% of Earth's sulphur should at present reside in the core, which implies that the residual silicate mantle should exhibit fractionated (34)S/(32)S ratios according to the relevant metal-silicate partition coefficients, together with fractionated siderophile element abundances. However, Earth's mantle has long been thought to be both homogeneous and chondritic for (34)S/(32)S, similar to Canyon Diablo troilite, as it is for most siderophile elements. This belief was consistent with a mantle sulphur budget dominated by late-accreted chondritic components. Here we show that the mantle, as sampled by mid-ocean ridge basalts from the south Atlantic ridge, displays heterogeneous (34)S/(32)S ratios, directly correlated to the strontium and neodymium isotope ratios (87)Sr/(86)Sr and (143)Nd/(144)Nd. These isotope trends are compatible with binary mixing between a low-(34)S/(32)S ambient mantle and a high-(34)S/(32)S recycled component that we infer to be subducted sediments. The depleted end-member is characterized by a significantly negative δ(34)S of -1.28 ± 0.33‰ that cannot reach a chondritic value even when surface sulphur (from continents, altered oceanic crust, sediments and oceans) is added. Such a non-chondritic (34)S/(32)S ratio for the silicate Earth could be accounted for by a core-mantle differentiation record in which the core has a (34)S/(32)S ratio slightly higher than that of chondrites (δ(34)S = +0.07‰). Despite evidence for late-veneer addition of siderophile elements (and therefore sulphur) after core formation, our results imply that the mantle sulphur budget retains fingerprints of core-mantle differentiation.

Metal–silicate partitioning of Pb and U: Effects of metal composition and oxygen fugacity

This study investigates the metal–silicate partitioning of Pb and U (DPb and DU, respectively) during core-mantle differentiation of terrestrial planets. For this, we equilibrated a molten CI-chondrite model mantle-composition with various Fe-rich alloys in the system Fe–C–Ni–Si–S in a multi-anvil over a range of P, T, fO2 (3–8GPa, 2073–2373K and from 1.5 to 5 log units below the iron-wüstite (IW) buffer). The chemical compositions of the run products and contents in Pb and U trace elements were determined using electron microbe and laser ablation inductively coupled plasma-mass spectrometry, respectively. We combined our new data set with pre-existing results in similar chemical systems to refine the major parameters controlling the DPb and DU values, based on thermodynamical calculations. We show: (i) a significant increase of the metal–silicate partition coefficients, from ∼1 to ∼50 for DPb and from ∼10−5 to 10−2 for DU with the addition of sulfur. (ii) At a fixed fO2, the presence of C in the metal favors U and Pb partitioning toward the silicate melt, when S produces the opposite trend. (iii) Temperature plays a non-negligible role in DPb value. Although the results suggest that the Pb depletion observed in the present-day bulk silicate Earth can be explained by core-mantle segregation at high-pressure in a deep-magma ocean, at conditions similar to that previously refined based on the behavior of many siderophile elements, such as Ni and Co, we will also discuss the possibility of late S-rich phases segregating to the core.