Information about the materials and conditions involved in planetary formation and differentiation in the early Solar System is recorded in iron isotope ratios. Samples from Earth, the Moon, Mars and the asteroid Vesta reveal significant variations in iron isotope ratios, but the sources of these variations remain uncertain. Here we present experiments that demonstrate that under the conditions of planetary core formation expected for the Moon, Mars and Vesta, iron isotopes fractionate between metal and silicate due to the presence of nickel, and enrich the bodies' mantles in isotopically light iron. However, the effect of nickel diminishes at higher temperatures: under conditions expected for Earth's core formation, we infer little fractionation of iron isotopes. From our experimental results and existing conceptual models of magma ocean crystallization and mantle partial melting, we find that nickel-induced fractionation can explain iron isotope variability found in planetary samples without invoking nebular or accretionary processes. We suggest that near-chondritic iron isotope ratios of basalts from Mars and Vesta, as well as the most primitive lunar basalts, were achieved by melting of isotopically light mantles, whereas the heavy iron isotope ratios of terrestrial ocean floor basalts are the result of melting of near-chondritic Earth mantle.

We apply an experimentally based thermodynamic model of Si+O saturation for the core to determine the saturation level of these elements under the conditions when the core formed. The model limits the bulk Si content of the core to between 0.4 and 3.1 wt% depending on the pressure, temperature, and oxygen content of the metal when it segregated from silicate. With knowledge of the core's Si content, the measured Si-30 content of the silicate Earth, and the experimentally determined metal-silicate fractionation factor, we can calculate the core's delta Si-30, which is between -0.92 to -1.36%. SiO2 cycled through the core and then released into the mantle might be trapped in inclusions in diamond formed in the lower mantle. These would be characterized by significantly lighter delta Si-30 values of -1.12 +/- 0.13 parts per thousand (1 sigma), compared to bulk silicate earth values of -0.29% and a potentially key indicator of mass transfer from the core to the mantle.

Iron isotopic compositions are demonstrably powerful tracers of foundational planetary processes, including crust and core formation. In many volcanic environments, however, geochemical vestiges of these processes are obscured by the effects of magmatic differentiation on Fe isotopic compositions. Recent decades have witnessed continued refinement of observational and experimental approaches to Fe isotope fractionation during silicate differentiation. In contrast, the influence of sulfide fractionation on Fe isotopic compositions in terrestrial environments is known only from theoretical approaches and limited experimental data for relatively siliceous magmatic systems. One reason for this may be that sulfide fractionation is difficult to definitively trace using traditional major and minor element variation patterns. We utilize well-characterized lavas and cumulate xenoliths from Piton de la Fournaise and Piton des Neiges, Reunion Island, that have previously been examined for their highly siderophile element (HSE: Os, Ir, Ru, Pt, Pd, Re) contents to investigate the effect of sulfide fractionation on Fe isotopes. The Fe isotopic compositions of the basalts range from delta Fe-56 values of 0.04 to 0.15 parts per thousand (average: 0.10 parts per thousand) and the compositions of the cumulate xenoliths range from delta Fe-56 values of -0.07 to 0.08 parts per thousand (average: 0 parts per thousand). In the absence of metal, HSE preferentially partition into sulfide phases, making them important tracers of sulfide segregation during magmatic differentiation. We find that commonly-observed co-variations between Fe isotopic compositions and major element oxide abundances are relatively underdeveloped for Reunion lavas. The correlation between Fe isotopic composition and MgO, for example, has a similar statistical significance to the correlation between Fe isotopic composition and Pd/Ir ratios, suggesting an important role of sulfides during Fe isotopic fractionation. After accounting for sulfide segregation, we determine that the parental magma Fe isotopic composition calculated for Piton de la Fournaise would be overestimated by 0.04 parts per thousand (within propagated error, 0.01-0.06 parts per thousand) when considering silicate differentiation alone. An analogous calculation for Kilauea Iki basalts, for which there is available Fe isotopic and HSE data, yields a somewhat smaller difference of 0.02 parts per thousand (0-0.03 parts per thousand). These differences may partially explain Fe isotopic compositions in other settings that could not previously be reconciled with a dominantly peridotitic and/or chondritic mantle source. This discovery may warrant discussion of the apparent decoupling between Fe and radiogenic isotopes in ocean island basalts, where the latter shows significant global variations and the former may show little or none. Our work highlights the need for additional constraints on the behavior of Fe isotopes during crustal recycling processes and reinforces the notion that consideration must be given to the effect of magmatic differentiation on Fe isotopic compositions. (C) 2018 Elsevier Ltd. All rights reserved.

Isotopic fractionation associated with diffusion in crystals is the most reliable means of understanding the origin of mineral zoning in igneous and metamorphic rocks. We have experimentally determined the relative diffusivities of iron isotopes in olivine as a function of crystallographic orientation, composition, and temperature. For two isotopes i and j of an element, the isotope effect for diffusion is parameterized as D-i/D-j = (m(j)/m(i))(beta), where beta is a dimensionless parameter, and D and m stand for diffusivity and mass, respectively. A series of single crystal diffusion couple experiments were conducted at an oxygen fugacity of QFM - 1.5 at temperatures of 1200, 1300, and 1400 degrees C. For the Fo(83.4)-Fo(88.8) composition pair, beta(Fe) is isotropic and a value of 0.16 +/- 0.09 can be used to describe diffusion along all major crystallographic axes in olivine. Based on our experiments and previously reported coupled Mg-Fe isotopic data, we also estimate beta(Mg) = 0.09 +/- 0.05 for this range of olivine composition. For the Fo(88.8)-Fo(100) composition pair, beta(Fe) becomes anisotropic with beta(Fe [100]) = 0.11 +/- 0.03, beta(Fe [010]) = 0.14 +/- 0.03 (both within error of the value measured for the Fo(83.4)-Fo(88.8) pair), and beta(Fe [001]) = 0.03 +/- 0.03. For Fo# between 83.4 and 100, beta(Fe [100]) and beta(Fe [010]) are thus independent of composition. The reason why beta(Fe) ([001]) transitions from similar to 0.16 to similar to 0.03 close to the Mg-endmember is unclear. Over the temperature range studied, a dependence of beta(Fe) on temperature was not resolved. General analytical expressions are introduced to calculate isotopic fractionation as a function of distance, time, beta, and the concentration contrast between the diffusing media for spherical, cylindrical, and planar geometries. (C) 2018 Elsevier Ltd. All rights reserved.

We have conducted high-pressure, high-temperature isotope exchange experiments between molten silicate and molten Fe-Si-C-alloys to constrain the effect of Si on equilibrium Fe isotope fractionation during planetary core formation. The values of Delta Fe-57(Metal-Silicate) at 1850 degrees C and 1 GPa determined by high-resolution MC-ICP-MS in this study range from -0.013 +/- 0.054 parts per thousand (2SE) to 0.072 +/- 0.085 parts per thousand with 1.34-8.14 atom % Si in the alloy, respectively. These results, although not definitive on their own, are consistent with previous experimental results from our group and a model in which elements that substitute for Fe atoms in the alloy structure (i.e., Ni, S, and Si) induce a fractionation of Fe isotopes between molten silicate and molten Fe-alloys during planetary differentiation. Using in situ synchrotron X-ray diffraction data for molten Fe-rich alloys from the literature, we propose a model to explain this fractionation behavior in which impurity elements in Fe-alloys cause the nearest neighbor atomic distances to shorten, thereby stiffening metallic bonds and increasing the preference of the alloy for heavy Fe isotopes relative to the silicate melt. This fractionation results in the bulk silicate mantles of the smaller terrestrial planets and asteroids becoming isotopically light relative to chondrites, with an enrichment of heavy Fe isotopes in their cores, consistent with magmatic iron meteorite compositions. Our model predicts a bulk silicate mantle delta Fe-57 ranging from -0.01 parts per thousand to -0.12 parts per thousand for the Moon, -0.06 parts per thousand to -0.33 parts per thousand for Mars, and -0.08 parts per thousand to B -0.33 parts per thousand for Vesta. Independent estimates of the delta Fe-57 of primitive mantle source regions that account for Fe isotope fractionation during partial melting agree well with these ranges for all three planetary bodies and suggest that Mars and Vesta have cores with impurity (i.e., Ni, S, Si) abundances near the low end of published ranges. Therefore, we favor a model in which core formation results in isotopically light bulk silicate mantles for the Moon, Mars, and Vesta. The processes of magma ocean crystallization, mantle partial melting, and fractional crystallization of mantle-derived melts are all likely to result in heavy Fe isotope enrichment in the melt phase, which can explain why basaltic samples from these planetary bodies have variable delta Fe-57 values consistently heavier than our bulk mantle estimates. Additionally, we find no clear evidence that Fe isotopes were fractionated to a detectable level by volatile depletion processes during or after planetary accretion, although it cannot be ruled out. (C) 2019 Elsevier B.V. All rights reserved.

Mercury, the Solar System's innermost planet, has an unusually massive core prompting speculation that the planet lost silicate after it formed. Using the unusually high sulfur and low iron composition of its surface and space geodetic constraints on its core composition, we show Mercury's chemistry to be compatible with formation in a larger planet at minimum 1.4-2.5 times Mercury's present mass and possibly 2-4 times its mass by similarity with other rocky Solar System bodies. To do this, we apply an experimentally determined metal-silicate partitioning model for sulfur to Mercury's silicate. The model is validated by applying it to Vesta, which, when evaluated at the conditions of Vestan self-differentiation, yields sulfur contents in its silicate in the range of HED meteorites. Mercury could have lost a substantial fraction of its rocky material through impacts or by being itself a remnant impactor. Independent of any stripping, because a significant amount of silicon resides in Mercury's core, silicate meteoritic debris from Mercury would likely be characterized by Si-30 isotopic enrichment >+ 0.10 parts per thousand relative to parent sources that could aid identification of a new meteorite class.

Laser-ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) is one of the most popular techniques for determining trace element concentrations in sulfides. Due to the lack of matrix-matched standards, standardization of sulfide analyses are usually based on silicate glass calibrant materials. Matrix effects during ns-LA-ICP-MS analyses of Fe-rich sulfides were quantified for many trace elements by comparison of elemental concentrations obtained by LA-ICP-MS and electron microprobe (EPMA) for many synthetic sulfides. The data was used to obtain the fractionation indices (F-i, the ratio between the EPMA- and LA-ICP-MS-determined concentrations of element i) for many elements while considering Fe, Cu and Ni as internal standards. The results show that significant (>15% RD) matrix effects arise during ns-LA-ICP-MS analyses of Ti, Zn, Ge, Se, Mo, Cd, In, Sb, Te, Pb, Bi in sulfides when using Fe as the internal standard. The use of Ni as an internal standard yields on average higher F-i values for most elements, resulting in more pronounced matrix effects for refractory elements and less so for volatile elements, relative to Fe. The use of Cu as an internal standard yields overall more significant matrix effects for volatile elements (i.e., lower F-i values). The F-i values for most elements remain constant with increasing concentrations, and matrix correction factors for these elements can therefore be applied across the ppm to wt% range. In agreement with previous observations for Fe-rich metals and silicate glasses, the magnitudes of the matrix effects for the various elements are strongly correlated with elemental volatility. This correlation was used to obtain a predictive model for describing F-i for Fe-rich sulfides. The results were used to assess the effects of matrix effects on calculated sulfide liquid-silicate melt partition coefficients derived from experiments. Matrix effects arising through the use of non-matrix-matched standards will result in significant discrepancies between measured and true partition coefficients, the extent mainly depending of the volatility of the element considered. Corrections on ns-LA-ICP-MS derived element concentrations therefore need to be performed to obtain true abundances in the absence of matrix-matched standards.