Search

Highly siderophile elements (HSE = Au, Re, and the Pt-group elements) are tracers of silicate / metal interactions during planetary processes. Since most core-formation models involve some state of equilibrium between liquid silicate and liquid metal, understanding the partioning of highly siderophile elements (HSE) between silicate and metallic melts is a key issue for models of core / mantle equilibria and for core formation scenarios. However, partitioning models for HSE are still inaccurate due to the lack of sufficient experimental constraints to describe the variations of partitioning with key variable like temperature, pressure, and oxygen fugacity. In this abstract, we describe a self-consistent set of experiments aimed at determining the valence of platinum, one of the HSE, in silicate melts. This is a key information required to parameterize the evolution of platinum partitioning with oxygen fugacity.

The variation of oxygen fugacity within inner solar system materials spans a range of nearly 15 orders of magnitude. Igneous and metamorphic rocks commonly contain a mineral assemblage that allows oxygen fugacity to be calculated or con-strained such as FeTi oxides, olivine-opx-spinel, or some other oxy-barometer. Some rocks, however, contain a limited mineral assemblage and do not provide constraints on fO2 using mineral equilibria. Good examples of the latter are orthopyroxenites or dunites, such as diogenites, ALH 84001, chassignites, or brachinites. In fact it is no surprise that the fO2 of many of these achondrites is not well known, other than being "reduced" and below the metal saturation value. In order to bridge this gap in our understanding, we have initiated a study of V in chromites in achondrite. Because the V pre-edge peak intensity and energy in chromites varies with fO2, and this has been calibrated over a large fO2 range, we can apply this relation to rocks for which we otherwise have no fO2 constraints.

Iron redox systematics of the high FeO shergottitic liquids are poorly known, yet have a fundamental control on stability of phases such as magnetite, ilmenite, and pyroxenes.

Martian magmas are known to be FeO-rich and the dominant FeO-bearing mineral at many sites visited by the Mars Exploration rovers (MER) is magnetite. Morris et al. proposed that the magnetite appears to be igneous in origin, rather than of secondary origin. However, magnetite is not typically found in experimental studies of martian magmatic rocks. Magnetite stability in terrestrial magmas is well understood, as are the stabilities of FeO and Fe2O3 in terrestrial magmas. In order to better understand the variation of FeO and Fe2O3, and the stability of magnetite (and other FeO-bearing phases) in martian magmas, we have undertaken an experimental study with two emphases. First, we determine the FeO and Fe2O3 contents of super- and sub-liquidus glasses from a shergottite bulk composition at 1 bar to 4 GPa, and variable fO2. Second, we document the stability of magnetite with temperature and fO2 in a shergottite bulk composition.

Martian magmas are known to be FeO-rich and the dominant FeO-bearing mineral at many sites visited by the Mars Exploration rovers (MER) is magnetite [1]. Morris et al. [1] propose that the magnetite appears to be igneous in origin, rather than of secondary origin. However, magnetite is not typically found in experimental studies of martian magmatic rocks [2,3]. Magnetite stability in terrestrial magmas is well understood, as are the stability of FeO and Fe2O3 in terrestrial magmas [4,5]. In order to better understand the variation of FeO and Fe2O3, and the stability of magnetite (and other FeO-bearing phases) in martian magmas we have undertaken an experimental study with two emphases. First we document the stability of magnetite with temperature and fO2 in a shergottite bulk composition. Second, we determine the FeO and Fe2O3 contents of the same shergottite bulk composition at 1 bar and variable fO2 at 1250 C, and at variable pressure. These two goals will help define not only magnetite stability, but pyroxene-melt equilibria that are also dependent upon fO2.

The metal-silicate partitioning behavior of molybdenum has been used as a test for equilibrium core formation hypotheses [for example, 1-6]. However, current models that apply experimental data to equilibrium core-mantle differentiation infer the oxidation state of molybdenum from solubility data or from multivariable coefficients from metal-silicate partitioning data [1,3,7]. Molybdenum, a multi-valent element with a valence transition near the fO2 of interest for core formation (approx.IW-2) will be sensitive to changes in fO2 of the system and silicate melt structure. In a silicate melt, Mo can occur in either 4+ or 6+ valence state, and Mo(6+) can be either octahedrally or tetrahedrally coordinated. Here we present X-ray absorption near edge structure (XANES) measurements of Mo valence in basaltic run products at a range of P, T, and fO2 and further quantify the valence transition of Mo.

Martian meteorites record a range of oxygen fugacities from near the IW buffer to above FMQ buffer [1]. In terrestrial magmas, Fe(3+)/ SigmaFe for this fO2 range are between 0 and 0.25 [2]. Such variation will affect the stability of oxides, pyroxenes, and how the melt equilibrates with volatile species. An understanding of the variation of Fe(3+)/SigmaFe for martian magmas is lacking, and previous work has been on FeO-poor and Al2O3-rich terrestrial basalts. We have initiated a study of the iron redox systematics of martian magmas to better understand FeO and Fe2O3 stability, the stability of magnetite, and the low Ca/high Ca pyroxene [3] ratios observed at the surface.

The moderately siderophile element molybdenum has been used as an indicator in planetary differentiation processes, and is particularly relevant to core formation [for example, 1-6]. However, models that apply experimental data to an equilibrium differentiation scenario infer the oxidation state of molybdenum from solubility data or from multivariable coefficients from metal-silicate partitioning data [1,3,7]. Partitioning behavior of molybdenum, a multivalent element with a transition near the J02 of interest for core formation (~IW-2) will be sensitive to changes in JO2 of the system and silicate melt structure. In a silicate melt, Mo can occur in either 4+ or 6+ valence state, and Mo6+ can be either octahedrally or tetrahedrally coordinated. Here we present first XANES measurements of Mo valence in basaltic run products at a range of P, T, and JO2 and further quantify the valence transition of Mo.

Igneous and metamorphic rocks commonly contain a mineral assemblage that allows oxygen fugacity to be calculated or constrained such as FeTi oxides, olivine-opx-spinel, or some other oxybarometer [1]. Some rocks, however, contain a limited mineral assemblage and do not provide constraints on fO2 using mineral equilibria. Good examples of the latter are orthopyroxenites or dunites, such as diogenites, ALH 84001, chassignites, or brachinites. In fact it is no surprise that the fO2 of many of these samples is not well known, other than being "reduced" and below the metal saturation value. In order to bridge this gap in our understanding, we have initiated a study of V in chromites in natural meteorite samples. Because the V pre-edge peak intensity and energy in chromites varies with fO2 (Fig. 1) [2], and this has been calibrated over a large fO 2 range, we can apply this relation to rocks for which we otherwise have no fO2 constraints.

A series of experiments with garnet and coexisting melt have been carried out across a range of oxygen fugacities (near hematite-magnetite (HM) to below the iron-wustite (IW) buffers) at 1.7 GPa to study the partitioning and valence of Cr and V in both phases. Experiments were carried out in a non end loaded piston cylinder apparatus, and the run products were analyzed with electron microprobe and xray absorption near edge structure (XANES) analysis at beamline 13-ID at the Advanced Photon Source of Argonne National Lab. The valence of vanadium and chromium were determined using the position and intensity of the Ka pre-edge peaks, calibrated on a series of Cr and Vbearing standard glasses. This technique has been applied to V and Cr in glasses and V in spinels previously, and in these isotropic phases there are no orientational effects on the XANES spectra (Righter et al., 2006, Amer. Mineral. 91, 1643-1656). We also now demonstrate this to be true for V and Cr in garnet. Also, previous work has shown that V has a higher valence in the glass (or melt) than in the coexisting spinel. This is also true for V in garnet-glass pairs in this study. Vanadium valence in garnets varies from 2.7 below the IW buffer to 3.7 near HM, and for coexisting glass it varies from 3.2 to 4.3. Vanadium valence measured in some natural garnets from mantle localities indicates V in the more reduced range at 2.5. Comparisons will be made between fO2 estimated from V valence and other methods for garnet-bearing mantle samples. In contrast, Cr valence measured in garnet and coexisting glass for all experimental and natural samples is 2.9- 3.0, suggesting that the valence of Cr does not vary within either phase across a large fO2 range. These results demonstrate that while V varies from 2+ to 3+ to 4+ in garnet-melt systems, Cr does not, and this will ultimately affect the partitioning behavior of these two elements in natural systems. Garnet/melt D(Cr) are between 12 and 17 across this range of fO2, whereas D(V) has the highest partition coefficient approx.3, near the IW buffer where the valence of V is almost entirely 3+.

Tungsten is important in constraining core formation of the Earth because this element is a moderately siderophile element (depleted ~ 10 relative to chondrites) and, as a member of the Hf-W isotopic system, it is useful in constraining the timing of core formation. A number of previous experimental studies have been carried out to determine the silicate solubility and metal-silicate partitioning behavior of W, including its concomitant oxidation state. However, results of previous studies are inconsistent on whether W occurs as W(4+) or W(6+). It is assumed that W(4+) is the cation valence relevant to core formation. Given the sensitivity to silicate composition of high valence cations, knowledge of the oxidation state of W over a wide range of fO2 is critical to understanding the oxidation state of the mantle and core formation processes. This study seeks to measure the W valence and change in valence state over the range of fO2 most relevant to core formation, around IW-2.

Tungsten is important in constraining core formation of the Earth because this element is a moderately siderophile element (depleted approx. 10 relative to chondrites) and, as a member of the Hf-W isotopic system, it is useful in constraining the timing of core formation. A number of previous experimental studies have been carried out to determine the silicate solubility and metal-silicate partitioning behavior of W, including its concomitant oxidation state. However, results of previous studies (figure 1) are inconsistent on whether W occurs as W(4+) or W(6+).

Spinel can be a significant host phase for V which has multiple oxidation states V(sup 2+), V(sup 3+), V(sup 4+) or V(sup 5+) at oxygen fugacities relevant to natural systems. The magnitude of D(V) spinel/melt is known to be a function of composition, temperature and fO2, but the uncertainty of the oxidation state under the range of natural conditions has made elusive a thorough understanding of D(V) spinel/melt. For example, V(sup 3+) is likely to be stable in spinels, based on exchange with Al in experiments in the CaO-MgO-Al2O3-SiO2 system. On the other hand, it has been argued that V(sup 4+) will be stable across the range of natural oxygen fugacities in nature. In order to build on our previous work in more oxidized systems, we have carried out experiments at relatively reducing conditions from the FMQ buffer to 2 log fO2 units below the IW buffer. These spinel-melt pairs, where V is present in the spinel at natural levels (approx. 300 ppm V), were analyzed using an electron microprobe at NASA-JSC and micro- XANES at the Advanced Photon Source at Argonne National Laboratory. The new results will be used together with previous results to understand the valence of V in spinel-melt systems across 12 orders of magnitude of oxygen fugacity, and with application to natural systems.

We have been developing an oxygen barometer based on the valence state of V (V(2+), V(3+), V(4+), and V(5+)) in solar system basaltic glasses. The V valence is determined by synchrotron micro x-ray absorption near edge structure (XANES), which uses x-ray absorption associated with core-electronic transitions (absorption edges) to reveal a pre-edge peak whose intensity is directly proportional to the valence state of an element. XANES has advantages over other techniques that determine elemental valence because measurements can be made non-destructively in air and in situ on conventional thin sections at a micrometer spatial resolution with elemental sensitivities of approx. 100 ppm. Recent results show that fO2 values derived from the V valence technique are consistent with fO2 estimates determined by other techniques for materials that crystallized above the IW buffer. The fO2's determined by V valence (IW-3.8 to IW-2) for the lunar pyroclastic glasses, however, are on the order of 1 to 2.8 log units below previous estimates. Furthermore, the calculated fO2's decrease with increasing TiO2 contents from the A17 VLT to the A17 Orange glasses. In order to investigate these results further, we have synthesized lunar green and orange glasses and examined them by XANES.

Spinel can be a significant host phase for V as well as other transition metals such as Ni and Co. However, vanadium has multiple oxidation states V(2+), V(3+), V(4+) or V(5+) at oxygen fugacities relevant to natural systems. We do know that D(V) spinel/melt is correlated with V and TiO2 content and fO2, but the uncertainty of the oxidation state under the range of natural conditions has made elusive a thorough understanding of D(V) spinel/melt. For example, V(3+) is likely to be stable in spinels, based on exchange with Al in experiments in the CaO-MgO-Al2O3-SiO2 system. On the other hand, it has been argued that V(4+) will be stable across the range of natural oxygen fugacities in nature. In order to gain a better understanding of D(V) spinel/melt we have equilibrated spinel-melt pairs at controlled oxygen fugacities, between HM to NNO, where V is present in the spinel at natural levels (approx. 300 ppm V). These spinel-melt pairs were analyzed using micro-XANES at the Advanced Photon Source at Argonne National Laboratory. The new results will be used together with spinel compositional data (Ti, V content) and oxygen fugacity, to unravel the effects of these variables on D(V) spinel/melt.

The determination of oxidation conditions for basaltic magmas derived by the melting of planetary mantles is critical to our understanding of the nature and evolution of planetary interiors. Yet, these determinations are compromised in terrestrial and especially extraterrestrial basalts by our analytical and computational methods for estimating oxygen fugacity (fO2). For example, mineralogical barometers (1, 2) can be reduced in effectiveness by subsolidus re-equilibration of mineral assemblages, inversion of mineralogical data to melt characteristics, and deviations of the natural mineral compositions from ideal thermodynamic parameters.

The Cl/Br ratios in fracture-filling materials in veins in Nakhla olivine was determined using x-ray microprobe (Br) and EDX (Cl) techniques. The Cl/Br ratio of 55 (standard deviation: 13) shows that the secondary altered material is pristine, extraterrestrial and akin to the Martian soil. Additional information is contained in the original extended abstract.

Microanalysis of Fe-3+/Total(Fe) in extraterrestrial samples is important due to sample size constraints of sample return missions. We compare Fe XANES spectra with Co optical spectra that predict valence electron levels based on 'Z1' model. Additional information is contained in the original extended abstract.

Oxidation state microanalyses for V in glass have been made by calibrating XANES spectral features with optical spectroscopic measurements. The oxidation state change with fugacity of O2 will strongly influence partitioning results.