Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Partition coefficient”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 91 records · Page 5
All sourcesNASA NTRSDOE OSTI
NASA NTRS1992

Experimental determination of the solubility of iridium in silicate melts: Preliminary results

Little is known of the geochemical behavior of iridium. Normally this element is taken to be chalcophile and/or siderophile so that during planetary differentiation processes, e.g., core formation, iridium is extracted from silicate phases into metallic phases. Experimental determination of the metal/silicate partition coefficient of iridium is difficult simply because it is so large. Also there are no data on the solubility behavior of iridium in silicate melts. With information on the solubility of iridium in silicate melts it is possible, in combination with experimental data for Fe-Ir alloys, to calculate the partition coefficient between a metallic phase and a silicate melt.

Borisov, Alexander↗
NASA NTRS1980

Laboratory studies of actinide metal-silicate fractionation

Actinide and Sm partition coefficients between silicate melt and several metallic phases have been measured. Under reducing conditions Si, Th, U and Pu can be reduced to metals from silicate melts and alloyed with a platinum-gold alloy. U and Pu enter a molten Pt-Si alloy with roughly equal affinity but U strongly partitions into the solid Pt. Th behaves qualitatively the same as Pu but is much less readily reduced than U, and Sm appears to remain unreduced. Experiments with Fe metal have shown that the partition coefficients of the actinides between Fe and silicate liquid are extremely low, suggesting a very low actinide concentration in planetary cores. Experiments show that platinum metals can efficiently fractionate actinides and fractionate actinides from lanthanides and this process may be relevant to the condensation behavior of these elements from the solar nebula. Pt-metal grains in Allende Ca-Al-rich inclusions appear to be U-poor, although the sub-class of Zr-bearing Pt metals may have high U contents.

Jones, J. H.↗
NASA NTRS1992

Experimental determination of U and Th partitioning between clinopyroxene and natural and synthetic basaltic liquid

Experimental measurements of U and the partition coefficients between clinopyroxene and synthetic and natural basaltic liquid are presented. The results demonstrate that crystal-liquid U-Th fractionation is fO2-dependent and that U in terrestrial magmas is not entirely tetravalent. During partial melting, the liquid will have a Th/U ratio less than the clinopyroxene in the source. The observed U-238 - Th-230 disequilibrium in MORB requires that the partial melt should have a U/Th ratio greater than the bulk source and therefore cannot result from clinopyroxene-liquid partitioning. Further, the magnitudes of the measured partition coefficients are too small to generate significant U-Th fractionation in either direction. Assuming that clinopyroxene contains the bulk of the U and Th in the MORB source, the results indicate that U-238 - Th-230 disequilibrium in MORB may not be caused by partial melting at all.

Latourrette, T. Z.↗
NASA NTRS2006

Aluminum Effect on the Calibration of the Eu Oxybarometer for Nakhlites

Recent studies of Martian meteorites have shown that Martian magma had wide range of oxygen fugacity. In order to estimate the redox state of the Martian crust and mantle, the partitioning coefficient of Eu in the shergottite pyroxene has been extensively studied. Nakhlite, a cumulate clinopyroxenite, is another important group of Martian meteorites. We have studied synthetic compositions (NT, NL, NJ, NJ2 and NJ4) in order to estimate the Nakhla parent melt composition. In our previous work, we calibrated the Eu oxybarometer for nakhlite using the NJ4 composition which we believe is the closest to the Nakhla parent melt. Consequently, we concluded that Nakhla may have crystallized under fairly reducing condition. However, we also found that the partition coefficient is strongly affected by Al contents in melt and in pyroxene. In this abstract we report the correlation between the D values and Al contents in melt and in pyroxene from our Nakhla experimental studies. Also we check the reliability of our previous calibration of Eu oxybarometer, because Rare Earth Elements (REEs) are easily affected by Al content in melt and in pyroxene and Nakhla pyroxene shows unusual patchy Al zoning.

Makishima, J.↗
NASA NTRS1993

Metal-silicate thermochemistry at high temperature - Magma oceans and the 'excess siderophile element' problem of the earth's upper mantle

Low-temperature metal-silicate partition coefficients are extrapolated to magma ocean temperatures. If the low-temperature chemistry data is found to be applicable at high temperatures, an important assumption, then the results indicate that high temperature alone cannot account for the excess siderophile element problem of the upper mantle. For most elements, a rise in temperature will result in a modest increase in siderophile behavior if an iron-wuestite redox buffer is paralleled. However, long-range extrapolation of experimental data is hazardous when the data contains even modest experimental errors. For a given element, extrapolated high-temperature partition coefficients can differ by orders of magnitude, even when data from independent studies is consistent within quoted errors. In order to accurately assess siderophile element behavior in a magma ocean, it will be necessary to obtain direct experimental measurements for at least some of the siderophile elements.

Capobianco, Christopher J.↗
NASA NTRS1986

Partitioning of K, U, and Th between sulfide and silicate liquids - Implications for radioactive heating of planetary cores

Experimental partitioning studies are reported of K, U, and Th between silicate and FeFeS liquids designed to test the proposal that actinide partitioning into sulfide liquids is more important then K partitioning in the radioactive heating of planetary cores. For a basaltic liquid at 1450 C and 1.5 GPa, U partitioning into FeFeS liquids is five times greater than K partitioning. A typical value for the liquid partition coefficient for U from a granitic silicate liquid at one atmosphere at 1150 C and low fO2 is about 0.02; the coefficient for Th is similar. At low fO2 and higher temperature, experiments with basaltic liquids produce strong Ca and U partitioning into the sulfide liquid with U coefficient greater than one. The Th coefficient is less strongly affected.

Murrell, M. T.↗
NASA NTRS2011

Effect of Sulfur on Siderophile Element Partitioning Between Olivine and Martian Primary Melt

Since olivine is a common early crystallizing phase in basaltic magmas that have produced planetary and asteroidal crusts, a number of experimental studies have investigated elemental partitioning between olivine and silicate melt [e.g., 1, 2, 3]. In particular, olivine/melt partition coefficients of Ni and Co (DNi and DCo) have been intensively studied because these elements are preferentially partitioned into olivine and thus provide a uniquely useful insight into the basalt petrogenesis [e.g., 4, 5]. However, none of these experimental studies are consistent with incompatible signatures of Co [e.g., 6, 7, 8] and Ni [7] in olivines from Martian meteorites. Chemical analyses of undegassed MORB samples suggest that S dissolved in silicate melts can reduce DNi up to 50 % compared to S-free experimental systems [9]. High S solubility (up to 4000 ppm) for primitive shergottite melts [10] implies that S might have significantly influenced the Ni and Co partitioning into shergottite olivines. This study conducts melting experiments on Martian magmatic conditions to investigate the effect of S on the partitioning of siderophile elements between olivine and Martian primary melt.

Usui, T.↗
NASA NTRS2010

High Pressure/Temperature Metal Silicate Partitioning of Tungsten

The behavior of chemical elements during metal/silicate segregation and their resulting distribution in Earth's mantle and core provide insight into core formation processes. Experimental determination of partition coefficients allows calculations of element distributions that can be compared to accepted values of element abundances in the silicate (mantle) and metallic (core) portions of the Earth. Tungsten (W) is a moderately siderophile element and thus preferentially partitions into metal versus silicate under many planetary conditions. The partitioning behavior has been shown to vary with temperature, silicate composition, oxygen fugacity, and pressure. Most of the previous work on W partitioning has been conducted at 1-bar conditions or at relatively low pressures, i.e. <10 GPa, and in two cases at or near 20 GPa. According to those data, the stronger influences on the distribution coefficient of W are temperature, composition, and oxygen fugacity with a relatively slight influence in pressure. Predictions based on extrapolation of existing data and parameterizations suggest an increased pressured dependence on metal/ silicate partitioning of W at higher pressures 5. However, the dependence on pressure is not as well constrained as T, fO2, and silicate composition. This poses a problem because proposed equilibration pressures for core formation range from 27 to 50 GPa, falling well outside the experimental range, therefore requiring exptrapolation of a parametereized model. Higher pressure data are needed to improve our understanding of W partitioning at these more extreme conditions.

Shofner, G. A.↗
NASA NTRS2015

Re-Evaluation of HSE DATA in Light of High P-T Partitioning Data: Late Chondritic Addition to Inner Solar System Bodies Not Always Required for HSE

Studies of terrestrial peridotite and martian and achondritic meteorites have led to the conclusion that addition of chondritic material to growing planets or planetesimals, after core formation, occurred on Earth, Moon, Mars, asteroid 4 Vesta, and the parent body of the angritic meteorites. One study even proposed that this was a common process in the final stages of growth. These conclusions are based al-most entirely on the 8 highly siderophile elements (HSE; Re, Au, Pt, Pd, Rh, Ru, Ir, Os), which have been used to argue for late accretion of chondritic material to the Earth after core formation was complete. This idea was originally proposed because the D(metal/silicate) values for the HSE are very high (greater than 10,000), yet their concentration in the terrestrial mantle is too high to be consistent with such high Ds. The HSE in the terrestrial mantle also are present in chondritic relative abundances and hence require similar Ds if this was the result of core-mantle equilibration. The conclusion that late chondritic additions are required for all five of these bodies is based on the chondritic relative abundances of the HSE, as well as their elevated concentrations in the samples. An easy solution is to call upon addition of chondritic material to the mantle of each body, just after core formation; however, in practice this means similar additions of chondritic materials to each body just after core formation which ranges from approximately 4-5 Ma after T(sub 0) for 4 Vesta and the angrites, to 10-25 Ma for Mars, to 35 to 60 Ma for Moon and perhaps the Earth. Since the work of there has been a realization that high PT conditions can lower the partition coefficients of many siderophile elements, indicating that high PT conditions (magma ocean stage) can potentially explain elevated siderophile element abundances. However, detailed high PT partitioning data have been lacking for many of the HSE to evaluate whether such ideas are viable for all four bodies. Re-cent partitioning studies have covered P, T, fO2, and compositional ranges that allow values to be predicted at conditions relevant to these five inner solar system bodies. Because the D(HSE) metal/silicate are lowered substantially at higher PT conditions and natural com-positions (FeNi metallic liquids and peridotites) it is natural to re-examine the role of core formation on the HSE patterns in a variety of inner solar system bodies. Here I will discuss other processes (including high PT core formation for Mars, Moon and Earth) that can produce the observed HSE patterns, and demonstrate that there are viable hypotheses other than the "one size fits all" hypothesis of late chondritic additions.

Righter, K.↗
NASA NTRS1994

The stability of hibonite, melilite and other aluminous phases in silicate melts: Implications for the origin of hibonite-bearing inclusions from carbonaceous chondrites

Phase fields in which hibonite and silicate melt coexist with spinel CaAl4O7, gehlenitic melilite, anorthite or corundum at 1 bar in the system CaO-MgO-Al2O3-SiO2-TiO2 were determined. The hibonites contain up to 1.7 wt% SiO2. For TiO2, the experimentally determined partition coefficients between hibonite and coexisting melt D(sub i)(sup Hib/L), vary from 0.8 to 2.1 and generally decrease with increasing TiO2 in the liquid. Based on Ti partitioning between hibonite and melt, bulk inclusion compositions and hibonite-saturated liquidus phase diagrams, the hibonite in hibonite-poor fluffy Type A inclusions from Allende and at least some hibonite from hibonite-rich inclusions is relict, although much of the hibonite from hibonite-glass spherules probably crystallized metasably from a melt. Bulk compositions for all of these CAIs are consistent with an origin as melite + hibonite + spinel + perovskite phase assembalges that were partially altered and in some cases partially or completely melted. The duration of the melting event was sufficient to remove any Na introduced by the alteration process but frequently insufficient to dissolve all of the original hibonite. Simple thermochemical models developed for meteoritic melilite and hibonite solid solutions were used to obtain equilibration temperatures of hibonite-bearing phase assemblages with vapor. Referenced to 10(exp -3) atm, hibonite + corundum + vapor equilibrated at approximately 1260 C and hibonite + spinel +/- melilite + vapor at 1215 +/- 10 C. If these temperatures reflect condensation in a cooling gas of solar composition, then hibonite +/- corundum condensed first, followed by spinel and then melilite. The position of perovskite within this sequence is uncertain, but it probably began to condense before spinel. This sequence of phase appearances and relative temperatures is generally consistent with observed textures but differs from expectations based on classical condensation calculations in that equilibration temperatures are generally lower than predicted and melilite initially condenses with or even after spinel. Simple thermochemical modes for the substitution of trace elements into the Ca site of meteoritic hibonites suggest that virtually all Eu is divalent in early condensate hibonites but that Eu(2+)/Eu(#+) decreases by a factor of 20 or more during the course of condensation primarily because the ratio is proportional to the partial pressure of Al, which decreases dramatically as aluminous phase condense. The relative sizes of Eu and Yb anomalies in meteoritic hibonites and inclusions may be partly due to this effect.

Beckett, J. R.↗
NASA NTRS2005

Influence of Silicate Melt Composition on Metal/Silicate Partitioning of W, Ge, Ga and Ni

The depletion of the siderophile elements in the Earth's upper mantle relative to the chondritic meteorites is a geochemical imprint of core segregation. Therefore, metal/silicate partition coefficients (Dm/s) for siderophile elements are essential to investigations of core formation when used in conjunction with the pattern of elemental abundances in the Earth's mantle. The partitioning of siderophile elements is controlled by temperature, pressure, oxygen fugacity, and by the compositions of the metal and silicate phases. Several recent studies have shown the importance of silicate melt composition on the partitioning of siderophile elements between silicate and metallic liquids. It has been demonstrated that many elements display increased solubility in less polymerized (mafic) melts. However, the importance of silicate melt composition was believed to be minor compared to the influence of oxygen fugacity until studies showed that melt composition is an important factor at high pressures and temperatures. It was found that melt composition is also important for partitioning of high valency siderophile elements. Atmospheric experiments were conducted, varying only silicate melt composition, to assess the importance of silicate melt composition for the partitioning of W, Co and Ga and found that the valence of the dissolving species plays an important role in determining the effect of composition on solubility. In this study, we extend the data set to higher pressures and investigate the role of silicate melt composition on the partitioning of the siderophile elements W, Ge, Ga and Ni between metallic and silicate liquid.

Singletary, S. J.↗
NASA NTRS1991

Partitioning of siderophile elements in the Fe-Ni-S system - 1 bar to 80 kbar

Partition coefficients for Au, Ni, P, and Ge between solid Fe-Ni metal and sulfur-bearing metallic liquids have been measured at 7, 27 and 80 kbar. These are the only such data for Au, P, and Ge at high pressure. Comparison of the present partitioning results to those obtained at 1 bar indicate that only the 80-kbar Ge data differ significantly from the 1-bar experiments. Thus, many low-pressure partitioning experiments in the Fe-Ni-S-P system may have applicability to the greater portion of the earth's upper mantle or, alternatively, the entire mantle of Mars.

Jones, J. H.↗
NASA NTRS1993

An experimental study of trace element partitioning between perovskite, hibonite and melt: Equilibrium values

The presence of perovskite (CATiO3) and hibonite (Ca Al12O19) within different regions of Calcium-, Aluminum-rich Inclusions (CAI) and the trace element concentrations of these minerals in each circumstance, constrain models of precursor formation, nebular condensation, the thermal history of inclusions with relict perovskite and hibonite, and the formation of the Wark-Lovering rim. At present mineral/melt partition coefficient data for hibonite are limited to a few elements in simple experimental systems, or to those derived from hibonite-glass pairs in hibonite/glass microspherules. Similarly, there is only limited data on perovskite D that are applicable to meteorite compositions. Apart from the importance of partitioning studies to meteorite research, D values also are invaluable in the development of thermodynamic models, especially when data is available for a large number of elements that have different ionic charge and radii. In addition, study of the effect of rapid cooling on partitioning is crucial to our understanding of meteorite inclusions. To expand our knowledge of mineral/melt D for perovskite and hibonite, a study was instituted where D values are obtained in both equilibrium and dynamic cooling experiments. As an initial phase of this study mineral/melt D was measured for major elements (Ca, Mg, Al, Ti, and Si), 15 rare earth elements (La-Lu) and 8 other elements (Ba, Sr, U, Th, Nb, Zr, Hf, and Ge) in perovskite and hibonite grown under equilibrium conditions, in bulk compositions that are respectively similar to Compact Type A (CTA) CAI and to a hibonite/glass microspherule. Experimental mixes were doped with REE at 20-50x chondritic (ch) abundances, Ba at 50 ppm, Sr, Hf, Nb, and Zr at 100 ppm and, U and Th at 200 ppm. Trace element abundances were measured with the PANURGE ion microprobe. Major element compositions were obtained by electron microprobe analysis.

Kennedy, A. K.↗
NASA NTRSDate not supplied

Elemental Partitioning Constraints on the Mineralogy of the Martian Mantle

Incompatible elements weakly partition into mantle mineral phases so precise measurements of incompatible element abundances in mantle melts have been used to infer mineralogy of mantle sources [1, 2]. In a recent study [3], we showed that several incompatible chalcophile elements (As, Tl, Pb) relative to REE were more compatible in martian igneous rocks than in terrestrial basalts. This observation can be accounted for by (1) the bulk partition coefficients of chalcophile elements being higher due to the presence of more residual sulfide in the martian mantle [3] and/or (2) different relative compatibilities in the martian mantle due to differences in mineralogy. In this study, the compatibility sequences of lithophile elements in martian and MORB mantles are revised to better represent relative partitioning behavior during mantle differentiation and the influence of mantle mineralogy is examined.

mantle geochemistry↗
NASA NTRS2017

Experimental Study into the Partitioning Behavior of Fluorine, Chlorine, Hydroxyl, and Sulfur (S2-) Between Apatite and a Synthetic Kreep Basalt Melt

The mineral apatite (Ca5 (PO4)3(F, Cl, OH)) is known for its ability to constrain the petrogenesis of the rock in which it is hosted and for its ubiquity throughout the Solar System, as it is found in lunar, martian, and terrestrial rocks alike (McCubbin et. al, 2015). The abundance of volatile elements, and for this particular study, the elevated abundance of sulfur (S2-) in high-Al basalt samples bearing apatite, could provide more insight for inquiries posed about the behavior of volatiles in lunar and martian magmatic systems (Boyce et. al, 2010). Oxygen fugacity will be an important parameter for these experiments, as the Moon, Mars, and Earth have different redox states (Herd, 2008). The objective of this experimental endeavor is to determine apatite-melt partition coefficients for the volatile elements (F-, Cl-, OH-, S2-) that make up the X-site (i.e., the typically monovalent anion site) in the mineral apatite in a lunar melt composition under lunar oxygen fugacity conditions approx.1-2 log units below the iron-wüstite buffer). All experiments will be conducted at NASA, Johnson Space Center in the High Pressure Experimental Petrology Laboratory. In order to conduct apatite-melt partition experiments with oxygen fugacity as an additional parameter, we will create a synthetic mix of the lunar KREEP basalt 15386, a sample retrieved during Apollo 15 that is believed to represent an indigenous volcanic melt derived from the lunar interior (Rhodes, J.M et. al, 2006). Other geochemically significant elements including C, Co, Ni, Mo, and rare earth elements will be included in the mix at trace abundances in order to assess their partitioning behavior without effecting the overall behavior of the system. The synthetic mix will then be loaded into a piston cylinder, an apparatus used to simulate high-pressure/high-temperature conditions of planetary interiors, and exposed to 0.5 GPa of pressure, the pressure observed in the upper mantle of the Moon, and heated to the melting temperature of the materials. To make sure crystals grow large enough for the necessary analyses, the sample will be kept at the crystallization temperature for 8 hours. This extended run time should also allow the sample to achieve a steady state which is necessary to accurately assess the partitioning of these elements between apatite and melt. The results from this experimental study will allow us to determine the fate of F-, Cl-, OH-, and S2- during the magmatic evolution of the Moon.

Turner, Amber↗
NASA NTRS1992

Permeability of lipid bilayers to amino acids and phosphate

Permeability coefficients for amino acid classes, including neutral, polar, hydrophobic, and charged species, were measured and compared with values for other ionic solutes such as phosphate. The rates of efflux of glycine, lysine, phenylalanine, serine and tryptophan were determined after they were passively entrapped in large unilamellar vesicles (LUVs) composed of egg phosphatidylcholine (EPC) or dimyristoylphosphatidylcholine (DMPC). The following permeability coefficients were obtained for: glycine, 5.7 x 10(-12) cm s-1 (EPC), 2.0 x 10(-11) cm s-1 (DMPC); serine, 5.5 x 10(-12) cm s-1 (EPC), 1.6 x 10(-11) cm s-1 (DMPC); lysine, 5.1 x 10(-12) cm s-1 (EPC), 1.9 x 10(-11) cm s-1 (DMPC); tryptophan, 4.1 x 10(-10) cm s-1 (EPC); and phenylalanine, 2.5 x 10(-10) cm s-1 (EPC). Decreasing lipid chain length increased permeability slightly, while variations in pH had only minor effects on the permeability coefficients of the amino acids tested. Phosphate permeability was in the range of 10(-12)-10(-13) cm s-1 depending on the pH of the medium. The values for the polar and charged amino acids were surprisingly similar to those previously measured for monovalent cations such as sodium and potassium, which are in the range of 10(-12)-10(-13) cm s-1, depending on conditions and the lipid species used. This observation suggests that the permeation rates for the neutral, polar and charged amino acids are controlled by bilayer fluctuations and transient defects, rather than partition coefficients and Born energy barriers. The results are relevant to the permeation of certain peptides into lipid bilayers during protein translocation and membrane biogenesis.

NASA Discipline Exobiology↗
NASA NTRS2016

Metal-Silicate-Sulfide Partitioning of U, Th, and K: Implications for the Budget of Volatile Elements in Mercury

During formation of the solar system, the Sun produced strong solar winds, which stripped away a portion of the volatile elements from the forming planets. Hence, it was expected that planets closest to the sun, such as Mercury, are more depleted in volatile elements in comparison to other terrestrial planets. However, the MESSENGER mission detected higher than expected K/U and K/Th ratios on Mercury's surface, indicating a volatile content between that of Mars and Earth. Our experiments aim to resolve this discrepancy by experimentally determining the partition coefficients (D(sup met/sil)) of K, U, and Th between metal and silicate at varying pressure (1 to 5 GPa), temperature (1500 to 1900 C), oxygen fugacity (IW-2.5 to IW-6.5) and sulfur-content in the metal (0 to 33 wt%). Our data show that U, Th, and K become more siderophile with decreasing fO2 and increasing sulfur-content, with a stronger effect for U and Th in comparison to K. Using these results, the concentrations of U, Th, and K in the bulk planet were calculated for different scenarios, where the planet equilibrated at a fO2 between IW-4 and IW-7, assuming the existence of a FeS layer, between the core and mantle, with variable thickness. These models show that significant amounts of U and Th are partitioned into Mercury's core. The elevated superficial K/U and K/Th values are therefore only a consequence of the sequestration of U and Th into the core, not evidence of the overall volatile content of Mercury.

Habermann, M.↗
NASA NTRS2016

Rare Earth Element Partitioning in Lunar Minerals: An Experimental Study

The partitioning behavior of rare earth elements (REE) between minerals and melts is widely used to interpret the petrogenesis and geologic context of terrestrial and extra-terrestrial samples. REE are important tools for modelling the evolution of the lunar interior. The ubiquitous negative Eu anomaly in lunar basalts is one of the main lines of evidence to support the lunar magma ocean (LMO) hypothesis, by which the plagioclase-rich lunar highlands were formed as a flotation crust during differentiation of a global-scale magma ocean. The separation of plagioclase from the mafic cumulates is thought to be the source of the Eu depletion, as Eu is very compatible in plagioclase. Lunar basalts and volcanic glasses are commonly depleted in light REEs (LREE), and more enriched in heavy REEs (HREE). However, there is very little experimental data available on REE partitioning between lunar minerals and melts. In order to interpret the source of these distinctive REE patterns, and to model lunar petrogenetic processes, REE partition coefficients (D) between lunar minerals and melts are needed at conditions relevant to lunar processes. New data on D(sub REE) for plagioclase, and pyroxenes are now available, but there is limited available data for olivine/melt D(sub REE), particularly at pressures higher than 1 bar, and in Fe-rich and reduced compositions - all conditions relevant to the lunar mantle. Based on terrestrial data, REE are highly incompatible in olivine (i.e. D much less than 1), however olivine is the predominant mineral in the lunar interior, so it is important to understand whether it is capable of storing even small amounts of REE, and how the REEs might be fractionatied, in order to understand the trace element budget of the lunar interior. This abstract presents results from high-pressure and temperature experiments investigating REE partitioning between olivine and melt in a composition relevant to lunar magmatism.

McIntosh, E. C.↗