Parent body histories recorded in Rumuruti chondrite sulfides: Implications for the onset of oxidized, sulfur‐rich core formation

Models of planetary core formation beginning with melting of Fe,Ni metal and troilite are not readily applicable to oxidized and sulfur‐rich chondrites containing only trace quantities of metal. Cores formed in these bodies must be dominated by sulfides. Siderophile trace elements used to model metallic core formation could be used to model oxidized, sulfide‐dominated core formation and identify related meteorites if their trace element systematics can be quantified. Insufficient information exists regarding the behavior of these core‐forming elements among sulfides during metamorphism prior to anatexis. Major, minor, and trace element concentrations of sulfides are reported in this study for petrologic type 3–6 R chondrite materials. Sulfide‐dominated core‐forming components in such oxidized chondrites (ƒO2 ≥ iron‐wüstite) follow metamorphic evolutionary pathways that are distinct from reduced, metal‐bearing counterparts. Most siderophile trace elements partition into pentlandite at approximately 10× chondritic abundances, but Pt, W, Mo, Ga, and Ge are depleted by 1–2 orders of magnitude relative to siderophile elements with similar volatilities. The distribution of siderophile elements is further altered during hydrothermal alteration as pyrrhotite oxidizes to form magnetite. Oxidized, sulfide‐dominated core formation differs from metallic core formation models both physically and geochemically. Incongruent melting of pentlandite at 865°C generates melts capable of migrating along solid silicate grains, which can segregate to form a Ni,S‐rich core at lower temperatures compared to reduced differentiated parent bodies and with distinct siderophile interelement proportions.


INTRODUCTION
The compositions and proportions of core-forming phases are controlled by the cosmochemistry of the nebular environments in which they formed (e.g., Lodders, 2003;Wood et al., 2019). Thus, a parent body that accreted in the presence of abundant oxygen and sulfur would ultimately produce a core that is compositionally distinct from one that is dominated by Fe,Ni metal. Among the meteoritic record, the effective concentrations of oxygen and sulfur (i.e., oxygen and sulfur fugacities, ƒO 2 and ƒS 2 , respectively) vary by orders of magnitude (e.g., Righter et al., 2016), and may have been broadly correlated with heliocentric distance due to evaporation and transport of volatile phases in the high-temperature regions of the innermost regions of the protoplanetary disk . As a result, much of the iron in many oxidized and sulfur-rich meteorite parent bodies partitioned into silicates, oxides, and sulfides after oxidation and sulfidation of nebular metals (e.g., Frost, 1991;Kerridge, 1976;Lauretta et al., 1998). Since these reactions occurred prior to the onset of partial melting and parent body differentiation, they have significant effects on the melting temperatures of minerals and the igneous evolution of the parent body (e.g., Tomkins et al., 2020).
A variety of geochemical methods have been used to investigate core formation. Notably, highly siderophile trace elements (HSEs) Re, Os, Ir, Ru, Rh, Pt, and Pd can be used to model metal-silicate segregation, metal fractionation, and core formation (e.g., Chabot & Jones, 2003), as these elements partition strongly into Fe, Ni metals (D metal/silicate ≥ 10 4 ; Walker, 2016) but will partition into pentlandite in the absence of chondritic metal (D pentlandite/silicate~1 0 3 ; Crossley et al., 2020). Consequently, siderophile elements may be used to model core formation on highly oxidized, sulfur-rich bodies, but only if their partitioning behaviors at the onset of melting are accurately characterized.
Rumuruti-type chondrites provide the ideal samples to study the distribution of siderophile trace elements in oxidized, sulfur-rich systems because they formed in high ƒO 2 (iron-w€ ustite [IW] À 1 to IW + 4; McCanta et al., 2008;Righter & Neff, 2007) and high ƒS 2 environments (2 orders of magnitude greater than the iron-troilite buffer, IT+2; Miller et al., 2017). Despite a dearth of Fe,Ni metal, R chondrites still contain chondritic abundances of siderophile trace elements (Isa et al., 2014) with the majority of HSEs held in pentlandite (Crossley et al., 2020). Many of the more than 250 R chondrites that have been identified are breccias composed of clasts that are variably metamorphosed, including petrologic types 3-6 and some hydrothermally altered samples (e.g., Bischoff et al., 2006Bischoff et al., , 2011Gattacceca et al., 2020). These breccias provide an opportunity to observe the effects of thermal alteration on opaque mineral assemblages and trace element systematics throughout metamorphism on the R chondrite parent body. The lack of exogenous materials among R chondrite breccias (Bischoff et al., 2006) further reduces the risk of comparing metamorphic effects among genetically unrelated materials that is inherent when comparing separate samples. R chondrites also share bulk chemical, mineralogic, and isotopic similarities with more thoroughly studied, reduced, and metal-bearing ordinary chondrites (Bischoff et al., 2011;Greenwood et al., 2000;Miller et al., 2017), providing an opportunity to isolate the effects of high ƒO 2 and ƒS 2 on core-forming mineral phases and compare sulfide-dominated to metal-dominated siderophile element partitioning behaviors during parent body metamorphism.
Published datasets are lacking for siderophile trace element systematics during the metamorphism of highly oxidized, sulfur-rich R chondrites. To investigate the equilibrium distribution of siderophile elements during metamorphism, major and trace element concentrations in sulfides and oxides are reported for two R chondrites, Northwest Africa 11304 and La Paz Icefield 04840. These samples include clasts of petrologic types 3-6, as well as hydrothermally altered materials (e.g., Gattacceca et al., 2020;Satterwhite & Righter, 2006). While siderophile element systematics among R chondrite sulfides share many similarities with those of ordinary chondrite metals during metamorphism, several distinct mechanisms relating to the high ƒO 2 and ƒS 2 of R chondrites result in the fractionation of Pt, W, Mo, Ga, Ge, and other siderophiles relative to elements with similar volatilities and nominal mineralogic affinities.

Sample Selection and Preparation
Two R chondrites were analyzed in this study: Northwest Africa (NWA) 11304 and La Paz Icefield (LAP) 04840,11. A probe mount of NWA 11304 (R3-6; Gattacceca et al., 2020) was used for in situ compositional analyses, and a thin section was subsequently made for petrographic investigation from the same section after destructive analyses were completed. In situ sulfide and oxide mineral chemistries were measured from a 50 lmthick section of LAP 04840 (R6; Satterwhite & Righter, 2006). The probe mount and thin section of NWA 11304 are from the personal collection of Dr. Richard Ash, and the polished thin section of LAP 04840 is from the Antarctic Meteorite Collection at NASA's Johnson Space Center.

Terrestrial Alteration
Under ideal circumstances, analyses of meteoritic metals and sulfides would be performed on minimally weathered materials. The only reported R chondrite fall is the type specimen Rumuruti (Schulze et al., 1994). Due to the scarcity of unweathered R chondrite samples and the destructive nature of analyses in this work, minimally weathered samples were chosen for analysis. NWA 11304's weathering grade is classified as W2 (moderate weathering; Gattacceca et al., 2020). Some reddening is present around the edges of the sample thin section and along cracks but textural evidence of extensive alteration is not present throughout the interior of NWA 11304 thin section, in backscatter electron (BSE) image maps, or in elemental maps. Most sulfides in this section of NWA 11304 lack any visible alteration greater than 20% of their surface area, which is more consistent with a W1 classification (Wlotzka, 1993). Similarly, LAP 04840 is classified with an Antarctic weathering grade A/B (Gattacceca et al., 2020), which is comparable to W1 (Wlotzka, 1993). The minimal weathering of sulfides in both samples provides suitable materials for analyses of siderophile element systematics during metamorphism and hydrothermal alteration.

Petrologic Type Classification
Petrologic types were determined following methods for ordinary chondrites in Huss et al. (2006) based on textural characteristics and the percent mean deviation (PMD) of FeO content in chondrule and matrix olivine (Table 1 and Table S1). Given that parameters from Huss et al. (2006) were originally established for ordinary chondrites, samples are also compared to descriptions for R chondrites of various petrologic types summarized in Bischoff et al. (2011).

Modal Mineralogy
Modal mineralogy for NWA 11304 was calculated by creating mineral classification maps (e.g., Beck et al., 2012;Crossley et al., 2020). Element maps were created from energy-dispersive X-ray spectroscopy (EDS) analyses at the Smithsonian Institution National Museum of Natural History Department of Mineral Sciences (SI) using a ThermoFisher energy dispersive X-ray detector attached to a FEI Nova NanoSEM 600. In total, 17 individual element maps were stacked into layers of a data cube using Environment for Visualizing Images (ENVI 4.4) image processing software. The final mineral classification map of NWA 11304 is provided in Fig. S1. Mineral modes were calculated by counting pixels for each class (Table S2). To estimate uncertainty for mineral modes, homogeneous grains of each mineral phase were identified by scanning electron microscopy and electron probe microanalysis (EPMA), then were examined for misclassified pixels in the mineral classification maps. Less than 10% of pixels in each of these grains were misclassified and can be attributed to noise in the original EDS data. Therefore, mineral mode uncertainties are estimated to be AE7 vol% for olivine, AE2 vol% for plagioclase, and less than AE1 vol% for other phases.
Step-  Type 3  Type 4  Type 5  Impact melts  Vol%   Whole sample  12  54  28  5  Olivine  66  63  62  64  Plagioclase  16  16  17  14  Low-Ca pyroxene  14  9  7  10  High-Ca pyroxene  3  6  4  5  Pyrrhotite  4  3  4  3  Pentlandite  3  3  4  4  Oxides  3  2  2  2  by-step instructions for this methodology are provided in supplementary materials. EDS and EPMA data necessary for constructing modal mineral maps of LAP 04840 could not be collected due to limited facility accessibility imposed by public health policies during the COVID-19 pandemic. For this sample, opaque minerals were identified via reflected light optical microscopy using an OMAX petrographic polarizing microscope, and digital images were assembled into a mosaic to calculate the volume percent of opaque phases through pixel counting by thresholding pixel values using ImageJ. Within opaque assemblages, phases were distinguished in digital images at higher magnification by their relative brightness in grayscale (pentlandite >pyrrhotite ≫ magnetite). Each phase was identified, masked in separate layers using Adobe Photoshop, and the proportion of pixels for phases in each layer was used to calculate the modal mineralogy of opaque assemblages in volume percent (Table S2).

Compositional Analyses
Major and minor element concentrations for mineral phases were collected via EPMA at SI using a field emission JEOL 8530F+. Operating conditions were set at 1 lm spot size, 10 kV, and 20 nA for all mineral phases. Natural and synthetic standards were used for calibration and were measured in duplicate before and after each analytical run. For LAP 04840, our measurements yielded systematically low values for Fe and S in sulfides and oxides due to spectrometer failure for those elements that occurred between calibration and the first analytical run while operating the instrument with limited facilities access due to public health policies. The troilite and chromite standards were therefore renormalized by their official compositions (Jarosewich, 2002, and references therein) and were used to correct the EPMA measurements. Our corrected measurements for Fe and S are within~2 atom% of the values reported for sulfides in McCanta et al. (2008). Detection limits, renormalizing calculations, and EPMA standards are provided in Tables S3 and S4.
Trace elements were measured in situ via laser ablation inductively coupled mass spectrometry (LA-ICP-MS) at the University of Maryland Plasma Laboratory using a New Wave Research UP-213 laser ablation unit attached to a ThermoFinnigan Element2 single collector ICP-MS. Samples were ablated into a stream of helium (approximately 0.6 L min À1 ), which was then mixed with Ar (approximately 0.9 L min À1 ) before introduction into the mass spectrometer. Laser ablation was done via spot analyses using a spot size varying from 15 to 80 lm, depending on the size of the grain ablated. The output of the laser was modified to ensure the fluence remained between 2 and 4 J cm À2 , typically an output of approximately 60% of the total available power. The repetition rate of the laser was 7 Hz for NWA 11304 and 10 Hz for LAP 04840. The change in repetition rate was due to instrument failure between analytical runs that disallowed the selection of 7 Hz during the analysis of LAP 04840. For siderophile and chalcophile element measurements, samples were standardized using iron meteorites Coahuila, Filomena, and NIST 610 alumino-silicate glass standard (described by Walker et al., 2008). All samples were cleaned with a brief, low-power pulse of the laser prior to the data gathering ablation. Gas background measurements were taken for 20 s prior to each sample analysis. LAMTRACE was used for data reduction (van Achterbergh et al., 2001). Upper limits for concentration were calculated by setting the detection limit for any given isotope measured as three times the mean measured gas background. Replicate analyses of standards yielded external precisions less than AE5% (2r) for all elements analyzed. Specific precisions and detection limits are reported in Table S5. Standards were analyzed in identical fashion to their respective samples. Due to the lack of a matrix-matched sulfide trace element standard, the use of multiple iron meteorites and basaltic glass standards provided a means to assess any matrix-related fractionation of elements (e.g., Steenstra et al., 2020;Sylvester, 2001) that may have occurred during laser ablation of unknown sulfides. Analytical runs were standardized using each of these materials, and the variance of concentrations for most siderophile elements in unknown measurements is typically <10%. For Mo, concentrations differed by as much as a factor of 2 between basaltic glass and iron meteorite standards. The values for iron meteorite standards are reported in this work, as the three isotopes of Mo ( 95 Mo, 97 Mo, and 98 Mo) are in good agreement for these standardizations, while basaltic glass standardization yields discordant values for each isotope. Consequently, our interpretations are limited to interelement ratios that exceed these uncertainties. Caution should be taken with the reported values for chalcophile elements like Se and Te, as these elements have been shown to fractionate in nonmatrix-matched standardizations of sulfide analyses (Steenstra et al., 2020). As such, we do not include interpretation of chalcophile elements in this study but provide the data for completeness. All data and standardizations are available in supplementary materials. Trace element analyses are biased toward sulfide and oxide grains larger than~20 lm due to the minimum useful diameter of the laser. As such, these measurements do not account for the trace element compositions of smaller matrix sulfides common in petrologic types 3 and 4.

Northwest Africa 11304
The textures in NWA 11304 are consistent with other breccias and the sample is composed primarily of R chondrite petrologic type 4 materials with several type 3 and 5 clasts, as well as two clasts that experienced extensive impact-induced melting of sulfides and silicates (Tables 1 and 2; Figure 1, and Table S2). Clast boundaries are distinguishable in BSE images, modal mineral maps, and by plane and cross-polarized light (Figures S1 and S2).

Petrologic Type 3
Petrologic type 3 materials ( Figure 1a) together comprise approximately 12 vol% of the sample. They are composed of 66 vol% olivine, 16 vol% feldspar, 14 vol% low-Ca pyroxene, 7 vol% sulfide, 3 vol% oxides, 3 vol% high-Ca pyroxene, and <1 vol% Ca phosphate. These clasts contain the greatest variability in olivine FeO content (PMD FeO = 22.7, Table 1), the lowest chondrule/ matrix ratios (0.7), and the smallest matrix silicate and sulfide grain sizes ((1 lm). These parameters are consistent with petrologic subtypes 3.6-3.8 for ordinary chondrites (Huss et al., 2006), although a direct comparison between the two groups is complicated by the differences in mineralogy and mineral chemistries. Both type I and type II chondrules are present in these clasts and include porphyritic olivine, porphyritic pyroxene, and porphyritic olivine-pyroxene textures ranging in size from 250 to 600 lm in diameter. The matrix is opaque in thin section and dominated by FeO-rich olivine with average Mg# = 46.7 (Mg# = Mg/Mg + Fe 9 100, in mole%), as well as pentlandite, pyrrhotite, and albitic plagioclase. Matrix sulfides are randomly dispersed, and larger sulfide nodules (typically 50-150 lm) are often associated with chondrules. Minor and accessory matrix phases include pyroxene (both low and high-Ca), Ca-phosphates, and magnetite. The type 3 clasts include the greatest abundance of low-Ca pyroxene contained in chondrules and chondrule fragments. High-Ca pyroxene is an accessory phase and is most commonly found in chondrule mesostases. Small subhedral magnetite grains (<10 lm) are present in type 3 clasts and are usually associated with sulfide nodules. The largest sulfide nodule (500 lm) in type 3 clasts appears to have been incorporated as a free-floating sulfide melt droplet, evidenced by its relatively circular shape, embayed boundaries, and quenched margin ( Figure 2). This nodule contains the largest concentration of magnetite in NWA 11304, constituting roughly half of the nodule's surface area in thin section. Subhedral magnetite grains (~10 lm) with smooth, rounded surfaces are clustered in the center of the sulfide nodule with intergranular pentlandite and pyrrhotite, similar to textures observed in some aqueously altered CM chondrites (Singerling & Brearley, 2020). Smaller magnetite grains (≤1 lm) populate the margins of this nodule. These textures are consistent with magnetite growth through oxidation of the sulfide nodule either with highly oxidized nebular gasses (i.e., ƒO 2~f ayalitemagnetite-quartz [FMQ]) or through reaction with the FIGURE 1. Representative backscatter electron (BSE) images of petrologic types among NWA 11304 clasts. This sample is a genomict breccia containing clasts of (a) petrologic type 3, (b) type 4, and (c) type 5. Sulfides (white) coalesce and coarsen, silicates (grays) homogenize, and chondrules and matrix grains recrystallize from types 3 to 5. The circular feature in the type 5 sulfide (c) is a laser ablation pit. Detailed petrographic descriptions for each petrologic type are presented in the text. surrounding matrix materials. However, the matrix grains directly in contact with the sulfide nodule display no obvious reaction textures and are consistent with matrix materials throughout the rest of the clast.
Major element compositions for type 3 sulfides are most consistent with equilibration at 500°C (Table 2; Figure 3), falling along tie lines between pentlandite and monosulfide solid solution (MSS) in the 500°C isothermal plane of Fe-Ni-S phase diagrams drawn from crystallization experiments (after Kitakaze et al., 2016).

Petrologic Type 4
Petrologic type 4 materials ( Figure 1b) dominate the majority of this section of NWA 11304 (54 vol%). Type 4 materials contain chondrule/matrix ratios (1.2) higher than type 3, but with fewer FeO-poor type I chondrules.
Chondrules are typically 0.5-1 mm in diameter, but can range up to 2 mm, and include porphyritic olivine, porphyritic olivine-pyroxene, porphyritic pyroxene, and cryptocrystalline pyroxene chondrule types. Olivine FeO variability in type 4 material (PMD FeO = 15.8) is less than FIGURE 3. Isothermal planes for Fe-Ni-S ternary phase diagrams after Kitakaze et al. (2016) for sulfides in NWA 11304 petrologic endmembers and LAP 04840 in atomic %. MSS = monosulfide solid solution, HPN = high-form pentlandite, and ɣ = awaruite. For NWA 11304, type 3 sulfides fall along tie lines at 500°C, and sulfides in petrologic type 4 and type 5 clasts are consistent with equilibration at 700°C. In LAP 04840 (type 6), pentlandite and pyrrhotite compositions vary over several atom percent, probably due to oxidation of sulfides during hydrothermal alteration. Fe/Ni ratios are most consistent with 700°C phase diagrams in Kitakaze et al. (2016). The inferred closure temperature for the sulfide system in LAP 04840 is within the range of temperatures calculated using Fe-Mg silicate-spinel thermometers (McCanta et al., 2008) and is reasonable for petrologic type 5 materials (Huss et al., 2006). (Color figure can be viewed at wileyonlinelibrary.com.) in type 3, indicating an approach to equilibrium compositions, although chondrule olivine has not fully equilibrated with matrix olivine (Table 1). Type 4 matrix materials are microcrystalline, with variable grain sizes for plagioclase (typically 10-30 lm). High-Ca pyroxene is twice as abundant as seen in type 3 material, although it is still mostly confined to chondrules with low-Ca pyroxene and/or chondrule olivine. Low-Ca pyroxene abundance decreases by approximately one-third from type 3 to type 4. Sulfides larger than 10 lm are more common than in type 3 materials but span a similar range of compositions. However, matrix sulfides (<10 lm) are still present in type 4 materials. Major element compositions from EPMA for sulfides in type 4 materials are consistent with higher equilibration temperature (up to 700°C; Table 2; Figure 3), but this temperature should be considered as an upper limit because analyses for type 4 sulfides are biased toward the largest grains. These petrographic and compositional characteristics are consistent with 3.8-4.0 petrologic types in ordinary chondrites and are consistent with descriptions of type 4 classifications for other R chondrites (Bischoff et al., 2011). Given the petrographic differences and less variable olivine FeO content in these materials compared to type 3 clasts, the R4 classification assigned to the majority of materials in NWA 11304 is accurate (Gattacceca et al., 2020).

Petrologic Type 5
In petrologic type 5 materials (28 vol%), chondrule and matrix olivine grains are homogenized (Mg# = 62, PMD FeO = 1) and most chondrules are texturally wellintegrated with recrystallized matrix materials ( Figure 1c). However, several relict chondrules are discernable in thin section. Plagioclase grains are typically 20-50 lm and form networks with high-Ca pyroxene (~50 lm) between large olivine grains (commonly 300-400 lm). Sulfides are most commonlỹ 100 lm but reach up to 650 lm in diameter, and small matrix sulfide grains are rare. Sulfides are typically composed of homogeneous subhedral pyrrhotite and pentlandite, with pentlandite dominating the largest grains. The boundaries between exsolved pentlandite and pyrrhotite are typically much smoother and more regular in type 5 materials than in more primitive clasts (Figure 4a), indicative of textural equilibration between the two phases near 700°C (Figure 3). Sulfide sizes and grain boundaries are consistent with coalescing and coarsening during metamorphism, similar to what has been observed for metals in ordinary chondrites (e.g., Kimura et al., 2008). Small HSE-rich grains of arsenides and sulfarsenides (generally~1 lm diameter) are found within all petrologic types, but the largest grains are found in type 5 (up to 10 lm diameter) and are most often in contact with sulfides. Calcium phosphate grain sizes are variable, typically 50-100 lm, but are larger on average than their type 3 and type 4 counterparts and are commonly associated with sulfide assemblages. The compositional and textural characteristics of this clast are consistent with petrologic type 5 classifications for both R chondrites (Bischoff et al., 2011) and ordinary chondrites (Huss et al., 2006).

Impact-Shocked and Melted Clasts
Two clasts in this section of NWA 11304 are strongly shocked and comprise approximately 5 vol% of the observable sample ( Figures S1 and S2). The less shocked clast contains some chondrules and mafic fragments of olivine and low-Ca pyroxene (up to Fo 91 and En 95 , respectively), but olivine is homogeneous (Mg# = 65 AE 0.5). The low-Ca pyroxene fragments display highly embayed disequilibrium textures that indicate destruction via reaction with matrix phases. Quenched melts of Krich feldspar and high-Ca pyroxene coexist with submicron FeO-rich olivine and sulfides interstitial to large olivine grains (Fig. S2). Potassium-rich feldspar is only observed in impact-melted clasts and in the mesostasis of porphyritic chondrules in type 3 and type 4 materials. Another clast appears to have been completely melted and rapidly cooled. This is evident in zoned olivine crystals (Mg# = 91-51, core to rim) with interstitial microcrystalline K-rich feldspar and high-Ca pyroxene, as well as the chaotic distribution of silicate microcrystals throughout a large, highly embayed sulfide assemblage (850 9 1700 lm) that is characteristic of impact melts seen in other R chondrites (e.g., figure 4c from Bischoff et al., 2011).

LaPaz Icefield 04840
LAP 04840 is classified as an R6 chondrite (Satterwhite & Righter, 2006), and of interest because it is one of the three hydrothermally altered R chondrites and is among the most oxidized endmembers of the group (log ƒO 2 = IW + 5; McCanta et al., 2008). In tandem with NWA 11304, it provides insight into the siderophile systematics across a range of R chondrite ƒO 2 conditions as well as an opportunity to investigate the effects of hydrothermal alteration on siderophile element distribution among sulfides and oxides (Table 3). Previous work provides a thorough description of nonopaque mineral phases in this sample (e.g., McCanta et al., 2008). Nickel-rich (5 wt%) pyrrhotite is the most common sulfide phase in LAP 04840. Most pyrrhotite grains have exsolution lamellae of pentlandite throughout (1-5 lm diameter), a common feature of chondritic Nirich sulfides that is interpreted to be the result of crystallization from liquid MSS after chondrule formation (e.g., Boctor et al., 2002;Brearley & Martinez, 2010;Harries & Langenhorst, 2013;Schrader et al., 2015Schrader et al., , 2016Schrader et al., , 2021Singerling & Brearley, 2020;Singerling et al., 2021;Zanda et al., 1995). Discrete grains of pentlandite and also mantle pyrrhotite grains are included within them. Most large magnetite grains (50-100 lm) are subhedral and associated with anhedral pentlandite and pyrrhotite. Smaller subhedral magnetite grains cluster along the margins of sulfide assemblages (Figure 4b). However, magnetite does not occur exclusively in contact with sulfides. Isolated grains of magnetite are usually smaller (~5-20 lm), although one large magnetite grain (150 lm) does not visibly contact any sulfides in this section.

Siderophile Trace Element Concentrations of Sulfides
Across all petrologic types in NWA 11304, most HSEs are concentrated in pentlandite (Table 4; Figure 5 and Fig. S3), regardless of whether pentlandite is isolated or is visibly in contact with pyrrhotite. In clasts of petrologic types 3 and 4, sulfide siderophile element abundances span a continuous range of values from subchondritic concentrations ((0.19 CI) in Ni-poor pyrrhotite to super chondritic concentrations (>109 CI) in the most Ni-rich pentlandite endmembers. The continuous range of HSE/Ni ratios among different sulfide phases likely reflects intermediate MSS in type 3 and type 4 clasts and/or fine-scale exsolution lamellae of pentlandite that cannot be resolved during laser ablation measurements. In contrast, EPMA measurements of Ni in type 4 sulfides are not in agreement with Ni concentrations measured by laser ablation, but instead tightly cluster into pentlandite and pyrrhotite endmember compositions. The discrepancy between EPMA and laser ablation measurements is likely due to the difference in spot size between analyses. In contrast to the 1 lm spot size of EPMA, laser ablation analyses in type 4 materials had a minimum diameter of 20 lm and a sampling depth of~10 lm. Consequently, trace element measurements sample fine-scale mixtures of pyrrhotite and pentlandite that are not sampled simultaneously by EPMA. Regardless, the absolute concentrations of trace elements in sulfides clearly reflect the exsolution and/or coarsening of pentlandite from Nirich pyrrhotite by petrologic type 5 with the majority of siderophile elements concentrated in pentlandite (K d pn/po 2-500) as the two phases coarsen and approach equilibrium ( Figure 5). However, W, Mo, Pt, Ga, Au, and Ge in pentlandite are depleted by orders of magnitude relative to concentrations in OC metals ( Figure 6). Concentrations vary among individual pentlandite grains, ranging from chondritic to subchondritic, but three grains in NWA 11304 contain superchondritic W concentrations that approach that of OC taenite (~59 CI; Kong & Ebihara, 1997). Olivine, pyroxene, and plagioclase in NWA 11304 also hold chondritic to superchondritic concentrations of W on average (Table S6).
Siderophile element concentrations for both pentlandite and pyrrhotite in LAP 04840 (Table 5; Figures 7 and 8, and Fig. S4) are similar to those of pentlandite in NWA 11304. The superchondritic abundances of most siderophile elements in pyrrhotite of LAP 04840 are probably due to the ubiquitous exsolution lamellae of pentlandite throughout pyrrhotite grains, which could not be avoided during laser ablation, given the larger volume of ablated material. Most HSE concentrations are~109 CI, apart from Pt, which is depleted by an order of magnitude relative to Os in pentlandite and by 2 orders of magnitude in pyrrhotite. Tungsten is also depleted by an order of magnitude relative to Os. In contrast to NWA 11304, LAP 04840 sulfides have superchondritic concentrations of Mo (5-99 CI) and Au (up to 7.89 CI) on average.

Siderophile Trace Element Concentrations in LAP 04840 Magnetite
Average siderophile trace element concentrations in LAP 04840 magnetite range from subchondritic (0.039 CI) to >209 CI (Table 5; Figure 8) but are highly variable among individual measurements (Figure 7). On average, most siderophile elements are present in roughly chondritic concentrations with several notable exceptions. Highly siderophile elements Re, Os, Ir, Ru, Pt, Rh, and Au are roughly chondritic (0.3-2.59 CI), while Pd is subchondritic (0.039 CI). Average concentrations of moderately siderophile elements range from subchondritic Cu (0.159 CI) to strongly superchondritic Ga and W (14 and 299 CI, respectively). Concentrations of W in magnetite span over 4 orders of magnitude, resulting in average W concentrations that are~1009 greater than coexisting sulfides (Fig. S4).

Metamorphism
Most siderophile elements follow a weak trend from subchondritic concentrations in pyrrhotite to superchondritic concentrations in pentlandite with intermediate trace element concentrations in Ni-rich pyrrhotite ( Figure 5). As pentlandite exsolves from pyrrhotite, coarsens, and approaches equilibrium from petrologic types 3 and 4 to type 5, trace element concentrations homogenize in each sulfide phase, typically with the highest concentrations in pentlandite.
While the majority of siderophile elements in R chondrites partition into pentlandite in proportions similar to OC metals (e.g., Campbell & Humayun, 2003;Kong & Ebihara, 1997), some trace elements defy their typical mineralogic affinities and/or volatilities based on nebular condensation temperatures (e.g., Lodders, 2003;Wood et al., 2019), indicating that siderophile trace element systematics in oxidized, sulfide-dominated Fe-Ni-S systems involve distinct geochemical mechanisms that deviate from metal-dominated counterparts.  (Isa et al., 2014), although NWA 11304 is only moderately weathered (W2; Gattacceca et al., 2020). Uncertainties for metal averages are <5% based on counting statistics (OC metal data are from Kong & Ebihara, 1997), and uncertainties for individual R chondrite measurements are smaller than their symbols. CI normalization from McDonough and Sun (1995). (Color figure can be viewed at wileyonlinelibrary.com.)

Anomalous Highly Siderophile Element Behaviors
Within sulfides, CI-normalized Pt concentrations are depleted by an order of magnitude relative to Os, Ir, Ru, Rh, and Pd ( Figure 6). Fractionation of Pt relative to other HSEs cannot be explained through nebular fractionation via evaporation or condensation (Lodders, 2003) nor through typical igneous fractionation on the parent body (e.g., Chabot & Jones, 2003;Day et al., 2012Day et al., , 2016Walker, 2016). Bulk measurements for other R chondrites, in contrast, have found chondritic concentrations of Pt (Isa et al., 2014). In NWA 11304, discrete Pt-rich arsenides and sulfarsenides (1-10 lm) are often found in association with sulfides in clasts of all petrologic types, whereas Pt-rich metal alloys are found in sulfides, within chondrules, and among matrix assemblages. These phases can explain the Pt depletions among sulfides and account for chondritic bulk compositions. Based on their mineralogic associations in NWA 11304, platinum-group metals are consistent with products of nebular condensation (e.g., Wood et al., 2019) that were incorporated into the R chondrite parent body during accretion. Sperrylite (PtAs 2 ), platarsite (PtAsS), irarsite (IrAsS), and other noble metal alloys found in NWA 11304 are not phases predicted to condense from the solar nebula, but instead probably crystallized directly from sulfides melted during chondrule formation (Helmy et al., 2013;Helmy & Bragagni, 2017;Miller et al., 2017). Throughout subsequent metamorphism, these phases may continue to grow as Pt and other siderophile elements diffuse from the surrounding sulfides, resulting in growth of platinum group element-rich phases up to the largest (~10 lm) grain of niggliite (PtSn) observed in the type 5 clast of NWA 11304.
Gold is virtually absent from most sulfides in NWA 11304. Previous work has identified a correlation between gold and weathering grade (Isa et al., 2014), which may imply that trace element concentrations within sulfides in NWA 11304 have been subjected to terrestrial alteration. However, this stands in contrast to the observed low weathering grade of NWA 11304 sulfides. The correlation between Re and Os concentrations among sulfides in NWA 11304 (Fig. S3) is also consistent with minimal terrestrial alteration, as Re is readily leached from sulfides during weathering (e.g., Hyde et al., 2014). Small (~1 lm) noble metal grains that are rich in Au, Ag, and Fe are detectable via EDS in the petrologic type 5 clast. These grains probably formed due to incompatibility with MSS crystallizing from liquid sulfide, similar to the petrogenesis of noble metal grains in terrestrial massive sulfide ores (e.g., Barnes et al., 2006). Noble metal grains, arsenides, and sulfarsenides have been identified in clasts of the only known R chondrite fall, Rumuruti (Schulze et al., 1994), providing further evidence that the atypical distribution of siderophile trace elements in R chondrites is not the result of terrestrial alteration.

Redox-Controlled Moderately Siderophile Element Distribution
The high ƒO 2 of R chondrites is manifest in the redox-sensitive moderately siderophile trace element (MSE) concentrations in sulfides and oxides. Average concentrations of W in NWA 11304 pentlandite are depleted relative to other refractory siderophile elements and compared to concentrations in ordinary chondrite metals ( Figure 6). W/Os ratios decrease inversely with redox state from metals in H, L, and LL to R chondrite sulfides (W/ Os = 1.4, 1.4, 0.96, and 0.17, respectively). The equal distribution of W among R chondrite sulfides and silicates is consistent with equilibration near the W-WO 2 oxidation buffer, which coincides with the IW buffer up to~1400°C at 1 bar (O'Neill & Pownceby, 1993) and is at the less oxidized end of the R chondrite ƒO 2 range (Righter & Neff, 2007). Similarly, the subchondritic average concentration of Mo in type 5 pentlandite follows an oxidation trend from OC metals to R chondrite sulfides where Mo/Ru decreases in sequence  (Hillgren, 1991;Holzheid et al., 1994). These interpretations are supported by the atomic Fe:S ratio of pyrrhotite in NWA 11304 (0.91), which is consistent with sulfide equilibration near the IW oxidation buffer (Schrader et al., 2021).

Comparison to Ordinary Chondrites
Siderophile element systematics among sulfides during R chondrite metamorphism share some similarities with observations for ordinary chondrite metals. Some variability in trace element concentrations remains among individual grains even at high petrologic type ( Figures 5  and 7), indicating that equilibration of siderophile elements is localized within individual grains (Gilmour & Herd, 2020). However, the two processes deviate regarding FIGURE 7. Log-log plots of CI-normalized Ni vs. HSE concentrations in LAP 04840 sulfides and oxides demonstrate distinct trace element partition during hydrothermal alteration when compared to anhydrous NWA 11304. The variable and high concentrations of HSEs in magnetite are most likely due to oxidation of sulfides during hydrothermal alteration, which fueled the rapid growth of magnetite and disrupted equilibrium between pentlandite and pyrrhotite. As a consequence, siderophile trace elements are distributed between the pentlandite and pyrrhotite in roughly equal proportions. Only measurements above detection limits are illustrated. All measurements are provided in supplementary data. CI normalization is from McDonough and Sun (1995).  (Kerridge, 1976;Lauretta et al., 1998;Singerling et al., 2021).

Hydrothermal Alteration
The high Cr 2 O 3 content in LAP 04840 magnetite could indicate that it formed after hydrothermal alteration of chromite, similar to what has been observed in some terrestrial hydrothermal systems (e.g., Holwell et al., 2017). Magnetite growth would have been further fueled through oxidation of Fe sulfides during hydrothermal alteration. Within magnetite-sulfide assemblages in LAP 04840, pyrrhotite contains fine-scale exsolution lamellae of pentlandite ( Figure 4b) consistent with latestage oxidation seen in terrestrial Fe sulfides during hydrothermal alteration (e.g., Terranova et al., 2018). This scenario is viable for LAP 04840 if the oxidizing hydrothermal fluid was introduced while the precursor was still hot (McCanta et al., 2008) and had already formed some amount of chromite through metamorphism (Isa et al., 2014). The smooth, rounded, and sometimes granoblastic textures of many magnetite grains in LAP 04840 are clearly distinct from the porous, irregular magnetite grains produced through lowtemperature aqueous alteration in many carbonaceous chondrites (e.g., Singerling & Brearley, 2020). Instead, the textures of magnetite in LAP 04840 are more similar to those found in CK chondrites (e.g., Righter & Neff, 2007), which further supports the hypothesis that magnetite in LAP 04840 formed either during or prior to metamorphism (McCanta et al., 2008), rather than through subsequent aqueous alteration.
Higher oxidation state also accounts for the higher Ni concentrations of LAP 04840 sulfides and silicates (McCanta et al., 2008) relative to those in NWA 11304 (e.g., Kerridge, 1976). As some fraction of Fe 2+ in pyrrhotite oxidized to Fe 3+ and partitioned into the growing magnetite, the residual sulfides became more Ni-rich, exsolving pentlandite lamellae from Ni-rich pyrrhotite upon cooling during retrograde metamorphism. This scenario can explain the textural differences between sulfides of LAP 04840 and NWA 11304, despite both containing assemblages with a type 5 and type 6 petrologic classification. Due to the late-stage growth of magnetite, LAP 04840 sulfides were not provided with the opportunity to coarsen throughout the duration of metamorphism.
In contrast to NWA 11304, LAP 04840 siderophile elements partition equally between pyrrhotite and pentlandite. This can be explained by the positive correlation observed between Ni and siderophile trace element concentrations in sulfides ( Figure 5); Ni-rich pyrrhotite and pentlandite in LAP 04840 are both capable of hosting superchondritic concentrations of siderophile elements. Magnetite also contains variable concentrations of siderophile trace elements that can span several orders of magnitude (Figure 7). The highly variable concentrations of siderophile elements in magnetite grains and similar distribution of trace elements between pentlandite and pyrrhotite are consistent with magnetite growth from the late-stage oxidation of pyrrhotite, yielding Ni-enriched pyrrhotite and magnetite that equilibrated with pre-existing, trace element-rich pentlandite.

Consequences for Core Formation Ages
Recently, multiple studies have used the Hf-W isotopic system to calculate the timing of metal-silicate segregation (i.e., core formation) in meteorite parent bodies (e.g., Kleine et al., 2009). These studies follow the assumption that lithophilic 182 Hf decays to siderophilic 182 W, with the change in mineralogic affinities between parent and daughter isotopes being used to calculate core formation ages by measuring excess 182 W (e.g., Lee & Halliday, 1995). However, due to its high oxygen fugacity, approximately 90% of W resides in the silicates of NWA 11304 by petrologic type 5 (Table S6). For similarly oxidized materials, the distribution of W would affect the ages calculated from the Hf-W isotopic system, which rely upon the siderophile affinity of W in ƒO 2 environments below IW-1 (D W metal-silicate = 2-600; Schmitt et al., 1989). Even for the comparatively reduced H chondrites (~IW-2; Kessel et al., 2004), Hf-W isotopic measurements yield disparate ages for metal and nonmetal fractions, interpreted as the result of a small fraction of W in silicates FIGURE 9. Schematic of redox-based differentiation pathways. Prior to melting, sulfide-dominated opaque assemblages composed of pentlandite and pyrrhotite coarsen and discretize throughout metamorphism, although equilibria between the two sulfides may be disrupted by magnetite growth during hydrothermal alteration. In contrast to metallic liquids in reduced systems, the anion/cation ratios of sulfide assemblages at high ƒO 2 permit the propagation of sulfide liquids via grain wetting prior to silicate melting. The observed distribution of siderophile elements between sulfide phases can be used to establish starting compositions for melting experiments to investigate the potential for sulfide-dominated core formation on oxidized parent bodies and search for potentially related iron meteorites via characteristic trace element abundances. Chondritic precursor examples are colored according to their isotopic grouping (blue = carbonaceous, red = noncarbonaceous), and demonstrate that oxidation state is not universally correlated with isotopic grouping, which is consistent with interpretations of chondritic sulfides (Schrader et al., 2021) but stands in contrast to conclusions drawn from some iron meteorites (e.g., Hilton et al., 2022). a Backscatter electron (BSE) image of a metal grain in the Guarana (H6) ordinary chondrite adapted from Reisener and Goldstein (2003). b Fe,Ni-FeS melting temperature and diagram for incipient melting of metal adapted from McCoy et al. (2006). Relative redox conditions of chondrite parent bodies are interpreted from Righter et al. (2016). (Color figure can be viewed at wileyonlinelibrary.com.) due perhaps to closure temperature effects or disequilibrium between metal and matrix phases (Archer et al., 2019). In highly oxidized, sulfur-rich chondrites that formed at or above the IW oxidation buffer, such complications with age calculations would be further exacerbated by the high concentration of W in silicates. A large excess 182 W may not develop in the silicate fraction due to high initial concentrations of W, resulting in age estimates for core-mantle segregation that are too old for the differentiated silicate fractions. Similarly, the retention of radiogenic 182 W in silicates could result in erroneously young age calculations for metals that crystallized from Fe-Ni-S melts. Some of these complexities may be mitigated with Hf-W isotopic measurements for R chondrites that demonstrate the initial distribution of W isotopes between sulfides and other phases prior to melting, similar to recent measurements reported for H chondrites (e.g., Archer et al., 2019). The effects of ƒO 2 on the Hf-W system may be reflected in the isotopic dichotomy of iron meteorites, as previous work has proposed that the CC isotopic group of iron meteorites crystallized within more oxidized parent bodies than the NC iron meteorites (Hilton et al., 2022). Investigations into the effects of ƒO 2 on the initial distribution of redoxsensitive W will facilitate calculation of the ages of core formation among iron meteorites that formed at high ƒO 2 .

Melting of Sulfide-Magnetite Assemblages
If LAP 04840 was hydrothermally altered as a result of impact mixing (McCanta et al., 2008), then the effects of hydrothermal alteration on its opaque mineral assemblage would not be directly applicable to core formation. However, evidence for hydrothermal alteration of R chondrites is limited to petrologic type R6 samples (Gattacceca et al., 2020). In the impact mixing scenario, evidence of hydrothermal alteration should be independent of petrologic type. Thus, the possibility remains that hydrothermal alteration is a thermally regulated, high-temperature metamorphic process that did not affect petrologic type 3 and type 4 regions of the R chondrite parent body. Regardless of the source of hydrothermal fluid, the oxidation of pyrrhotite to magnetite further disturbs the distribution of siderophile elements prior to core formation by increasing the Ni concentration in the remaining pyrrhotite and enhancing its capacity for hosting siderophile trace elements. Given the range of reported ƒO 2 for R chondrites from IWÀ1 to IW + 5 (~FMQ; McCanta et al., 2008;Righter & Neff, 2007;Righter et al., 2016), a coordinated analysis of oxygen fugacities and sulfide/oxide ratios across the R chondrite group would help to quantify the range of initial ƒO 2 conditions and the effects on opaque mineral proportions and their siderophile trace element systematics.

Meteoritic Evidence for Oxidized, Sulfide-Dominated Core Formation
Trace element concentrations within the sulfides of the most oxidized primitive achondrites, that is, brachinites, are consistent with loss of a Fe-Ni-S liquid (Crossley et al., 2020;Nehru et al., 1983). Currently, there are no corresponding sulfide-dominated meteorites reported in the Meteoritical Bulletin (Gattacceca et al., 2020) akin to a proposed Fe-Ni-S core. However, anomalous Ni-rich iron meteorites like Oktibbeha County may be derived from the minor metal fraction that can crystallize from S-rich Fe-Ni-S liquids (Wasson et al., 1980), whereas the corresponding solidified sulfides did not survive, perhaps due to mechanical weakness (Kracher & Wasson, 1982).
The products of oxidized, sulfur-rich core formation may be identified among the meteoritic record, similar to how magmatic iron meteorites sample the cores of planetary embryos (e.g., Haack & McCoy, 2007). If Nirich irons like Oktibbeha County are products of highly oxidized, sulfide-dominated core formation, their trace element chemistries must be reflective of the parent sulfide liquids. In contrast to typical HSE partitioning between liquid and solid Fe,Ni metal (e.g., Chabot & Jones, 2003;Malvin et al., 1986;Walker, 2016), if HSEs fractionate into melting pentlandite, a R chondrite-like precursor would lose effectively all of its HSE content to the first melts produced at 865°C, and siderophile trace elements would continue to follow Ni during the crystallization of awaruite from a sulfide melt. Consequently, Ni-rich metals should still contain superchondritic concentrations of HSEs. Given the incompatibilities of Au and Pt with pentlandite (e.g., Barnes et al., 2006), awaruite may be correspondingly enriched in these elements relative to other HSEs. Depletions in redox-sensitive W and Mo relative to other MSEs may also be characteristic of awaruite crystallization from a highly oxidized, sulfidedominated liquid, as these elements would be sequestered in silicate phases during liquid sulfide segregation. Further investigation into the petrogenesis of Ni-rich iron meteorites will require experimental distribution coefficients for siderophile elements in pentlanditepyrrhotite-awaruite systems.

Application to Upcoming NASA Missions
NASA's Psyche mission is currently planned to launch between 2023 and 2024 and will orbit the M-type asteroid 16 Psyche in 2029-2030. Once believed to be an exposed metallic planetary core, recent density estimates (3.4-4.1 g cm À3 ) fall below the densities of Fe,Ni metals (~8 g cm À3 ; Smyth & Mccormick, 1995), requiring alternative explanations for its formation (Elkins-Tanton et al., 2020). Fe,Ni sulfides have densities (4-5 g cm À3 ) that are much closer to estimates for 16 Psyche. The possibility that 16 Psyche is a core that is composed primarily of sulfides rather than metals is among potential explanations for its estimated low density (Bercovici et al., 2022;Elkins-Tanton et al., 2020) and high radar albedo measurements (e.g., Shepard et al., 2017).
Generally, oxidized bodies are assumed to produce smaller cores than reduced counterparts, as the oxidation of Fe metal to FeO leads to enhanced Fe incorporation into silicates, which decreases the volume of iron available to participate in core formation (e.g., Rubie et al., 2015). However, the initial concentration of S complicates this assumption through the formation of Fe-sulfides (e.g., Bercovici et al., 2022). The highly oxidized, sulfur-rich R chondrites contain 5-8 vol% sulfides (Bischoff et al., 2011), similar to the proportions of metals in L and LL chondrites (e.g., Jarosewich, 1990). Thus, Psyche could be a sulfide-dominated core if it formed through efficient segregation of 8 vol% sulfides from an R chondrite-like precursor with a minimum diameter of~560 km (Table S7). The addition of 10 vol% silicates and < 15 vol% void space would fully account for Psyche's low bulk density.
The Psyche mission will be able to evaluate whether 16 Psyche's surface is consistent with an oxidized, sulfidedominated core via the onboard gamma ray and neutron spectrometer, which will be capable of measuring average Fe, Ni, and S concentrations. Equilibrium Fe-Ni-S phase diagrams (e.g., Figure 10) show that the phases crystallizing from a Fe-Ni-S liquid at temperatures below 865°C will be a mixture of pentlandite and awaruite (Fe 0.4 Ni 0.6 ) AE MSS (Kitakaze et al., 2016). Fe/Ni ratios for the segregated sulfide liquid and its major components (awaruite and pentlandite) are ≤1, whereas Fe/Ni ratios of metal/troilite assemblages and potential immiscible sulfide liquids are >1 in more reduced magmatic irons (e.g., Goldstein et al., 2009) and the troilite-dominated cores predicted to form after complete melting of oxidized carbonaceous chondrites (Bercovici et al., 2022).

CONCLUSIONS
Sulfide-dominated core-forming components in oxidized chondrites (ƒO 2 ≥ IW) follow metamorphic pathways that are distinct from reduced, metal-bearing counterparts. While metamorphic alterations of R chondrite sulfides share some textural and geochemical affinities with ordinary chondrite metals, the melts generated from oxidized sulfides can migrate more readily. In addition, the behaviors of several key siderophile trace elements among oxidized sulfides deviate from reduced metallic systems, which lead to compositionally distinct products of oxidized, sulfidedominated core formation that can be identified in meteorites and asteroids.
In summary, specific conclusions from this work include the following: • R chondrite sulfide assemblages of pentlandite and pyrrhotite coarsen throughout metamorphism in a process analogous to the metamorphism of metal and sulfide assemblages in ordinary chondrites. • The low melting temperature of pentlandite and the wetting properties of sulfide liquids could result in silicate-sulfide fractionation and possibly core formation in oxidized, S-rich bodies at lower temperatures (865-1150°C) before the onset of silicate melting. FIGURE 10. Fe-Ni-S ternary (in atom%) for oxidized and reduced cores. The red field is the compositional range for components of reduced cores and the purple field represents compositions that can crystallize from Fe-Ni-S liquid. The liquid Fe-Ni-S produced at 875°C in oxidized assemblages crystallizes pentlandite, MSS, Ni-rich metals (Ni atom% 55-60), whereas reduced bodies can only crystallize Ni-poor MSS and metals with Ni concentrations up to 36 atom%. The gamma ray and neutron spectrometer on NASA's Psyche spacecraft will analyze Fe, Ni, and S. If asteroid 16 Psyche has a Fe/Ni ratio of <1, then that would provide strong evidence that it is the core of a highly oxidized and sulfur-rich parent body. Magmatic iron compositions are bulk values for grouped irons in Goldstein et al. (2009). Subsolidus stability fields are drawn from Kitakaze et al. (2016). Oktibbeha County is from Wasson et al. (1980). (Color figure can be viewed at wileyonlinelibrary.com.) • Siderophile trace element concentrations in R chondrite pentlandite are typically superchondritic (~109 CI), but W, Mo, Pt, Ga, and Ge are depleted by factors of 2-25 relative to elements with similar volatilities. Platinum forms discrete arsenide and sulfarsenide phases, while the remaining W, Mo, Ga, and Ge partition into silicates and oxides. The complexity of siderophile trace element systematics among oxidized sulfides stands in contrast to their typical partitioning behaviors among metals. • The bulk sulfide cores of oxidized, sulfur-rich parent bodies may reflect Mo, W, Ga, and Ge depletions of their precursor sulfides, and Ni-rich metals that crystallize within them. These metal fractions may be represented by anomalous iron meteorites like Oktibbeha County. • The sequestering of W into silicate fractions can result in erroneously young isotopic age calculations for silicate assemblages and overestimates of ages for the core forming fraction in highly oxidized parent bodies. • Calculated densities for asteroid 16 Psyche are much closer to the densities of sulfides than to densities of Fe,Ni metals. The Psyche mission's gamma ray and neutron spectrometer will be able to measure a suite of elements that includes Fe, Ni, and S. Fe/Ni ratios ≤1 would provide strong evidence that Psyche is an oxidized, sulfide-dominated core.

SUPPORTING INFORMATION
Additional supporting information may be found in the online version of this article.
Data S1. Supporting data. Table S1. Major element oxides (wt%) for nonopaque mineral endmembers listed in Table 1. Table S2. Modal mineralogy calculations. Table S3. EPMA corrective factors. Table S4. Detection limits for EPMA measurements and standards used for respective mineral phases. Table S5. External precision (2r, relative % deviation) of standard measurements and detection limits used for LA-ICP-MS data processing. Table S6. Average siderophile and chalcophile trace element concentrations (ppm) in non-opaque mineral phases. Table S7. Calculations for size estimate of Psyche's precursor parent body. Fig. S1. Mineral classification map of NWA 11304. Fig. S2. Diagram of clasts in NWA 11304 and their assigned petrologic types. Fig. S3. Individual CI-normalized measurements for siderophile and chalcophile trace elements versus Ni in NWA 11304 sulfides. Fig. S4. Individual CI-normalized measurements for siderophile and chalcophile trace elements versus Ni in LAP 04840 sulfides and oxides.