¹Shunpei Yokoo,^1,2Kei Hirose
Geochimica et Cosmochimica Acta (in Press) Open Access Link to Article [https://doi.org/10.1016/j.gca.2024.06.027]
¹Department of Earth and Planetary Science, The University of Tokyo, Hongo, Tokyo 113-0033, Japan
²Earth-Life Science Institute, Tokyo Institute of Technology, Meguro, Tokyo 152-8550, Japan
Copyright Elsevier

Recent seismological studies of the Martian core revealed its relatively low density, suggesting the presence of large amounts of light elements including oxygen and carbon in addition to sulfur. In order to reveal crystallizing solids in the light-element-rich core of Mars, we performed high-pressure melting experiments on Fe-S-O-C alloys at 26–49 GPa using a laser-heated diamond-anvil cell. The liquidus phase relations in the Fe-S-O-C system were determined based on textural and chemical characterizations of recovered samples. The results show that Fe-S-O-C liquids crystallize FeO or Fe₃C in the presence of small amounts of O or C in liquids. Accordingly the liquidus fields of Fe₃S and Fe₂S are small, and the quaternary eutectic point is found to be close to the Fe-Fe₃S binary eutectic point. Under Martian core conditions, S-rich liquids have low liquidus temperatures to crystallize FeO or Fe₃C compared to S-poor liquids. The pressure dependence of liquidus temperatures suggests that crystallization of Mars’ core starts at the center upon cooling. According to the FeO-Fe₃C cotectic surfaces and their liquidus temperatures, an FeO and/or Fe₃C inner core is predicted unless the Martian core remains entirely liquid.

¹Hiroharu Yui et al. (>10)
Geochimica et Cosmochimica Acta (in Press) Open Access Link to Article [https://doi.org/10.1016/j.gca.2024.06.016]
¹Department of Chemistry, Tokyo University of Science, Tokyo 162-8601, Japan
Copyright Elsevier

The surface chemistry of pyrrhotites from intact particles directly collected from asteroid (162173) Ryugu was investigated by micro-Raman spectroscopy. The Raman peak characteristic to pyrrhotite was observed at around 115 cm−1 in Ryugu pyrrhotites, similar to freshly cleaved surfaces of terrestrial pyrrhotites. Additional Raman bands centered at around 220, 275, and 313 cm−1 with broadened features were also detected from the Ryugu pyrrhotites. The set of Raman bands at 220 and 275 cm−1 was assigned to typical Fe-S stretching vibrations of ν2 (225 cm−1) and ν1 (275 cm−1). These bands are not clearly observed in bulk crystals of pyrrhotite but appear in its nanoparticulate phase. These bands are ordinarily seen in amorphous monosulfides that formed under low oxygen fugacity (fO2) conditions in nature, indicating that the structural alteration of pyrrhotite surfaces occurred heterogeneously on the nanoscale under low fO2 conditions. Further, the Raman band at 313 cm−1 was attributed to a characteristic tetrahedral bonding of Fe(III) in the lattice of FeII1-3xFeIII1-2xS, followed by the local breakdown of the crystal lattice structures from planar bonding with Fe(II). In addition, some areas of the Ryugu pyrrhotite grains showed corroded structures with iridescence. Furthermore, assemblages of magnetite particles were also preferentially observed on small areas of the likely-dissolved pyrrhotite crystals in phyllosilicate matrices. These characteristic features in the Raman spectra and in corroded structures of Ryugu pyrrhotites record changes in the local environmental conditions via aqueous alteration. The corrosion of pyrrhotite crystals followed by the preferential formation of magnetite particles by asteroidal water it the likely product of dissolution of Fe(II) from the pyrrhotite surface and its oxidative precipitation in microchemical environments on the Ryugu parent body.

¹Mingming Zhan,¹Kohei Fukuda,²Michael J. Tappa,³Guillaume Siron,²William O. Nachlas,⁴Makoto Kimura,¹Kouki Kitajima,²Ann M. Bauer,¹Noriko T. Kita
Geochimica et Cosmochimica Acta (in Press) Link to Article [https://doi.org/10.1016/j.gca.2024.06.018]
¹WiscSIMS, Department of Geoscience, University of Wisconsin–Madison, Madison, WI 53706, USA
²Department of Geoscience, University of Wisconsin–Madison, Madison, WI 53706, USA
³Laboratoire Chrono-Environnement, Université de Franche-Comté, UMR 6249, 25000 Besançon, France
⁴National Institute of Polar Research, Meteorite Research Center, Midoricho 10-3, Tachikawa, Tokyo 190-8518, Japan
Copyright Elsevier

The mechanism of gas-melt interactions and the compositions of precursors are key to understanding the formation of chondrules. To shed light on the two enigmas, we studied the petrography, chemistry, and oxygen isotopes of six Al-rich chondrules (ARCs, five glassy and one plagioclase-bearing) in unequilibrated ordinary chondrites (OCs, petrologic subtype: 3.05). The plagioclase-bearing ARC was also investigated with Al-Mg chronology. Elemental zonation and inter-element correlations in glassy mesostasis of two ARCs indicate the condensation of gaseous Mg, SiO, Fe, and Na onto chondrule melt. The plagioclase-bearing ARC appears to display internal mass-independent oxygen isotope fractionation with δ18O increasing following the order of mineral crystallization, suggesting partial oxygen isotope exchange with ambient gas during crystallization. Oxygen isotopes of the six ARCs are distributed along a mixing line of slope = 0.99 ± 0.05, which intersects with calcium-aluminum-rich inclusions (CAIs), consistent with a small portion of OC type IA chondrules, but deviates from other OC ferromagnesium chondrules (FMCs) towards higher δ17O, suggesting that OC ARCs and some IA chondrules were established by interactions between CAI-like melts and 16O-poor ambient gas, rather than simply remelting solid mixtures of CAI and FMC materials.

All ARCs have unfractionated refractory lithophile element patterns with bulk concentrations ranging from ∼7 × CI to ∼15 × CI, indicating ∼ 30–100 % of CAI-like materials in their precursors. Their bulk compositions are linearly evolved toward the Mg: SiO ∼ 3:2 to 2:1 (in atomic) apex, consistent with adding gaseous Mg and SiO to the chondrule bulk via gas–melt interactions. The back-calculated compositions of the recycled CAI-like materials closely overlap with pyroxene-anorthite-rich CAIs, suggesting that extensive interactions between the melt of pyroxene-anorthite-rich CAI-like materials and ambient gas could make OC ARCs. The Al-Mg age of the plagioclase-bearing ARC is ∼2.2 Ma after CAIs, similar to typical OC FMCs, suggesting that the refractory component arrived in the OC reservoirs at the end of the chondrule heating events.

^1,2Qingshang Shi et al. (>10)
Geochmica et Cosmochimica acta (in Press) Link to Article [https://doi.org/10.1016/j.gca.2024.06.011]
¹State Key Laboratory of Geological Processes and Mineral Resources, Frontiers Science Center for Deep-time Digital Earth, China University of Geosciences, Beijing 100083, China
²School of Geophysics and Information Technology, China University of Geosciences, Beijing 100083, China
Copyright Elsevier

To investigate the chemical variation during space weathering of young mare basalts, here we report elemental, radiogenic Sr-Nd and stable Fe-Mg-Ca isotopic data of Chang’E-5 sieved soils and breccias. From the coarse fraction to the fine one, the sieved soils display increasing Al2O3 (10.34 wt%–13.36 wt%) and Sr (248 ppm–307 ppm) but decreasing FeO (23.50 wt%–20.22 wt%), MgO (6.88 wt%–5.78 wt%), FeO/Al2O3 (2.27–1.51) and MgO/Al2O3 (0.67–0.43). The contents of rare earth elements (except Eu) and high field strength trace elements do not vary with particle size but correlate with P2O5 contents. Given the limited contribution from contamination by meteorites and exotic materials ejected far away from the landing site, these elemental variations can be explained by differential comminution and distribution behaviors of plagioclase and mesostasis phases. These sieved soils yield a Sm-Nd isochron age (1.84 ± 0.83 Ga) comparable to that of basaltic clasts obtained by U-Pb dating (∼2.0 Ga). However, their Rb-Sr isotopic system is disturbed as indicated by their relatively homogeneous 87Sr/86Sr (0.701425–0.701592) despite variable Rb/Sr (0.017–0.028). These results suggest the Sm-Nd isotopic system is more robust to impact disturbance during space weathering compared to the Rb-Sr isotopic system. Given that the bulk soil still plots on the 2.03 Ga Rb-Sr reference isochron from the pristine plagioclases in CE-5 basalts, this disturbance did not affect the Rb-Sr isotopic system on the bulk scale. The CE-5 bulk soil has higher Mg# (33.6), 87Rb/86Sr (0.06) and present-day 87Sr/86Sr (0.701542) than the mean composition of reported basaltic clasts (Mg#: ∼28; 87Rb/86Sr: ∼0.038; 87Sr/86Sr: ∼0.700941), possibly implying that the bedrocks in CE-5 landing site consist of multiple magma pulses. The δ56Fe (0.122 ± 0.002 ‰ to 0.199 ± 0.008 ‰) and δ26Mg (−0.204 ± 0.016 ‰ to −0.109 ± 0.006 ‰) of sieved CE-5 soils increase with decreasing particle sizes but their δ44/42Ca (0.38 ± 0.04 ‰ to 0.44 ± 0.02 ‰) are relatively homogeneous. Mass balance modelling indicates that differential comminution has limited influence on the Fe-Mg-Ca stable isotopic compositions. We further dismiss the role of solar-wind sputtering, as Ca and Mg are more susceptible to sputtering and thus would be expected to show larger isotope fractionations compared to Fe, which is inconsistent with the observations. Free evaporation may explain the elevated δ56Fe and δ26Mg in fine fractions at given very limited depletion in FeO and MgO. The observed positive correlation between δ56Fe and δ26Mg, however, is much steeper than the slope expected for free evaporation, indicating also other mechanisms (e.g., Fe-Mg inter-diffusion). Since the CE-5 soil has a unique composition compared with Apollo and Luna soils, the chemical differentiation identified in this study provides new insights for establishing a connection between the chemistry and reflectance spectral properties of lunar soil. Our combined Fe-Mg-Ca isotopic study also provides a paradigm to distinguish the role of solar-wind sputtering and impact evaporation, and shows that the inter-particle diffusion process may be an important mechanism for the isotope fractionation among lunar soil components.

^1,2Yankun Di,²Qing-Zhu Yin,³François L.H. Tissot,^1,4Yuri Amelin
Geochimica et Cosmochimica Acta (in Press) Open Access Link to Artile [https://doi.org/10.1016/j.gca.2024.06.012]
¹Research School of Earth Sciences, Australian National University, Acton, ACT 2601, Australia
²Department of Earth and Planetary Sciences, University of California Davis, Davis, CA 95616, USA
³The Isotoparium, Division of Geological and Planetary Sciences, California Institute of Technology, Pasadena, CA 91125, USA
⁴Korea Basic Science Institute, Building 202, 162 YeonGu DanJi-ro, Ochang, Cheongwon, Cheongju, Chungbuk 28119, Republic of Korea
Copyright Elsevier

We introduce a new isotope chronological model in which the natural mass-dependent isotopic fractionation effects of the radioactive (“parent”) and radiogenic (“daughter”) elements are systematically and rigorously considered. Using this model, we show that internally-normalized radiogenic isotopic ratios, commonly determined for daughter elements such as Sr, Nd, Cr, Ni, Hf, W, and Os, are dependent on the extent of natural isotopic fractionation of the daughter and parent elements at the time of system closure. This dependence indicates that (1) in two samples derived from the same isotopically homogeneous source at the same time and with identical radiogenic ingrowth over time, the present-day internally-normalized radiogenic isotope ratios would be different if they were initially fractionated to different degrees, and (2) if different internally-normalized radiogenic isotopic ratios are observed for two co-genetic objects, the difference between them would include contributions from both radiogenic ingrowth and natural isotopic fractionation. Consequently, the isochron dating equations employed in traditional chronological studies will yield inaccurate results when significant natural isotopic fractionation are present among the studied samples. Modified isochron equations that can be used to retrieve correct chronological information from isotopically-fractionated samples are presented. These theoretical considerations are applied to the 87Rb–87Sr, 147Sm–143Nd, and 146Sm–142Nd isotope systems of calcium–aluminium-rich inclusions (CAIs), a set of samples that have undergone significant natural Sr, Nd, and Sm isotope fractionation during their formation. The large natural Sr isotope fractionation (up to ca. 5.3 ‰ for 88Sr/86Sr) in fine-grained CAIs can generate analytically well-resolvable biases (>120 ppm) in the internally-normalized 87Sr/86Sr ratios and lead to significant scatters of their 87Rb–87Sr isochron (in conjunction with scatters induced by open-system disturbances). The 87Rb–87Sr systems of coarse-grained CAIs, on the contrary, are essentially not affected by natural Sr isotopic fractionation due to their much subdued fractionation degrees, resulting in a more robust isochron. Similarly, the large natural Nd (up to ca. 4.0 ‰ for 146Nd/144Nd) and Sm (up to ca. 7.1 ‰ for 152Sm/148Sm) isotopic fractionation in fine-grained CAIs can induce significant scatters of the 147Sm–143Nd isochron if the natural fractionation followed the kinetic or power law, and 146Sm–142Nd isochron if the natural fractionation followed the equilibrium, Rayleigh, or power law. This implies that when studying radioactive isotope systems in objects whose daughter and parent elements can undergo significant isotope fractionation in nature, accompanying stable isotope analyses are necessary for accurate chronological interpretations.

¹Le Zhang,¹Cheng-Yuan Wang,²Hai-Yang Xian,¹Jintuan Wang,¹Yan-Qiang Zhang,³Zhian Bao,¹Mang Lin,¹Yi-Gang Xu
Geochimica et Cosmochimica Acta (in Press) Link to Article [https://doi.org/10.1016/j.gca.2024.06.019]
¹State Key Laboratory of Isotope Geochemistry and CAS Centre for Excellence in Deep Earth Science Guangzhou Institute of Geochemistry, Chinese Academy of Sciences, Guangzhou 510640, China
²CAS Key Laboratory of Mineralogy and Metallogeny/Guangdong Provincial Key Laboratory of Mineral Physics and Materials, Guangzhou Institute of Geochemistry, Chinese Academy of Sciences, Guangzhou 510640, China
³State Key Laboratory of Continental Dynamics, Department of Geology, Northwest University, Xi’an 710069, China
Copyright Elsevier

Impact glasses are abundant in the lunar regolith, and Mg isotopes have the potential to trace components from various lunar crustal reservoirs, which have recently been shown to exhibit large Mg isotopic fractionations. However, it remains unclear whether Mg isotopic fractionation occurs during the formation of impact glasses. In this study, we report in situ Mg isotopic and elemental compositional data for agglutinatic glasses returned by the Chang’e 5 mission and obtained using the laser ablation split stream technique. Vesicular textures, Fe–Ni alloys, tiny Fe droplets, and high Ni contents suggest the studied agglutinatic glasses had an impact origin. The agglutinatic glasses exhibit large Mg isotopic fractionation, with δ26Mg values ranging from −1.36 ‰ to −0.01 ‰. The lack of correlations between δ26Mg values, Ni contents, and ratios between volatile and relatively refractory elements (K/La, Rb/Sr, and Ce/Pb) indicate the addition of a meteoritic component and evaporation was not the major process responsible for the measured Mg isotopic variations. In fact, the MgO profiles and correlations between δ26Mg and MgO, Na2O, Sc, Sr, CaO/Al2O3, and δEu reflect Mg isotopic fractionation caused by Mg diffusion from a region with high Mg contents (i.e., more melted pyroxene) to one with lower contents (i.e., more melted plagioclase). Diffusion modeling shows that the duration of diffusion was less than a fraction of a second. Our results indicate that chemical diffusion can produce large Mg isotopic fractionation in impact glasses on a scale of at least tens of microns, and that isotopic fractionation driven by chemical diffusion needs to be considered when the Mg isotopic compositions of impact glasses are used to identify different lunar rock reservoirs.

¹José A. P. M. Devienne,¹Thomas A. Berndt,²Wyn Williams,¹Shichu Chen
Journal of Geophysical Research (Planets) Link to Article [https://doi.org/10.1029/2023JE008268]
¹Department of Geophysics, School of Earth and Space Sciences, Peking University, Beijing, PR China
²School of GeoSciences, The University of Edinburgh, Edinburgh, UK
Published by arrangement with John Wiley & Sons

An increasing amount of evidence suggests that the tetrataenite-bearing cloudy zones (CZ) in iron and stony-iron meteorites can preserve magnetic records of ancient magnetic activity of their parent bodies over solar system timescales. Tetrataenite islands in the CZ are nanometer-sized (<200 nm) crystals that usually form through ordering from precursor taenite islands upon extremely slow cooling through 320°C. Recent micromagnetic models have shown that such precursor taenite islands form highly thermally stable single-domain (SD) or single-vortex states (SV). In this work we employ a 3D finite element multi-phase micromagnetic modeling to show that tetrataenite inherits the magnetic remanence of taenite precursor when it forms over underlying SD states. When taenite forms SV states, however, tetrataenite resets the precursor magnetization and records a new remanence through chemical ordering at 320°C. We further assess the thermal stability of tetrataenite islands. We show that in cases where tetrataenite inherits the domain states of its precursor taenite, the origin of the remanence can be up to ∼105 years older than previously thought in fast-cooled meteorites, and ∼1–≳6 Myr in slowly cooled meteorites. It indicates, therefore, that different regions across slowly cooled CZ record distinct stages of planetary formation.

¹E. Caminiti,²C. Lantz,³S. Besse,²R. Brunetto,⁴C. Carli,⁵L. Serrano,⁶N. Mari,²M. Vincendon,¹A. Doressoundiram
Icarus (in Press) Link to Article [https://doi.org/10.1016/j.icarus.2024.116191]
¹LESIA, Observatoire de Paris, Université PSL, CNRS, Sorbonne Université, Université de paris, 5 place Jules Janssen, 92195 Meudon, France
²Institut d’Astrophysique Spatiale, Université Paris- Saclay, CNRS, 91400 Orsay, France
³European Space Agency (ESA), European Space Astronomy Centre (ESAC), Camino Bajo del Castillo s/n, Villanueva de la Cañada, 28692 Madrid, Spain
⁴IAPS-INAF, Via Fosso del Cavaliere, 100, 00133 Rome, Italy
⁵Independent researcher, 660001 Pereira, Colombia
⁶Department of Earth and Environmental Sciences, University of Pavia, 27100 Pavia, Italy
Copyright Elsevier

The surface of Mercury is subject to space weathering that complicates remote sensing data analysis. We present an experimental study performed on Mercury volcanic surface analogues to provide a better constraint on spectral alterations induced by solar wind. We used 20 keV He+ with fluences up to 5 × 1017 ions/cm2 to simulate ion irradiation reaching the surface. Terrestrial ultramafic lava already identified as good analogues for Mercury were used: a boninite, a basaltic komatiite and a komatiite. Spectra were acquired in the visible to mid-infrared (VMIR) wavelength range, between 0.4 and 16 μm. Spectral alterations induced by irradiation are observed. In the visible to near-infrared (VNIR) samples show an exponential darkening, a reddening and a flattening of spectra. Above a certain irradiation dose (1 × 1017 ions/cm2 in our conditions), the darkening reaches a plateau while the reddening and flattening do not show any definable trend. In the mid-infrared (MIR) we observe a shift of Reststrahlen bands towards longer wavelengths (≤0.42 μm). The Christiansen feature is shifted towards longer or shorter wavelengths according to the irradiation dose (≤0.2 μm). The spectral alteration is closely influenced by the composition. As Mercury’s surface is compositionally heterogeneous, the degree of spectral alteration varies on the planet and putatively participates in the heterogeneous spectral properties of the surface. This work provides ground-truth data for future ESA-JAXA-BepiColombo observations. The alteration of VMIR spectral features induced by ion irradiation simulated in the laboratory will be used for future SIMBIO-SYS (Spectrometer and Imaging for MPO BepiColombo Integrated Observatory SYStem) and MERTIS (Mercury Radiometer and Thermal Infrared Spectrometer) data analysis.

^1,2,3Yingnan Zhang,^1,2,3Liping Qin
Earth and Planetary Science Letters 641, 118807 Link to Article [https://doi.org/10.1016/j.epsl.2024.118807]
¹CAS Key Laboratory of Crust-Mantle Materials and Environments, University of Science and Technology of China, Hefei 230026, China
²Deep Space Exploration Laboratory, Hefei 230088, China
³CAS Center for Excellence in Comparative Planetology, Hefei 230026, China
Copyright Elsevier

Thermal metamorphism for asteroids influences their structures and chemical compositions. As the only CCs with the petrologic type of 3-6, their metamorphis history and heat sources are unclear. The isotopic composition of molybdenum (Mo) was used as an indicator of oxidation state to investigate the oxidation and thermal metamorphic history of CK chondrites. CK chondrites are characterized by positively fractionated Mo isotopic composition relative to other chondritic groups, and the degree of enrichment in heavy Mo isotopes in CKs generally decreases with Mo content. Combined with numerical simulations for hexavalent Mo evaporation at elevated temperatures and thermodynamic calculations of the valence transformation of Mo, the Mo isotopic characteristic of CKs is proven in concordance with oxidative Mo lost during thermal metamorphism. This evaporation loss highlights the importance of the parent body process on the MVE depletion. Our results also require oxidation of CKs and subsequent thermal metamorphism to have occurred after the disintegration of the CK parent body. A viable heat source in this scenario could be solar radiation, and the oxidization may perform by the water accretion in orbits.

¹Birgit Schulz,¹Christian Vollmer,^2,3Jan Leitner,⁴Lindsay P. Keller,^5,6Quentin M. Ramasse
Geochimica et Cosmochimica Acta (in Press) Open Access Link to Article [https://doi.org/10.1016/j.gca.2024.06.013]
¹Institut für Mineralogie, Universität Münster, Corrensstr. 24, 48149 Münster, Germany
²Institut für Geowissenschaften, Ruprecht-Karls-Universität Heidelberg, Im Neuenheimer Feld 234-236, 69120 Heidelberg, Germany
³Max-Planck-Institut für Chemie, Hahn-Meitner-Weg 1, 55128 Mainz, Germany
⁴XI3, Astromaterials Research and Exploration Science (ARES) Division, NASA Johnson Space Center, Houston, TX 77058, United States
⁵SuperSTEM Laboratory, SciTech Daresbury Campus, Daresbury WA4 4AD, UK
⁶School of Chemical and Process Engineering, School of Physics and Astronomy, University of Leeds, Leeds LS2 9JT, UK
Copyright Elsevier

GEMS (glass with embedded metal and sulfides) are the dominant carrier of amorphous silicates in anhydrous interplanetary dust particles (IDPs) and one of the most suitable materials to study early solar system processes. Amorphous silicates in 105 GEMS from eight IDPs were analyzed regarding texture and chemical composition to reassess GEMS formation theories and genetic relationships to amorphous silicate material in meteorites. Petrography of bulk IDPs was investigated to understand GEMS’ relationships to other IDP components. Furthermore, carbon and nitrogen isotopic compositions were measured. Nearly all GEMS are aggregates of several subgrains with variable amount of nanophase inclusions and different Mg- and Si-contents, while single GEMS are rare. The subgrains within aggregates are typically surrounded by one or more carbon rims with high density. The chemical compositions of GEMS amorphous silicates are subsolar for all major element/Si ratios but exhibit wide heterogeneity. This is not influenced by silicon oil from the capturing process of IDPs as assumed before, as a penetration of the silicon oil is excluded by high resolution EELS (electron energy loss spectroscopy) areal density maps of silicon. Furthermore, low Fe-content in GEMS amorphous silicates shows that these are not altered by aqueous activity on the parent body as it is the case for amorphous silicate material in primitive meteorites. The subsolar element/Si ratios and the wide chemical heterogeneity point to a non-equilibrium fractional condensation origin either in the early solar nebula or in a circumstellar environment and are not in agreement with homogenization via sputtering in the ISM. The close association with carbon around GEMS subgrains and as double-rims around GEMS aggregates argue for a multi-step aggregation after formation of the smallest GEMS subgrains in the ISM or the early solar nebula. Carbon acting as matrix material connecting GEMS and other IDP components has lower areal density as seen from carbon EELS areal density maps and isotopic anomalies varying at the nanometer scale, pointing to different origins and processing of materials to varying extent or at changing temperatures.

To balance GEMS’ subsolar element/Si ratios, a supersolar component in IDPs was assumed to account for the overall chondritic composition of bulk IDPs. Nevertheless, our bulk IDP analyses revealed subsolar, but variable, element/Si ratios for complete particles as well, depending on type and amount of mineral phases in each particle. Pyroxenes in the investigated particles can occur as elongated euhedral crystals, but are overall rare. The dominant crystalline fraction in the investigated IDP samples are equilibrated aggregates (EAs) that show the same chemical compositions as GEMS, indicating that the EAs are recrystallized GEMS grains and formed after GEMS formation as secondary phases.