GEOCHEMICAL JOURNAL
Online ISSN : 1880-5973
Print ISSN : 0016-7002
ISSN-L : 0016-7002
ARTICLE
Ryugu particles found outside the Hayabusa2 sample container
Aiko NakatoShiori InadaShizuho FuruyaMasahiro NishimuraToru YadaMasanao AbeTomohiro UsuiHideto YoshidaTakashi MikouchiKanako SakamotoHajime YanoYayoi N. MiuraYoshinori TakanoShinji YamanouchiRyuji OkazakiHirotaka SawadaShogo Tachibana
Author information
JOURNAL OPEN ACCESS FULL-TEXT HTML
Supplementary material

2022 Volume 56 Issue 6 Pages 197-222

Details
Abstract

The Hayabusa2 spacecraft explored C-type near-Earth asteroid (162173) Ryugu and returned asteroidal materials, collected during two touchdown operations, to the Earth as the first sample from carbonaceous-type asteroid. The sample container, in which ~5 g of Ryugu sample was enclosed, was safely opened in the clean chamber system with no severe exposure to the terrestrial atmosphere. In the course of preparation operation of the sample container, two dark-colored millimeter- to sub-millimeter-sized particles were found outside the sealing part of the sample container. Because they look similar to the Ryugu particles inside the sample container, the particles were named as Q particles (Q from questionable). In this study, we investigated Q particles (Q001 and Q002) mineralogically and petrographically to compare them with potential contaminants (the ablator material of the reentry capsule and fine sand particles at the capsule landing site), Ryugu sample, and CI chondrites. The Q particles show close resemblance to Ryugu sample and CI chondrites, but have no evidence of terrestrial weathering that CI chondrites experienced. We therefore conclude that the Q particles are originated from Ryugu and were expelled from the sample catcher (sample storage canister) in space prior to the enclosure operation of the sample catcher in the sample container. The most likely scenario is that the Q particles escaped from the sample catcher during the retrieval of the sample collection reflector, which was the necessary operation for the sample container closing.

Introduction

Samples returned from C-type near-Earth asteroid (162173) Ryugu by the Hayabusa2 spacecraft (Tsuda et al., 2022) are in close relation to CI chondrites (Yokoyama et al., 2022; T. Nakamura et al., 2022; Noguchi et al., 2022; Okazaki et al., 2022a; Yabuta et al., 2022; Naraoka et al., 2022; Hopp et al., 2022; Moynier et al., 2022; Barosch et al., 2022; E. Nakamura et al., 2022; Ito et al., 2022), which are the chemically most primitive meteorites with the highest abundances of volatiles among meteorites found on the Earth. The total amount of returned sample was ~5 grams, which was more than the minimum mission requirement (0.1 g) (Tachibana et al., 2014; Sawada et al., 2017; Okazaki et al., 2017) and is currently the largest returned from deep space.

These samples were collected by shooting of a 5-gram projectile made of tantalum inside the sampler horn (Sawada et al., 2017; Thuillet et al., 2019) at two different surface locations of Ryugu (Morota et al., 2020; Kikuchi et al., 2020, 2022; Tachibana et al., 2022; Terui et al., 2022; Yoshikawa et al., 2022). The firing of the projectile and ejection of surface particles beneath the sampler horn were confirmed in both touchdown operations (Tachibana et al., 2022; Yoshikawa et al., 2022). Successful touchdown operations resulted in collection of about three and two grams of samples at the first and second touchdown sites, respectively. The second touchdown was made ~20 meter north from the spacecraft-made artificial crater (Arakawa et al., 2020) made by a small carry-on impactor (Saiki et al., 2013), and collection of sub-surface materials as ejecta of the artificial impact was expected. The collected samples were separately stored in two chambers of the sample catcher without significant mixing (Tachibana et al., 2022). After two touchdown operations in February and July 2019, the sample catcher was transported into the sample container inside the reentry capsule and was sealed with a metal-to-metal sealing system (Sawada et al., 2017; Okazaki et al., 2017).

The reentry capsule landed in the Woomera Prohibited Area in South Australia on December 6, 2020 (Nakazawa et al., 2022; Yamada and Yoshihara, 2022). Volatile components inside the sample container in Australia was extracted and analyzed with the GAs Extraction and Analyses system (GAEA) on the next day of the capsule landing (Miura et al., 2022; Okazaki et al., 2022b), and the sample container was safely transported to the curation facility in JAXA (Sagamihara, Japan) after ~57 hours from the capsule landing. At the curation facility, the sample container was first attached to the container opening device, which was designed to open the container in the clean chamber system developed for initial description of Hayabusa2 returned samples (Abe, 2021; Yada et al., 2022; Pilorget et al., 2022). The outer part of the sample container lid, which was pressed with a load of ~2700 N through 12 presser springs for metal-to-metal sealing (Sawada et al., 2017; Okazaki et al., 2017), was taken away with keeping the load constant using the container opening system. During this operation before installation of the sample container into the clean chamber, the sample container was cleaned several times to remove any particles present outside of the metal-to-metal sealing.

The sample container was opened in the vacuum part of the clean chamber system to take the sample catcher out on December 14, 2020. When the sample container was opened, fine dark powder was found at the bottom of the sample container. This implies that a part of Ryugu samples may have spilled out from the sample catcher during the proximity observation at Ryugu, in the return cruise phase, or after landing of the reentry capsule. If samples were expelled in space prior to the container closing, there is a possibility that Ryugu particles are present outside the sample container sealing.

In this study, we studied two millimeter- to sub-millimeter-sized particles found outside of the sample container during cleaning before the container installation into the clean chamber, and report the finding of Ryugu particles that are likely to have spilled out from the sample catcher before or during the container closing operation in space.

Methods

Cleaning of the sample container

To remove the outer part of the sample container lid (Fig. 1), the load from presser springs to the inner lid was gradually released with adding the load from jigs of the container opening device to keep the total load of ~2700 N. After the inner load was fully pressed by the container opening jigs, the outer lid was lifted up and the gap between the inner lid and the sample container was cleaned using clean room polyurethane swabs (Fig. 1). Two millimeter- to sub-millimeter-sized particles were removed from the gap (Fig. 2), and both particles were black in hue like surface boulders and pebbles on Ryugu (Sugita et al., 2019) or carbon-phenolic ablator of the reentry capsule (Yamada and Yoshihara, 2022). They looked different from soil collected at the capsule landing site with authorization from the Australian Government and other potential contaminants archived during the sampler system development (Takano et al., 2020; Sakamoto et al., 2022). We named these two grains ‘Q particles’ (Q001 and Q002) because they are likely to be those collected at Ryugu but questionable. Since their finding, they were kept stored in the nitrogen desiccator cabinet in the ISAS clean room, except for during sample preparation and analysis, to avoid unnecessary oxidation and/or hydration in the air.

Fig. 1.

Schematic illustration of the Hayabusa2 sample container (left) and a photo of the sample container attached to the vacuum flange for connection to the clean chamber system (right). Two Q particles were found outside the metal seal.

Fig. 2.

A dark particle found in the gap between the inner lid and the sample container during cleaning. An exposure of the enlarged inset image was artificially adjusted to emphasize the presence of grain.

Observation of particles with optical and scanning electron microscopy

The color, size, and shape of Q001 and Q002 were first examined using a stereo microscope (NIKON AZ100), and a high magnification map and an omni focal image were taken using an optical microscopy with an electric control stage (NIKON SMZ1270) equipped on the Curation Clean Chamber CC4-2 (Abe, 2021). Their detailed surface features were observed with a field emission scanning electron microscope (FE-SEM; Hitachi SU6600) in the clean room at the JAXA curation facility. For FE-SEM observation, the particles were partly fixed in low melting point Bi-Pb-Sn alloy in a nitrogen filled glove box. The observation was made in low vacuum mode (60 Pa) at the accelerating voltage of 10 kV and a beam current of ~0.3 nA with no coating of conductive material such as carbon. Qualitative compositional analysis of constituting materials was made with energy dispersive X-ray spectroscopy (EDS) attached to the FE-SEM. After the surface observation, each grain was turned over, re-fixed in the glove box, and observed the back surface. The FE-SEM observation in the clean room was also performed for soil particles collected at the capsule landing site in the Woomera Prohibited Area, Australia.

The particle Q001 was mounted in epoxy, polished down to 1 micron using alumina wrapping film and diamond paste with tap water, and coated with carbon. The polished surface was observed with FE-SEM (Hitachi NX2000 and JEOL JSM 7000F) at the JAXA curation facility and at University of Tokyo. The observations by NX2000 and JSM 7000F were made at the accelerating voltage of 15 kV and the probe current of 0.16 nA and 0.51 nA respectively.

Electron microprobe analysis

Elemental mapping analysis was performed for the polished section of Q001 to identify mineral phases with a field-emission electron microprobe (FE-EPMA; JEOL JXA-8530F) at Department of Earth and Planetary Science, University of Tokyo. The electron beam current of 60 nA with an accelerating voltage of 15 kV was used for mapping as in T. Nakamura et al. (2022), where Ryugu particles were analyzed with FE-EPMA.

Quantitative compositional analyses of minerals were performed at different conditions for different phases using well-characterized natural and synthetic standards (T. Nakamura et al., 2022). Phyllosilicates were analyzed at the accelerating voltage of 15 kV and a beam current of 12 nA. The electron beam was broadened to be 1 μm in diameter to reduce the beam damage. The counting times were 10 seconds at the peaks. The detection limits were K: 125, Mg: 138, Na: 148, Si: 158, Ca: 158, Al: 193, P: 205, Mn: 229, Cr: 229, S: 237, Fe: 241, Ti: 263 and Ni: 293 (ppm). Analysis of sulfides were made with a focused beam current of 20 nA at the accelerating voltage of 15 kV. The signal counting time at the peaks was set at 20 seconds. The correlation of Co Ka counts affected by overlapping Fe Kb peaks were considered and corrected. The detection limits were Si: 58, S: 77, P: 80, Cr: 108, Mn: 131, Ni: 166, Fe: 192, Co: 192, Cu: 230 and Zn: 371 (ppm). To avoid a beam damage on carbonates, the electron beam (15 keV and 3 nA) was scanned over a 2 × 2 μm square, and the counting times at the peaks were 10 seconds. The detection limits were Ca: 324, Si: 333, Mg: 365, Pb: 460, Sr: 545, Fe: 739, Ba: 753 and Mn: 770 (ppm). For phosphates, the accelerating voltage was 15 kV and the beam current was 6 nA with 1 μm beam size. The counting times were 10 seconds at the peaks. The detection limits were Cl: 142, Mg: 161, Na: 188, K: 195, Al: 204, Ca: 234, Si: 246, P: 308, Cr: 320, Fe: 326, Mn: 330, S: 336, Ti: 360 and F: 660 (ppm). Anhydrous oxides were analyzed with the focused beam at the accelerating voltage of 12 kV and the beam current of 30 nA. The counting times were 30 seconds at the peaks. The detection limits were Al: 23, Mg: 64, Si: 65, Na: 67, Ca: 77, Cr: 125, Ti: 137, Mn: 202 and Fe: 206 (ppm). ZAF correction was employed to calculate elemental abundances except for anhydrous phase analysis (12 kV and 30 nA), for which employed a Phi-Rho-Z method (Armstrong, 1995).

For comparison, we made quantitative compositional analyses of the phyllosilicate matrix of Orgueil CI chondrite using the electron microprobe with the same analytical conditions.

Elemental maps (Fe, Ni, and S) of a couple of iron sulfides in Q001 were additionally obtained with JEOL JCM-7000 NeoScope Benchtop SEM equipped with EDS. The accelerating voltage of 15 kV and a beam current of ~1 nA were applied for mapping.

Results

Optical microscopy and scanning electron microscopy of grain surface

Both Q001 and Q002 contained a small fraction of materials with high reflectance, some of which showed flattened or hexagonal morphology, and transparent materials in dark matrix (Fig. 3). The particles are thus mixtures of materials with different optical properties, implying that they are not fragments of the ablator (Fig. 4). The dark nature of Q001 and Q002 appears also different from fine sand and soil at the capsule landing site (Fig. 4).

Fig. 3.

Stereomicrographs of Q particles (Q001 and Q002) taken from two opposite directions. All the scale bars represent 0.5 mm.

Fig. 4.

Forebody heat shield of the Hayabusa2 sample return capsule coated with carbon-phenolic ablator material (left) and fine sand and soil at the capsule landing site (right).

Scanning electron microscopy of the surface with EDS found that both particles (Fig. 5) consist mainly of magnesian silicates, leading us to conclude that they are not ablator materials. Both Q001 and Q002 contain iron sulfide and iron oxide with characteristic morphology. Iron sulfide is often present as a hexagonal plate (Fig. 6a, b). On the surface of Q002, iron sulfide showing surface modification texture was observed (Fig. 6b). The iron oxide shows various morphologies; framboids, plaquette, or spherulites (Fig. 6c–f). These morphologies of iron sulfide and iron oxide are similar to those typically observed in Ryugu grains (T. Nakamura et al., 2022; Yokoyama et al., 2022; Noguchi et al., 2022; E. Nakamura et al., 2022; Ito et al., 2022).

Fig. 5.

Back-scattered electron images of Q001 and Q002. The observation was made in low vacuum mode (60 Pa) without coating of conductive material.

Fig. 6.

Back-scattered electron images of hexagonal-shaped iron sulfides at the surface of Q001 (a) and Q002 (b). Iron oxides show a variety of morphologies such as plaquette (c, d), framboids (d, f), and spherulites (e). The observation was made in low vacuum mode (60 Pa) without coating of conductive material.

Carbonate-like and phosphate-like components were also observed on the surfaces of both particles (Fig. 7). Carbonate-like components contained Ca and Mg with high signal of C, suggesting that they were dolomite (CaMg(CO3)2). Phosphate-like components show a possible variation of cation concentrations (Fig. 7). Some were enriched in Ca, while others were enriched in Mg, Fe, and Na. Dolomite, Ca-rich phosphate (apatite), and Mg-, Fe-, Na-rich phosphate are also observed in Ryugu particles (T. Nakamura et al., 2022; Yokoyama et al., 2022; Noguchi et al., 2022; E. Nakamura et al., 2022).

Fig. 7.

Back-scattered electron images of carbonate-like and phosphate-like materials observed on the surface of Q001 and Q002.

The matrix consisting of magnesian silicates and the presence of iron sulfide, iron oxide, carbonates and phosphates in Q001 and Q002 strongly suggest their close relation to Ryugu particles and that they are different from the soil particles at the capsule landing site that consist mainly of silica, most likely quartz (Huntley et al., 1993), accompanied with Al- or Fe-bearing minerals (Fig. 8).

Fig. 8.

A back-scattered electron image of fine sand particles at the capsule landing site.

Scanning electron microscopy and elemental mapping of the polished surface of Q001

Scanning electron microscopy and elemental mapping of the polished section of Q001 show that the particle consists mainly of magnesian silicate-rich matrix containing iron sulfides (Ni-poor and Ni-rich), carbonates (Ca-rich and Ca-poor), oxides (mainly iron oxide), and Ca-rich phosphates (Figs. 9 and 10). Many cracks are observed but no large minerals are fractured by cracks, indicating that Q001 did not experience severe shock. These features show a good agreement with “the major lithology” of Ryugu particles (T. Nakamura et al., 2022). We note that some cracks (Fig. 9) may have formed during polishing due to the fragile nature of Q001.

Fig. 9.

A back-scattered electron image of the polished cross section of Q001.

Fig. 10.

X-ray elemental maps of (a) Si, (b) Fe, (c) Ca, and (d) S of Q001.

Matrix magnesian silicates and coarse magnesian silicates

The compositions of the matrix components (Table 1) show that the matrix is primarily composed of magnesian silicates. The total concentrations of cations as oxides are much smaller than 100 wt% (70–90 wt%) for all the analyzed points. These total concentrations are consistent with the reported concentrations of phyllosilicates in Ryugu particles (T. Nakamura et al., 2022), suggesting that the Q001 matrix also consists of magnesian phyllosilicates. The matrix compositions show a variety of sulfur concentrations, which are given as SO3 in Table 1. This is likely to reflect the presence of nano- to submicron-sized sulfide particles within the X-ray generation volume of the matrix components (Fig. 11a, b), and the composition data in Table 1 should represent that of a mixture of magnesian phyllosilicates and iron-nickel sulfide, which is discussed below. The atomic ratios of matrix components are summarized in Table S1.

Table 1. Elemental compositions of matrix components (wt%)
# SiO2 TiO2 Al2O3 MgO FeO MnO CaO Na2O K2O SO3 P2O5 Cr2O3 NiO Total
007 30.09 0.034 2.24 21.27 9.84 0.295 2.55 0.103 0.019 3.14 0.031 0.45 1.07 71.13
008 30.01 0.070 1.79 19.69 12.36 0.074 0.81 0.042 0.014 8.81 0.055 0.49 1.25 75.47
009 34.31 0.019 2.37 22.31 9.22 0.136 1.21 0.173 0.039 2.84 0.042 0.42 0.79 73.87
010 32.41 0.051 1.71 19.62 20.86 0.080 0.26 0.089 0.036 11.42 0.045 0.63 3.41 90.62
011 31.90 0.052 1.85 20.93 14.33 0.077 0.46 0.067 0.025 9.19 0.063 0.64 2.81 82.38
012 28.63 n.d. 1.67 18.51 11.39 0.050 0.60 0.071 0.012 7.46 0.033 0.47 2.33 71.22
013 27.37 0.064 1.67 17.91 13.81 0.061 0.30 0.030 0.009 10.89 0.006 0.44 2.32 74.89
014 28.61 n.d. 1.94 19.23 14.64 0.018 0.70 0.033 0.031 13.47 0.220 0.43 3.54 82.85
015 30.78 0.068 1.85 21.63 11.79 0.049 0.15 0.176 0.024 7.00 0.050 0.43 1.25 75.24
017 33.05 0.036 2.19 21.40 11.96 0.110 0.15 0.112 0.007 6.29 0.019 0.41 1.07 76.80
019 33.07 n.d. 2.12 21.75 11.15 0.081 0.52 0.053 0.024 6.00 0.053 0.62 1.70 77.15
022 32.47 0.08 2.08 21.64 10.20 0.058 0.32 0.130 0.024 3.68 0.017 0.23 0.33 71.26
023 33.71 0.02 2.51 20.96 13.27 0.008 0.61 0.056 0.046 7.62 0.072 0.44 0.91 80.25
025 29.89 0.029 1.78 19.74 13.16 0.077 0.67 0.166 0.038 7.85 0.033 0.68 1.64 75.74
026 32.50 0.073 2.05 21.31 12.56 n.d. 0.49 0.065 0.011 7.16 0.032 0.56 2.41 79.23
028 30.11 0.029 1.91 20.17 13.51 0.026 0.61 0.048 0.045 9.94 0.131 0.46 2.55 79.54
030 37.67 0.037 2.30 25.32 7.86 0.037 0.45 0.077 0.038 0.36 0.057 0.42 0.09 74.72
033 32.49 0.056 2.10 21.86 9.55 0.037 0.83 0.055 0.039 2.74 0.224 0.40 1.27 71.66
034 28.84 0.135 1.90 20.55 18.84 0.016 0.42 0.008 0.020 15.66 0.060 0.47 2.14 89.05
Average 31.47 0.045 2.00 20.83 12.65 0.068 0.64 0.082 0.026 7.45 0.065 0.48 1.73 77.53
2.47 0.033 0.24 1.62 3.15 0.065 0.53 0.050 0.013 3.89 0.061 0.11 0.97 5.61

n.d.: Not detected.

Fig. 11.

Back-scattered electron images of matrix ((a) and (b)) and coarse magnesian silicates ((c) and (d)) observed in the cross section of Q001. Open circles show the locations with ID numbers, where elemental compositions were obtained with electron microprobe (Tables 1 and 2). Po: pyrrhotite.

The coarse magnesian silicates are commonly rounded or irregularly shaped grains or aggregates of several to several ten μm in size (Fig. 11c, d). They also have distinct outlines, which makes them morphologically distinguishable from the matrix components. The elemental compositions of coarse magnesian silicates are listed in Table 2. The total concentrations of cations as oxides are lower 100 wt% (77–83 wt%) for all the analyzed points, indicating that they are also phyllosilicates. The layered texture (Fig. 11c, d) also suggests that they are phyllosilicates. The atomic ratios of coarse magnesian silicates are summarized in Table S2.

Table 2. Elemental compositions of coarse magnesian silicates (phyllosilicates) (wt%)
# SiO2 TiO2 Al2O3 MgO FeO MnO CaO Na2O K2O SO3 P2O5 Cr2O3 NiO Total
016 38.60 n.d. 3.71 22.90 10.75 n.d. 0.56 0.201 0.077 0.12 0.018 0.42 0.103 77.47
018 42.96 n.d. 4.14 25.07 6.68 0.018 1.33 0.014 0.034 0.08 0.050 0.64 0.009 81.04
020 43.51 0.022 4.02 26.43 6.64 0.041 1.38 0.034 0.035 0.09 0.062 0.86 0.064 83.18
021 41.54 0.047 3.59 23.34 6.69 0.026 1.40 0.048 0.037 0.09 0.129 0.46 0.063 77.45
024 41.90 0.004 3.70 23.83 6.23 n.d. 0.51 0.014 0.032 0.06 0.023 0.49 n.d. 76.79
027 40.19 0.033 3.51 24.09 7.08 0.041 1.13 0.042 0.036 0.70 0.039 0.60 0.157 77.65
029 39.54 0.033 3.20 24.07 6.85 n.d. 0.68 0.179 0.048 0.04 0.004 0.65 0.089 75.37
031 40.19 n.d. 4.00 25.45 6.75 0.029 1.27 0.036 0.047 0.03 0.044 0.56 0.088 78.50
032 43.54 0.011 3.95 25.32 6.71 0.015 1.16 0.075 0.052 0.23 0.040 0.46 0.116 81.68
Average 41.33 0.017 3.76 24.50 7.15 0.019 1.05 0.071 0.044 0.16 0.045 0.57 0.077 78.79
1.80 0.018 0.30 1.14 1.37 0.017 0.36 0.070 0.014 0.21 0.036 0.14 0.050 2.58

n.d.: Not detected.

The compositional range of fine and coarse phyllosilicates are shown in the Mg-Fe-(Si + Al) ternary diagram (Fig. 12), which will be discussed later comparing with those reported for phyllosilicates in Ryugu and CI chondrites (Yokoyama et al., 2022; Tomeoka and Buseck, 1988).

Fig. 12.

The ternary compositional diagram of matrix phyllosilicates and coarse phyllosilicates in Q001 (Mg-Fe-(Si + Al)). The ideal Mg-Fe, solid solution lines of serpentine and saponite (Tomeoka and Buseck, 1988) are also shown. (a) The compositions of phyllosilicates in Ryugu (Yokoyama et al., 2022) and matrix of Orgueil CI chondrite (this study) are also plotted. (b) The enlarged diagram of (a). The compositions of coarse and fine phyllosilicates in Orgueil (Tomeoka and Buseck (1988)) are also shown in addition to the data in (a).

Sulfides

The elemental compositions of sulfides (Table 3) shows that they can be classified into low-Ni and high-Ni iron sulfides.

Table 3. Elemental compositions of low-Ni sulfide (wt%)
# Si S P Fe Co Zn Ni Cu Mn Cr Total
057 0.17 37.30 n.d. 57.17 0.051 0.039 0.98 0.003 0.008 0.035 95.75
059 0.04 39.14 0.001 58.91 0.041 0.036 0.81 0.024 0.004 0.053 99.06
060 0.10 37.36 n.d. 57.63 0.045 n.d. 0.73 0.011 n.d. 0.051 95.92
067 0.05 39.89 n.d. 58.21 0.036 n.d. 0.75 0.020 n.d. 0.025 98.98
068 0.03 38.64 n.d. 58.49 n.d. n.d. 0.70 n.d. 0.006 0.031 97.90
069 0.06 38.63 0.009 58.43 0.060 0.063 0.82 0.005 n.d. 0.025 98.10
070 0.04 40.05 0.015 58.59 0.023 0.018 0.82 n.d. n.d. 0.021 99.58
071 0.08 38.51 n.d. 57.65 0.028 0.011 0.84 0.012 0.010 0.046 97.19
072 0.09 36.33 0.006 57.00 0.028 n.d. 0.80 n.d. n.d. 0.042 94.29
073 0.03 40.23 n.d. 58.83 n.d. n.d. 0.72 n.d. n.d. 0.054 99.85
074 0.53 39.83 0.007 56.79 0.037 n.d. 0.68 n.d. 0.001 0.040 97.92
075 0.08 38.93 0.003 57.56 0.042 n.d. 0.80 0.019 0.010 0.061 97.50
076 0.10 39.67 n.d. 56.17 0.006 n.d. 0.70 0.002 0.004 0.051 96.71
077 0.04 39.61 n.d. 59.16 0.031 0.044 0.80 n.d. 0.019 0.043 99.75
078 0.11 40.32 n.d. 58.19 0.012 0.001 0.74 0.002 0.022 0.049 99.45
079 0.02 39.27 n.d. 59.54 0.055 n.d. 0.76 n.d. 0.007 0.057 99.71
080 1.02 33.99 n.d. 53.36 n.d. n.d. 0.73 n.d. 0.008 0.063 89.17
081 0.04 39.42 n.d. 59.45 0.054 0.068 0.81 n.d. 0.013 0.055 99.90
082 0.02 39.83 n.d. 59.39 0.019 0.008 0.72 n.d. n.d. 0.043 100.03
083 0.06 39.54 n.d. 57.67 0.006 n.d. 0.79 n.d. 0.015 0.051 98.12
084 0.04 40.55 n.d. 59.27 0.040 n.d. 0.83 n.d. 0.027 0.048 100.80
085 0.05 38.90 n.d. 58.27 0.051 n.d. 0.81 n.d. 0.013 0.044 98.14
086 0.04 39.83 n.d. 58.38 0.016 n.d. 0.68 0.035 0.039 0.475 99.50
087 0.06 39.65 n.d. 59.04 0.009 0.015 0.73 0.031 n.d. 0.039 99.58
088 0.07 39.69 n.d. 58.63 0.053 n.d. 0.82 n.d. 0.004 0.062 99.33
089 0.20 41.45 n.d. 58.74 n.d. n.d. 0.70 n.d. 0.019 0.058 101.16
090 0.15 40.20 n.d. 57.33 0.049 n.d. 1.43 n.d. n.d. 0.092 99.25
091 0.04 39.59 0.008 58.44 0.019 n.d. 0.82 n.d. 0.004 0.032 98.95
092 0.08 38.83 0.004 58.10 0.062 n.d. 0.86 n.d. 0.003 0.026 97.97
093 0.07 39.87 0.002 58.44 0.033 0.009 0.78 n.d. n.d. 0.051 99.26
094 0.04 39.56 0.005 58.93 0.010 n.d. 0.72 n.d. n.d. 0.034 99.28
095 0.10 38.97 n.d. 56.92 0.023 0.010 0.71 0.015 0.006 0.024 96.79
058* 0.22 37.60 n.d. 49.43 0.58 n.d. 9.66 0.007 n.d. 0.050 97.55
Average 0.12 39.13 0.002 57.82 0.046 0.010 1.06 0.006 0.007 0.059 98.25
0.19 1.38 0.004 1.92 0.098 0.019 1.55 0.010 0.009 0.076 2.22

*A grain showing an intermediate Ni content and is not included to obtain the averaged composition and its standard deviation. n.d.: Not detected.

Low-Ni iron sulfide is predominant over high-Ni iron sulfide and is ubiquitously present in the particle. They typically occur as lath or hexagonal shapes ranging from several to 40 μm (typically ~10 μm) in size (Figs. 11 and 13). The metal/sulfur ratio of low-Ni iron sulfide ranges from ~0.82 to ~0.91 with the average of 0.86 ± 0.02 (1σ, n = 32). The metallic elements considered here are listed in Table 3 except for Si, of which signals may have come from the surrounding phyllosilicate matrix. This indicates that low-Ni iron sulfides in Q001 are mostly pyrrhotite with the composition close to Fe7S8. The Ni content is highly homogeneous among individual pyrrhotite grains with an average value of 0.8 ± 0.1 wt% and the maximum of 1.4 wt% (Table 3).

Fig. 13.

Back-scattered electron images of pyrrhotite (Po) and pentlandite (Pn) in the cross section of Q001. The data ID in Tables 3 and 4 are also indicated.

Four sulfide grains analyzed in this study were high-Ni iron sulfides (Table 4). One of them has a ~20 μm lath shape, similar to typical pyrrhotite grains in the particle (Fig. 12d). The other three grains are relatively small (<10 μm) (Fig. 13). The average metal/sulfur ratio of these high-Ni iron sulfides is 1.14 ± 0.03 (except for #56 with a low total concentration), which is identical to the value of stoichiometric pentlandite (Fe, Ni)9S8 (metal/sulfur = 1.125). The metallic elements considered here are listed in Table 4 except for Si that may include the signals from the surrounding phyllosilicate matrix.

Table 4. Elemental compositions of high-Ni sulfide (wt%)
# Si S P Fe Co Zn Ni Cu Mn Cr Total
055 0.07 30.57 0.001 28.17 1.69 n.d. 34.04 0.107 0.038 0.031 94.71
056 0.17 23.12 0.006 26.94 1.71 n.d. 33.81 0.041 0.032 0.039 85.86
064 0.18 31.09 n.d. 28.12 1.64 n.d. 31.94 0.037 0.014 0.067 93.10
066 0.09 31.70 n.d. 28.73 1.84 n.d. 34.63 0.095 0.003 0.041 97.12
096* 0.06 32.33 n.d. 28.35 1.88 n.d. 35.02 0.033 0.018 0.049 97.75
Average 0.11 29.76 0.001 28.06 1.75 0.00 33.89 0.063 0.021 0.045 93.71
0.06 3.77 0.003 0.67 0.10 0.00 1.19 0.035 0.014 0.014 4.77

*Duplicated analysis of the grain #55. n.d.: Not detected.

Grains consisting of fine intermixtures of low-Ni and high-Ni sulfide were also found (Fig. 14), one of which showed an intermediate Ni concentration between pyrrhotite and pentlandite (#58; Table 3) mostly likely because the emission of characteristic X-rays occurred from both pyrrhotite and pentlandite.

Fig. 14.

Intergrown texture of pyrrhotite (Ni-poor region) and pentlandite (Ni-rich region) in the cross section of Q001. (a) #57. (b) #58.

Atomic abundances of elements in all sulfides analyzed are listed in Tables S3 and S4, respectively.

Carbonates

There are some grains (10–45 μm in size) containing Ca and Mg with the total oxide concentration of ~52 wt% (Table 5). These grains show more intense C Kα X-ray signals in the EDS spectra than other phases. The total concentration of ~52 wt% and the excess of C Kα signals show that they are dolomite (CaMg(CO3)2) containing small amounts of FeCO3 and MnCO3 components ((Ca0.43–0.48Mg0.40–0.45Mn0.05–0.07Fe0.03–0.08)CO3). Most of dolomite grains show irregular morphologies with many cleavage cracks (Fig. 15a–c), and they often coexist with magnetite (see below) showing various morphologies such as framboids, plaquette, and spherulites (Fig. 15b, c). The dolomite grains with higher Fe contents (#37, #49, and #51) than others may be due to the characteristic X-ray emission from associated magnetite (#51; Fig. 15c).

Table 5. Elemental compositions of dolomite (wt%)
# MgO SrO SiO2 CaO BaO FeO MnO PbO Total
036 18.58 0.006 0.20 26.33 n.d. 2.41 4.72 0.104 52.35
037 18.48 n.d. 1.39 24.53 n.d. 4.69 4.68 0.076 53.85
038 18.43 0.029 0.25 26.17 0.075 2.39 5.09 n.d. 52.44
039 17.84 n.d. 0.19 27.32 n.d. 2.33 5.09 n.d. 52.78
040 18.28 n.d. n.d. 26.43 0.018 2.69 4.83 n.d. 52.24
041 17.77 n.d. 0.17 27.31 0.037 2.36 5.06 n.d. 52.70
042 17.65 n.d. 0.07 27.76 0.117 2.50 4.90 0.009 53.01
043 17.63 0.100 0.32 26.79 0.086 2.11 5.12 n.d. 52.15
044 18.92 0.105 0.39 26.75 0.017 2.52 4.55 0.085 53.33
045 19.30 n.d. 0.38 26.48 n.d. 2.71 4.79 0.066 53.72
046 18.54 0.006 0.07 27.14 n.d. 1.97 4.71 0.038 52.53
047 18.61 n.d. 0.59 26.56 0.049 2.28 4.64 n.d. 52.73
048 17.61 n.d. 0.19 26.10 n.d. 2.32 4.51 n.d. 50.72
049 18.06 0.055 0.16 25.98 0.092 5.72 4.33 0.019 54.42
050 18.70 n.d. 0.17 26.96 n.d. 2.27 4.56 n.d. 52.64
051 15.54 n.d. 0.09 25.52 n.d. 4.75 4.53 n.d. 50.42
052 18.73 n.d. 0.10 27.10 n.d. 2.83 3.92 0.038 52.71
053 18.88 n.d. 0.59 26.03 n.d. 2.38 4.95 n.d. 52.83
Average 18.20 0.020 0.30 26.51 0.027 2.85 4.72 0.024 52.64
0.83 0.036 0.32 0.75 0.039 1.05 0.31 0.035 0.96

n.d.: Not detected.

Fig. 15.

Back-scattered electron images of dolomite (Dol) (a–c) and breunnetite (Bre) (d) in the cross section of Q001. The data ID in Tables 3 and 4 are also indicated. Dolomite grains are often associated with magnetite framboids, plaquette, or spherulites ((b) and (c)).

A single Ca-poor carbonate grain was found in Q001 (Fig. 15d). It has an elongated rhombic shape of ~ 50 μm in size, larger than typical dolomite grains. A couple of point analyses of this grain (Table 6) suggests the composition of (Ca0.01, Mg0.66, Mn0.14, Fe0.19)CO3, i.e., breunnerite.

Table 6. Elemental compositions of breunnerite (wt%)
# MgO SrO SiO2 CaO BaO FeO MnO PbO Total
035 27.49 n.d. 0.12 0.69 0.037 14.88 10.88 0.099 54.21
054* 28.27 0.007 n.d. 0.39 0.013 14.18 10.52 0.089 53.47
Average 27.88 0.004 0.06 0.54 0.025 14.53 10.70 0.094 53.84
0.55 0.005 0.09 0.21 0.017 0.50 0.26 0.007 0.52

*Duplicated analysis of the grain #35. n.d.: Not detected.

Atomic abundances of elements in carbonates are listed in Tables S5 and S6. The compositional ranges of dolomite and breunnerite are shown in Figs. 1618, details of which will be discussed below.

Fig. 16.

The ternary compositional diagram of dolomite in Q001 (MgCO3-CaCO3-(Mn, Fe)CO3). The dolomite compositions in CI chondrites (Endreß and Bischoff, 1996) are also shown for comparison.

Fig. 17.

Compositional correlations of dolomite in Q001. The dolomite compositions in CI chondrites (Endreß and Bischoff, 1996) are also shown for comparison.

Fig. 18.

The ternary compositional diagram of breunnerite in Q001 (MgCO3-MnCO3-FeCO3). The breunnerite compositions in CI chondrites (Johnson and Printz, 1993) are also shown for comparison. The symbols are the same as those in Figs. 17 and 18.

Magnetite

Iron oxide grains show the total weight percent smaller than 100 wt% when evaluated as ferrous iron oxide (FeO) (Table 7), suggesting that they are magnetite (Fe3O4) containing ferric iron. As mentioned above, magnetite shows a variety of morphologies (Figs. 6, 15b, c, and 19), among which framboidal magnetite appears to be the most abundant. Framboidal magnetite are the aggregates of submicron-to-micron-sized crystals, which often show isometric trapezohedron shapes (Fig. 19). Framboidal magnetite is present as a subrounded or lath-like form in the matrix and as an irregular-shaped cluster within cracks (Fig. 19). Spherulites showing a radiating fibrous interior structure are also common (Fig. 19). Their typical size is several to 10 μm and the exterior is often smooth and well-rounded. Plaquettes, less common than framboids and spherulites, are typically several μm in size and consist of multiple stacking lens-shaped plates (several hundred nm in thickness). Some magnetite coexist with carbonates (Fig. 15), and close association of magnetite with different morphologies is also observed (Fig. 19a). Atomic abundances of elements in magnetite are listed in Table S7.

Table 7. Elemental compositions of magnetite (wt%)
# Al2O3 MgO Na2O SiO2 CaO TiO2 FeO MnO Cr2O3 NiO Total
107 0.029 0.030 0.013 0.06 0.055 0.008 89.51 0.025 0.035 0.174 89.93
108 0.035 0.070 0.028 0.48 0.036 0.149 86.70 0.008 0.129 0.02 87.66
109 n.d. 0.025 n.d. 0.14 0.031 0.01 89.24 0.099 0.036 0.091 89.67
110 0.009 n.d. n.d. 0.06 0.043 0.043 89.92 0.008 0.118 0.185 90.38
111 0.022 0.017 0.024 0.05 0.039 n.d. 89.94 n.d. 0.092 0.079 90.26
112 0.025 0.019 n.d. 0.06 0.071 n.d. 89.98 0.046 0.055 0.022 90.28
113 0.025 0.064 0.006 0.25 0.059 0.055 88.24 0.07 0.133 0.054 88.95
114 0.002 0.064 0.028 0.39 0.047 0.116 86.79 0.083 0.075 0.014 87.60
115 0.017 0.045 0.045 0.06 0.016 n.d. 89.37 n.d. 0.037 n.d. 89.59
116 n.d. 0.176 0.041 0.18 0.26 0.020 88.18 0.17 0.033 0.050 89.11
117 n.d. 0.082 n.d. 0.19 0.126 n.d. 86.55 0.128 0.143 n.d. 87.23
118 n.d. 0.053 0.015 0.23 0.15 0.057 89.66 0.087 0.023 0.13 90.41
119 0.052 0.045 0.017 0.21 0.021 n.d. 86.21 n.d. 0.144 0.073 86.78
120 0.005 n.d. 0.006 0.10 0.06 n.d. 88.95 0.029 0.080 0.119 89.35
Average 0.016 0.049 0.016 0.18 0.072 0.033 89.51 0.054 0.081 0.072 89.09
0.016 0.045 0.015 0.13 0.066 0.048 1.40 0.054 0.045 0.061 1.26

n.d.: Not detected.

Fig. 19.

Back-scattered electron images of magnetite (Mag) with various morphologies.

Other oxides

Five oxide grains have compositions distinct from magnetite. They are several μm in size and show angular or irregular shapes. Although the analyzed elemental compositions show overlaps with surrounding matrix phyllosilicates (#102 and #106), two grains are identified as ilmenite FeTiO3 (Fig. 20a) containing small amounts of MgO and MnO, and three other grains as chromian spinel (Fig. 20b) with various amounts of FeO, MgO, and MnO (Table 8 and Table S8). Mn-rich chromian spinel (up to 22 wt% of MnO) was reported in serpentinite of Bou-Azzer ophiolite, Morocco (Gahlan and Arai, 2007). The formation of the Mn-rich spinel grains was discussed in the context of hydrothermal alteration of magmatic chromian spinel by Mn that was leached out from olivine and was mobile in hydrothermal fluid. The observed Mn-rich spinel in Q001 may thus also be the evidence of hydrothermal activity of Ryugu’s parent planetesimal.

Fig. 20.

Back-scattered electron images of (a) ilmenite (Ilm), (b) chromian spinel (Cr-Sp), and (c, d) phosphate (Phos).

Table 8. Elemental compositions of ilmenite and chromian spinel (wt%)
  # Al2O3 MgO Na2O SiO2 CaO TiO2 FeO MnO Cr2O3 NiO Total
ilmenite 102* 0.46 5.23 0.101 6.20 0.107 24.19 30.27 0.89 0.29 0.19 67.93
103 n.d. 0.35 0.066 0.84 0.094 48.56 39.71 1.52 0.28 0.13 91.54
Cr-spinel 104 2.75 1.80 0.038 0.52 0.083 0.788 33.30 0.38 55.7 0.21 95.53
105 0.03 4.46 n.d. 0.27 0.038 0.001 8.68 15.77 68.3 0.05 97.63
106* 0.59 15.37 0.023 18.47 1.468 0.086 9.39 3.81 51.7 0.83 101.73

n.d.: Not detected.

*May be overlapped with the surrounding silicate matrix.

Phosphates

Six phosphate grains are analyzed in Q001. There are more phosphate-like P-containing grains in the matrix, but precise compositional analyses were unavailable because of their small sizes. The analyzed grains are several μm in diameters and sometimes occur as aggregates of multiple grains or adjacently to dolomite or magnetite (Fig. 19c, d). The grains are Ca-rich and contain little F and Cl (Table 9 and Table S9), suggesting that they are hydroxyapatite.

Table 9. Elemental compositions of phosphate (wt%)
# SiO2 TiO2 Al2O3 FeO MnO MgO CaO Na2O K2O Cr2O3 P2O5 SO3 F Cl Total
001 1.60 0.017 0.14 1.17 0.34 1.72 47.02 0.42 n.d. n.d. 35.9 0.12 0.70 0.019 88.80
002 0.22 0.053 0.09 1.03 0.40 0.22 54.33 0.48 0.022 0.026 40.8 0.07 0.72 n.d. 98.15
003 2.63 0.061 0.12 2.13 0.30 1.50 49.16 0.50 n.d. 0.027 36.9 1.07 0.45 n.d. 94.64
004 1.75 n.d. 0.14 1.49 0.21 1.49 51.33 0.47 n.d. n.d. 40.2 0.42 0.71 0.004 97.96
005 2.10 0.052 0.19 1.55 0.32 1.61 50.38 0.41 n.d. 0.042 37.1 0.29 0.30 n.d. 94.26
006 6.01 0.053 0.52 1.97 0.29 4.30 43.19 0.51 n.d. 0.044 32.9 1.34 0.99 0.017 91.66
Average 2.38 0.039 0.20 1.56 0.31 1.81 49.23 0.46 0.004 0.023 37.3 0.55 0.65 0.007 94.24
1.95 0.025 0.16 0.43 0.06 1.34 3.82 0.04 0.09 0.019 2.9 0.53 0.24 0.009 3.62

n.d.: Not detected.

Discussion

All the observation and analysis results suggest the close relationship of Q001 and Q002 particles with Ryugu particles (e.g., Yokoyama et al., 2022; T. Nakamura et al., 2022), and we here focus on detailed comparison between Q particles and Ryugu. Because Q001 and Q002 particles are almost identical to one another in the surface mineralogy (Figs. 3, 57), the comparison with Ryugu samples is made especially for Q001, which was mineralogically and petrographically studied using the polished section. We also compare the Q particles with CI chondrites, which are a group of chondrites that are most similar to Ryugu (e.g., Yokoyama et al., 2022; T. Nakamura et al., 2022; E. Nakamura et al., 2022; Ito et al., 2022).

Sulfides

The presence of both low-Ni iron sulfide (pyrrhotite) and high-Ni iron sulfide (pentlandite) and their sizes and morphologies in Q001 is consistent with the observation of sulfides in Ryugu samples and CI chondrites (Figs. 13 and 14) (T. Nakamura et al., 2022; Berger et al., 2016; Schrader et al., 2021).

The metal/sulfur atomic ratio of pyrrhotite in Q001 ranges from 0.82 to 0.91 with an average of 0.86 ± 0.02 (Table S3). This metal/sulfur ratio is consistent with that of pyrrhotite in Ryugu (T. Nakamura et al., 2022) and in CI chondrites (Berger et al., 2016; Schrader et al., 2021). The Ni concentration in pyrrhotite (0.8 ± 0.1 wt% with the maximum of 1.4 wt%) is similar to or slightly lower than those of Orgueil and Alais reported in previous studies (Bullock et al., 2005) and falls within the range of Ryugu (up to 2 wt%; T. Nakamura et al., 2022).

The presence of pyrrhotite with a modified surface texture observed on the surface of Q002 (Fig. 6b) may suggest its link to Ryugu grains. Morphological alteration of space weathered pyrrhotite has been identified in Itokawa particles and lunar soils (Matsumoto et. al., 2020, 2021a), and the similar texture has been reported for Ryugu grains as the evidence of space weathering at the surface of Ryugu (Matsumoto et al., 2021b; Noguchi et al., 2022; Nakato et al., 2022). The modified surface texture of the pyrrhotite on Q002 well resembles those reported for Ryugu grains. Further investigation is surely required to conclude that the pyrrhotite on Q002 is similar to those observed on space-weathered Ryugu particles.

Pentlandite in Q001 contain 27.5 ± 0.7 atom% Ni on average (except for #56 with a low total concentration), which is consistent with previously reported pentlandite compositions of CI chondrites (Berger et al., 2016). The Ni/Fe atomic ratio, ranging from 1.1 to 1.2, agrees with that of pentlandite in Ryugu (Ni/Fe = 1–1.2) (T. Nakamura et al., 2022). The Co content of pentlandite ranges from 1.6 to 1.9 wt%, which is comparable to pentlandite in CI chondrites (1.1, 1.9, and 3.1 wt% for Alais, Tonk, and Ivuna, respectively; Bullock et al., 2005). The Co/Ni atomic ratio of pentlandite grains ranges from 4.93 × 10–2 to 5.36 × 10–2, which is comparable to the Co/Ni abundance ratios of CI chondrites (4.64 × 10–2) and the Sun’s photosphere (5.37 × 10–2) (Lodders, 2021). Considering the CI-like (Sun-like) bulk elemental composition of Ryugu (Yokoyama et al., 2022), the Co/Ni ratio of pentlandite suggests that Co and Ni behaves in the same manner during formation of pentlandite on the Ryugu’s parent body.

Cubanite (CuFe2S3), sphalerite (ZnS), and daubréelite (FeCr2S4) are minor sulfides occurring in Ryugu (Yokoyama et al., 2022; T. Nakamura et al., 2022) and in CI chondrites (Brearley and Jones, 1998; Berger et al., 2011; Rubin and Ma, 2017), but none of them was found in Q particles.

Carbonates

Dolomite is clearly the most abundant carbonate phase in Q001 as in Ryugu (Yokoyama et al., 2022; T. Nakamura et al., 2022; E. Nakamura et al., 2022; Ito et al., 2022) and CI chondrites (Johnson and Printz, 1993; Endreß and Bischoff, 1996). Their size and occurrence are also similar to those in Ryugu and CI chondrites (T. Nakamura et al., 2022; Endreß and Bischoff, 1996). The compositional range of dolomite in Q001 shows good agreement with the majority of dolomite in CI chondrites (Figs. 16 and 17) and is also consistent with dolomite in Ryugu (T. Nakamura et al., 2022).

Ryugu and CI chondrites contain low-Ca carbonate (breunnerite), which is less common but tends to be larger (often rhombic in shape) than dolomite (Yokoyama et al., 2022; T. Nakamura et al., 2022; Johnson and Printz, 1993; Endreß and Bischoff, 1996). One low-Ca carbonate (breunnerite) grain occurs in Q001, which is ~50 μm in size with an elongated rhombic shape (Fig. 15d). The composition of breunnerite in Q001 is similar to the CI breunnerite with the high MnCO3 content (Fig. 18) (Johnson and Printz, 1993). The MnCO3 content of breunnerite in Q001 is also close to the highest Mn concentration of breunnerite in Ryugu (MnO 1–10 wt%; T. Nakamura et al., 2022).

Calcium carbonate is a rare carbonate mineral in CI chondrites and Ryugu (Johnson and Printz, 1993; Endreß and Bischoff, 1996; T. Nakamura et al., 2022; E. Nakamura et al., 2022). In Ryugu, it occurs in the less altered lithology, where anhydrous silicates are present (T. Nakamura et al., 2022). No Ca-carbonate was found in the polished section of Q001 most likely because no less altered lithology is present in the section.

Magnetite, ilmenite, chromian spinel, and phosphates

Magnetite in Q001 and Q002 show a wide variety of morphologies (framboids, spherulites, and plaquettes; Fig. 19), well consistent with those observed in both CI chondrites and Ryugu (Kerridge et al., 1979; Hua and Buseck, 1998; Yokoyama et al., 2022; T. Nakamura et al., 2022; E. Nakamura et al., 2022; Ito et al., 2022).

Ilmenite and chromian spinel containing MnO have been found as minor phases in CI chondrites and Ryugu (Endreß and Bischoff, 1993; Alfing et al., 2019; T. Nakamura et al., 2022). Q001 also contains these accessary minerals, some of which are enriched in MnO (Fig. 20, Table 8 and Table S8). Ca-phosphate is also a minor mineral in CI chondrites and Ryugu (Endreß and Bischoff, 1994; Morlok et al., 2006; Alfing et al., 2019; T. Nakamura et al., 2022; Tack et al., 2022; E. Nakamura et al., 2022), which is present in Q001 (Fig. 20 and Table 9). Mg-Na phosphate was reported in Ryugu (T. Nakamura et al., 2022; E. Nakamura et al., 2022), and a similar material was found on the surface of Q002 (Fig. 7).

These observations of minor phases also support the similarity between the Q particles, Ryugu, and CI chondrites.

Matrix and coarse phyllosilicates

The chemical compositions of matrix and coarse phyllosilicates in Q001 are plotted in the ternary elemental composition diagram (Mg–Fe–(Si + Al); Fig. 12). The compositions of phyllosilicates in Ryugu (Yokoyama et al., 2022), matrix of Orgueil CI chondrite (this study; Table S10), and coarse and fine phyllosilicates in Orgueil reported by Tomeoka and Buseck (1988) are also shown for comparison.

The compositional range of the matrix of Q001, showing different degrees of Fe enrichment, is consistent with those of Ryugu’s phyllosilicates (Yokoyama et al., 2022), the Orgueil matrix (this study), and fine phyllosilicate of Orgueil in Tomeoka and Buseck (1988) (Fig. 12).

The coarse phyllosilicates in Q001 (Fig. 11) show morphological similarity to coarse phyllisilicates in Orgueil (Tomeoka and Buseck, 1988). They are depleted in Fe with the average Mg/(Mg + Fe) ratio of ~0.86 (~0.87 if #106 is not included; Table 2). This Mg/(Mg + Fe) ratio is slightly higher but consistent with that of coarse phyllosilicates in Orgueil (~0.85; Tomeoka and Buseck, 1988) (Fig. 12). The coarse phyllosilicates in Q001 show enrichment of saponite component compared to the matrix (Fig. 12), which is also consistent with the compositional range of the majority of coarse phyllosilicates in Orgueil (Tomeoka and Buseck, 1988).

Tomeoka and Buseck (1988) proposed that fine phyllosilicates formed by further alteration of coarse phyllosilicates that was also responsible for the formation of ferrihydrite. However, the present observation and other studies (e.g., Yokoyama et al., 2022; T. Nakamura et al., 2022; Noguchi et al., 2022; Ito et al., 2022) indicates that ferrihydrite is a terrestrial product and that fine phyllosilicates contain fine iron sulfides. T. Nakamura et al. (2022) observed that the least altered lithology of a Ryugu grain C0002 contain micrometer-sized olivine and pyroxene grains and GEMS (glass with embedded metal and sulfides)-like amorphous silicate particles. This implies that the micrometer-sized olivine and pyroxene could have been precursors of coarse phyllosilicates, while GEMS-like amorphous silicate was altered into fine phyllosilicates with nanometer-sized opaque phases and that the multi-stage alteration process (Tomeoka and Buseck, 1988) is not required to explain coarse and fine phyllosilicates.

Difference between Q001 and CI chondrites

The mineralogy and petrography of Q001 well agree with those of Ryugu (Yokoyama et al., 2022; T. Nakamura et al., 2022; E. Nakamura et al., 2022; Ito et al., 2022) and CI chondrites (Brearley and Jones, 1998 and references therein). The clear difference between Q001 and CI chondrites is the absence of sulfates and ferrihydrite in Q001, both of which are commonly observed in CI chondrites (Tomeoka and Buseck, 1988). Although only the surface observation and EDS analysis were made on Q002, it seems to contain no sulfates and ferrihydrite either. It has been suggested that sulfates (at least partly) in CI chondrites are alteration products after their fall on the Earth (Gounelle and Zolensky, 2001, 2014). Because Ryugu sample does not contain sulfates and ferrihydrite in spite of its close resemblance to CI chondrites, sulfates and ferrihydrite in CI chondrites are most likely to be terrestrial products (Yokoyama et al., 2022; T. Nakamura et al., 2022). Therefore, Q001 and Q002, which resemble to Ryugu sample and CI chondrites but contain no sulfates and ferrihydrite, are most likely grains returned from Ryugu. The absence of sulfates and ferrihydrite in Q particles suggests that those minerals do not form in the typical Ryugu lithology through the short-duration exposure to the air and polishing using water as lubricant as in other Ryugu grains (e.g., Yokoyama et al., 2022; T. Nakamura et al., 2022), but through the long-term exposure to the Earth’s surface environment (Gounelle and Zolensky, 2001, 2014).

The principal component analysis (PCA) of the elemental compositions of the Q001 matrix phyllosilicates provides supporting evidence of the Ryugu origin of Q particles. The PCA analysis is a statistical technique to analyze the inter-relationships among multiple variables using a smaller number of variables (principle components that are described as a mixture of original variables), and we used Microsoft Excel PCA add-in in this work. The PCA was applied to six components (Mg, (Si + Al), (Ca + Na + K), Fe, Ni, S) of the elemental abundance data of the Q001 matrix. Because the matrix consists of phyllosilicates (saponite + serpentine), Si and Al were considered to be a single component occupying tetrahedral sites of phyllosilicates (Fig. 12), and Ca, Na, and K to be a single component existing as interlayer cations of saponite. The PCA shows that the first principal component (PC1) is (Mg, Si + Al, Ca + Na + K, Fe, Ni, S) = (0.572, 0.595, 0.0464, –0.319, –0.101, –0.452), which explains 94% of data variability. The obtained PC1 corresponds to addition/subtraction of Fe-Ni sulfides ((Fe + Ni)/S = 0.93; (Fe0.76, Ni0.24)0.93S) to magnesian phyllosilicates (Fig. 21), which is consistent with FE-SEM observation of the matrix, where tiny iron sulfide particles are found within fine phyllosilicates in Q001. The ubiquitous presence of nano- to sub-micron-sized pyrrhotite and pentlandite is also reported in the phyllosilicate matrix of Ryugu (T. Nakamura et al., 2022). If sub-micron-sized iron sulfides in the Q001 matrix have the same average compositions as larger pyrrhotite and pentlandite grains in Q001 (Tables 3 and 4), the contributions of pyrrhotite and pentlandite to the iron sulfide component added to/subtracted from fine phyllosilicates is roughly 3:1. The Mg/(Si + Al) ratio of the PC1 is 0.96, which is slightly lower than the solar Mg/(Si + Al) ratio of 1.04 (Lodders, 2021). This lower ratio is most likely due to the incorporation of Mg into carbonates such as breunnerite and dolomite. The area fractions of phyllosilicate, breunnerite, and dolomite in two sections of major lithology of C0002 grain (T. Nakamura et al., 2022) are ~0.86–0.87, ~0.01–0.015, and ~0.02–0.03. Assuming that the densities of serpentine, saponite, breunnerite, and dolomite are given by ~2.6, 2.3, 3.0, and 2.8 g/cm3 and that serpentine and saponite are present as a 1:1 molar ratio in Q001 (Fig. 12; Tomeoka and Buseck, 1988), the distribution of Mg between carbonates (Mgcarb) and phyllosilicates (Mgphyllo) is estimated to be Mgcarb/(Mgcarb + Mgphyllo) = 0.94–0.96. Considering heterogeneous distribution of carbonates in Ryugu samples (T. Nakamura et al., 2022; Tack et al., 2022; E. Nakamura et al., 2022), this estimate is well consistent with the Mg distribution to PC1 relative to the solar ratio (0.96/1.04 ≈ 0.92).

Fig. 21.

The ternary compositional diagram of matrix phyllosilicates of Q001 and of Orgueil (Fe-S-(Si + Al)). The Q001 matrix data shows a linear trend toward (Fe0.76, Ni0.24)0.93S (a closed circle on the Fe-S side), while the Orgueil matrix data is more scattered with a trend toward the Fe apex.

The same principal component analysis of Orgueil matrix data (Table S11) yields the PC1 of (Mg, Si + Al, Ca + Na + K, Fe, Ni, S) = (0.633, 0.478, –0.0258, –0.573, –0.0436, –0.201) that explains ~81% of data variability. The second principal component (PC2) of (Mg, Si + Al, Ca + Na + K, Fe, Ni, S) = (–0.0563, 0.661, –0.0343, 0.626, 0.0736, –0.401) covers ~13% of data variability. Because the PC1 for Orgueil matrix shows a higher Fe/S ratio than that for Q001 matrix, the PC1 for Orgueil matrix may correspond to addition/subtraction of Fe-rich and S-poor component rather than iron sulfides (Fig. 21). The Fe-rich and S-poor component is likely to be ferrihydrite that exists in the fine phyllosilicate in CI chondrites (Tomeoka and Buseck, 1988). Likewise, the PC2 for Orgueil matrix may be attributed to terrestrial weathering to form Mg-, Ca-, Na-sulfates that are also present in the matrix of CI chondrites (Tomeoka and Buseck, 1988). The S-rich PC2 component compensates not fully but partly for the S-poor nature of PC1 because sulfur forms coarser sulfates as observed in CI chondrites (e.g., Gounelle and Zolensky, 2014).

We thus conclude that the Q particles, found outside the Hayabusa2 sample container, are originated from asteroid Ryugu. They were expelled from the sample catcher in space prior to the container closing (Fig. 22), and placed in the gap between the sample container and the inner rid (Fig. 1). Evacuation of the reflector component of the sampler horn, which changed the direction of particles for sample collection, was made just before the container closing. Along with the reflector component, Q particles may have come out from the sample catcher. The finding of Ryugu particles outside the Hayabusa2 sample container will be useful lessons learned for future sample return missions to design the sample collection and the sample canister sealing system to fulfill scientific requirements.

Fig. 22.

A possible cause of escape of Q particles in space. The particles may have been expelled during the reflector evacuation operation for the container closing.

Conclusions

Two black-colored particles, found outside the Hayabusa2 sample container before the container opening, were investigated to identify their origin. Both particles, named as Q particles (Q001 and Q002), consist mainly of magnesian silicates with various mineral grains such as sulfides and oxides, which is distinct characteristics from the ablator material used for the reentry capsule and sand particles at the capsule landing site. Detailed mineralogical and petrographical investigation of the polished cross section of Q001 showed that its matrix consisted of magnesian phyllosilicates and contained iron sulfides (pyrrhotite and pentlandite), magnetite with a variety of morphologies, and carbonates (dolomite and breunnerite). It also contained phosphates, chromian spinel, ilmenite as accessory minerals. These mineralogical and petrographical characteristics of Q particles closely resemble those of Ryugu’s typical lithology and CI chondrites. There are clear differences between Q001 and CI chondrites. Q001 does not contain sulfates and ferrihydrite, commonly observed minerals in CI chondrites (Ryugu samples do not contain them either). The variation of matrix elemental composition of Q001 is explained by mixing of phyllosilicate and sub-micron-sized iron sulfides, while that of CI chondrites is not explained by such a mixing but probably by the presence of ferrihydrite and sulfates in the fine-grained matrix.

Based on a series of evidence, we conclude that Q particles are Ryugu grains that were expelled from the sample catcher in space before the sample container closing.

Acknowledgments

The authors express our gratitude to Trevor Ireland (University of Queensland) for helping the reentry capsule retrieval operation team as an international witness at Woomera, Australia. The authors thank Takuma Maruyama and Kentaro Hatakeda for their assistance during the sample container transportation from Australia to Japan and the preparation of the sample container opening. Valuable comments and suggestions from Makoto Kimura and Michael Zolensky, which improved the manuscript significantly, and the editorial handling by Kentaro Terada are deeply appreciated. This work was partly supported by JSPS Kakenhi grant (20H05846 for ST and RO).

References
 
© 2022 by The Geochemical Society of Japan

Copyright © 2022 The Geochemical Society of Japan. This is an open access article distributed under the terms of the Creative Commons BY (Attribution) License (https://creativecommons.org/licenses/by/4.0/legalcode), which permits the unrestricted distribution, reproduction and use of the article provided the original source and authors are credited.
https://creativecommons.org/licenses/by/4.0/legalcode
feedback
Top