A history of mild shocks experienced by the regolith particles on hydrated asteroid Ryugu

Micrometeorites, a possible major source of Earth’s water, are thought to form from explosive dispersal of hydrated chondritic materials during impact events on their parental asteroids. However, this provenance and formation mechanism have yet to be directly confirmed using asteroid returned samples. Here, we report evidence of mild shock metamorphism in the surface particles of asteroid Ryugu based on electron microscopy. All particles are dominated by phyllosilicates but lack dehydration textures, which are indicative of shock-heating temperatures below ~500 °C. Microfault-like textures associated with extensively shock-deformed framboidal magnetites and a high-pressure polymorph of Fe–Cr–sulfide have been identified. These findings indicate that the average peak pressure was ~2 GPa. The vast majority of ejecta formed during impact on Ryugu-like asteroids would be hydrated materials, larger than a millimetre, originating far from the impact point. These characteristics are inconsistent with current micrometeorite production models, and consequently, a new formation mechanism is required. Signatures of impact-induced shocks are found on Ryugu returned particles. Observations show that they happened at moderate temperatures (~500 °C) and pressures (~2 GPa) and did not dehydrate the particles substantially, suggesting that bigger meteoroids, rather than micrometeoroids, provide Earth with hydrated minerals.

Micrometeorites, a possible major source of Earth's water, are thought to form from explosive dispersal of hydrated chondritic materials during impact events on their parental asteroids. However, this provenance and formation mechanism have yet to be directly confirmed using asteroid returned samples. Here, we report evidence of mild shock metamorphism in the surface particles of asteroid Ryugu based on electron microscopy. All particles are dominated by phyllosilicates but lack dehydration textures, which are indicative of shock-heating temperatures below ~500 °C. Microfault-like textures associated with extensively shock-deformed framboidal magnetites and a high-pressure polymorph of Fe-Cr-sulfide have been identified. These findings indicate that the average peak pressure was ~2 GPa. The vast majority of ejecta formed during impact on Ryugu-like asteroids would be hydrated materials, larger than a millimetre, originating far from the impact point. These characteristics are inconsistent with current micrometeorite production models, and consequently, a new formation mechanism is required.
Hypervelocity collisions between small bodies were a frequent occurrence in the early Solar System 1 , resulting in a variety of outcomes such as brecciation, cratering, accretion, magma generation and degassing 2 . These processes have been recorded as a wide variety of physical and chemical modifications to asteroidal and planetary surface materials, including brittle/plastic deformation structures, solid-state phase transformations, recrystallization, melting and vaporization 2 . Therefore, planetary material scientists have been enthusiastically investigating shocked meteorites, which are thought to be derived from asteroids, the Moon and Mars 3 , to understand the nature of the impact events and related processes that have taken place throughout the Solar System's history [4][5][6] . For instance, peak pressure and its duration decoded from Article https://doi.org/10.1038/s41550-023-01947-5 not show planar fractures, as found in experimentally and naturally shocked olivines 2 (Fig. 1b,c).
Shock-induced melt veins and melt pockets are absent in all particles. However, particle C0014 exhibits rare, thin, straight veins less than 70 µm in length and less than 5 µm in width. In one of these veins, an aggregate of spherical magnetite particles has a thin lens shape, and the aggregate is terminated on both sides by elongated phyllosilicates along the veins (Fig. 2a,b). The phyllosilicate vein contains neither vesicles nor Fe-Ni-S nanoglobules, which are known to represent dehydration/melting products 37 . The chemical composition of the phyllosilicates is nearly the same as that of the fine-grained phyllosilicate matrix (Fig. 2c). Along the same direction as the vein, another framboidal magnetite aggregate is deformed in simple shear (Fig. 2a,b).
Four Ryugu particles (A0002, A0037, C0009 and C0068) were further examined at a higher spatial resolution by TEM. The ultrathin sections extracted from the above particles consist mainly of phyllosilicates containing Fe sulfides and Fe oxide grains ( Fig. 3a and Supplementary Figs. 1 and 2). Coarse-grained phyllosilicate aggregates of several micrometres in size occur with feathery textures embedded in fine-grained phyllosilicate matrices. The high-resolution TEM image and selected-area electron diffraction (SAED) patterns clarified that the phyllosilicates are intergrown serpentine and saponite, with interlayer spacings of 0.7 and 1.1 nm, respectively (Fig. 3b). Portions of the fine matrix phyllosilicates are poorly crystalline, showing diffraction rings with d spacings of 0.45, 0.25 and 0.15 nm (Fig. 3c). The chemical compositions of coarse-and fine-grained phyllosilicates in particles A0037 and C0068 have an Mg/(Mg + Fe) atomic ratio of 0.84 ± 0.03 (1σ) (Fig. 3d and  shocked meteorites provide constraints on past collisional velocities and also, the sizes of asteroids 7,8 . To link such impact event parameters to well-defined Solar System environments, knowledge of the relevant source regions is required. However, at present, there are only limited numbers of returned samples available for detailed study. Shock effects have been reported in lunar rocks returned by the Apollo missions 9,10 , cometary nucleus materials returned by the Stardust mission 11,12 , and the surface particles of the S-type asteroid Itokawa returned by the Japan Aerospace Exploration Agency ( JAXA) Hayabusa mission. In Itokawa particles, shock-induced defect structures (crystal lattice dislocations), melt splash, and microcraters were first identified in olivine grains 13,14 . The shock effects of the entire Itokawa rock could not be fully evaluated, despite some attempts based on crystallographic and spectroscopic analyses of olivine and plagioclase 15,16 . This is because most of the returned particles consist of single grains or aggregates of several grains only <50 µm in size.
Shock effects related to hydrated asteroids are of particular interest in planetary sciences as such asteroids are thought to be one of the major sources of extraterrestrial dust particles 17 . The bulk of extraterrestrial material arriving on Earth comes in the form of micrometeorites in the size range of 50-500 µm 18 , with an estimated flux of ~30,000 tons per year 19 . Most unmelted and partially melted micrometeorites are similar in chemistry and mineralogy to the matrices of CI (Ivuna-type), CM (Mighei-type) and Tagish Lake-type carbonaceous chondrites, which are porous aggregates composed mainly of hydrous phyllosilicates 20,21 . Such materials are likely to be explosively pulverized to become micrometeorites by the vaporization of volatile components, like H 2 O, during shock heating and therefore, are unlikely to survive as meteorites 22 . This hypothesis was further confirmed by comparing the petrology and mineralogy of recovered samples of anhydrous CV (Vigarano-type) and hydrated CM chondritic meteorites using laboratory shock experiments 23,24 .
JAXA's Hayabusa2 mission has provided an opportunity to directly evaluate the shock metamorphism of hydrated asteroidal materials. The Hayabusa2 spacecraft successfully brought back to Earth approximately 5.4 g of total material from the C-type asteroid 162173 Ryugu in early December 2020 25 . The returned materials include Ryugu's surface particles collected by the first touchdown (chamber A particles) and the second touchdown near the artificial impact crater (chamber C particles) formed by the small carry-on impactor (SCI) 25 . The present study aims to evaluate the degree of shock metamorphism of Ryugu's surface materials by scanning electron microscopy (SEM) and transmission electron microscopy (TEM) and to assess the hypothesis that massive dust production is driven by dehydration during impact processing on small hydrated Solar System bodies, including Ryugu.
The Ryugu particles examined in this study initially appeared unshocked; however, we found some characteristic features related to shock metamorphism. The observed bulk particles have irregular fractures, especially in those examples that are particularly carbonate rich 27,32 . Subparallel fractures, which have been reported in experimentally shocked hydrated CM chondrites 37 , were not observed (Fig. 1a). Another possible shock-related feature observed in SEM is the brittle deformation of olivine grains. In particle C0009, more than 10 grains of isolated olivine were identified in the phyllosilicate matrix 28,32 . The olivine grains (<50 µm) partly exhibit irregular fractures but do Notably, a unique Fe sulfide grain was observed in A0002 (Fig. 4a). This grain is euhedral and 1.2 µm in size with a pure FeCr 2 S 4 composition (Fig. 4b), and it is located within the fine-grained phyllosilicate matrix ( Supplementary Fig. 2a). This type of sulfide is known as daubréelite and has a cubic spinel structure. However, SAED patterns from the grain are indexed only with a monoclinic FeCr 2 S 4 phase known as a mineral zolenskyite, which was recently discovered in an enstatite chondrite 38 (Fig. 4c,d and Methods).

Discussion
Progressive petrographic and mineralogical transitions are used to classify meteorites and terrestrial impact crater rocks 2 . Brittle deformation features in bulk particles and their constituent minerals are criteria related to relatively low-shock grades in chondritic meteorites (shock stages S1-S3) 2,37 . This classification proposes that subparallel fractures in bulk meteorites and planar fractures in olivine grains are indicators of peak pressures above 15 (ref. 37) and 5 GPa (ref. 2), respectively. Hence, the absence of these types of fractures suggests that the peak pressure for Ryugu particles was lower than 5 GPa.
In the present study, we report a new constraint on peak pressure using a microfault-like vein found in particle C0014. The vein lacks any melting microtextures that are commonly observed in strongly shocked meteorites, such as vesicles and Fe-Ni-S nanoglobules 6,37 . The latter is evidence of a quenched mixture of immiscible silicate and metal sulfide melts. Therefore, this feature in C0014 is likely a fine-grained lithic vein formed by brittle cataclastic deformation. The elongation and offset of framboidal magnetites also suggest that the vein is likely a microfault produced by shock metamorphism. This study attempts to evaluate shock-induced faulting as an analogous process to the faulting that causes earthquakes on Earth. The stresses applied to such faults can be expressed by the following equation based on a fault mechanics analysis 39 : where τ, ΔT, C, d, ρ and D are the frictional shear stress, temperature increase, heat capacity of bulk rock, fault thickness, density of bulk rock and displacement along the fault, respectively. The frictional shear stress is converted into the mean stress P m ([σ 1 + σ 3 ]/2: σ 1 > σ 2 ≈ σ 3 ) using the following equation 40 : Article https://doi.org/10.1038/s41550-023-01947-5 where µ denotes the friction coefficient. Here, ΔT corresponds to localized temperature increase only at the microfault zone. We assume that the maximum ΔT of ~1,100 K is between the lowest Ryugu surface temperature (~300 K) 41 and the solidus temperature of hydrated carbonaceous chondrites corresponding to the upper temperature limit of frictional heating without melting (~1,400 K) 42 . The friction coefficient µ was set to 0.1 as a typical value in the high-speed friction of serpentine-clay-rich rocks 43,44 . C (865 J kg −1 K −1 ) 31 and ρ (1.79 g cm −3 ) 31 values are assumed to be those of typical Ryugu particles. D (48.9 µm) and d (5.3 µm) are estimated values from the observed dimensions of a deformed spherical framboidal magnetite aggregate (details are in Methods). By inserting the above parameters into equations (1) and (2), the upper bound of the mean stress, which is approximated as the peak pressure, is roughly 2 GPa (Fig. 5 and Methods). The particle C0014 could be an ejecta from the artificial impact crater made by the SCI. It could be suggested that the microfault-like veins were produced by shock deformation during the SCI cratering operation. However, the possibility of microfaulting being caused by the impact of the SCI can be ruled out because only <0.003 vol% of SCI ejecta would experience >2 GPa of pressure based on shock physics calculations (Supplementary Figs. 8-10).
Submicrometer-scale observations also provide evidence concerning the peak temperature and pressure experienced by Ryugu particles. Even at the TEM scale, we could not find any dehydration and melting textures of Mg-Fe phyllosilicates, such as abundant vesicles and the formation of fine Fe sulfide/oxide particles embedded in Si-rich amorphous silicate, which were reported in an experimentally shocked hydrated chondrite 37 . This suggests that the heating temperature did not exceed ~500 °C, which corresponds to the dehydration temperature of Mg-Fe-serpentine 45 (Supplementary Fig. 4). As recently reported, the stabilities of aliphatic carbon-rich organics and cubanite (CuFe 2 S 3 ) found in Ryugu materials also provide constraints on the upper temperature bound to 30 °C (ref. 27) and 210 °C (ref. 32), respectively. The detailed shock-induced thermal history of Ryugu's materials is uncertain since the cooling process after shock heating depends on the impactor/targe body sizes. Aliphatic carbon and cubanite, which are stable at low-temperature conditions, could be retained if the duration of high-temperature shock heating was very short (for example, seconds to hours) and fluctuated extensively. In contrast, phyllosilicates as the most abundant component are susceptible even within microseconds of shock heating and are dehydrated at >580 °C (ref. 24). Therefore, the peak temperature constraint from Mg-Fe-serpentine is more reliable. Previous hydrocode impact simulations of serpentine-rich materials were optimized to investigate the peak pressure and peak temperature relations of both CM carbonaceous chondrites (dominated by Mg-Fe phyllosilicates) and CI chondrites 46 . When the shock impedance of the CI-like Ryugu materials is adapted from that of CM chondrites, the upper bound of the peak pressure for the Ryugu materials is ~5 GPa (Supplementary Fig. 4).
The discovery of a unique Fe sulfide phase (zolenskyite) in particle A0002 strongly constrains the peak pressure. The phase equilibrium of FeCr 2 S 4 has been investigated experimentally 47 ( Supplementary  Fig. 5). Daubréelite is stable at ambient pressure and room temperature. At 700 °C, a disordered hexagonal NiAs-type phase was stable above ~4 GPa, and the transition pressure increased with decreasing temperature. The study also reported that an ordered monoclinic NiAs-type phase (zolenskyite) was recovered at 520 °C and 5.5 GPa. Although the precise phase relations of the hexagonal and monoclinic phases have not been clarified in detail, the higher density of zolenskyite (4.09 g cm −3 ) 38 compared with daubréelite (3.83 g cm −3 ) 48 suggests that zolenskyite is a high-pressure phase of FeCr 2 S 4 . Considering the maximum shock-heating temperature of ~500 °C for Ryugu particles as a whole, which was deduced from the survival of phyllosilicates, daubréelite would have transformed into zolenskyite in A0002 close to 2 GPa.
Recent high-resolution TEM analysis reported that some surfaces of Ryugu particles have space weathering features caused by solar wind irradiation and/or (micro-)meteorite bombardment 49 . Amongst them, the outermost frothy layer is interpreted as the melting products of the phyllosilicates-rich matrix by the latter process on Ryugu's surface. The shock effects are limited to near-surface portions of <1 µm in thickness; therefore, they do not represent the shock features of the whole particle. In the present study, we conclude that the average peak pressure of the Ryugu particles is ~2 GPa, corresponding to an impact velocity of ~1 km s −1 , based on petrological and mineralogical shock-related features (Supplementary Figs. 4 and 6); it should be also noted that the estimated peak pressure could have some uncertainty.
The occurrence of hydrothermal minerals, such as phyllosilicates, dolomite, and framboidal magnetite, indicates that the Ryugu surface materials experienced extensive aqueous alteration, as documented for hydrated chondrites 34,50 . However, the size of Ryugu (<1 km) is too small to maintain internal heat caused by 26 Al decay long enough for aqueous alteration to form hydrothermal minerals 51 , indicating that this asteroid must have originated from a larger precursor parent body. 53 Mn-53 Cr dating of carbonates in Ryugu particles suggests that the carbonates formed within 1.8 million years after Ca-Al-rich inclusion formation 29 , much earlier than estimated in previous studies 30,33 . This implies that the large precursor body needs to be broken into pieces smaller than 20 km in diameter earlier than or shortly after the carbonate formation age; otherwise, the amount of 26 Al in the large body at that time ( 26 Al/ 27 Al = ~10 −5 ) would yield enough heat to cause dehydration of hydrous minerals and then, melting subsequent to aqueous alteration 29 . Therefore, the present Ryugu, as a rubble-pile hydrated asteroid 52 , likely formed by the aggregation of fragments from a large impact event on the precursor body before the onset of any extensive thermal metamorphism in this larger body. Alternatively, the Ryugu materials originated following an impact on the CI/CM-like surface layer of a large differentiated precursor body during prograde thermal metamorphism as previously proposed for the asteroid Ceres 53 .
Recent impact experiments and hydrocode simulations 54 modelled the collision of a 20-km-diameter impactor asteroid and a 100-km-diameter target asteroid, both of which are made of hydrated and porous material, thus simulating Ryugu-like materials. Impact-induced volatile release during the collision was limited to only 2-4 wt% of the impactor mass, even at a typical impact velocity (6-7 km s −1 ) in the main asteroid belt 55 . Summarizing the theoretical and present observational results, most of the masses of collided asteroids would be merely mechanically disrupted without dehydration to form the rubble-pile Ryugu body. Only a small mass near the impact point on a Ryugu precursor body may have been extensively heated above the dehydration temperature of Mg-Fe serpentine and pulverized to comprise partly/completely dehydrated and submillimeter-sized dust particles, along with vapour release, as previously predicted by shock experiments 23,24 .
The complete lack of dehydration textures and mineral features in Ryugu particles observed in the present study demonstrates that hydrated asteroids preserved their water as hydroxyl in phyllosilicates throughout the impact events that they experienced. This study also suggests that the production of micrometeorites (<500 µm in size 18 ) due to shock heating-induced volatilization during such breakup events would be limited to the vicinity of the impact point, and its total amount is much smaller than previously expected 23 . Hence, the vast majority of hydrous materials from CI-like asteroids would come to the Earth as meteorites (>500 µm in size) rather than micrometeorites. This finding appears to be at odds with the fact that most unmelted and partially melted micrometeorites collected on the Earth's surface have genetic relationships with hydrated CI, CM, and Tagish Lake-like carbonaceous chondrites 20,21 . A possible explanation to reconcile this discrepancy is the breakup of hydrous meteorites due to aerodynamic heating when entering Earth' atmosphere 56,57 . However, solar wind noble gases retained in most of micrometeorites 58 suggest that those particles kept their original sizes when entering Earth's atmosphere.
Recent remote-sensing observations of the hydrated asteroid Bennu by National Aeronautics and Space Administration's OSIRIS REx spacecraft found that particles (centimetre scale or less) were being ejected repeatedly from the surface of the asteroid 59 . We hypothesize that the cracking of surface rocks of the hydrated asteroids due to thermal fracturing is an alternative mechanism to produce a large number of micrometeorites. Further analyses of the returned samples from Bennu will provide more detailed insights into how hydrated asteroids preserve their water and how this water can then be delivered to the Earth.

SEM
Ryugu's surface particles collected by the Hayabusa2 spacecraft were recovered from the reentry capsule and transported to the JAXA Curation Facility, Japan, without terrestrial atmospheric exposure 25 . After The numbers beside each curve represent temperature increase (Kelvin) at the microfault zone. The observed displacement along the fault is estimated to be 48.9 µm. Although average peak temperature is estimated to be below ~500 °C based on non-dehydration of Mg-Fe-serpentine 45 , a much higher temperature is expected within shear zones due to frictional heating. Given that the upper bound of the peak temperature is ~1,100 °C considering the solidus temperature of bulk carbonaceous chondrite 42 , the peak pressure in the Ryugu particle is estimated to be below ~2 GPa.
Article https://doi.org/10.1038/s41550-023-01947-5 cataloguing and preliminary analysis at the JAXA curation facility, eight particles up to 4.1 mm in size were allocated to the phase 2 curation Kochi team for in-depth investigations. An airtight sample transport vessel with a sample capsule pack made of sapphire glass and stainless steel 60 was used to avoid terrestrial contaminants during sample transportation among institutes. The particles were polished under dry conditions to avoid the elution of any material from the surface during polishing. The polished surface of each sample was examined using a scanning electron microscope ( JEOL JSM-7100F) equipped with an EDS (Oxford Instruments AZtec Energy) to obtain an overview of the mineralogy and textures of the samples using high-resolution imaging at the National Institute of Polar Research, Japan. The details of the sample transfers, processing and SEM observations are described in Ito et al. 27 and Yamaguchi et al. 32 .

Sample preparation using a focused ion beam instrument
Approximately 150-to 200-nm-thick sections of Ryugu particles were prepared using a focused ion beam (FIB) instrument (Hitachi High-Tech SMI4050) at the Kochi Institute for Core Sample Research (Kochi), Japan Agency for Marine-Earth Science and Technology. All the sections were extracted from unprocessed broken pieces of original particles immediately after removal from airtight sample transport vessels, which were filled with purified N 2 gas. The pieces were mounted on carbon tape and transported to the FIB chamber. After the deposition of tungsten protection layers, regions of interest (up to ~25 × 25 µm 2 ) were cut out and thinned using a Ga + ion beam at an accelerating voltage of 30 kV and then, finalized at 5 kV and a probe current of 40 pA to minimize surface ion-damage layers. Subsequently, the ultrathin sections were mounted on scaled-up Cu grids (Kochi grid 60 ) using a micromanipulator equipped with an FIB.

TEM
Eleven FIB sections with and without scanning transmission X-ray microscopy-near edge X-ray absorption fine structure spectroscopy (STXM-NEXAFS) and high spatial resolution secondary ion mass spectrometry (NanoSIMS) analyses were examined using a transmission electron microscope ( JEOL JEM-ARM200F) operated at an accelerating voltage of 200 kV at Kochi, Japan Agency for Marine-Earth Science and Technology. Microtextural observations were performed by bright-field TEM and high-angle annular dark-field scanning TEM. Mineral phases were identified using SAED and lattice-fringe imaging (high-resolution TEM), and chemical analyses were performed using EDS with a 100-mm 2 silicon drift detector and the JEOL Analysis Station 4.30 software. For quantitative analyses, the intensities of the characteristic X-rays of each element were measured using a fixed acquisition time of 30 s, beam scan area of ~100 × 100 nm 2 , and beam current of 50 pA in the scanning TEM mode. The (Si + Al)-Mg-Fe ratios of the phyllosilicates were determined using experimental thickness-corrected k factors obtained from a natural pyrope-almandine garnet standard.

Pressure estimation by microfault analysis
In the present peak pressure estimation by fault mechanics calculations, shear stress on the microfault plane was calculated by equation (1) in the text using the following physical properties for Ryugu particles: 865 J kg −1 K −1 at 298 K for heat capacity C and 1.79 g cm −3 for density ρ (ref. 31). As for temperature increase ΔT, none of decomposition features were recognized in the phyllosilicates in association with the microfault-like vein observed by SEM. This may suggest that the phyllosilicates experienced temperatures below 500 °C 45 . However, in contrast to the entire particle heated to the mean shock temperature, the locally heated portions near the microfault should be cooled more rapidly by thermal conduction to the relatively cold surrounding materials. If the dehydration kinetics of phyllosilicates is sluggish, it is likely that the phyllosilicates were not decomposed even above 500 °C during transient heating by faulting. Therefore, the upper temperature limit of microfault was set to the solidus temperature of the bulk Ryugu particles (~1,100 °C) in the present calculations.
To estimate the displacement D of the microfaulting, we assumed a simple shear model for sphere-ellipsoid deformation. Shear strain γ is expressed by the following equation: where r 0 is the radius of the predeformed spherical object. The shear strain γ was calculated using the following equation: where a and b are the long and short axes of an elliptically deformed object, respectively 61 . The radius r 0 (9.5 µm) of an aggregate of original spherical framboidal magnetite was estimated by measuring the total area of the elongated framboidal magnetite aggregate in an observed backscattered electron image taken from particle C0014 using ImageJ software. The long axis a (45.2 µm) and short axis b (5.3 µm) of the framboidal magnetite were directly measured by the backscattered electron image. From these measured parameters and equations (3) and (4), the displacement D of the microfault was estimated as 48.9 µm. Finally, the mean stress P m was calculated using the frictional shear stress τ from equations (1) and (2), as shown in the text. Although fault mechanics analysis and shock physics calculations are not simply compared with each other, the mean stress P m ([σ 1 + σ 3 ]/2: σ 1 > σ 2 ≈ σ 3 ) in the former analysis 40 determined by equations (1) and (2) is comparable with peak pressure termed as the average stress ([σ 1 + 2σ 3 ]/3: σ 1 > σ 2 ≈ σ 3 ) in the latter analysis 62 . Recent shock physics calculations using the iSALE code clarified peak pressure and deviatoric stress history in shock metamorphism for the particles in granite as a model target material 62 . The largest component of deviatoric stress is maximized around the timing of peak pressure. In addition, the differential stress becomes a maximum by the time when pressure drops to several tens of percent of the peak pressure. A microfault is likely to be the most displaced immediately after the end of compression phase (that is, at the beginning of decompression phase) because the normal stress to the fault starts to decrease.

Phase identification of zolenskyite
Single-crystal electron diffraction patterns from an FeCr 2 S 4 grain were indexed by the zolenskyite unit cell with the space group C2/m and lattice parameters a = 1.284 nm, b = 0.344 nm, c = 0.594 nm and β = 117° (ref. 38).

Data availability
All data needed to evaluate the conclusions are present in the paper and Supplementary Information. They will also be put on the Japan Aerospace Exploration Agency Data Archives and Transmission System (https://www.darts.isas.jaxa.jp/curation/hayabusa2) after a 1-year proprietary period. Source data are provided with this paper.