Reassigning CI chondrite parent bodies based on reflectance spectroscopy of samples from carbonaceous asteroid Ryugu and meteorites

The carbonaceous asteroid Ryugu has been explored by the Hayabusa2 spacecraft to elucidate the actual nature of hydrous asteroids. Laboratory analyses revealed that the samples from Ryugu are comparable to unheated CI carbonaceous chondrites; however, reflectance spectra of Ryugu samples and CIs do not coincide. Here, we demonstrate that Ryugu sample spectra are reproduced by heating Orgueil CI chondrite at 300°C under reducing conditions, which caused dehydration of terrestrial weathering products and reduction of iron in phyllosilicates. Terrestrial weathering of CIs accounts for the spectral differences between Ryugu sample and CIs, which is more severe than space weathering that likely explains those between asteroid Ryugu and the collected samples. Previous assignments of CI chondrite parent bodies, i.e., chemically most primitive objects in the solar system, are based on the spectra of CI chondrites. This study indicates that actual spectra of CI parent bodies are much darker and flatter at ultraviolet to visible wavelengths than the spectra of CI chondrites.

and  λc  are the reflectance of the continuum and the measured reflectance at the wavelength λc, respectively.Finally, the width of the band or component is considered as the Full Width at Half-Maximum (FWHM).The number of components and the wavelengths at the points of contact with the continuum were adjusted for each spectrum.The quality of the fitting model was evaluated by the difference between the modeled and measured spectra in the wavelength range considered, i.e. the residue.The fitting model with the lowest residue was considered the best model.No constraints were applied to the position and amplitude of the components.The errors on the fit model and the calculated band parameters are determined using the bootstrap statistical method.This necessitates the fit and calculation of the parameters a large number of times after addition of a small fluctuation on the data.Here 500 iterations of the bootstrap are performed.This leads to a list of 500 models, thus 500 series of band and component parameters.The distributions of these parameters allow the consideration of their center as the best fit, and their FWHM as the corresponding error.The fluctuations added to the data for the bootstrap method should correspond to the measurement error.In this case, we considered the measurement error as the standard deviation of the residue.EMG fitting was performed without all information on the measurements and the samples in order to remove bias regarding possible links between spectra of the same sample or environmental conditions.Ultraviolet reflectance spectroscopy Ultraviolet (UV) reflectance spectra (0.20-0.40 µm in wavelength) of heated and unheated Orgueil powder samples were measured using a UV spectrometer equipped with a deuterium light source (H2D2 L11799, Hamamatsu Photonics K.K.) and a Maya2000 PRO spectrometer (Ocean Optics, Inc.) at Tohoku University.The measurements were performed in ambient conditions and with an incidence angle of 30° and an emission angle of 0°.Samples were placed in a stainless-steel dish and covered by a dark aperture made of black flock paper to avoid effects of stray light.A quartz powder was used as a standard material and the reflectance ratio of a sample to the quartz powder was calculated in the following way: where ′  is the reflectance of a sample to the quartz powder,  (30°,0°) and  (30°,0°) are the measured intensities of a sample and quartz powder at (i, e)=(30°, 0°), respectively,  (15°,15°) and  (15°,15°) are the measured intensities of the quartz powder and Vacuum UV (VUV) mirror (Edmund Optics) at (i, e)=(15°, 15°), respectively, and  400 is the reflectance ratio at 400 nm of the quartz powder measured by FTIR at (i, e)=(30°, 30°) to that measured by the UV spectrometer (i, e)=(15°, 15°).Note that VUV mirror exhibits >85% and >80% reflectance at >220 and 170-220 nm, respectively.The reflectance values of a sample were calculated using a wavelength calibration standard (Labsphere, ID: CSTM-WCS-MC, a whitish disk made of the oxide of a rare earth element) and 99% Spectralon (Labsphere, ID: HL-TXH-025-A, RELAB) in the following way: where   is the reflectance of a sample and  (30°,0°) ,  −−(30°,0°) , and  99%(30°,0°) are the measured intensities of the quartz powder, the wavelength calibration standard, and the 99% Spectralon, respectively.The absolute reflectance of 99% Spectralon from 200 to 400 nm was assumed as 1 because the absolute values provided by Labsphere only covers longer wavelengths than 250 nm.A blackout curtain covered the system during measurements to avoid stray lights.UV spectra were normalized by the average reflectance at 390-400 nm.
Grain size effects on reflectance spectra of Orgueil CI chondrites Reflectance spectra of Orgueil samples sieved to different size fractions were measured to investigate grain size effects on the spectral features.The Vis-NIR reflectance spectra of Orgueil with a maximum grain size of 2000, 512, and 155 µm and the chip sample are characterized by an intense UV drop-off feature, a small shoulder at ~0.5 µm attributed to Fe oxides/hydroxides, an absorption band at 1.95 µm due to hydrated minerals, including gypsum, and an absorption band at ~2.3 µm due to Mg-OH (39).The chip sample shows the brightest and bluest spectrum at visible wavelengths, whereas the powder samples with a maximum grain size of 2000 µm have the lowest reflectance and blue-sloped spectra (Fig. S12A and B).The samples with a grain size of <512 and <155 µm show redder and brighter spectra than those with a grain size of <2000 µm.
The metal-OH absorption band depth at 2.71 µm due to Mg-rich phyllosilicates, which is a distinct spectral feature of Orgueil, increases, relative to the adjacent H2O absorptions, with decreasing grain size (Fig. S12C).The 2.71 µm OH band remains distinct from the H2O bands at ~2.8-3.1 µm, allowing its strength to be accurately determined.The MIR region exhibits a reflectance peak due to phyllosilicates (~9.8 µm) (28) and small peaks due to sulfates (8.65 and 8.90 µm) (49,58), none of which exhibit significant changes due to grain size when they are scaled (Fig. S12D).
Surface shape model based on X-ray computing tomography Synchrotron X-ray computed tomography (SR-CT) data of grains was obtained at BL20XU Spring-8 after spectroscopic measurements at Tohoku University, and was used to investigate how a surface was oriented with respect to incident and emittance angles for different spectral measurements.The details of SR-CT analytical procedures are described in supplementary materials of (13).A microscopic shape model of the surfaces from which reflectance spectra were measured was produced using CT data based on a picture of each coarse grain taken before spectral measurement to clarify the sample orientation and the footprint of incident light on the sample surface.CT data using binning of 8 and 4 voxels of original CT data with 0.85 micron/voxel of spatial resolution for C0002 and the other coarse grains, respectively.
A normal vector at each sample surface gridded by micron-scale was calculated to evaluate the relationship (or correlation) between sample surface tilt or orientation and the phase angle used for reflectance spectral measurements.Figure S14 shows sample surfaces irradiated by incident light as a projection of a point in the direction perpendicular to the plane of the incident and emittance angle (y-z plane).Black and blue arrows indicate the direction of the incident angle (30°), the mean value of the normal vector to each small area (micrometer-scale) on the surface of the sample, respectively.Red arrows indicate the specular reflection directions calculated from the angles of incidence and the normal vector of the sample assuming that the sample has a specular surface.The red arrows of A0064_settingA (Fig. S14B) and A0067_settingA (Fig. S14D) point toward the detector, suggesting that anomalous effects by specular reflection are intense when the irradiated surfaces are flat enough to produce a mirror-like surface.For instance, the red arrow in A0067_settingA (Fig. S14D) points closer to the direction of the detector (0°) than that in A0067_setting B (Fig. S14E), which suggests that reflectance spectra measured from A0067_settingA (Fig. S14D) should be affected more strongly by specular reflection from the surface than that from A0067_setting B (Fig. S14E).
Synchrotron-based X-ray diffraction (S-XRD) Fine particles (<155 µm in size) picked up from matrix grains of heated and unheated Orgueil samples were analyzed to identify their mineral phases by Synchrotron X-ray diffraction analysis using a Gandolfi camera at undulator beam line 3A in the Photon Factory, High Energy Accelerator Research Organization, Japan.For this experiment, each particle was glued using glycol phthalate to the top of carbon fiber that stands on a thin glass tube.X-rays at a wavelength of 2.17 Å were applied to the samples, and the exposure time was 13-20 minutes depending on the sample size.

Scanning transmission X-ray microscopy (STXM) and µ-XRF X-ray absorption near edge structure (XANES) analysis
To investigate average Fe 2+ /Fetotal in typical phyllosilicates in a sample, scanning transmission Xray microscopy and µ-XRF X-ray absorption near edge structure (XANES) analysis at Fe Kedge using a 1.0 µm × 1.0 µm X-ray beam was performed in air-tight conditions for a particle of Orgueil (~150 µm in size) heated at 300 °C for 50 hours in a reduced condition.A particle of the unheated Orgueil was also analyzed in the same way for comparison.Detailed analytical procedures are described in (13).
Elemental analysis for carbon contents Carbon contents of unheated and 500 °C-heated Orgueil were measured using an Elemental Analyzer (EA), FLASH 2000 (Thermo Scientific) at Tohoku University.Powder samples (~2 mg in weight) were analyzed three times.S5.

Fig. S14. A cross-sectional view of each coarse sample during spectral measurements.
A plane parallel to the plane formed by the incidence and emergence angles is shown, and values perpendicular to this plane are projected onto this plane.A sample top surface is indicated by gray dots.The black, blue, and red arrows show the direction of the incident light, the mean value of the normal vector to each small area on the surface of the sample, and the specular reflection direction that is expected from the angle of incidence and the normal vector of the sample assuming the sample were a specular surface.The angle size of a specular reflection direction (a red arrow) is determined by the angle between the incidence angle (a black arrow) and the normal vector (a blue arrow).Each panel indicates an individual Ryugu grain set for a measurement.Sample sizes shown in the panels are scaled to fit in each panel.(A) A0026 with a rugged surface, (B) A0064 with a flat surface, (C) the same grain as (B) but in a different setting, (D) A0067 with a flat surface, (E) the same grain as (D) but in a different setting, (F) A0094 with a flat surface, (G) C0002 with a rugged surface, (H) the same grain as (G) but rotated 120 degrees to the right about the z-axis, (I) the same grain as (G) but rotated 240 degrees to the right about the z-axis, (J) C0025 with a rugged surface, (K) C0046 with a rugged surface, and (L) C0076 with a rugged surface.The Ryugu grains A0067 and A0094 have a thin amorphous layer covering the uppermost part of phyllosilicates, indicating space weathering (15).The shift of RB peak position towards longer wavelengths is consistent with the results of helium ion irradiation experiments on CI chondrites that simulated space weathering (54).Ryugu TD1 and TD2 samples were collected from the undisturbed area and the area containing excavated materials on the asteroid, respectively (10,11,36).

Fig. S19
. Reflectance spectra of a black sheet (Fineshut-Kiwami).The spectrum was measured at incidence, emergence, and phase angles of 30°, 0°, and 30° in the principal plane, respectively by a bruker FTIR at Tohoku University.*The presence of minerals is expressed as "✓".Each mineral is abbreviated as: serp: serpentine, sap: saponite, gyp: gypsum, mag: magnetite, pyr: pyrrhotite, dol: dolomite, ol: olivine, wü: wüstite, tae: taenite, ka: kamacite, and px: pyroxene.†The heat stage is assigned based on the nomenclature for XRD patterns of naturally-heated carbonaceous chondrites proposed by (23).‡All the data were obtained from a small particle after exposure to air, and gypsum in the 300 °C-heated samples are expected to have been hydrated by atmospheric interaction after heating and FT-IR measurements.§With small peaks.6) † Effective band width of the ONC-T band filters (6) ‡ The starting wavelength of the ul-band defined in this study is at a longer wavelength than that of the ONC-T band filters due to the wavelength limits of the FT-IR measurement.

Fig. S1 .
Fig. S1.Vis-NIR reflectance spectra (0.4-3 µm).(A and B) Ryugu samples, (C) carbonaceous chondrite chip samples, (D) experimentally-heated Orgueil powder samples, and (E) carbonate powder samples.All the spectra are normalized at 0.55 µm (v-band) and shifted arbitrarily.Gray lines indicate 0.55 µm.Some spectra exhibit a concave-up toward UV wavelengths probably due to specular reflection from a sample surface, whose intensities may depend on the sample orientation.Some Ryugu samples show a "shoulder" at 1 µm, or the turning point that connects the rapid reflectance increase from ~0.5 to 1 µm and the slight reflectance increase to NIR wavelengths.The presence of the 1-µm shoulder may also depend on the sample orientation because Ryugu grains A0064 and A0067 show 1-µm shoulder features with different intensities.Low signal levels at ~0.38-0.5 µm in wavelength may affect the spectral data, particularly resulting in an abrupt change in slope towards the UV region.

Fig. S3 .
Fig.S3.A synchrotron X-ray diffraction pattern of a matrix grain from LAP 04721 CR2 chondrite.The matrix grain consists mainly of phyllosilicate (serpentine) and magnetite.

Fig. S4 .
Fig. S4.(A) The v-band reflectance and (B) the reflectance ratio of preheated and heated Orgueil powder and Ryugu TD1 and TD2 samples.A low reflectance ratio of the ul-band (0.40 µm) reflectance to the v-band (0.48 µm) reflectance indicates a deep UV drop-off feature.

Fig. S5 .
Fig. S5.Spectral parameters of Ryugu samples compared to hydrated carbonaceous chondrites.(A) The band depth and the maximum absorption position of the 2.7-µm absorption band.(B) The v-band reflectance and the reflectance ratio of ul-band and v-band.(C) The v-band reflectance and the spectral slope from b-band to x-band.(D) The peak position of Christiansen Feature and Reststrahlen Band of Si-O stretching.

Fig. S6 .
Fig. S6.An EMG fitted result of the 2.7-µm absorption band of the Ryugu grain C0002.Four components are assigned and the maximum absorption positions of the components are located at 2.71, 2.91, and 3.07 µm.

Fig. S13 .
Fig. S13.Vis-NIR reflectance spectra of Tagish Lake C2 chondrite with various grain sizes obtained (A) at RELAB and (B) Tohoku University.No large Vis-NIR spectral variations are observed in the plots depending on the sample forms of Tagish Lake meteorite.The spectral data shown in (A) can be found in Reflectance Experiment Laboratory (RELAB) database with IDs summarized in TableS5.

Fig. S15 .
Fig. S15.Spectral variations in relation to specular reflection direction.(A) v-band reflectance, (B) b-x slope, (C) 2.7-µm band depth, and (D) the ul-band/b-band reflectance ratio as a function of expected angles of specular reflection direction based on the investigation as shown in Fig. S14.Red and blue dots indicate the spectral data obtained from a flat surface (A0064, A0067, and A0094) and a rugged surface (A0026, C0002, C0025, C0046, and C0076), respectively.The dots of the same grains measured with different orientations twice (i.e., A0064 and A0067) or three times (i.e., C0002) are connected by lines with each other.

Fig. S17. The maximum absorption positions of the 2 . 7 -
Fig. S17.The maximum absorption positions of the 2.7-µm band of Ryugu samples determined by Exponentially Modified Gaussian fitting(42).The gray line indicates 2.710 µm.The Ryugu coarse grain A0067 has an absorption band at a longer wavelength than the other samples.

Fig. S18 .
Fig. S18.Positions of Christiansen feature and Reststrahlen Band of Ryugu samples.TD1 and TD2 Ryugu samples are marked in red and blue, respectively.A magnified view of Fig. 9D.The Ryugu grains A0067 and A0094 have a thin amorphous layer covering the uppermost part of phyllosilicates, indicating space weathering(15).The shift of RB peak position towards longer wavelengths is consistent with the results of helium ion irradiation experiments on CI chondrites that simulated space weathering(54).Ryugu TD1 and TD2 samples were collected from the undisturbed area and the area containing excavated materials on the asteroid, respectively(10,11,36).

Reflectance spectra of the mixture of montmorillonite and carbon lamp black (A) at Vis-IR wavelengths and (B) at MIR wavelengths.
All the spectra are normalized at 2.595 µm and shifted arbitrarily.Gray lines indicate 2.71 and 3.05 µm.Purple lines indicate 3.40, 3.95, 4.20, 4.58, 5.50, 6.40, and 7.05 µm for carbonate features.Orange lines indicate 5.83 µm for the carbonyl peak.Reflectance and absorption features of montmorillonite become low and faint, respectively, by adding even less than a few percent of carbon black.Masking effects due to carbon black are more intense at Vis-NIR wavelengths than at MIR wavelengths.The spectral data can be found in Reflectance Experiment Laboratory (RELAB) database with IDs summarized in Table S5.Orgueil Vis-IR spectra measured from a whitish surface and a surface after the removal of whitish particles.Gray lines indicate 0.5 µm for a ferrihydrite feature and 8.45, 8.60, and 8.92 µm for sulfate peaks.(C) MIR spectra of ferrihydrite, Mg sulfates (kieserite and hexahydrite), Ca sulfates (anhydrite and gypsum), Ca sulfate in Paris (CM) (58), and a whitish surface and sulfate-removed surface of Orgueil (CI).Gray lines indicate 8.45, 8.60, and 8.92 µm for peaks of sulfates in Orgueil.

Table S6 . Summary of the experimentally-heated Orgueil samples investigated
Analyses performed are indicated by "✓".Blank cells indicate not analyzed.* The FT-IR measurements of reflectance spectra.† The powder samples were sieved to a grain size of <155 µm.‡ The grain size of <2000 and <512 µm were investigated.§ The two surfaces were characterized: a surface of Orgueil chip A after removing white particles by hand and a sulfate-rich surface of Orgueil chip A. Start point of the band (nm) 384.9 ‡ 466.2 532.6 581.9 685.3 835.4 916.7 End point of the band 414.5 490.8 562.2 594.2 712.3 877.3 970.8 * Effective wavelength of the ONC-T band filters with respect to solar spectrum (