Photometry of (162173) Ryugu and its artificial crater as inferred by Hayabusa2/ONC images

ABSTRACT

1 INTRODUCTION

The JAXA/Hayabusa2 mission rendezvoused with the (162 173) Ryugu Cb-type asteroid from 2018 June 27^th to 2019 November 13^th (Watanabe et al. 2017; Tachibana et al. 2021) at a distance of 20 km or lower and sent four rovers on its surface. During this period, Hayabusa2 performed two touchdown and sampling operations, i.e. TD-1 (2019 February 21^th) and TD-2 (2019 July 11^th), as well as an artificial impact experiment. Ryugu was observed by the Optical Navigation Camera (ONC; Kameda et al. 2017), while its composition was assessed by the Near Infrared Spectrometer (NIRS3, Iwata et al. 2017).

Ryugu is a rubble-pile, top-shaped asteroid, with a polar and an equatorial radius of about 450 m (Watanabe et al. 2019) and 500 m (Sugita et al. 2019), respectively. Its albedo is overall low, with an average value of 0.040 ± 0.005 (Tatsumi et al. 2020).

The artificial impact experiment occurred on 5^th April 2019 in the Small Carry-on Impactor (SCI) mission stage, by shooting a 2 kg bullet to the Ryugu surface at a speed of 2 km/s (Arakawa et al. 2020). The artificial crater area is named as Omosubi–Kororin crater, is centred at about 8°N 301°E and extends between 7°–10°N latitudes and 299°–303°E longitudes. The crater has a maximum depth of 2.7 m and includes boulders larger than 60 cm, even if boulders of different size (from tens of centimeters to meters) were driven away in the surrounding areas (Arakawa et al. 2020). The experiment exposed the Ryugu’s subsurface, providing the possibility of studying it in great detail. Composition and spectral signatures of the crater area are similar to the surroundings, although with a slightly larger magnesium phyllosilicates abundance and bluer spectral slope, probably due to fresher material, less exposed to solar wind irradiation (Kitazato et al. 2021; Tatsumi et al. 2020; Galiano et al. 2022). Although Ryugu particles get finer due to space weathering and therefore coarser particles are expected in fresh terrains (Yumoto et al. 2024), finer particles are instead found within the Omosubi–Kororin area (Arakawa et al. 2020; Ogawa et al. 2022), since dark and red fine particles rose from subsurface as consequence of the artificial impact. This is also evidenced by NIR spectra, showing absorption bands (centred at 2.7 and 2.8 microns) deepening, reflectance darkening and local slope reddening (Galiano et al. 2022).

Photometric functions, i.e. spectral parameters behaviour as a function of illumination and viewing angles, are a powerful tool to study optical and physical properties of a planetary surface. By means of the Hapke model, Tatsumi et al. (2020) obtained a phase function reproducing the Ryugu’s observations (including ground-based ones) not corrected for shape model and a phase function modelling the images corrected for shape model. They found an average geometric albedo of 0.04. Pilorget et al. (2021) applied the Hapke model to NIR spectra provided by NIRS3, revealing photometric properties uniformity across the Ryugu surface.

Longobardo et al. (2022) also studied the NIR Ryugu photometry by applying an empirical model based on NIRS3 statistical analysis. They found that the Ryugu’s phase function is similar to other dark asteroids (e.g. Ceres and Bennu), while an anomalous band depths and phase reddening behaviour has been found and ascribed to dark albedo and to particle smoothness, respectively.

Preliminary studies focused on photometric properties of the artificial crater area have suggested a steeper phase function (Honda et al. 2022) and probably a rougher surface (Yokota et al. 2022).

This work aims at studying the photometric properties of the artificial impact area before and after the impact, in order to highlight difference between surface (observed before the impact) and subsurface (exposed after the impact) of Ryugu. The same empirical method applied on NIRS3 data is applied to retrieve firstly the average Ryugu phase function and then the SCI area phase functions in the visible range.

Details on the used ONC data set are given in section 2, while the method is described in Section 3. Results are presented in Section 4 and discussed in Section 5. Finally, conclusions are summarized in Section 6.

2 DATA

After its arrival to Ryugu, the Hayabusa2 spacecraft underwent three BOX-B operations (i.e. tour observations from an altitude of 20 km over the Ryugu surface) and one BOX-C operation (i.e. from a 5 km altitude). The first touchdown was followed by a new BOX-C operation and then by the Crater Search Operation (CRA-1, from a 1.7 km altitude), lasted from 20^th to 22th March 2019, which was preparatory to the SCI experiment. The SCI mission stage lasted from 3^rd to 6^th April 2019. Then, a new Crater Search Operation (CRA-2) took place from 23th to 25^th April. The second touchdown occurred after a few low-altitude observations. Finally, two BOX-B and three BOX-C operations, as well as the MINERVA lander deployment, were performed before the departure from Ryugu.

The ONC (Honda et al. 2023) includes one telescopic (ONC-T) and two wide-angle cameras (ONC-W1 and ONC-W2). While the two wide-angle cameras are panchromatic, ONC-T includes a wheel with seven bandpass filters, ranging from 390 to 950 nm. The instrument FOV is 6.27° × 6.27° and the spatial resolution is 0.1 mrad/pixels. This means that during the BOX-B operations ONC-T imaged the Ryugu surface with a resolution of 2 m/pixels, which improved to 0.17 m during the CRA stages.

To study the Ryugu photometry and compare the obtained results with other asteroids, we selected the entire data set acquired in the v band (i.e. 549 nm), a wavelength where most of disc-resolved asteroid phase curves are available, during the CRA-1, SCI and CRA-2 mission stages: this data set includes about 40 million pixels from 38 images, with a phase angle ranging approximately from 15° to 40°.

To obtain the Omosubi–Kororin phase curve, we selected from the data set above all pixels comprised between 4.5°–10°N latitudes and 299°–305°E longitudes. However, the phase angle interval of this data set slightly changes from before to after the impact, being 15°–25° before the impact and 20°–30° after the impact. To minimize the influence of the phase angle range, we repeated the analysis on a different data set, including 16 images (8 before and 8 after the impact) acquired throughout the entire mission at similar spatial resolution: this set of images (listed in Yokota et al. 2022) allowed to cover a similar and wider phase angle range (i.e. 0–35°) both before and after the impact. Moreover, the crater area was restricted to 7°–10°N latitudes and 303°–305°E longitudes, to minimize the influence of crater surroundings. This new data set includes about 70 thousand pixels (35 thousand before and 35 thousand after the impact).

3 METHOD

We applied the semi-empirical procedure based on a statistical analysis of the data set, already exploited for other small bodies such as Vesta (Longobardo et al. 2014), Lutetia (Longobardo et al. 2016), 67P/Churyumov-Gerasimenko (Longobardo et al. 2017), and Ceres (Longobardo et al. 2019). To obtain the photometrically corrected reflectance, we considered the v band radiance factor R and applied the following steps:

• Retrieval of the equigonal albedo to remove the topography influence (i.e. incidence i and emission e angles). There are many disc functions in literature suitable for this aim, but photometry studies on other dark small bodies (e.g. Longobardo et al. 2017; 2019) demonstrated that the Akimov parameterless disc function (Shkuratov et al. 2011) provides generally the best results. Therefore, we applied this disc function and then verified its goodness by assessing the equigonal albedo behaviour as a function of incidence and emission angles. The Akimov disc function is defined as follows:

  $$\begin{eqnarray} D\left( {\beta ,\gamma ,\varphi } \right) = \mathrm{ cos}\frac{\varphi }{2}\cos \left[ {\frac{\pi }{{\pi - \varphi }}\left( {\gamma - \frac{\varphi }{2}} \right)} \right]\frac{{{{\left( {\mathrm{ cos}\beta } \right)}}^{{\raise0.7ex\rm{\varphi } \!{\left/ {\vphantom {\varphi {\pi - \varphi }}}\right.-} \!\rm{{\pi - \varphi }}}}}}{{\mathrm{ cos}\gamma }}, \end{eqnarray}$$

  where |$\varphi $| is the phase angle, |$\gamma = {\mathrm{ arctan}}\frac{{\ cos i - \ cos e\ cos \varphi }}{{\ cos e\ sin \varphi }},$| is the photometric longitude, |$\beta = \textit{arccos}\frac{{\cos e}}{{\cos \gamma }}$| is the photometric latitude. The equigonal albedo was calculated as R/D, where R is the radiance factor.

• Retrieval of the median equigonal albedo at each phase angle, by considering phase angle bins 1° wide. Phase angle bins without observations are not considered in the following step.

• Least squares fit of median equigonal albedo as a function of phase angle. In this case, due to the narrow phase angle interval, a linear fit is sufficient to model the phase function.

The comparison of small body phase functions is generally based on the definition of two photometric parameters, i.e. the reflectance at a 30° phase angle (R30) and the phase function steepness between 20° and 60° (PCS₂₀₆₀): the parameters are calculated on radiance factor curves, in order to allow a fast and easy comparison with phase functions available in literature. In this work, due to the narrow phase angle interval available, photometric parameters were redefined as R20 (i.e. reflectance at 20° phase angle) and PCS₁₅₄₀ (i.e. phase curve steepness between 15° and 40°) and recalculated on disc-resolved phase functions of Vesta dark, average and bright terrains (Longobardo et al. 2014), Steins (Jorda et al. 2008), Annefrank (Hillier, Bauer & Buratti 2011), Eros (Li et al. 2004), Gaspra (Helfenstein et al. 1994), Lutetia (Longobardo et al. 2016), Mathilde (Clark et al. 1999) and Ceres (Longobardo et al. 2019).

For photometric analysis of the artificial crater area, we applied the same method and compared the phase function before and after the impact. Due to the smaller data set considered, uncertainties on phase curves and on photometric parameters are larger. To minimize uncertainties, only phase angle bins having at least 100 observations were considered and reflectance median values having a statistical error larger than 1 per cent were discarded. The reliability of photometric functions comparison was maximized by calculating the phase curve steepness in three different intervals, i.e. 0°–30° (PCS₀₃₀), 10°–30° (PCS₁₀₃₀) and 0°–10° (PCS₀₁₀).

4 RESULTS

The equigonal albedo obtained by applying the Akimov disc function is quite independent of incidence (Fig. 1) and emission angles.

Figure 1.

Equigonal albedo in the v band obtained by applying the Akimov parameterless disc function as a function of incidence angle (all phase angles between 15° and 40° are considered).

Open in new tabDownload slide

For incidence angles lower than 60°, the median equigonal albedo variations are lower than 2 per cent, well below the acceptability threshold of 4 per cent, corresponding to a 0.001 variation. Between 60° and 70°, a residual decreasing trend is observed, corresponding to 8 per cent variation. Nevertheless, data with incidence angles above 60° are not statistically significant and hence do not affect the retrieved phase function. Therefore, in the following steps we considered the Akimov-corrected equigonal albedo.

The Ryugu phase function obtained between 15° and 40° phase angle is shown in Fig. 2. Its slope is –(4.6 ± 0.1)·10^–4 deg^-1.

Figure 2.

Ryugu’s median equigonal albedo as a function of phase angle.

Open in new tabDownload slide

To retrieve the photometric parameters, we considered the phase function calculated on radiance factor values measured at incidence and emission angles lower than 30°: this selection allows us to minimize the i and e influence on photometric parameters. The obtained R20 is 0.01 (for a homogeneous comparison with other asteroids, it was rounded at the second digit), while PCS₁₅₄₀ is (45 ± 2) per cent. PCS₁₅₄₀ as a function of R20 is shown in Fig. 3. Uncertainties are not shown for clarity: errors on R20 are less than 0.001 and therefore are included within the symbol size, while error bars on PCS are about 2 per cent.

Figure 3.

Phase Curve Steepness (PCS) between 15° and 40° phase angles as a function of radiance factor at 20° phase angle. Circles include asteroids belonging to the same taxonomic class. C-type asteroids (small circles) are Mathilde and Ceres, S-type asteroids (crosses) include Gaspra, Eros and Annefrank, Lutetia is the only X-type asteroid (triangle), the achondritic asteroids are V-type (asterisks), i.e. Vesta bright, average and intermediate terrains, and E-type (square), i.e. Steins. Uncertainties are not shown for clarity: errors on R20 are less than 0.001 and therefore are included within the symbol size, while error bars on PCS are about 2 per cent.

Open in new tabDownload slide

The asteroids belonging to the same taxonomic class are generally in the same area of the scatterplot, but two exceptions arise, i.e. the Vesta dark terrains (V-type, having a larger PCS than other V- and E-type asteroids) and Ryugu (C-type, having a flatter PCS than Mathilde and Ceres).

The Omosubi–Kororin phase function calculated on the CRA-1, SCI and CRA-2 data set has a slope of –(3.8 ± 0.4)·10^–4 deg^–1 before the impact and –(3.3 ± 0.3)·10^–4 deg^–1 after the impact.

The phase curves obtained by considering the entire phase angle range (second analysis) are shown in Fig. 4.

Figure 4.

Omosubi–Kororin phase curves before (black symbols and curve) and after (red symbols and curve) the impact. For colours, refer to the web version of the paper.

Open in new tabDownload slide

In both cases, phase function was modelled by a 3^rd order polynomial:

Phase function parameters and PCS calculated for different phase angle intervals before and after the impact are reported in Table 1.

5 DISCUSSION

5.1 Ryugu phase function

The retrieved phase function slope (i.e. –4.6 ± 0.1·10^–4 deg^–1) is consistent with the phase function derived from NIRS3 data (i.e. –4.1 ± 0.1·10^–4 deg^–1; Longobardo et al. 2022) within two times the error. This result is noteworthy, given the differences in spectral range, spatial resolution and number of observations (which can influence uncertainty in statistical analysis). However, the consistency likely reflects Ryugu’s similar albedo in both the visible and near infrared ranges (Sugita et al. 2019; Pilorget et al. 2021), suggesting comparable optical properties.

5.2 Ryugu versus other asteroids

In the R20 versus PCS₁₅₄₀ scatterplot, shown in Fig. 3, the Vesta dark regions and Ryugu are ungrouped, differently than other asteroids. The Vesta dark regions are a mixture of HED (achondrites) and carbonaceous chondrites (i.e. the main components of C-type asteroids) (Palomba et al. 2014) and therefore in the R20-PCS scatterplot they locate between achondritic and C-type asteroids (Longobardo et al. 2016).

Otherwise, the Ryugu phase function is flatter than other C-type asteroids, despite a similar R20. We calculated PCS₁₅₄₀ on the disc-resolved phase function retrieved by applying the Hapke model on ONC images (Tatsumi et al. 2020), obtaining even in this case 45 per cent. Therefore, this result is independent of the applied methodology. Nevertheless, the PCS₁₅₄₀ retrieved on the disc-integrated phase function by Tatsumi et al. (2020), which considers radiance factors not corrected for the shape model, is 60 per cent, closer to other C-type asteroids. This indicates that spatial resolution does play a role in the phase function steepness.

Longobardo et al. (2016) studied the spatial resolution role on bright asteroid (i.e. S-type and V-type) disc-resolved phase functions: at increasing spatial resolution, phase function becomes steeper because albedo heterogeneities are highlighted. This has been observed on both Vesta (Longobardo et al. 2014) and Ida (Helfenstein et al. 1996).

The reverse behaviour is observed on Ryugu, where increasing spatial resolution leads to a flatter phase function. In the case of a dark asteroid such as Ryugu, albedo spatial distribution is more homogeneous and therefore no phase function steepening is observed. On the contrary, phase function can steepen because of shadowing due to morphological and topographic variations. These variations are obviously minor if we consider smaller and smaller areas (i.e. when improving spatial resolution), and this may explain the phase function flattening on a dark asteroid. The same behaviour is observed on other dark asteroids, i.e. Ceres (Longobardo et al. 2019) and Bennu (Golish et al. 2021).

Basing on all asteroid disc-resolved observations and phase functions retrieved so far, we found that PCS variations are observed for spatial resolutions better than about 100 m, even if this threshold could be not the same for all asteroids but depending on their size.

5.3 Omosubi–Kororin phase function

When only CRA-1, SCI, and CRA-2 data are considered, the phase function of the artificial crater area seems slightly flatter after the impact, even if within the uncertainty. However, this result has two issues: (1) the large uncertainity due to the narrow phase angle range (only 10° width); (2) the different phase angle interval before (i.e. 15°–25°) and after (i.e. 20°–30°) the impact, which makes the phase function steeper before the impact because of the higher opposition effect influence (despite its minor role at these phase angles).

For this reason, the analysis was repeated by considering a wider data set and the same phase angle range before and after the impact. The two new phase functions still agree within uncertainties (Table 1), indicating small variations between surface (observed before the impact) and subsurface (exposed after the impact). In particular, the phase function steepness slightly increases after the impact. Provided that this steepening is within the uncertainty and therefore we cannot discard that it is not real, the reasons for the observed behaviour can be the following:

• Particle size. In the case of coarser particles, phase function is steeper (Hapke 1981). This would mean that the impact would have exposed coarser particles from the subsurface. This is in line with the particle evolution on Ryugu, which get finer due to space weathering (Yumoto et al. 2024), but it is not consistent with the presence of finer dust expected in the artificial crater because of its raise in the first, transient stages after the impact (Kadono et al. 2020; Wada et al. 2021). Therefore, it is unlikely that particle size can justify the observed phase functions.

• Roughness and filling factor. A larger roughness, from microscopic up to pixel scale (i.e. 20 meters), could steepen the phase function (Hapke 1981). A rougher surface inside the crater would be confirmed by the post-impact increase of the Hapke roughness parameter found by Yokota et al. (2022). A lower particle filling factor would be related to both larger porosity and roughness and can be revealed by a lower opposition width (Hapke 1993). The latter is indeed observed in the post-impact phase function (Fig. 4), giving a further indication that a roughness increase and a filling factor decrease can justify the observed photometric behaviour. The lower subsurface porosity (and larger roughness, once exposed) would agree with observations of other Ryugu impact craters (Sakatani et al. 2021) and with Ryugu formation models, predicting that subsurface includes remnants of porous and primitive (i.e. less heated and compacted) planetesimals (Neumann et al. 2014). Nevertheless, a lower porosity/roughness variation is expected within the artificial crater with respect to other impact craters, because the Omosubi–Kororin area shows a smaller difference of both thermal inertia (Sakatani et al. 2021) and boulders size frequency distribution (Grott et al. 2020) between crater interior and exterior. The small roughness variations would justify the pre- and post-impact phase functions consistency within errors.

6 CONCLUSIONS

The statistical analysis of the ONC data set acquired during the CRA-1, SCI, and CRA-2 stages of the Hayabusa2 mission revealed that:

• the Akimov disc function is able to remove the topography effects on Ryugu;

• the Ryugu’s phase functions in the visible and near-infrared ranges are similar within two times the errors, in line with the similar albedo and optical properties in the two spectral intervals, as already observed also for the phase reddening (Longobardo et al. 2022);

• the disc-resolved phase function of Ryugu is flatter than other C-type asteroids (Ceres and Mathilde), while the disc-integrated one better reproduces the dark asteroid phase functions.

The latter result highlights the role of spatial resolution in the phase function retrieval. While an improving spatial resolution leads to a phase function steepening on bright asteroids, due to albedo heterogeneities, the opposite behaviour is observed on dark asteroids due to the reduced shadowing role.

The artificial crater area’s phase function study based on the same data set above is poorly reliable due to the different phase angle range between before and after the impact, so the analysis was repeated on a different data set covering the entire phase angle range up to 35°. This study revealed a slight phase function steepening and narrowing after the impact. A larger particle size is a possible explanation for that, but it is not in agreement with spectral and morphological considerations. The most plausible interpretation is that the impact exposed a rougher and more porous terrain, as suggested by Ryugu formation models (Sakatani et al. 2021). Nevertheless, it is important to note that the variation in the phase function after the impact falls within the statistical uncertainty, likely due to the minimal change in roughness. Therefore, this result can be further validated by considering a larger data set (to minimize errors) or by cross checking with other photometric studies adopting different methods.

ACKNOWLEDGEMENTS

The Hayabusa2 spacecraft was developed and built under the leadership of Japan Aerospace Exploration Agency (JAXA), with contributions from the German Aerospace Center (DLR), and the Center National d’´Etudes Spatiales (CNES), and in collaboration with NASA, Nagoya University, University of Tokyo, National Astronomical Observatory of Japan (NAOJ), University of Aizu, Kobe University, and other universities, institutes, and companies in Japan. We also thank the engineers who contributed to the success of Hayabusa2 operations.

This work is founded by the Italian Space Agency (ASI) and it has been developed under the agreement 2022–12-HH-0.

DATA AVAILABILITY

Hayabusa2/ONC data are available at https://sbnarchive.psi.edu/pds4/hayabusa2/hyb2_onc/.

REFERENCES