OTES data analysis of spectroscopic variation due to temperatures change between day and night offer a previously unexplored method to infer mineralogical composition of Bennu’s surface. The consistency of OTES spectra with those of low petrologic type CM, CR and CI chondrites (Hamilton et al. 2019, 2021, 2022), which reflect a combination of various aqueous altered silicates, sulfides, oxides, carbonates, organics, and other mineral phases in the meteorite matrix, indicates that an intense and pervasive aqueous alteration was experienced by the asteroid’s parent body (Kerridge 1985, Miyamoto and Zolensky 1994).
A dependence of the RB intensity of minerals and meteorites with temperature was observed in laboratory (Poggiali et al 2021), inferred by thermal properties of meteorites (Opeil et al 2020) and is attributed to the composition of the sample. The laboratory results showed that highly hydrated meteorites and minerals show a bell shape trend in RB area and intensity when the temperature increases, with a maximum at about 250 to 300 K (Poggiali et al 2021). This phenomenon is linked to the unusual non-linear and negative thermal expansion of layered structure of phyllosilicates (Opeil et al 2020) that dominate the mineralogy of the matrix of low petrologic type carbonaceous chondrite meteorites (Buseck et al 1993) where the content of the intra- or inter-layer OH− group in the mineral structure prevail, as for the hydrated silicates of the serpentine group. This behavior is likely a consequence of the electrostatic properties of the interaction between OH− and OH–O with the surrounding metals in the serpentine group minerals, which is responsible for the anomalous contraction of crystal structure at low temperature (Balan et al 2010, Bish and Johnson 1993). On the contrary, the lower the abundance of hydroxyl incorporated in the silicate structure, the more monotonically the RB intensity varies (Poggiali et al 2021). Such behavior is attributed to the lessening presence of OH in the silicate structure. Minor amounts of the order of or less than tenths of percentage of hydrous component in olivine and pyroxene are thus not sufficient to induce an anomalous behavior of RB variation at low temperature. This allows the trend of the RBs with temperature to be more linear. Furthermore, Poggiali et al. (2021) observed in the laboratory that the Christiansen feature (CF) peak positions shift towards higher wavenumbers when the temperature of these meteorite samples decreases. Hence, by comparing laboratory measurements with OTES emissivity analysis, we formulate the hypothesis that Bennu’s surface is made of fragments with different levels of hydration.
Before totally attributing the changes observed on Bennu to differences in the composition of the surface we must consider other possible factors that could induce the spectroscopic variation. A plausible cause of spectroscopic variation between daytime and nighttime observations is the presence of spatially variable thin dust deposits covering the surface, as identified in OTES spectra (Hamilton et al., 2021). Even if the surface of Bennu lacks a regolith containing abundant fines (DellaGiustina et al 2019), the presence of thin deposits of dust < 50 µm thick cannot be excluded based on thermal modeling (Rozitis et al 2020). The lower thermal inertia of such dust could be responsible for a different equilibrium temperature with respect to the rocks underneath. However, the influence of such thin layers of dust on the temperatures of regions with low and high thermal inertia is negligible (Rozitis et al 2020). The contribution to the Bennu’s temperatures from a hotter dust layer during daytime, and from a cooler dust layer at night, is only about ± 10 K. Because the thermal inertia of Bennu is dominated by the porosity (Rozitis et al. 2020, Cambioni et al. 2021), thermophysical modeling has shown that a surface with thermal inertia similar to that of Bennu reaches equilibrium temperature in less than one hour (Delbo et al. 2015).
Thus, changes of RB intensities among observations at different temperatures and times of day cannot be easily justified by the presence of material with different thermal inertia or a thin layer of dust. Instead, what we identified is linked to differences in mineralogical composition of rocks. Specifically, the inhomogeneities are due to degree of hydration of the mineral present on Bennu surface as observed in laboratory experiments (Poggiali et al 2021, Opeil et al 2020). Laboratory measurements revealed that CM2 meteorites shows negative thermal expansion (NTE) at temperatures found on Bennu (Opeil et al 2020). This implies that phyllosilicates with high water hydroxyl content, such those found on CM, CI, and CR meteorites, expand and contract more than anhydrous or poorly hydrated minerals within a temperature range of few tens of degrees around 235 K, where the minimum NTE occurs. The magnitude of such fast expansion and contraction could be responsible of breakup of boulders producing particle ejecta as observed on Bennu (Lauretta et al. 2019).
The two distinct behaviors observed in laboratory experiments (Poggiali et al 2021) can therefore be identified within the temperature range observed on Bennu when plotting the values of RB intensities of a specific area (covered by a limited number of spectra) with respect to each average temperature determined by the individual spectra (see Fig. 4). In the laboratory, a wider range of temperature was investigated from 50 K to 350 K (Poggiali et al 2021). However, despite the smaller temperature range observed on Bennu between 230 and 330 K, the two trends are clearly recognizable. The regions colored in blue (Figs. 3 and 4) show a flat trend corresponding to the maximum of the bell shape observed in the laboratory and thus with a high content of hydrated minerals. The regions colored in green are, on the other side, associated with a decreasing trend linked to the presence to anhydrous (or poorly hydrated) minerals.
Detailed multivariate analysis showed very small variation in the depth of the 2.74 µm band, which implies a uniform distribution of hydrated phyllosilicates on Bennu (Barucci et al 2020), with global average distribution of H2O/OH in the range of 0.80–0.91 wt% (Praet et al 2021). However, the presence of a very limited amount of hydrated minerals, of only 1% in mass, and with grain size of a few tens of micrometers covering grains of anhydrous silicates is enough to profoundly modify the near infrared spectrum. The intense band at 2.74 µm characteristic of hydrated silicates dominate the spectrum that would otherwise not have been observed if only anhydrous silicates were present. In contrast, the mid-infrared spectrum is significantly less affected by the presence of hydrated fine grains and the spectroscopic characteristics of anhydrous minerals continue to be dominant (Poggiali et al. 2023).
Despite the ubiquitous spectroscopic presence on Bennu of hydrated minerals (Hamilton et al 2019), we found that the temperature-dependent variation of mid-infrared features disentangled regions on the surface characterized by higher content of hydrated minerals and regions characterized by the presence of anhydrous minerals. Differences in hydration level of carbonaceous asteroids were inferred from the study of CM and CI meteorites (McAdam et al 2015) but never observed in a single asteroid.
The presence of two distinct mineralogical components that dominate the rock composition, places new constraints on the origin and evolutionary history of Bennu. The highly hydrated component found on Bennu would have originated from its undifferentiated primitive parent body with a chemical and mineralogical composition similar to that found as major component of carbonaceous chondrites. This component would not have undergone intense transformation processes during the impact that destroyed the parent body and thus allowing the hydrated component to remain intact or less altered. The material rich in hydrated silicates would indicate the least modified component which would continue to retain the most primitive characteristics of the parent body. However, in the impact scenario in which Bennu originated from a single catastrophic disruption of a single large parent asteroid with size > 100 km (Walsh et al. 2013, Bottke et al. 2015, Michel et al. 2020), the asteroid would be formed from a mix of materials from the surface and the interior of the parent body made of fragments that experienced variable heating temperatures and compaction (Michel et al. 2020).
Portion of carbonaceous chondrite meteorites subjected to thermal metamorphism were analyzed (Hanna et al. 2020, Tonui et al. 2014, Nakamura 2005). Some CM and CI chondrites contain anhydrous minerals that are thought to have formed due to dehydration processes caused by the high temperature (Nakamura 2005, Nakato et al. 2008 and references therein). Laboratory analyses showed that samples of nontronite, montmorillonite, chlorite, kaolinite, prehnite, and serpentine at temperatures > 600°C decompose, resulting in phase transformation and loss of volatile components (Fairén et al. 2010). Due to dehydroxylation, saponite above 800°C undergoes a phase transition becoming enstatite (Che et al. 2011). Thus, during the disrupting impact, a fraction of the parent body hydrated silicates lose H2O and OH− resulting from the high temperature achieved during the shock leading their decomposition into anhydrous forsterite and enstatite.
Furthermore, impacts would have played a role in creating defects and vacancies in the anhydrous silicates, providing active sites where hydrogen from the solar wind can react and be implanted on Bennu, giving rise to siloxane. Experimental data demonstrated that 5 keV protons are easily implanted in olivine giving rise to hydroxyl species responsible for the 2.7-µm band, which saturates at fluence of ~ 2⋅1017 H+/cm2 (Schaible and Baragiola 2014). In this latter scenario, anhydrous minerals like olivine, possibly originated in a different parent body, if present on the surface would be modified by space weathering induced by the solar wind, which would give rise to the formation of hydroxyl groups in layers as deep as a few microns. Considering the proton solar flux of 2⋅108 H+ cm− 2 s–1, the OH skin-layer responsible for the 2.7-µm band would have enough time to be formed during the 10–65 Myr lifetime of Bennu (Bierhaus et al. 2022). Despite the heterogeneity of the material observed on Bennu, the evidence that the position of the peak of the OH band is stable at 2.74 ± 0.01 µm over the entire surface, would suggest a second generation process such as space weathering involving the entire asteroid. However, this hypothesis must be confirmed by a detailed laboratory study on the returned samples. In any case, the mid-infrared wavelengths at each EQ would concern the entire crystalline structure of anhydrous silicates rather than just the external proton irradiated skin. Thus, the RB intensity trends with temperature observed in this work would have been driven by the crystalline lattice of silicates with low hydration. This consideration can lead to the hypothesis that Bennu could be the results of the collision of two different type of asteroids with different composition, partially mixed by the re-accumulation and homogenized on the surface by space weathering action. This kind of scenario was suggested to explain the nature of meteorites Kaidun (Zolensky and Ivanov 2003) and Almatha Sitta (Bishoff et al 2010), which are a peculiar case, to date two of the few examples of meteorites that contain a mix of materials that come from many different asteroids. The parent body from which Kaidun originated accumulated very different materials from collisions of bodies of the main belt which underwent various levels of physical processing, heating, shock, fusion, and aqueous alteration leading to a variety of lithologies in the meteorite, mainly carbonaceous and enstatite chondrites of the types EH3–5, EL3, CV3, CM1-2, and R (Zolensky and Ivanov 2003). Furthermore, the heterogeneous composition of Lutetia has been proposed to result from the collision of two distinct objects made of carbonaceous and enstatite chondrites, respectively (Barucci et al 2012). From this perspective, Bennu would be an asteroid of second or even third generation, where the catastrophic disruption of the parent bodies caused a family of fragments with different levels of hydration.
To evaluate the statistical significance of the RB ratio variation across the Bennu surface observed in Fig. 3, we took the RB ratio for every facet for each of the EQs divided by EQ2 and plotted the distribution of values on a standard histogram. Then we use Phyton SciPy library curve fit tool to identify the best Gaussian to reproduce data distribution (Extended Data Fig. 6). We runed D’Agostino test, which refused the hypothesis that the distributions were Gaussian with Pvalues on the order of magnitude of 10− 16 (to be unable to reject hypothesis Pvalues need to be greater than 0.05). Then, we iteratively reduced the amplitude of the best fit using an elbow analysis, as shown in Extended Data Fig. 7. Once found and removed the gaussian distribution, we were able to isolate the statistical significance component shown in Fig. 5. The statistical analysis revealed that Bennu surface is made of at least of 9% of anhydrous or less hydrated materials coming alternatively from different parent bodies or being thermally shocked during the impact. This result, obtained by a new independent method, indicates that Bennu is a heterogeneous asteroid composed of large boulders that may have different origins or have undergone different evolutionary histories, making Bennu a possible source of falling heterogeneous meteorites such as Kaidun and Almatha Sitta.