Characterization of asteroid analogues by means of emission and reflectance spectroscopy in the 1- to 100-µm spectral range

Solar system small bodies (SSSB) are remnants of the process that led to the formation of the solar system from the solar nebula. Information on their composition and physical properties is a key to understand the formation and evolution of our solar system. Asteroids and comets may contain the molecular precursors (organic molecules such as amino acids) to the origin of life on Earth and could possibly be the source of water that formed the Earth’s oceans.

When electromagnetic radiation of a specified range of wavelengths passes through matter, energy is absorbed at some of the wavelengths and not others, by a variety of processes. By observing which wavelengths a molecular species absorbs, and to what extent it absorbs them, we can gain information about the nature of the electronic, vibrational and rotational transitions that a molecule is able to undergo and thus information about its structure. The infrared region of the electromagnetic spectrum, for instance, contains energies corresponding to the vibrational frequencies of bonds in organic molecules (stretching and bending) (Farmer 1974).

The near-infrared (NIR) region from 1 to 5 μm contains primarily information on overtones and combinations for phyllosilicates, most sorosilicates, hydroxides, some sulfates, amphiboles, carbonates, soil water and organic matter; the thermal-infrared (TIR) region from 5 to 100 μm contains mostly information on Si–O lattice vibrations for silicates (quartz, feldspars, clay), other than mafic, carbonate mineral group and organic compounds (Chabrillat et al. 2013).

To fully understand the mineralogical composition of SSSB and the implications to solar system formation and the nature of water in the universe, it is important to investigate and integrate data over the whole spectral range from 1 to 100 µm.

In the last years, Earth- and space-based spectroscopic infrared observations have helped to constrain the chemical and mineralogical composition of several small bodies.

During its cruise to the comet 67P/Churyumov–Gerasimenko (Heinemann et al. 2013), the ESA Rosetta spacecraft performed flybys of asteroid 21 Lutetia and 2867 Steins (Glassmeier et al. 2007). Vernazza et al. (2011) show spectral similarities of Lutetia to E chondrites in the Vis–NIR spectral range. Combining together the information available from ground observations, space observations and Rosetta data, Barucci et al. (2012) concluded that spectral variations on Lutetia’s surface are explained by different surface compositions and textures. Furthermore, Lutetia differs from all asteroids visited from spacecraft missions, as well as asteroids observed from ground. Its surface is composed of materials similar to chondrites: Some southern regions are similar to carbonaceous chondrites (like CV, CO, CK), while the northern hemisphere shows more similarity to a mixture of enstatite and carbonaceous chondrites.

Rosetta’s flyby of Steins was too short and distant to get enough data to understand its surface composition. Telescopic observations with the infrared spectrograph (IRS) of the Spitzer Space Telescope concluded that Steins emissivity spectrum is similar to the enstatite achondrite meteorites and to the mineral enstatite, confirming the rare E-type classification already suggested on the basis of ground-based spectral and polarimetric observations (Barucci et al. 2008).

The NASA Dawn mission, launched in 2007 to study two proto-planets of the asteroid belt Vesta and Ceres (Russell et al. 2007), is currently in orbit around the dwarf planet Ceres.

Ceres is a very complicated object, which experienced complex internal chemistry processes that may have extended to the surface. Spectral observations identified hydrated phases (clays, silicates, salts) leading to a possible comparison with water-rich carbonaceous chondrite material (McCord and Sotin 2005, Rivkin et al. 2011).

Vesta (a volatile-depleted body), largely recognized as the source of the howardite–eucrite–diogenite (HED) meteorites, thought to be the result of collisions of several proto-planet bodies with Vesta (Burbine et al. 2001). A weak hydration signature at 3 μm was detected from telescopic observations (Rivkin et al. 2006; Hasegawa et al. 2003), but not confirmed by later observations. The VIR imaging spectrometer on Dawn detected OH (hydroxyl) bands on regions of Vesta’s surface (De Sanctis et al. 2012). The distribution of the hydration band on Vesta is likely related to differences in the mineralogy of its surface materials and/or the presence of exogenous OH-bearing materials. The analysis of samples returned by the Hayabusa spacecraft has confirmed the LL chondrite composition of Itokawa (Nakamura et al. 2011), with the identification of olivine and pyroxene as major components, followed by plagioclases. Carbonaceous materials among the particles collected by the Hayabusa spacecraft sample catcher are discussed in Uesugi et al. (2014) and Kitajima et al. (2015).

A new light on surface composition of SSSB will come from two future missions. JAXA launched in 2014 the Hayabusa2 mission to collect a sample from a C-type asteroid 1999 JU3 and bring it back to Earth (Saiki et al. 2013). This class of asteroids is thought to contain more organics and water, linking them closely to the origin and distribution of volatiles (see also King et al. 2015). The mission features a near-infrared spectrometer (Iwata et al. 2014), a mid-infrared imager (Okada et al. 2012) and a radiometer with four narrow-band spectral channels on the lander (Grott et al. 2013).

Several previous studies have been devoted to building databases of VNIR (visible near infrared)-TIR spectra of meteorites and analogues. Among them, Salisbury et al. (1991) measured reflectance (mostly bi-conical) spectra in air of many meteorites in the 2.5- to 13-µm spectral range. Emissivity (in air) spectra for large grain size samples (710–1000 µm) in the 5- to 45-µm spectral range were reported by Christensen et al. (2000). Moroz et al. (2006) concentrated on hydration features of some meteorites, acquiring bi-conical reflectance between 2.5 and 14 µm. The ASTER spectral library (Baldridge et al. 2009) contains hemispherical reflectance of natural and man-made materials measured in air in the 0.4- to 15-µm spectral range. The USGS spectral library described in Clark et al. (2007) contains reflectance measurements in dry-nitrogen atmosphere for the spectral range 0.2–25 µm (and often extended to 150 µm). The NASA RELAB database contains reflectance spectra of meteorites in the 0.3- to 2.5-µm spectral range (Gaffey 1976). Cloutis and colleagues (cf. Cloutis et al. 2013 and references therein) collected reflectance spectra of meteorites in the spectral range from 0.3 to 2.5 µm. Maturilli et al. (2008) and Maturilli and Helbert (2014) measured emissivity of minerals and rocks for several grain size ranges in purged air between 2.5 and 50 µm. Morlok et al. (2010, 2012) measured mid-infrared (5–25 µm) transmission/absorption spectra of meteorites. This list is not complete but features a representative selection of the typical spectral libraries that are available for the interpretation of remote sensing infrared measurements of asteroid surface. Despite the high scientific value of existing spectral libraries, for the most direct comparison with radiation emitted from an airless SSSB a spectral database of emissivity measurements acquired with a high signal-to-noise ratio in vacuum over a large spectral range and for small grain sizes is needed.

At PEL we conducted spectroscopic measurements of several meteorites and other analogues for asteroid surfaces. Here we present emissivity and reflectance of seven fine-powdered surface analogues, which were measured under vacuum in the 1- to 100-µm spectral range. These data form an important set of asteroid analogues that will directly support the data analysis of ongoing and future missions. It can also help to plan future missions and provides important input for thermal models of small bodies (cf. Davidsson et al. 2015).

Figures 1, 2, 3, 4, 5, 6 and 7 show the spectral measurements for the asteroid analogue materials we selected for our sample set. For each figure we split the spectra in 2 panels, for a better visualization of the data. In the left panel we show emissivity measured in vacuum and under purging and 1-reflectance measured under vacuum in the NIR and MIR (medium infrared) spectral range (1–16 µm). In the right panel we show emissivity and 1-reflectance measured in vacuum in the thermal-infrared (TIR) spectral range (16–100 µm). An additional upper X-axis reports wave numbers (cm⁻¹) to facilitate the plot comprehension. In a recent paper we show that bi-conical reflectance spectra of samples measured at PEL under purging or under vacuum conditions are identical (Maturilli et al. 2015), but the measurements under vacuum are preferable because they are free of atmospheric spectral features.

Emissivity and 1-R (R is bi-conical reflectance) show systematic differences under vacuum, which is already extensively discussed in the literature (cf. Salisbury et al. 1994). However, it is important to note that in a powdered sample heated in vacuum, vertical thermal gradients develop and are mostly influenced from the intensity of vacuum, and from the material thermal conductivity, that is very strongly controlled by particle size because of its effect on porosity (Salisbury et al. 1994). Also, it is not surprising that our 1-R spectra are much similar to emissivity measurements taken under purging environment (where the sample temperature is vertically homogeneous, and also the Kirchhoff’s law is valid) and differ significantly from emissivity measurements taken under vacuum (where thermal gradients arise in the sample), as already pointed out in Salisbury et al. (1994). The vacuum intensity in our external chamber is “medium” (typical values between 30 and 0.001 mbar) and is not fully comparable with the high vacuum of outer space. Therefore, only from the point of view of vacuum, we expect higher thermal gradients on asteroid surface than in our sample cups. On the other hand we chose the smallest standard grain size range used at PEL (<25 µm). Previous studies (e.g., Logan et al. 1973) show that most of the particles in this range fall below the 2-µm grain size range, which results in an increase of the vertical temperature gradient in the sample, because the sample porosity is one of the key factors driving the intensity of thermal gradient. We can therefore conclude that our instrumental setup and sample selection are reasonably good approximations reproducing asteroid surficial emissivity spectra.

We discard data below 4 µm because of too low signal-to-noise ratio, due to the low temperature used to heat the samples in this experiment. In the spectra measured under purging some features of atmospheric gas absorption (cf. CO₂ band at 15 µm) are still present, due to slight imperfections in the air purging mechanism.

Carbonaceous chondrites display large chondrules of Mg-rich olivine, which provides the most prominent bands of the Allende sample spectra (Fig. 1). The lack of well-shaped O–H stretching vibration bands between 2 and 3 µm suggests a low-grade alteration of the sample (Salisbury and Hunt 1974; Miyamoto 1988). Lack of H–O–H bending vibration at 6.1 micron in the spectra under vacuum precludes the presence of surface adsorbed water in the sample. Most of the vibrational bands observable between 8 and 16 µm appear strongly reduced in the vacuum emissivity spectrum with respect to the purged emissivity, due to thermal gradients in the sample (Logan and Hunt 1970) and to the weakening (disappearance in high vacuum) of the transparency feature in vacuum (Salisbury et al. 1991). On the other hand some overtones below 7 µm are more prominent with respect to the other spectra shown in Fig. 1. The deep emissivity minimum (comparable to the transparency feature of olivine) at 12.5 µm suggests a dominant surface scattering and confirms the fine grain size of the sample (Salisbury and Walter 1989; Mustard and Hays 1997).

Millbillillie meteorite spectra (Fig. 2) confirm the small grain size range of the sample by the enhanced transparency feature around 12.4 µm. The Christiansen feature (CF) broadens due to the combination of different silicates as both Ca-rich pyroxene and plagioclase are present, whereas the weak Reststrahlen bands below 16 µm appear obliterated in the vacuum emissivity spectrum, as already shown in Salisbury et al. (1991). Spectra of achondrites such as Millbillillie show strong bands at 19.8 µm and 23.9 µm that can be due to orthopyroxene and Ca-rich pyroxene (Morlok et al. 2012). Similar results arise from the Murchison meteorite spectra in Fig. 3. The broadness and lack of structure of the 10-μm features in the spectra of Murchison may result from the overlapping of Si–O absorptions of various phyllosilicates composing the matrices of CM chondrites (Moroz et al. 2006). The vacuum emissivity spectrum of Murchison reveals the weak emissivity maximum (reflectance minimum) near 10.5 μm that is typical of olivine. Another confirmation of olivine content in Murchison is found with the strong band at 19.6 µm (Morlok et al. 2010). The band at 22.5 micron is typical for serpentine phyllosilicate, while the band at 27 µm can be found in graphite and montmorillonite.

Typical phyllosilicates band around 9.5–10 µm, well shaped both in montmorillonite and in serpentine spectra (Figs. 6, 7), appears strongly suppressed in Murchison measurements, whereas bands around 3 and 6.5 µm, characteristic of hydrated phases (Morlok et al. 2010), are still prominent.

Low-Ca Mg-rich orthopyroxenes are the main constituent of E-type asteroids and—in combination with olivine and iron–nickel metals—of ordinary chondrites. Spectra of a natural sample of enstatite (Fig. 4) display a distinct CF at 8.4 µm and a broad Reststrahlen band around 9.3 µm (Salisbury et al. 1991) as well as an additional broad Reststrahlen feature between 12 and 14 µm (that is shifted to longer wavelengths in the spectra of high-Ca pyroxenes). The characteristic 18-µm band (Barucci et al. 2008) is detectable in our enstatite sample. In the emissivity spectrum of our enstatite sample we can recognize a distinctive band around 39 µm together with a very broad feature above 43 µm centered around 53 µm. The enstatite band around 97 µm (a similar band is present in the serpentine spectrum too) can be seen in both our Allende and Millbillillie meteorites.

Thanks to the variability of techniques and wavelength extension of our measurements, for each of the samples in our dataset it is possible to concurrently recognize multiple features previously described in literature by other authors. The set of spectral measurements for asteroid analogues we are presenting can also help in the planning of future missions and provides important input for thermal models of small bodies (cf. Davidsson et al. 2015). Specifically, such models solve the heat conduction equation in order to estimate the temperature in the regolith as function of time and depth below the surface. The upper boundary condition of this equation contains a term for thermal radiative heat loss to space that depends on the surface temperature and the integrated hemispherical emissivity E. The latter parameter is the ratio between the energy loss at all wavelengths of an actual surface to that of an ideal blackbody at the same temperature. In thermophysical modeling it is not uncommon that this parameter is set to unity or 0.9, for lack of better alternatives. The current measurements offer the possibility to calculate E for specific minerals and temperature intervals for application in thermophysical models. This is important for minerals such as enstatite and phyllosilicates, which have low emissivity at wavelengths below 8 μm. Failure of a model to account for this property will result in incorrect surface temperatures, particularly when surfaces are being heated above 350 K (80 °C).

The growing number of sophisticated missions to small bodies in the solar system requires dedicated laboratory work to maximize the science return of remote sensing observations and support sampling site selection.

The set of analogues presented here is a first step toward this goal. It provides high signal-to-noise spectra for a range of potential surface materials obtained with a wide spectral coverage encompassing all currently flying and planned NIR and TIR instruments. Already this small starting set allows a number of interesting conclusions.

The compositional information that can be obtained in the TIR (especially if we start around 5 microns) is highly complementary to the NIR. On the other hand most (but not all) of the spectral information is contained in the wavelength range <30 microns. Only very few analogues show spectral features between 30 and 100 microns, and they all have a very low spectral contrast. This is interesting for trade-offs concerning instrument design. In general, it seems that spectral coverage beyond 50 microns does not justify massive increases in the complexity (and therefore cost and mass) for a TIR instrument. Nevertheless, if the spectrometer target has areas that experience polar night, they will be very cold and will only emit radiation at very long wavelengths. If these can be measured, it provides important information about, e.g., thermal inertia at large depths. This was done for comet 67P (Biver et al. 2015) that had temperatures of 30 K in the polar night region. The radiation was being observed at cm wavelengths with the MIRO instruments (that was the only instrument available that could cover this longer wavelengths region), though a spectrometer covering the 100-micron region would have been optimal.

As expected, the small grain sizes used in this study result in very low spectral contrast. However, it is clear from the spectra shown here that even the low spectral contrast of a very fine-grained regolith would be enough to identify major phases. Given that Itokawa actually showed a much coarser grain size distribution (particle size ranges between 30 and 180 µm, Tsuchiyama et al. 2011), it is clear that mid-infrared observations are a highly diagnostic tool for the study of the surface of asteroids. The most likely thermal inertia for 1999JU3 ranges between 200 and 600 J m⁻² s^−0.5 K⁻¹, about a factor of 2 lower than the value for 25143 Itokawa (Mueller et al. 2010). This indicates that the surface lies somewhere between a thick-dust regolith and a rock/boulder/cm-sized, gravel-dominated surface like that of 25143 Itokawa. Based on the obtained laboratory results TIR instruments would also make a valuable addition for future missions to targets with very fine-grained surfaces.

Furthermore, our spectral measurements allow to accurately estimate the emissivity for specific minerals, which can be used for application in thermophysical models.