Ultraviolet spectral reflectance of carbonaceous materials

Icarus 307 (2018) 40–82

Contents lists available at ScienceDirect

Icarus journal homepage: www.elsevier.com/locate/icarus

Ultraviolet spectral reflectance of carbonaceous materials Daniel M. Applin a,∗, Matthew R.M. Izawa a,b, Edward A. Cloutis a, Jeffrey J. Gillis-Davis c, Karly M. Pitman d, Ted L. Roush e, Amanda R. Hendrix f, Paul G. Lucey c a

Department of Geography, University of Winnipeg, 515 Portage Avenue, Winnipeg, Manitoba R3B 2E9, Canada Institute for Planetary Materials, Okayama University, 827 Yamada, Misasa, Tottori 682-0193, Japan c Hawaii Institute of Geophysics and Planetology, University of Hawaii, 2525 Correa Road, Honolulu, HI 96822, USA d Space Science Institute, 4750 Walnut Street, Suite 205, Boulder, CO 80301, USA e NASA Ames Research Center, Moffett Field, California 94035-0001, USA f Planetary Science Institute, 1700 East Fort Lowell, Suite 106, Tucson, AZ 85719-2395, USA b

a r t i c l e

i n f o

Article history: Received 11 November 2017 Revised 24 January 2018 Accepted 8 February 2018 Available online 16 February 2018 Keywords: Carbon Graphite Ultraviolet Reflectance spectroscopy Airless bodies

a b s t r a c t A number of planetary spacecraft missions have carried instruments with sensors covering the ultraviolet (UV) wavelength range. However, there exists a general lack of relevant UV reflectance laboratory data to compare against these planetary surface remote sensing observations in order to make confident material identifications. To address this need, we have systematically analyzed reflectance spectra of carbonaceous materials in the 20 0–50 0 nm spectral range, and found spectral-compositional-structural relationships that suggest this wavelength region could distinguish between otherwise difficult-to-identify carbon phases. In particular (and by analogy with the infrared spectral region), large changes over short wavelength intervals in the refractive indices associated with the trigonal sp2 π –π ∗ transition of carbon can lead to Fresnel peaks and Christiansen-like features in reflectance. Previous studies extending to shorter wavelengths also show that anomalous dispersion caused by the σ –σ ∗ transition associated with both the trigonal sp2 and tetrahedral sp3 sites causes these features below λ = 200 nm. The peak wavelength positions and shapes of π –π ∗ and σ –σ ∗ features contain information on sp3 /sp2 , structure, crystallinity, and powder grain size. A brief comparison with existing observational data indicates that the carbon fraction of the surface of Mercury is likely amorphous and submicroscopic, as is that on the surface of the martian satellites Phobos and Deimos, and possibly comet 67P/Churyumov–Gerasimenko, while further coordinated observations and laboratory experiments should refine these feature assignments and compositional hypotheses. The new laboratory diffuse reflectance data reported here provide an important new resource for interpreting UV reflectance measurements from planetary surfaces throughout the solar system, and confirm that the UV can be rich in important spectral information. © 2018 Elsevier Inc. All rights reserved.

1. Introduction The ultraviolet (UV) wavelength range (λ = ∼10–400 nm) has historically been largely used to study planetary atmospheres and gas production ratios from comets (e.g., Krasnopolsky et al., 2004). The application of this wavelength range to planetary surface reflectance spectra is relatively underexplored and underutilized, but interest and applications have been rapidly increasing, because a number of planetary missions have carried and will carry instruments with sensors covering at least some portion of this spectral

∗

Corresponding author. E-mail addresses: [email protected] (D.M. Applin), [email protected] (M.R.M. Izawa), [email protected] (E.A. Cloutis), [email protected] (K.M. Pitman), [email protected] (T.L. Roush), [email protected] (A.R. Hendrix). https://doi.org/10.1016/j.icarus.2018.02.012 0019-1035/© 2018 Elsevier Inc. All rights reserved.

range, including the following spacecraft and satellites: Mariners 6, 7, 9, and 10, Voyagers 1 and 2, Galileo, the International Ultraviolet Explorer (IUE), Cassini, the Hubble Space Telescope, the MErcury Surface Space ENvironment GEochemistry and Ranging (MESSENGER), the Mars Atmosphere and Volatile Evolution Mission (MAVEN), Mars Express, Dawn, Rosetta, New Horizons, and the Mercury Planetary Orbiter (MPO). There exists a general lack of UV laboratory reflectance data that is relevant to conducting comparative analysis of planetary surface reflectance spectra. The limited studies that have been performed in this wavelength region generally find the far-UV to near-UV (FUV-NUV) spectrum to be a potentially powerful and an important spectral range for detection and discrimination of minerals and volatile compounds on airless bodies (e.g., Wagner et al., 1987; Cloutis et al., 2008). Hence, recent efforts (e.g., Cloutis et al., 2015; Liu et al. 2016) are focusing on building new FUV-NUV laboratory datasets for analyses

D.M. Applin et al. / Icarus 307 (2018) 40–82

of the Moon, comets, asteroids, Mercury, Phobos, Deimos, Iapetus, and other airless bodies. In the context of carbonaceous materials, which appear to be present on a number of planetary surfaces, we expect that these materials exert a strong influence on UV diffuse reflectance spectra due to their wide wavelength range opacity and highly absorbing behavior (Cloutis et al., 2011a, 2012). 1.1. UV detection and discrimination of carbonaceous material on planetary bodies Ultraviolet spectral features have been identified in observational spectra for several bodies expected to contain significant surficial carbon fractions and which are not highlighted in laboratory analyses of inorganic materials. For example, Parker et al. (2002) and Li et al. (2009) demonstrated that the surface of Ceres exhibits unusual UV spectral behavior, with reflectance increasing with decreasing wavelengths toward λ = 150 nm. To date, no other planetary surface has shown this spectral behavior. This was confirmed by Hendrix et al. (2016a) using data from the Hubble Space Telescope STIS instrument, finding that a peak in reflectance exists near λ ∼ 160 nm, and they ascribed it to a spectrally local increase in volume scattering due to a highly concentrated (∼85–90%) ∼0.5 μm thick veneer of a material similar to demineralized anthracite termed “graphitized carbon” by Hendrix et al. (2016a, 2016b). This material was noted to not contribute to the near-UV–visible spectral shape. Bertaux et al. (2011, 2016, pers. comm., 2016), using Mars Express Spectroscopy for Investigation of Characteristics of the Atmosphere of Mars (SPICAM) data, showed that UV reflectance spectra of Phobos exhibit a sharp peak in reflectance near λ ∼280 nm with an I/F value at or below ∼1% depending on phase angle, but were unable to attribute this feature to a specific cause. They suggested that an associated minimum in reflectance near λ = 220 nm could be due to an absorption band in solid, dehydrogenated coronenelike polycyclic aromatic hydrocarbons (PAHs), citing work from Duley and Lazarev (2004), who measured absorption spectra of plasma-deposited carbon in gases and vacuum from XeCl laser ablation of graphite. Duley and Lazarev (2004) found that absorption spectra of the plasma-deposited carbon in gas that had been subsequently heated exhibited an extinction maximum to occur near λ = 220 nm, and showed that this is analogous to previous calculations of the spectroscopic properties of corenene-like gas phase molecules. Bertaux et al. (2016) also showed that the Deimos spectrum may exhibit a weak peak at λ ∼260 nm, when ratioed against the spectrum of Phobos. Roettger and Buratti (1994) compiled all measured reflectance spectra from the IUE of 45 asteroids, co-added and recalibrated these spectra for maximum signal-to-noise (S/N) and photometric accuracy, and presented and analyzed these spectra. They concluded that many of the 45 asteroids exhibited spectral slopes shortward of λ = 260 nm that were consistent with organic material, but that S/N was not sufficient to be conclusive. Stern et al. (2015) presented the first extreme UV (EUV) and FUV reflectance spectra of a cometary nucleus, 67P/Churyumov– Gerasimenko. The spectra show an overall I/F of ∼0.02, a peak in reflectance near λ ∼ 90 nm, minima near λ ∼ 170 nm, from which they concluded that the spectra are consistent with carbon and tholin-like materials, and possible contribution from Rayleigh scattering. Hendrix et al. (2016b), using HST/FOS data (from Noll et al., 1997) and Cassini UVIS data (from Hendrix and Hansen, 2008), found that the dark leading hemisphere of Iapetus may exhibit a peak in reflectance near λ = 190–205 nm, and suggested that this feature could be consistent with reflectance peaks in matrix-isolated phases (spectra of pentacene and acenaphthylene were shown) of polycyclic aromatic hydrocarbons (PAHs), as computed photo-absorption cross sections produced by Malloci et al.

41

(2004) showed absorption maxima near this wavelength for pentacene. Hendrix et al. (2016b) also indicated that many condensed PAH films show an absorption peak near λ = 200 nm, citing work from Steglich et al. (2010), who showed that thin film transmission spectra of coronene exhibit strong extinction features near λ = 300 nm and 200 nm. Steglich et al. (2010) also showed that film transmission spectra of two solvent extracts of soot exhibited extinction features near λ = 200 nm, and that the feature moves to longer wavelengths as PAHs become larger. Hendrix et al. (2016b) indicated that mixtures of many aromatic molecules will generally lead to the presence of only the π band near λ = 200 nm, citing Papoular et al. (1993) who showed that as coal rank decreases, the π –π ∗ band weakens, widens, and shifts to longer wavelengths from λ = 250 nm. The presence of organic materials on the surface of both of Iapetus’ hemispheres was previously suggested from infrared spectral data (Cruikshank et al., 2008, 2014), who found partially resolved, overlapping absorption bands near λ = 3.30 μm and ascribed them to aromatic and aliphatic C–H bonding. 1.2. Detection of carbonaceous material on planetary bodies using albedo The evidence for the widespread presence of carbonaceous material on other planetary bodies is both direct and implied from reflectance spectra collected at longer wavelengths. Much of this evidence points to the fact that carbonaceous materials have an exponential impact on the spectroscopic properties of assemblages within which they are contained, due to the strong nonlinear effect of “dark” (i.e. highly absorbing) materials in the presence of multiple scattering in powders. High S/N disk-averaged UV spectra from the surface of Mercury have previously been acquired with the Mercury Atmospheric and Surface Composition Spectrometer (MASCS) instrument on the MESSENGER spacecraft (McClintock et al., 2008). They found that Mercury’s reflectance spectrum relative to that of the lunar farside exhibits lower albedo and a progressive increase in relative reflectance with decreasing wavelength from λ = 1400 nm to λ ∼ 320 nm, with an abrupt downturn at λ = 300 nm, with a suggestion that the UV feature could be due to oxygen-metal charge transfers in Fe2+ - and/or Ti4+ -bearing silicates. A non FeO-bearing material, which would be opaque over a wide spectral range, is required for the darkening and suppression of silicate nearinfrared (NIR) absorption bands in reflectance spectra collected from the surface of Mercury (e.g., Robinson and Lucey, 1997). Low iron (Weider et al., 2015; Peplowski et al., 2011) and titanium (Peplowski et al., 2011) abundances do not support the hypothesis that opaque oxides of these elements contribute substantially to Mercury’s low and variable surface reflectance. As such, the presence of carbon at concentrations of multiple wt.% on Mercury has been deduced on the basis of this and elemental data acquired by the MESSENGER spacecraft (Syal et al., 2015; Peplowski et al., 2016), but the source and specific phase is not yet uniquely resolved, because there are not diagnostic features in the VNIR. Trang et al. (2017) were able to successfully model the VNIR reflectance spectra of Mercury by adding wt.% concentrations of nanophase and microphase iron and amorphous carbon to a silicate matrix. Diverse carbonaceous phases are present at up to only the few percent level in the various classes of carbonaceous chondrite meteorites, yet the spectral properties of carbonaceous chondrites are very strongly influenced by those of the carbonaceous fraction (e.g., Cloutis et al., 2011a, 2011b). Carbonaceous phases are also important constituents of some differentiated meteorites, such as ureilites and certain iron meteorites (e.g., Brearley and Jones, 1998; Mittlefehldt and Lindstrom, 1998; Cloutis et al., 2010; Cloutis et al., 2012), but they also do not show direct evidence of specific car-

42

D.M. Applin et al. / Icarus 307 (2018) 40–82

bonaceous phases (i.e. absorption bands) in their VNIR reflectance. Carbonaceous chondrites, ureilites, and some iron meteorites are generally linked to various classes of dark asteroids, largely on the basis of low albedos and spectral slopes (DeMeo and Carry, 2013). Very high concentrations of carbon occur in some interplanetary dust particles (IDPs) (Rietmeijer, 1998; Flynn et al., 2003), and such particles represent an important delivery mechanism of carbon to planetary surfaces. For example, Fries et al. (2017) model the influx of IDPs to the surface of Phobos, suggesting its low albedo could be due to the presence of IDP-delivered carbon at the low wt.% level. Carbonaceous materials are known to be present on Mars from direct sampling (Steele et al., 2012; Freissenit et al., 2015). Fries et al. (2017) suggest that the refractory carbon on the surface of Phobos is spectrally observable and causes the low albedo because of the probable absence of the strong oxidation processes that are thought to form carboxylates in the martian regolith from IDP and meteoritic carbon (e.g., Benner et al., 20 0 0). The presence of carbonaceous materials on comets is known from direct sampling (Sandford et al., 2006), and spectroscopic evidence (Quirico et al., 2016). Reflectance spectra of cometary nuclei have been shown to be similar to D-type asteroids, which also suggest a strong carbon concentration. This has led to suggestions that some Jupiter trojans can be dead comets (e.g., Perna et al., 2017). The newly discovered interstellar asteroid 1I/2017 U1 (‘Oumuamua) and the nucleus of comet 67P (Meech et al., 2017; Ye et al., 2017) show spectral slopes similar to D-type asteroids. Dark exogenous material, likely related to carbonaceous chondrites, is present in many areas on Vesta (Reddy et al., 2012), and similar materials are inferred to comprise the crust of Ceres (Schaefer et al., 2016; McSween et al., 2017). Whereas carbonaceous materials are known or suspected to be present on a number of planetary surfaces, in many cases their precise nature has not been determined. That the UV spectral region may be useful for detecting and characterizing such materials is a motivation for this study.

•

1.3. Strong absorption bands and related spectral features Large changes over small wavelength intervals in the complex refractive indices of a material, n (real component of the complex refractive index) and k (imaginary component of the complex refractive index), occur due to strong absorption bands (k  0.1), which is most often termed anomalous dispersion because n will decrease with decreasing wavelength within the band (Figs. 1a and 1b). Overtones and combinations typically do not cause anomalous dispersion, such as in the NIR range of water ice for example (supplementary material). Anomalous dispersion in absorption bands causes the appearance of specific features in reflectance, emittance, and transmittance spectra. In this paper, we adopt the same terms used in IR spectroscopy (reststrahlen band, Fresnel peak, Christiansen feature, transparency feature, and transition minimum), because the values of the complex refractive indices underlying both the IR and UV features can be similar, despite the very different physical causes of absorption (i.e., vibrational dipole moments vs. electronic excitons, charge transfers, and interband transitions). The rationale for applying terminology largely developed for IR spectroscopy is detailed in the following section. •

Reststrahlen band: In MIR reflectance spectroscopy, anomalous dispersion near strong absorption leads to a reststrahlen band. As described in Hapke (2012) and throughout the literature, reststrahlen bands, “residual rays” in German, are the phenomenon by where incident light is coherently reflected from the middle of strong infrared absorption bands, where the refractive indices are highest, rather than being absorbed. The strong absorption bands that occur in the UV are not

•

called reststrahlen bands in the literature—presumably due to the evolution of the term residing in vibrational spectroscopy. The physical and spectral effects of anomalous dispersion in the UV and IR can be identical, however. Many researchers have used reflectance spectroscopy in the UV to study electronic absorption bands in materials, which typically cause reststrahlen bands instead of typical absorption band troughs (Hartman and Nelson, 1957a; Hartman et al., 1957b; Loh 1964; Arakawa and Williams, 1968; Chandrasekharan and Damany 1968); Haensel et al., 1969; Koch and Skibowski, 1971; Sigel 1971; Rubloff 1972; Kaneko et al., 1983; French et al., 1998; Tan et al., 2005). For the purposes of this paper, we will call the actual feature of a peak shape in reflectance the Fresnel peak, which is typically what the water ice O–H stretching Fresnel reflectance peak is referred as in remote sensing literature, rather than reststrahlen band (e.g., Clark et al., 2012). We will refer to the band as the entire absorption envelope that contains the Christiansen-like features, Fresnel peaks, transition minima, and absorption band wings. Fresnel peak: A Fresnel peak occurs when the centers of absorption bands have very high k, causing coherent reflectance from the grain surfaces. The maximum is generally offset to the short wavelength side of the maximum in k (Figs. 1a and 1b). The shape of a peak (increase in reflectance with increasing wavelength, followed by a reversal after the maximum in reflectance) over a short wavelength range occurs due to a rapid change in n from the Christiansen frequency (where n = 1; not to be confused with the Christiansen feature) over a short wavelength interval (Figs. 1a and 1b), while the areal density of induced electron displacement mechanisms, generally dipole moments (vibrational) or excitons/charge transfers (electronic), increases monotonically with k, and following Maxwell’s electromagnetic equations, so does the subsequently generated backscatter (Hapke 2012). Sufficiently high k (>0.1) causes the entry of light through the grain boundary to be inhibited, photons that do enter are absorbed, and volume scattering therefore becomes insignificant. Because these spectral ranges are opaque, and reflectance is due to surface scattering, the reduction of powder grain size will often reduce spectral contrast of the Fresnel peak, or peak height over the continuum, due to increased scattering between the grain surfaces before emergence (e.g., Salisbury and Wald 1992; Lyon 1965). Fresnel peak refractive indices are often similar to those found over wide wavelength ranges in metals, and therefore exhibit similar spectral properties. As shown in Figs. 1f and 1g, Fresnel peaks can exhibit very strong surface scattering when the refractive indices are sufficiently high. Christiansen feature: First, the reader should be aware that IR emission spectra are not perfectly analogous to UV reflectance spectra, but we include some references in addition to the reflectance experiments that highlight the important insights that can be gained by comparing and contrasting spectral behaviors in these two “strong” surface scattering wavelength domains. Many Fresnel peaks have a Christiansen feature on the short wavelength side of the peak and offset on the short wavelength side of the Christiansen frequency (n = 1; described below). In IR spectroscopy, the Christiansen feature refers to a maximum in emissivity, or minimum in reflectance (following Kirchhoff’s law), associated with this short-wavelength transition minimum of a strong absorption band. The Christiansen feature in reflectance does not occur exactly where n = 1, because in general, even small particles of materials are not transparent at this wavelength, as it is still very close to a very strong absorption band where k is significant (e.g., Hapke 2012). As such, the exact wavelength position of the minimum in reflectance can be dependent on the minimum or change in k. The offset is minimal but measurable, and the position of the feature is

D.M. Applin et al. / Icarus 307 (2018) 40–82

43

Fig. 1a. Ordinary-ray refractive indices of graphite from Djurišic´ and Li (1999) compared to a derived Fresnel reflectance spectrum. Ordinary-ray represents the typical spectral properties of graphite because it forms sheets that are weakly bonded together by Van de Waal forces. Extraordinary-ray refractive indices are plotted in the supplement. The average Fresnel reflectance at normal incidence of s- and p-polarizations was computed with Eqs. (4.36) and (4.3)7 from Heavens et al. (1992) and Eqs. (15) and (16) from Lawrence and Lucey (2007), and is normalized for clarity. At normal incidence, ignoring the polarization distinction will yield identical results.

Fig. 1b. Same as Fig. 1a, but over the UV wavelength range. As shown, Fresnel peak maxima are typically offset to shorter wavelengths than the corresponding maximum in k.

44

•

•

•

D.M. Applin et al. / Icarus 307 (2018) 40–82

still largely determined by n. This is highlighted in Figs. 1f and 1g, which show that when k is relatively high, a true Christiansen feature will not occur because relatively large amounts of surface scattering occur. However, a minimum in reflectance will still be present. Christiansen features are widely used in remote sensing of planetary surface emittance spectra; they are often described as the most useful feature in IR remote sensing of fine grained surface emission, and have been rather intensely studied with both emittance and reflectance spectroscopy. Christiansen frequency: In addition to use in reflectance spectroscopy, the Christiansen effect, where n = 1 and where zero refraction occurs, is also observed and used in transmittance spectroscopy, from where the term originated (Christiansen 1884; Prost 1973; Carlon 1979; Clarke 1968; Cloupeau and Klarsfe1d, 1973). Because the Christiansen frequency is where n = 1, it has been used, in various materials, to create optical filters for isolating narrow wavelength transmittance, including at UV wavelengths (e.g., Minkoff and Gaydon 1946; Sinsheimer and Loofbourow 1947; Wojak et al., 1987). Transparency feature: A transparency feature is a maximum in reflectance, or minimum in emittance that appears most prominently in IR reflectance spectra of fine powders. They are generally located between absorption bands in wavelength and are attributed to enhanced volume scattering due to low k in the absence of absorption bands (e.g., Salisbury and Walter, 1989). This is generally the cause of non-opaque mineral powders decreasing in brightness with increasing grain size in the VNIR and between absorption bands in the IR, since more absorption occurs as the mean optical path of the ray entering each grain and in the powder increases before emergence. The opposite can generally be observed for opaque powders where k > ∼0.1, where total surface scattering can increase with increasing grain size of a powder due to a reduction in scattering between grain surfaces before emergence, while volume scattering remains insignificant. Following this, the VNIR of nonopaque powders could be accurately described as a spectrally wide transparency feature with superimposed weak absorption bands, or a volume scattering regime. The refractive index n is typically near ∼1.5 at VNIR wavelengths of silicates and at wavelengths between IR Fresnel peaks. As shown in Figs. 1f and 1g, significantly less surface scattering will occur with these refractive indices compared to those that are typical of metals, Fresnel peaks, UV–Vis bands of oxides, and materials with a high refractive index n such as diamond and sapphire. Transition minimum: When strong absorption bands are surrounded by non-opaque spectral regions (i.e., the surrounding spectral regions are transparency features, or volume scattering regions), a minimum in reflectance can also occur on the long wavelength side of the Fresnel peak. This is the long wavelength transition minimum, where surface scattering begins to strongly dominate over volume scattering with decreasing wavelength, and where the Fresnel peak begins to rise in reflectance with decreasing wavelength. The reflectance values of this feature are generally higher than the Christiansen feature on the short wavelength side of the Fresnel peak because n is generally higher, but this offset breaks down if the absorption band is superimposed on a continuum, which is common. Christiansen features can be considered the short wavelength transition minimum. The offset gives reststrahlen bands the characteristic asymmetric, first derivative-like spectral appearance, as described by Hapke (2012). The position of this feature is also variable, especially with grain size, as it is dependent the amount of surface scattering of the powder at each wavelength, which changes significantly with grain size. Or in other words, as the Fresnel peak maximum reflectance decreases with decreasing grain size, the transition minimum

can move to shorter wavelengths as the peak progressively disappears and takes on the appearance of a typical absorption band. 1.3. Previous UV spectroscopic studies of carbonaceous phases The UV spectral properties of carbon were reviewed by Hendrix et al. (2016a), and are also summarized in the astrophysics literature (cf. Henning et al., 1999; http://www.mpia.de/HJPDOC/ carbon.php). Most UV spectral studies of carbonaceous materials have been conducted on polished or cleaved surfaces or pressed pellets of very fine powders in fully specular geometries for calculations of refractive indices via modified Kramers–Kronig relations, as matrix isolated transmittance of gas and solid particles, via transmission of thin-film vacuum-deposited compounds, and as computationally-derived spectra via many methods, but generally not by diffuse reflectance spectra of powders as is commonly used in planetary science. To expand on the review by Hendrix et al. (2016a), we discuss other reflectance experiments (pellets, powders) involving carbonaceous materials, in both diffuse and specular geometries. Gilbert (1960) measured the reflectance spectra of graphite and coals from λ = 230–900 nm using a fused silica plate as a reflectance standard, concluding that the π –π ∗ absorption band in graphite led to a peak at λ = 260 nm. The position of this peak was noted to not be precise due to the uncalibrated reflectance standard—most reflectance standards have an absorption edge in this spectral region, affecting peak position when ratioed. He discussed the cause of the peak as being due to anomalous dispersion from an electronic transition, citing work from Friedel and Breger (1959), who showed that coals do not exhibit the peak, but exhibit decreasing reflectance with decreasing wavelength. These results were consistent with work by HumphreysOwen and Gilbert (1958), who also found a peak in reflectance in the graphite which they studied. That work was expanded upon by McCartney and Ergun (1967), who reviewed and measured the complex refractive indices of graphite and coals in the NUV. They found that a well-defined peak in reflectance near λ = 250 nm occurs in graphites and high ranking coals (coal rank is based on calorific values and fixed-carbon content; anthracite is the highest ranking coal), due to the π –π ∗ transition. In lower rank coals, they found that n of all samples decreased strongly with decreasing wavelengths towards the FUV, but the peak was suppressed. Papoular et al. (1993, 1995) measured UV reflectance spectra of pressed pellets of extremely fine graphites and demineralized coal powders in specular geometries, and showed that the minimum near λ = 160 nm moves to shorter wavelengths with increasing aromaticity, and that spectral contrast of the λ = 250 nm peak increases and the peak narrowed with increasing aromaticity. Note that in their work, Papoular et al. (1995) indicated that the spectral properties of these samples measured in specular geometries compared to the diffuse spectra collected with a PTFE-coated integrating sphere (directional-hemispherical reflectance) are insignificantly different. For reference, we show the spectra presented by Papoular et al., (1995) in Figs. 1h and 1i, as they demonstrate clearly the spectral properties of graphite and carbon throughout the UV spectral range. This is also demonstrated by Taft and Philipp (1965), who showed that the minimum and π –π ∗ peak of glassy carbon g occurs at longer wavelengths than that of graphite. Glassy carbon has imperfectly understood physical characteristics, fullerene-related structure, non-graphitic or graphitizing but similar to strained graphite layer models metallic and vitreous, and probably pure sp2 (e.g., Ergun 1976; Robertson 1986). Greenaway et al. (1969) measured reflectance spectra of cleaved pyrolytic graphite surfaces, and found good agreement with those from Taft and Philipp (1965), showing the peak near λ = 250 nm

D.M. Applin et al. / Icarus 307 (2018) 40–82

and minimum near λ = 160 nm. Williams and Arakawa (1972) measured reflectance spectra of glassy carbon, which show a maximum in k at λ = 270 nm, but the sampling intervals are nearly 30 nm apart. Khare et al. (1984, 1993) measured UV reflectance spectra of multiple “tholin” samples (organic solids derived from gas irradiation), which exhibit similar spectral behavior to some of the spectra described above, with a peak in reflectance near λ = 80 nm; some of their samples also exhibit maxima in k at λ = 122 nm and 207 nm. Yokoyama et al. (1990) measured the reflectance spectra of a-C:H thin films and found two peaks near λ = 400 nm and 80 nm, ascribing them to π –π ∗ and σ –σ ∗ transitions, respectively. Chang and Charalampopoulos (1990) found the maximum in k of various soots occurred near λ = 260 nm. In addition to the spectral properties of these carbon phases, Kameta et al. (2002) show that the peak in FUV photoabsorption spectra measured for methane gas occurs near λ∼92 nm, and moves to shorter wavelengths as the gaseous alkanes becomes larger, with ethane at λ∼87 nm and both butane and propane near λ∼78 nm. From their review and spectral modeling, Hendrix et al. (2016b) concluded that progressive graphitization of organic material occurs due to hydrogen loss via multiple mechanisms, which leads to a minimum in reflectance at λ = 220 nm and an increase in reflectance between λ = 10 0 and 20 0 nm. They applied these results to analysis of reflectance spectra of airless bodies in the solar system, specifically Iapetus and Ceres. In their follow-on work with newly acquired observational spectra of Ceres, Hendrix et al. (2016a) indicated that irradiation, ion bombardment, and heating of carbon causes an absorption minimum at λ = 80 nm, a peak in absorption near λ = 210–220 nm causing a reflectance minimum, and that graphitization due to hydrogen loss causes a FUV-bump in reflectance near λ = 160 nm, both consistent with their observational spectra of Ceres. Hendrix et al. (2016a) indicated that this peak in reflectance at λ = 160 nm is especially dramatic for graphite, citing work from Papoular et al. (1996), who summarized carbon spectral properties, showing that the σ –σ ∗ transition and the π –π ∗ transition absorption maxima (reflectance peak) occur at λ∼80 nm and λ∼250–400 nm range, respectively. Hendrix et al. (2016a) suggested that other carbonaceous species do not exhibit the FUV-bum’ feature; only those that are graphitized develop the FUV-bump, citing earlier work by Hendrix et al. (2016b). Hendrix et al. (2016a) also modeled this feature to occur in glassy carbon. Glassy carbon has imperfectly understood physical characteristics, fullerene-related structure, non-graphitic or graphitizing but similar to strained graphite layer models metallic and vitreous, and probably pure sp2 (e.g., Ergun 1976; Robertson 1986). Hendrix et al. (2016a) attributed the FUV-bump to a spectrally local increase in volume scattering at λ = 160 nm due to the local minimum in k. We also find that Duley (1984) highlighted that the minimum in k for the amorphous carbon refractive indices they produced was near k = ∼0.4 at λ = 140 nm, and suggested this could cause an increase in albedo of single particles at grain radii much smaller than the wavelength. The very high concentration of carbon estimated by Hendrix et al. (2016a) for the reflectance spectra from the surface of Ceres was presumably required due to the extreme non-linear effects of the spectral mixing of transparency features with dark opaque materials, of which almost all expected planetary surface materials are in the FUV. For example, Wagner et al. (1987) showed that most minerals and meteorites (including H, L, C chondrites, HEDs, an ureilite, an aubrite, and martian meteorites) exhibit reflectance near or below 5% in the FUV with the exception of nearly Fe-free silicates, and these samples were only sieved to the <150 μm fraction—fine grained opaque powders would be much darker. Indeed, olivine of presumably Fa50 , exhibits k∼1 in the FUV (Huffman, 1975). This potential phenomenon of UV volume scattering peaks from submicroscopic carbon powders in reflectance spectra will be the subject of follow-

45

on laboratory measurements and mixing models involving multiple scattering scenarios. However, given presumed requirements for observation of this potential phenomenon in remote sensing scenarios, which likely requires only submicroscopic particles of 100 s of nm grain diameter, high concentration of carbon of probably >>∼95%, and possibly anthracitic composition, we do not focus on this within this paper. In addition to our laboratory spectra presented here, we have compiled many available carbon refractive indices and reflectance spectra from the literature, and the spectra can be found in the supplement. The ordinary-ray refractive indices of graphite derived by Djurišic´ and Li (1999) and the computed and normalized Fresnel reflectance are plotted in Figs. 1a and 1b. This visually demonstrates that the cause of the Fresnel peak and its shape of a peak in reflectance over a short wavelength interval is the coincident occurrence of very high k with rapid changes in n (also see Figs. 1f and 1g). The actual cause of the maximum reflectance value is described by the Fresnel equations. The effect of grain size on these features has yet to be fully investigated, which was another motivation for this study. As described above, particle size distributions can have small effects on the position of Christiansen featureless in the IR, whereas decreasing grain size causes decreasing spectral contrast of Fresnel peaks in the IR. We expect similar behavior with carbon powders in the UV. 1.4. Effects of the structure and bonding of carbon on electronic spectra The structure and bonding environments of carbon are wellstudied. Numerous studies exist that describe carbon in detail (e.g., Robertson 1986). The following is a brief summary of this, and how it relates to our laboratory experiments and the spectra present in the literature. Carbon has multiple bonding coordinations with itself. Carbon at sp3 sites has all of the valence electrons in tetrahedrally-coordinated hybridized orbitals, which form σ bonds with each other, ideally at 109.5°. In carbon sp2 sites, three of the four electrons are in trigonally-coordinated hydridized orbitals which form σ bonds ideally at 120°, and the fourth electron lies in a non-hybridized pz orbital coordinated orthogonally to the σ bonding plane. The pz orbital forms π bonds with adjacent pz orbitals, which are conjugated, and weaker than σ bonds. At sp1 sites, two of the electrons form σ bonds that are ideally planar at 180°; the other two are left in non-hydridized py and px orbitals, orthogonal to the σ bond planes, and form π bonds. A σ –σ bond is a single bond and π –π bonds are conjugated/non-localized. Diamond consists of pure sp3 , tetrahedrally-bonded carbon. Graphite consists of hexagonal layers of sp2 sites, with each sheet weakly bonded by van der Walls forces. Alkane carbon is saturated, singly bonded, tetrahedral, and sp3 (e.g., ethane). Alkene carbon is unsaturated, doubly bonded, and sp2 (e.g., ethylene). Alkyne carbon is unsaturated, triply bonded, and sp1 (e.g., acetylene). There is a large variety of amorphous carbon phases, which have various ratios of sp2 /sp3 sites, and therefore various ratios of σ and π bonding environments. A helpful representation of this can be seen as a ternary diagram between sp2 , sp3 , and H in Fig. 1 from Ferrari and Robertson, 20 0 0. To summarize the data presented in Fig. 1, and Section 1.3 in the context of carbon structural and bonding characteristics, Fresnel peaks from electronic transitions appear to occur near λ∼85 nm (σ –σ ∗ ) and λ∼250 nm (π–π∗ ), with minima in reflectance (Christiansen feature-like) shortward of λ = 50 nm and near λ = 160–190 nm. All of these features appear to move to longer wavelengths with increasing amorphization, presumably due to the mean bonding distance decreasing. These features also appear to broaden with amorphization, presumably due to a larger

46

D.M. Applin et al. / Icarus 307 (2018) 40–82

distribution of bonding distances. Diamond has entirely tetrahedral sp3 sites, and therefore only exhibits the σ –σ ∗ Fresnel peak, which appears to be situated at λ = 100 nm, being strong and narrow, and is presumably at a longer wavelength since there is no trigonal sp2 contribution and because the sp3 -bonded electrons are closer together than in sp2 and sp1 bonds. Graphite has purely sp2 sites, and therefore both Fresnel peaks appear strong and narrow due to a small distribution of bonding distances while broadening with decreasing graphitic ordering (e.g., Papoular et al., 1995). Various other solid carbon phases have varying ratios of sp2 /sp3 and crystallinity, and therefore varying ratios of σ –σ ∗ and π –π ∗ peak strengths and widths. Some carbon phases, such as tholins and kerogen, appear to exhibit multiple σ –σ ∗ peaks that appear strongly similar to pure amorphous tetrahedral carbon (Figs. 1d and 1e). Amorphous tetrahedral carbon is mostly sp3 , but does include some sp2 (see Fig. 1 from Ferrari and Robertson, 20 0 0), and therefore exhibits some π –π ∗ absorption, unlike diamond which is pure sp3 . Some hydrogenated and amorphous carbon phases may exhibit multiple π –π ∗ peaks in the λ = 200–250 nm range, presumably due to various site occupancies and possible sp1 contribution, which should be explored further. Coals, with the exception of anthracite, do not exhibit an observable π –π ∗ peak near λ = 250 nm, but the σ –σ ∗ peak near λ = 90 nm and the minimum near λ = 170 nm are present. As coal rank decreases, the σ –σ ∗ peak broadens and moves to slightly longer wavelengths, presumably due to decreased graphitic ordering and increasing sp3 /sp2 . The FP near λ = 250 nm appears to decrease in contrast with decreasing graphitization, such as in amorphous carbons, but the Fresnel peak near λ = 85 nm and minima in reflectance near λ = 160–180 nm remain observable, as they are caused by the σ –σ ∗ transition. In general, the minimum at λ = 160 nm is strongly affected by the π –π ∗ transition from trigonal sp2 sites, and although it remains observable as a transition minimum in many amorphous carbons and sp3 dominant carbon, its contrast with the surrounding spectrum decreases strongly due to the absence of the strong increase in reflectance towards the λ = 250 nm Fresnel peak with increasing wavelength. Impure carbon (on the order of inter-atomic bonding, not to be confused with high purity carbon in intimate mixtures), such as that found in many coals, should retain strong absorption towards the FUV due to decreasing n with decreasing wavelength, but not exhibit a significant Fresnel peak near λ = 250 nm. Some soot optical constants presented by Zubko et al. (1996) appear to not show significant σ –σ ∗ or π –π ∗ peaks, which is worthy of further investigation, but is likely due to the mathematical treatment of the measurements. Gas phase alkanes appear to show the σ –σ ∗ transition, in photoabsorption spectra, at ∼92 nm for methane, with peak wavelength decreasing with increasing carbon number. This could be of interest for outer solar system exploration as observational and laboratory data continue to accumulate. Our setup for this study only measures wavelengths as low as λ = 200 nm, but we can observe this behavior associated with the CF absorption edge and λ = 250 nm FP. 2. Methods Cloutis et al. (2008) reviewed previously employed experimental procedures in detail and results from initial investigations of some of these materials, and showed that using the appropriate experimental procedure (standard correction) is critical for acquiring reliable UV reflectance spectra, especially at wavelengths less than λ = 250 nm, which are needed to interpret observational data. Prior to the study by Cloutis et al. (2008), the most developed and calibrated methods were those by Hapke, Wagner, and coworkers (detailed in Wagner et al., 1987). Our reflectance spectra of the same meteorites and minerals are closely matched (supplement).

For comparison, we include their specific experimental procedure in the appendix (supplement), and also include all of the digital spectra from the Wagner et al. (1987) UV atlas in the supplementary material. 2.1. Sample descriptions We have measured the UV diffuse reflectance spectra from

λ = 200 to 500 nm of a diverse suite of carbonaceous materials (Table 1), which includes coals and oil sands with a variety of elemental compositions (Table 2), coal and oil sand toluene-insoluble fractions and extracts, carbonaceous chondrites, natural and synthetic graphites, and highly condensed carbonaceous materials (anthraxolite, albertite, asphaltum, gilsonite, and shungite). 2.2. Current ultraviolet reflectance spectral data acquisition We have expanded the spectral survey of expected planetary surface materials reported by Cloutis et al. (2008) to include a wider array of carbonaceous phases. UV diffuse reflectance spectra from λ = 20 0 to 50 0 nm were measured with an Ocean Optics (Dunedin, Fl) Maya20 0 0 PRO—a symmetric crossed Czerny– Turner miniaturized spectrometer equipped with an HC-1 grating and 50 μm entrance slit, giving an effective sampling interval between 0.48 nm at λ = 200 nm and 0.46 nm at λ = 400 nm, and a spectral resolution of ∼1.85 nm throughout. Sample illumination was provided by an Analytical Instrument Systems Inc. Mini-DT(deuterium)A light source with a 30 W deuterium lamp fed through a bifurcated fiber optic bundle consisting of six illumination fibers surrounding a central pick-up fiber feeding into the detector array. This assembly consisted of 400-μm diameter “solarization-resistant” SiO2 fibers. A collimating lens is attached to the light house in order to focus more light into the fibers. This lens was manually focused to maximize signal through the full spectral range. The fiber optic bundle was used in normal incidence and we used an integration time of 500 ms and averaged 700 individual spectra (or higher depending on required signal to noise to obtain high quality spectra). A figure showing the highest S/N acquired with this setup on Spectralon® and lampblack is available in the supplement. Measurements for each sample were made by first acquiring a dark current spectrum, a reference spectrum on Spectralon® , followed by measurement of the sample. The average value of the optically masked pixels is subtracted from all collected spectra to provide a more consistent dark current correction. All three measurements were made using the same viewing geometry, integration time, and number of averaged spectra. The reference and target were both placed at the same distance from the end of the fibers bundle (∼5 mm). The fibers have a 25.4° field of view, and this working distance provides optimal signal over the entire wavelength range. The most appropriate definition of the instrument configuration may be a partially off-axis biconical arrangement. Due to the arrangement of the incident and pickup “cones”, the majority of reflectance is diffuse, but the specular component is not eliminated. We have measured a UV-enhanced aluminum mirror to be ∼300% brighter in fully specular geometries, and to exhibit measureable reflectance in non-specular geometries due to minor surface asperities. Additionally, VNIR reflectance measurements performed on carbon lampblack (LCA101; <0.021 μm grain size), amorphous carbon paints, and amorphous carbon (GRP200; <0.03 μm grain size), indicate that the differences in reflectance in specular and non-specular geometries are not strong (Supplementary Material). The paints exhibit an increase in overall reflectance and minor spectral reddening in the specular mode. Carbon lampblack LCA101 and nanophase amorphous carbon GRP200 decreased in reflectance in a specular geometry, presumably due to behavior similar to a “Rayleigh absorber” because

D.M. Applin et al. / Icarus 307 (2018) 40–82

47

Table 1 Sample characteristics; carbonaceous materials—all powders. Sample

Locality

Description

COAL12 COAL16 COAL17 COAL19 COAL22 COAL32 COAL37 LCA101 GRP200 SHU102 ANT101E GRP101 GRP102

Sub-bituminous coal High volatility A bituminous High volatility A bituminous Medium volatility bituminous Anthracite High volatility bituminous Low volatility bituminous Lampblack Amorphous graphite; <30 nm particle size; 99.5% purity Shungite, >99% Anthraxolite Graphite Graphite

GRP105 GRP101E ANT101 OSFL03 OILS10 TAR07 OK-5E

Highvale, AB, Canada Phalen Mine, NS, Canada [Phalen, raw coal] Phalen/Lingan Mine, NS, Canada [Phalen/Harbour seam] Quintette, BC, Canada Mt. Klappan, BC, Canada Illinois #6 seam, IL, USA Pocahontas #3 seam, VA, USA Synthetic (Johnson Matthey, #14237A) Synthetic (MTI Corporation) Russia Chelmsford, Balfour County, ON, Canada Ratnapura District, Sri Lanka synthetic "amorphous" graphite (Johnson Matthey, #10130A), −300 mesh, 99.5% pure Spitzkoppe, Namibia Ratnapura District, Sri Lanka Chelmsford, Balfour County, ON, Canada Athabasca oil sand drill core, AB, Canada Near Rifle, CO, USA Fort McMurray, AB, Canada Carter Co., OK, USA

CTE101 ASP101 JO-9E

Prepared by Jeff Bell, University of Hawaii Seefeld, North Tyrol, Austria 31.08EN, 35.32EE, Jordan

the particle sizes are much smaller than the Mie size parameter of 1, and therefore exhibit strong forward scattering from diffraction and a relative absence of Fresnel reflection, possibly causing increased absorption. The data presented by Wagner et al. (1987) are in specular geometries only; yet strongly agree in spectral shape and overall reflectance to those collected in this study. The spectra were corrected to “absolute reflectance” by ratioing the sample spectra to a calibrated Spectralon® 99% diffuse reflectance standard corrected to a deep ultraviolet (DUV) mirror. A theoretical spectrum of the mirror was provided by Edmund Optics, which was adjusted only slightly based on standardless measurements at phase angles below ∼10° to assure minimal phase reddening. The Spectralon® standard was cleaned by following the procedure developed by Stiegman and Bruegge (1993). The standard was first cleaned by N2 airbrushing, then by vacuum-baking at 10−6 Torr and 90 °C for 48 h, followed by cooling down at ambient pressures in a N2 atmosphere. This facilitates the removal of organic impurities which could undergo photochemical reactions when exposed to the UV light source and cause spectral degradation of the standard. We have previously observed extreme UV-induced spectral degradation of various Spectralon® standards when this procedure was not used (Mann et al., 2016), due to the polymerization of the trace organics. The DUV mirror correction removes irregularities (increasing absorption below ∼250 nm) in the Spectralon® standard. Additionally, this correction is specific to each experimental run, rather than the single measurement of Spectralon® standards used by most laboratories for VNIR reflectance. Peak positions are derived from the data used in this paper by use of visual inspection of the local maxima abd minima. 3. Results and interpretations 3.1. Pure carbon phases and purity of carbon phases All varieties of pure carbon studied here (listed in Table 1) exhibit a Fresnel peak shortward of λ = 300 nm (examples shown in Figs. 2a–2e). The position of the peak maximum changes with grain size and degree of amorphization, as discussed in more detail below. All of the samples also exhibit decreasing reflectance with decreasing wavelength towards the FUV, which is caused by

3R-Graphite Toluene-insoluble fraction Anthraxolite Oil sand; 12.03% bitumen Oil shale Athabasca oil sand Type II kerogen extract from the Upper Cretaceous San Miguel Formation sandstone bed Coal tar extract Asphaltum Extract from natural asphalt

the decrease of both n and k towards the π –π ∗ Christiansen-like feature and/or σ –σ ∗ transition minimum (e.g., McCartney and Ergun, 1967; Fig. 1), and the various studies listed in the above review of carbon reflectance. Fresnel peaks can be degraded by the presence of chemical impurities (e.g., H, O, N, S); for example, anthracitic coal (COAL22), even though it is dominated by carbon, 77.2 wt.% C, does not exhibit a Fresnel peak near λ ∼ 241 nm, because the carbon is significantly impure at the inter-atomic level so that the areal density of overlapping sp2 orbitals is insufficient to cause a peak over the continuum. In other words, this is consistent with the breakdown of the π –π ∗ resonance in the carbon structure, and therefore inhibition of the electronic transition that causes the anomalous dispersion. The submicroscopic diamond (DIA101) powder does not show a π –π ∗ sp2 peak (Fig. 2e), as expected, because diamond is pure tetrahedrally sp3 -bonded carbon. The decrease in reflectance with decreasing wavelength beginning at λ ∼230 nm, by comparison with the refractive indices plotted in Fig. 1c, indicates that this is where the “absorption edge” associated with the sp3 σ –σ ∗ transition begins. A comparison of these refractive indices suggests the diamond sp3 Fresnel peak should begin to rise in reflectance with decreasing wavelength near λ ∼ 170 nm, where k surpasses 0.1 and increases strongly with decreasing wavelength. Therefore, the λ ∼ 170 nm minimum can be considered the long wavelength transition minimum of the sp3 σ –σ ∗ transition in diamond. In coals, n can remain similar to that of graphites (e.g., McCartney and Ergun 1967), but the suppression of the π –π ∗ electronic transition (lack of increase in k) causes the disappearance of the sp2 peak in reflectance. This is also suggested by Papoular et al. (1995), who note that the σ –σ ∗ transition is present in all carbon, while the π –π ∗ will only occur if enough sp2 bonds are present. This also explains why Khare et al. (1984, 1993) observed this peak in tholin reflectance spectra, as tholins are dominated by amorphous sp3 carbon (see Fig. 1d). This indicates that both the σ –σ ∗ Fresnel peak and the Christiansen-like feature should remain identifiable features in spectra of low rank coals/impure carbon phases, but the observation of the π –π ∗ Fresnel peak near would indicate the presence of carbon with significant sp2 concentration. This should not be confused with high purity carbon samples losing the Fresnel peak in intimate mixtures with other materials. In

48

D.M. Applin et al. / Icarus 307 (2018) 40–82

Fig. 1c. Refractive indices of diamond and amorphous carbons. Annotations highlight absorption features and spectral properties. Refractive indices from Philipp and Taft (1962; 1964); Rouleau and Martin (1991), and Zubko et al., 1996.

Fig. 1d. Refractive indices of amorphous carbons and ‘Tholins’. Annotations highlight absorption features and spectral properties. Refractive indices from Hagemann et al., 1975; Larruquert et al., 2013; Khare et al., 1993; and Tuminello and Clark 1994b.

D.M. Applin et al. / Icarus 307 (2018) 40–82

49

Table 2 Selected sample compositional information. Coals COAL12

COAL16

COAL17

COAL19

COAL22

COAL32

COAL37

6.1 10.5 33.7 49.7

1.5 28.5 26.4 43.5

0.8 24.1 29.5 45.6

0.4 9.6 22.8 67.2

3.1 15.6 6.9 74.4

8.0 14.3 36.9 40.9

0.7 4.7 18.5 76.1

Ultimate analysis (d.b.) (wt.%) C 69.7 H 4.1 N 1.1 S 0.19 13.7 Oa H/C ratio 0.71 0.35 Toluene soluble fractionb

58.6 3.8 1.2 2.07 5.4 0.78 0.46

63.2 4.1 1.3 1.89 5.2 0.78 0.53

80.9 4.0 1.0 0.4 4.1 0.59 0.65

77.2 2.2 1.0 0.42 3.1 0.34 0.04

77.7 5.00 1.16 4.83 13.5 0.77 0.63

91.1 4.44 1.27 0.66 2.5 0.58 1.83

LCA101 0.11

Proximate analysis (wt.%) Moisture Ash Volatiles Fixed Carbon

Non-coals Toluene soluble fractionb N C H S O Toluene-soluble fractionb N C H S O

GRP101 0.53

GRP102 0.10

GRP103 0.00

TAR07c 13.0 0.5 93.2 10.5 4.7 4.7

OK-5Ec 17.1 0.9 78.9 9.1 2.6 8.5

JO-9E 8.5 0.9 75.9 8.5 8.8 5.9

ASP101c 1.34 0.34 72.6 9.7 7.1 10.3

ANT101c 0.08 0.7 92 1.1 0.1 3.7

OILS10c 0.04 2.0 80.4 10.3 0.7 9.3

n.a. not available or not determined. Coal analyses from Kalkreuth et al. (1991) and Vorres (1993). Additional sample information for OK-5E and JO-9E available in Hosterman et al. (1990). a by difference b Weight % c Analyses performed at the University of Alberta.

such a scenario, the contrast of the Fresnel peak should decrease with decreasing carbon content, but the pure carbon phase would retain the peak in reflectance. As a result, there is a significant difference between intrinsically impure carbon at the molecular level and high purity carbon in low concentration. To demonstrate this experimentally, we have made an intimate mechanical mixture of graphite and a lunar mare analogue (PSA003). Figs. 3a and 3b show that the Fresnel peak of graphite at λ ∼ 241 nm is readily observable in an intimate mechanical mixture at 2.54 wt.% in the JSC-1A lunar mare regolith simulant (PSA003). Although particles of graphite have been identified in aliquots of lunar soils (Steele et al., 2010; Thomas-Keprta et al., 2014), bulk geochemical analyses of lunar samples indicate carbon concentration at the ppm level (e.g. Holland et al., 1972; Moore et al., 1973). As such, the lunar surface should be an excellent UV reflectance standard for the detection of carbon compounds in low concentration mixed with silicate minerals, such as the surface of Mercury. 3.2. Effect of amorphization Although our sample suite is limited because of the diversity of possible carbon-rich materials, it does appear that the Fresnel peak maximum wavelength increases with increasing degree of amorphization, or decreasing long range ordering (Table 3; Figs. 2c and 2d). This is seen by comparing peak positions of several graphites (GRP101, GRP102, GRP105, GRP106, PIG402, GRPGDgr), which occur mainly near λ = 241 nm, to those of amorphous carbons: lampblack (LCA101; 272 nm), shungite (SHU102; 251 nm), amorphous carbon GRP200 (253 nm), and amorphous carbon GRPGDam (262 nm). This difference in peak positions is consistent with previous investiga-

tions and reviews of carbon phases (Taft and Philipp 1965; Papoular et al., 1995; Papoular et al., 1996; Fig. 1). It likely indicates that the overall π –π ∗ bonding distance decreases and that there is a wider distribution of bonding distances with increasing amorphization, causing the π –π ∗ peak to broaden and move to longer wavelengths. Papoular et al. (1995) also show this to occur with the σ –σ ∗ peak observed in coals—increasing broadening of the peak with decreasing coal rank. Progressive pulsed laser irradiation of a graphite sample also leads to progressive amorphization (Figs. 4a). The VNIR reflectance spectra of these samples are not significantly different (Gillis-Davis et al., 2015), but the UV reflectance spectra,show progressive amorphization by comparison with the other samples studied here. This is indicated by a progressive loss in contrast of the π –π ∗ Fresnel peak, and the progressive broadening and movement of this peak to longer wavelengths. The change in these parameters appears to be systematic with the number of laser pulses, or total irradiation energy, which may lend itself to measuring the relationship between space weathering and exposure age of graphitic materials, such as for ureilite parent bodies, or the surfaces of Mercury, Phobos, and Deimos. 3.3. Grain size effects We investigated grain size effects on UV reflectance spectra of carbonaceous materials, as it is likely that grain size differences will affect features such as Fresnel peak intensity, as indicated in the voluminous IR literature, due to variations in overall multiple scattering between grain surfaces before emergence. Ultraviolet spectra of various grain size fractions of COAL16, a high volatility bituminous coal, are shown in Figs. 5a and 5b. Because this material is compositionally complex and impure, only 58.6% carbon

50

D.M. Applin et al. / Icarus 307 (2018) 40–82

that has significant H, O, and N, it does not show well-defined peaks or absorption bands, but it can still be used to probe general reflectance changes of a impure carbon-rich silicate material with grain size in the UV region. The low overall reflectance of the 250–500 μm diameter size fraction is likely due to abundant surface asperities, acting as “Rayleigh absorbers” (e.g. Hapke, 2012). With increasing grain size, reflectance increases with decreasing wavelength below λ∼240 nm. Such an effect is expected due to the transition from volume-scattering dominated behavior to surfacescattering dominated behavior near λ∼300 nm in silicates. This is characteristic of silicates and other materials whose spectra are dominated by Fe2+/3+ –O charge transfers, while this sample contains ∼58% C and various silicate minerals (Table 2). This transition indicates where they turn opaque in wavelength, and is on the long wavelength wing of the Fe2+/3+ –O absorption bands. The spectra of COAL16 exhibit a peak near λ∼230 nm, but it may be a combination of a Fresnel peak, the long wavelength wing of a silicate Fe2+ –O charge transfer near 260 nm, and the strong absorption towards the Christiansen-like feature in the carbonaceous fraction. A survey of silicate mineral UV reflectance spectra shows that an interband (between the Fe2+ –O and Fe3+ –O charge transfers) reflectance peak near λ = 230 nm occurs mainly in serpentinites and some Fe-bearing silicates (e.g., olivine, pyroxene: Cloutis et al., 2008, 2015). The largest grain size of COAL16 appears to show that the Fe2+ –O charge transfer becomes a Fresnel peak in reflectance at very large grain sizes. The spectral behavior of this silicate-rich coal sample also indicates that large amounts of impure carbon with significant volatile fractions (H, O, and N) do not completely mask the silicate spectral properties in the NUV. We also investigated the effects of grain size on graphite (GRP106) and amorphous carbon/shungite (SHU102), shown in Figs. 5c–5f. All of these aliquots were thoroughly wet sieved and washed with both distilled water and ethanol to remove clinging fines and soluble impurities. The smallest fractions presumably contain all soluble impurities (solutions were evaporated at 90 °C and ∼600 Pa pressure), which are expected to contribute minimally because these are high-purity samples. GRP106 shows that for graphite powders, reflectance strongly increases with increasing grain size (Figs. 5c and 5d), which is expected for an opaque powder due to the decrease in multiple scattering between grain surfaces before emergence, or that relatively more light is reflected from only one surface rather than being directed onto multiple surfaces before emergence. After normalizing these spectra at λ = 399 nm, it can be seen that the spectral contrast of the π –π ∗ sp2 Fresnel peak decreases with decreasing grain size. The change in spectral contrast only becomes strong for the smallest grain size fraction measured (sieve diameter <20 μm). Additionally, the wavelength position of the Fresnel peak remains static except at the smallest grain size, where it moves to λ = 248 nm. Similar results are shown for the amorphous carbon/shungite samples (Figs. 5e and 5f), although the overall reflectance does not appear to change strongly with different grain size fractions. As with graphite, the spectral contrast of the π –π ∗ sp2 Fresnel peak seems to only be significantly reduced at the smallest grain size fraction measured, and the Fresnel peak position moves to slightly longer wavelengths for only the smallest grain size fraction, moving from λ = 251 nm to λ = 265 nm. Some of this effect may be due to soluble impurities, which requires further investigation. The Fresnel peak remains observable in both nanophase lampblack (LCA101; <0.021 μm grain size) and amorphous carbon (GRP200; <0.03 μm grain size) (Figs. 2c and 2d). GRP200 is identified as amorphous carbon via Raman spectroscopy, however X-ray diffraction indicates there may be minor components of quartz and possible trace minerals; see the supplementary material. To summarize the grain size effects, decreasing grain size of graphite and amorphous carbon powders generally leads to decreasing overall reflectance due to an increase in the total amount

of scattering between grain surfaces, or that less light is reflectance only once, before emergence. Decreasing grain size also leads to a reduction in the spectral contrast of the trigonal sp2 π –π ∗ Fresnel peak in graphite and amorphous carbon, and appears to move the peak position to longer wavelengths. These results are broadly consistent with the spectral behavior of absorption bands in the IR spectral region. The effects of grain size on the σ –σ ∗ peak are not clear and need to be studied. Our measurements do not extend to these wavelengths, and grain size series of powders at these wavelengths have not been published. 3.4. Thermal processing effects We investigated how thermal processing of organics, which is expected to lead to a loss of more volatile organic compounds and possible increase in overall aromaticity, is expressed in UV reflectance spectra. UV spectra were collected on aliquots of a <45 μm fraction of COAL19, a medium-volatility bituminous coal, that were heated at 450 °C in air for up to three days (Figs. 6a and 6b). Overall reflectance generally increases with increasing heating time, likely due to loss of carbon, which is the primary darkening agent. The carbon is likely lost as CO and CO2 , but evolved gas analysis was not used in this study. Also shown is the progressive increasing reflectance with decreasing wavelength of the spectra up to a 6-hour heating time, which is accompanied by a 77.50 wt.% loss. Our interpretation of this is the progressive loss of carbonaceous phases, which causes a progressive loss in the absorption edge towards the Christiansen-like feature. This is supported by the change in this slope with H/C and O/C content of coals, described in Section 3.6. These spectra appear very similar to the vitrain reflectance spectra presented by Gilbert (1960), which show the same spectral trend with carbon content of the coal vitrain samples. It is also possible that the increase in reflectance at the lower wavelengths could be due to the formation of magnetite and/or hematite. With further heating, the spectra progress toward a characteristic silicate-dominated UV reflectance spectrum, which is strongly red-sloped with a somewhat skewed, concave feature with a minimum near λ = 260 nm. This likely indicates that the Febearing silicates contribute progressively more to the overall spectrum as carbon concentration decreases. This is consistent with the fact that some clay minerals are spectrally similar to the heated coal spectra, e.g. montmorillonite (Cloutis et al., 2015). 3.5. Compositional effects Previous studies have shown that for carbonaceous materials such as graphite and amorphous carbon, small changes in composition and structure can have an outsized effect on the derived refractive indices (Henning, 1965; Yasinsky and Ergun, 1965; McCartney and Ergun, 1967; Ergun, 1968; Lochmuller et al., 1981; Gilbert, 1960; Ergun et al., 1967; Foster and Howarth, 1968). Here, the UV spectral variability of coals with different compositions (Table 3) was also explored (Figs. 7a and 7b). The COAL16 spectrum appears to be dominated by silicates, as evidenced by its higher overall reflectance, volume-scattering behavior λ > 300 nm, strong Fe2+ –O absorption feature near λ = 260 nm, and relatively positive spectral slope. Strong absorption into the FUV should increase with decreasing H/C ratio due to increasing contrast of the Christiansenlike feature, which is confirmed by these samples. Coal samples 12, 16, 17, and 32 have the highest H/C ratios (Table 2), and are among the flattest spectra in the <250 nm region. Figs. 7c and 7d indicate that this spectral slope, in coals of similar grain size distribution, is a function of carbon content. The λ = 260 nm Fe–O absorption band is present in many of the spectra, most evident in the COAL16 spectrum, as well as an apparent peak near λ = 230 nm, which is likely the peak between the Fe3+ –O and Fe2+ –O features. These results are quite similar to those presented by Gilbert (1960), who

D.M. Applin et al. / Icarus 307 (2018) 40–82

51

Table 3 Samples that exhibit a UV Fresnel reflectance peak. Sample number

Sample type

Grain size (μm)

Fresnel peak max. (nm)

LCA101 GRP200 ANT101E ANT101 ANT101 SHU102 SHU102 SHU102 SHU102 SHU102 SHU102 SHU102 GRP101 GRP101 GRP101E GRP102 GRP105 GRP106 GRP106 GRP106 GRP106 PIG402 GRPGDam GRPGD40 GRPGD20 GRPGD17 GRPGD15 GRPGD12 GRPGD10 GRPGD5 GRPGDgr OSFL03 JO-9E

Lampblack Amorphous graphite Anthraxolite insoluble fraction—amorphous carbon Anthraxolite—amorphous carbon Anthraxolite—amorphous carbon Shungite—amorphous carbon Shungite—amorphous carbon Shungite—amorphous carbon Shungite—amorphous carbon Shungite—amorphous carbon Shungite—amorphous carbon Shungite—amorphous carbon Graphite Graphite Graphite insoluble fraction Graphite Graphite Graphite Graphite Graphite Graphite Graphite Amorphous carbon Graphite—laser zapped Graphite—laser zapped Graphite—laser zapped Graphite—laser zapped Graphite—laser zapped Graphite—laser zapped Graphite—laser zapped Graphite Oil sand Asphaltite

<0.021 <0.030 <45 Unsorted—large grains Slab Unsorted <25 20–32 32–45 45–61 61–90 90–150 45–90 <37 <37 <45 >45 <20 20–45 45–90 >90 Stick 10–20 <45 <45 <45 <45 <45 <45 <45 <45 Unsorted Unsorted

272 254 250 239 231 254 265 251 251 251 251 251 244 243 241 244 243 248 241 242 241 241 262 257 246 246 245 243 242 241 240 207, 242, 276a , 310 206a , 241, 274a , 310

a

very weak

shows vitrain coal reflectance spectra have strong absorption toward the FUV with increasing carbon content. 3.6. Carbonaceous chondrites and ureilites To understand how thermal metamorphism affects carbonaceous chondrite UV reflectance spectra, we measured the UV reflectance spectra of heated Murchison CM2 samples that were prepared and previously analyzed by Hiroi et al. (1993) in the VNIR (Figs. 8a and 8b). The 5377 (unheated) sample is from a different piece of the meteorite, and is of a different grain size range, and therefore may not be truly representative of the spectral properties of the starting unheated material. In general, we find that overall reflectance increases with increasing temperature. As with the heated COAL19bk samples, Murchison powder becomes brighter with decreasing wavelength in the UV with increasing heating temperature, followed by a reversal at higher temperatures (∼700 °C), likely due to increasing loss of carbon and formation of magnetite and/or hematite. The increasing spectral contribution of silicates with increasing temperature is evidenced by a peak near λ = 230 nm emerging near 700 °C. As with the coal spectra, our interpretation is that this is an interband feature between the longer wavelength Fe2+ –O absorption feature and the drop-off towards the Christiansen-like feature from carbon phases plus silicate Fe3+ –O absorption feature, rather than a typical Fresnel peak. This is consistent with increasing spectral contribution from Febearing silicates and Fe-oxide phases. We also examined UV reflectance spectra of Parr bomb-heated samples of the Jbilet (dry-only) and Murchison (wet and dry) CM2 carbonaceous chondrites (Figs. 9a–9c). We found that dry-heated powdered samples of Jbilet and Murchison (<45 μm, heated in sealed containers to 200 °C for 111 days) appear more spectrally

similar to unheated saw-cut fines and other CM2 powders. This may be due to the formation of hematite from heating and grinding, as the spectral shape looks broadly similar to hematite (e.g., Fig. S1). Murchison heated in a 1:1 ratio with deionized water appears much brighter towards the far-UV than the dry-heated sample, and is spectrally more consistent with the thermally processed Murchison samples from Hiroi et al. (1993). This may indicate that hydrous alteration removes carbon via CO2 formation at much lower temperatures than dry alteration. The absorption feature near λ = 260 nm may be mostly or entirely due to Fe2+ –O. This spectrum of wet-heated Murchison also appears to more like hematite and NWA801, which is contaminated by significant quantities of Fe-oxides (red visible color). We further examined how composition and grain size of carbonaceous chondrites may affect their UV reflectance spectra, as well as to look for evidence of contributions from their carbonaceous phases. Our samples included roughened slabs, laboratoryproduced powders (<150 μm grain size) and saw-cut fines (generally <5 μm grain diameter) (Figs. 9a–9c). Sections measured on CH and CB slabs were situated in regions where the least amount of metal was observed. The spectra from the slabs are more likely to contain Fresnel peaks, since the grains are generally larger than powders and our methods do not eliminate the specular angles. The Moss CO3.6 carbonaceous chondrite slab spectrum does contain a peak at λ∼242 nm, which would be consistent with spectral contributions from a crystalline carbon sp2 -bearing component. The NWA 2210 and SaU 290 CH3 slab spectra appear to have a very weak and broad peak situated near λ∼205 nm, which may also indicate spectrally active carbonaceous components (e.g., tetrahedral amorphous carbon or kerogen-like). Despite being highly oxidized and visibly red due to terrestrial weathering, the NWA 801 CR2

52

D.M. Applin et al. / Icarus 307 (2018) 40–82

Fig. 1e. Refractive indices of amorphous carbons and ‘Tholins’. Annotations highlight absorption features and spectral properties. Refractive indices from unpublished work by Bishun Khare; Khare et al., 1990; and Khare et al., 1994.

powder is spectrally somewhat brighter than the other chondrites with decreasing wavelength in the UV and exhibits possible peaks near λ∼250 nm and λ∼205 nm. This is likely the contribution of hematite (see Fig. S1). A number of other samples show this evidence for hematite (Fig. 9c), and they are all either dry-heated or saw-cut fines, which may indicate significant oxidation occurred. The peak near λ∼205 nm can be characteristic of amorphous tetrahedral sp3 -bearing materials, like kerogen (Fig. 1e), but contributions from other phases cannot be ruled out. We do not observe any evidence for a graphite or amorphous carbon π –π ∗ in the λ = 240–300 nm range, which is expected given the low carbon concentration and type IV kerogen-like composition. We find that CV and CO chondrites can be easily distinguished from CI and CM based on their albedo, which is likely indicative of their total carbon contents (Fig. 9d). Spectral slopes in the 200– 400 nm range, however, do not seem to be indicative of any classification (Fig. 9e). Shorter wavelengths show strong promise for detecting and classifying carbonaceous chondrites. The normalized spectra in Fig. 9e show that CI and CM chondrites show evidence of amorphous sp3 carbon, or kerogen-like, with peaks that look similar to those of kerogen, Murchison IOM, and tetrahedral amorphous carbon, as seen in Fig. 1d. CV and CO chondrites show a relatively stronger peak near 160 nm, which is likely indicative of a silicate band (Wagner et al., 1987), and is probably masked to a greater degree by carbon in the CI and CM chondrites. As shown in Figs. 3a and 3b, lunar materials are excellent UV reflectance standard when searching for carbon because of their extremely low carbon concentration. We have divided the carbonaceous chondrite spectra from Wagner et al. (1987) by that the mean of 10 lunar samples in order to investigate this further (Fig. 9f). We find that graphite (below 95 nm) and possibly diamond (near 100 nm) become readily observable in the Goalpara

ureilite, with relatively strong σ –σ ∗ peaks. Most of the carbonaceous chondrites appear to show evidence of tetrahedral amorphous carbon, or kerogen-like, showing the characteristic σ –σ ∗ peaks. We also find that the ∼160 nm silicate peak is likely due to olivine, since it is not readily observable in lunar samples, but present in some ordinary chondrites and CV and CO chondrites. It appears that the FUV reflectance provides a mechanism for identifying and classifying carbonaceous chondrite-like airless bodies. These results should be explored in much greater detail with further laboratory work. 3.7. Other carbon-bearing geological materials To further complement our analysis, we measured UV reflectance spectra of a suite of oil sands, oil shales, and their toluene-insoluble fractions (Figs. 10a and 10b). These materials are of interest in studying the carbon fraction of airless bodies, and have been suggested as analogues to the carbon fractions from airless bodies by numerous publications (e.g., Moroz et al., 1998). The spectra of the oil sands and some of their toluene-insoluble fractions appear very much like silicates in the UV; i.e., reflectance increasing strongly with increasing wavelength (Fig. 10a). Oil sands and oil shales exhibit a weak peak near λ = 220–230 nm, which can be interpreted as an interband feature between the longer wavelength Fe2+ –O charge transfer and the lower wavelength carbon Christiansen feature and/or Fe3+ –O charge transfer. The OILS10 and its toluene-insoluble fraction, OILS10e, both show this feature, but the latter is brighter with decreasing wavelength than the unextracted material, likely due to the removal of spectrally redsloped organic components. Bitumen-rich materials, like OSFL03 and the JO-9E type II kerogen exhibit relatively brighter spectra with decreasing wavelength and have similar peaks situated near

D.M. Applin et al. / Icarus 307 (2018) 40–82

53

Fig. 1f. Demonstration of the relationship between proportional scattering intensity from grain boundaries of single large spherical particles at normal incidence, with refractive indices. A careful comparison of this plot with Figs. 1a and 1b demonstrates how reflectance spectra are affected by high k and changes in n from the Christiansen frequencies (n = 1). These calculations assume no absorption from surface asperities. See Eqs. (4.17) and (5.37) from Hapke (2012) for discussion. These calculations highlight the different scattering domains.

Fig. 1g. Same as Fig. 1f, with a linear Y-axis. A careful comparison of this plot with Figs. 1a and 1b demonstrates how reflectance spectra are affected by high k and changes in n from the Christiansen frequencies (n = 1). These calculations highlight the different scattering domains.

54

D.M. Applin et al. / Icarus 307 (2018) 40–82

Fig. 1h. Ultraviolet reflectance of carbon samples, modified from Papoular et al. (1995). The samples in this study were pressed pellets of very fine powders, with the exception of anthracite, which was a powdered sample.

Fig. 1i. Same as Fig. 1h., but normalized to unity at 400 nm.

D.M. Applin et al. / Icarus 307 (2018) 40–82

Fig. 2a. Reflectance spectra of various graphite powders. Small, sharp peaks in reflectance are due to emission lines in the deuterium light source.

Fig. 2b. Same as Fig. 2a, but normalized to unity at 399 nm.

55

56

D.M. Applin et al. / Icarus 307 (2018) 40–82

Fig. 2c. Ultraviolet reflectance spectra of various amorphous carbon powders.

Fig. 2d. Same as Fig 2e, but normalized to unity at 399 nm.

D.M. Applin et al. / Icarus 307 (2018) 40–82

Fig. 2e. Ultraviolet reflectance spectra of submicroscopic diamond powder.

Fig. 3a. Ultraviolet reflectance spectra of an intimate mechanical mixture of graphite (GRP101; <37 μm) and a lunar analogue powder (PSA003; <45 μm).

57

58

D.M. Applin et al. / Icarus 307 (2018) 40–82

Fig. 3b. Ultraviolet reflectance spectra of the intimate mechanical mixture from Fig 3a ratioed against the reflectance spectrum of the pure lunar analog powder (PSA003; <45 μm).

λ = 207, 242, 276 nm, and very weakly at λ = 310 nm. As mentioned, the λ∼205 nm feature can be characteristic of kerogen-like materials, and poly-HCN measured by Khare et al. (1993) exhibits a maximum in k at λ = 207 nm. Some of these features may be due to graphitic components, especially the feature at λ = 242 nm (compare to Fig. 1b), while the best candidates for the feature near λ = 310 nm are PAHs, which can exhibit a wide variety of π–π∗ transition positions in the NUV (supplementary material), and their characterization will be part of follow-on work. Note that a feature near λ = 207 nm is observed in many of the carbonaceous chondrite slabs. This may be due to both an increased Fresnel peak reflectivity from larger grains, and the presence of kerogen-like components. The OK-5E sandstone extract spectrum appears like many of the coals, relatively featureless and sloping down towards the Christiansen features below λ = 200 nm. The asphaltum spectra do not show increasing absorption towards the far-UV, due to the relatively low concentration of carbon, as expected (Table 2; Fig. 10c). Gilsonite and albertite show some evidence of carbon peaks and downslopes towards the CF. 4. Application to remote sensing of airless bodies Recent studies have suggested that multiple wt.% of carbon is responsible for the overall low albedo and possible suppression of silicate features of mercurian surface reflectance spectra. Syal et al. (2015) estimated that delivery of 50 times as many carbon-rich micrometeorites per unit surface area relative to the lunar surface resulted in 3–6 wt.% carbon in the form of graphite, amorphous carbon, or diamond. Peplowski et al. (2016) reported that the MESSENGER measurements of thermal-neutron counts and reflectance spectra of the darkest units reveal 1–3 wt.% more carbon than the surrounding terrains. They suggest this carbon is the remnant of igneous graphite mixed with volcanic materials via im-

pact processes. Refractive indices of amorphous carbon have been used to accurately model the reflectance spectra from the surface of Mercury (Trang et al., 2017). Previously measured high S/N diskaveraged reflectance spectra of Mercury divided by those of the lunar farside show increasing relative reflectance with decreasing wavelength from λ = 1400 nm to ∼320 nm, with an abrupt downturn in reflectance towards the FUV (Fig. 11). As previously discussed, the lunar surface is an excellent UV reflectance standard in the search for carbon, given the extremely low carbon values measured in returned Apollo samples. A comparison of the Mercury spectral data to those presented here and in the literature indicates that the ratioed spectra look significantly similar to the UV– Vis-NIR spectrum of our lampblack LCA101 sample (Fig. 11). These reflectance spectra may indicate that the bulk carbon on surface of Mercury is low in graphitic sp2 fractions, and extremely amorphous and/or fine grained if the peak in reflectance is due to a π–π∗ transition. If the peak in reflectance is not due to a π–π∗ transition, the downturn in reflectance towards the FUV may still indicate that the carbon could be cubic diamond with an absorption edge from the σ –σ ∗ transition. Due to the probable requirement for this carbon to be a darkening agent, this indicates that the carbon is likely not diamond, it is likely to be amorphous sp2 bearing carbon which consistent with the results by Trang et al. (2017). Further, we have shown that pulsed laser weathering of graphite powders causes progressive amorphization (Fig. 4a). As such, if graphite is found on Mercury (whether igneous or introduced), it must be recently emplaced as it would progressively space weather to amorphous phases. A peak in reflectance, situated near λ∼280 nm, has been observed in spectra of the surface of both Phobos and Deimos (Bertaux et al., 2011, 2016, pers. comm.., 2017; Figs. 12a and 12b). Such a peak in addition to such low reflectance values, by com-

D.M. Applin et al. / Icarus 307 (2018) 40–82

59

Fig. 4a. Ultraviolet reflectance spectra of laser-irradiated graphite, normalized to unity at 399 nm. Note that the Fresnel peak broadens and shifts to longer wavelengths with increasing irradiation time.

Fig. 5a. Ultraviolet reflectance spectra of several grain size fractions of sample COAL16 powder, a high volatility A bituminous coal with a significant silicate fraction.

60

D.M. Applin et al. / Icarus 307 (2018) 40–82

Fig. 5b. Same as Fig 5a, but normalized to unity at 399 nm.

Fig. 5c. Ultraviolet reflectance spectra of various grain sizes of the graphite GRP106. All samples were thoroughly wet sieved with distilled water and ethanol and these solutions were evaporated from the sieve diameter <20 μm fraction.

D.M. Applin et al. / Icarus 307 (2018) 40–82

61

Fig. 5d. Same as Fig 5c, but normalized to unity at 399 nm.

Fig. 5e. Ultraviolet reflectance spectra of various grain sizes of the shungite/amorphous carbon samples SHU102. All samples were thoroughly wet sieved with distilled water and ethanol and these solutions were evaporated from the sieve diameter <25 μm fraction.

62

D.M. Applin et al. / Icarus 307 (2018) 40–82

Fig. 5f. Same as Fig 5e, but normalized to unity at 399 nm.

Fig. 6a. Ultraviolet reflectance spectra of COAL19, a medium volatility bituminous coal, heated to 450 °C in air for various lengths of time. Note that COAL19bk24 was heated at 410 °C.

D.M. Applin et al. / Icarus 307 (2018) 40–82

Fig. 6b. Same as Fig 6a, but normalized to unity at 399 nm.

Fig. 7a. Ultraviolet reflectance spectra of various coals with different compositions. See Tables 2 and 3 for sample descriptions and compositional information.

63

64

D.M. Applin et al. / Icarus 307 (2018) 40–82

Fig. 7b. Same as Fig 7a, but normalized to unity at 399 nm.

Fig. 7c. Relationship between the absorption edge and H/C ratio in coals.

D.M. Applin et al. / Icarus 307 (2018) 40–82

65

Fig. 7d. Relationship between the absorption edge and O/C ratio in coals.

Fig. 8a. Ultraviolet reflectance spectra of heated fractions of a homogenized Murchison sample. The unheated sample is not from the same piece of Murchison, but it included for comparison. Heated samples were prepared and analyzed with VNIR reflectance by Hiroi et al. (1994).

66

D.M. Applin et al. / Icarus 307 (2018) 40–82

Fig. 8b. Same as Fig 8a, but normalized to unity at 399 nm.

Fig. 9a. UV reflectance spectra of some carbonaceous chondrites. Spots shot on the Bencubbin and Moss slabs were areas with the least amount of observable metal. See Tables 2 and 3 for sample descriptions and compositional information.

D.M. Applin et al. / Icarus 307 (2018) 40–82

Fig. 9b. Same as Fig. 9a, but a different sample.

Fig. 9c. Samples from Figs. 9a and 9b, but normalized to unity at 399 nm.

67

68

D.M. Applin et al. / Icarus 307 (2018) 40–82

Fig. 9d. Ultraviolet reflectance spectra of the Goalpara ureilite and the Orgeuil CM2 chondrite, from Wagner et al. (1987).

parison to all of the current UV databases, is only consistent with fine grained amorphous carbon; lampblack LCA101, in particular. A comparison to the spectral data presented here indicates that the surface of both Phobos and Deimos probably contain spectrally significant concentrations of amorphous carbon. Our laboratory results indicate that as grain size decreases into the submicroscopic domain, the Fresnel peak position moves to longer wavelengths with decreasing grain size for both graphitic and amorphous carbon. As such, we cannot rule out nanophase graphite as an explanation for the peaks found in the spectra from Phobos and Deimos. Such laboratory experiments are required. Additional VNIR measurements show Phobos to exhibit spectra similar to main belt D-type asteroids and Jupiter trojans (Pajola et al., 2013). Reflectance spectra from the surface of Ceres has shown a peak in reflectance near λ ∼ 160 nm (Parker et al., 2002; Li et al., 2009; Hendrix et al., 2016a; Figs. 13a–13c). Hendrix et al. (2016a, 2016b) assigned this to a transparency feature from a ∼0.5 μm thick veneer of graphitized carbon at a concentration of ∼85–90%, which was noted to not contribute to the near-UV–visible spectral shape. Carbon transparency in this wavelength region can be calculated for slabs of different thicknesses (supplementary material). We find that for graphite, amorphous carbon, and glassy carbon, they only become transparent at these wavelengths below a grain diameter that departs geometric optics, or that the wavelength is larger than the particle size. The application carbon transparency at these wavelengths to remote sensing assumes that there is no multiple scattering in the particulate surface, that the transmission of the single particle in the forward direction is analogous to volume scattered light from a powder, that there is no internal scattering, and that total surface scattering of powders is not a function of grain size distribution. It is unclear if any carbon powder can exhibit a transparency feature near λ ∼ 160 nm, since the reflectance of opaque powders can be strongly dependent on the grain size

distribution due to the relative amount of scattering between grain surfaces before emergence, and that the number of times a photon passes through a grain in a powder rather than a single grain will be higher than 1 before emergence at wavelengths that become non-opaque at sufficiently small grain sizes. The spectral properties of nanophase carbon powders are likely complicated by diverging from the geometric optical regime, or when the Mie size parameter x = π D/λ becomes lower than about unity, which occurs for particles roughly smaller than ∼100 nm in the UV wavelength range (supplementary material). Whether any carbon powders can strongly increase in reflectance at λ ∼ 160 at any grain size and if such an increase in reflectance can be detected in intimate mixtures with geologic materials will be the subject of follow-on work involving laboratory experiments and radiative transfer modeling. There are a number of other materials that could potentially contribute to the reflectance peak observed in the spectra from the surface of Ceres. For example, a number of ordinary chondrites measured by Wagner et al. (1987) appear to show a Fresnel peak situated near 160 nm that is likely due to olivine (Fig. 13c). This peak is also weakly present in CV and CO chondrites (Figs. 9e, 9f, and 13d).There is also some indication the small carboxylic acids exhibit the C=O–OH transition at these wavelengths (Barns and Simpson, 1963; Nagakura et al., 1964; Pradeep and Rao, 1989; Limao-Vieria et al., 2006). In Fig. 13d we show that the reflectance spectra of the CO3 Warrenton and CV3 Allende powders look similar to those from Ceres, including the peak in reflectance near 160 nm. In these powders, this peak is likely indicative of olivine. As such, the surface of Ceres may exhibit spectral properties similar to metamorphosed carbonaceous chondrites that are significantly darkened. A Fresnel peak at UV wavelengths would likely have lower detection limits than the very weak Fe2+ bands exhibited by olivine. Olivine on Ceres could be from unreacted pri-

D.M. Applin et al. / Icarus 307 (2018) 40–82

Fig. 9e. Ultraviolet reflectance spectra of the Goalpara ureilite and the Orgeuil CM2 chondrite, from Wagner et al. (1987).

Fig. 9f. Ultraviolet reflectance spectra of the Goalpara ureilite and the Orgeuil CM2 chondrite, from Wagner et al., (1987).

69

70

D.M. Applin et al. / Icarus 307 (2018) 40–82

Fig. 10a. Ultraviolet reflectance spectra of oil sands, oil shales, and toluene-insoluble fractions. See Tables 2 and 4 for sample descriptions and compositional information.

Fig. 10b. Same as Fig 10a, but normalized to unity at 399 nm.

D.M. Applin et al. / Icarus 307 (2018) 40–82

71

Fig. 10c. Reflectance spectra of various other carbon-rich samples.

Fig. 11. Reflectance spectra from the surface of Mercury compared to lampblack. The observational spectra are from disk-averaged reflectance measurements. The lunar spectra used for the ratio are of the farside. Observational spectra were digitized from the source listed in the figure.

72

D.M. Applin et al. / Icarus 307 (2018) 40–82

Fig. 12a. Reflectance spectra from the surface of Phobos. Observational spectra were digitized from the sources listed in the figure, Bertaux et al. (2016) and Fraeman et al. (2012). The radiance factors of the two measurements by Bertaux et al. (2016) are directly comparable because the whole observation of orbit 756 at a distance of 158 km was averaged.

Fig. 12b. Reflectance spectra from the surface of Phobos and Deimos. Observational spectra were digitized from the sources listed in the figure. The radiance factors of the two measurements by Bertaux et al. (2016) are directly comparable because the whole observation of orbit 756 at a distance of 158 km was averaged.

D.M. Applin et al. / Icarus 307 (2018) 40–82

73

Fig. 13a. Reflectance spectra from the surface of Ceres. Observational spectra were digitized from the sources listed in the figure, De Sanctis et al. (2017); Thangjam et al. (2018); and Hendrix et al. (2016a). The spectra from Hendrix et al. (2016a) were normalized to the Framing Camera data presented by Thangjam et al. (2018) at the 438 nm channel.

Fig. 13b. Same as Fig. 13a, but over a longer wavelength range. Observational spectra were digitized from the sources listed in the figure.

74

D.M. Applin et al. / Icarus 307 (2018) 40–82

Fig. 13c. Same as Figs. 13a and 13b, but over a different wavelength range and compared to reflectance spectra of the Khanpur LL5 ordinary chondrite. Observational spectra were digitized from the sources listed in the figure.

Fig. 13d. Same as Figs. 13a and 13b, but over a different wavelength range and compared to reflectance spectra of the Warrenton CO3 chondrite and Allende CV3 chondrite. Observational spectra were digitized from the sources listed in the figure.

D.M. Applin et al. / Icarus 307 (2018) 40–82

75

Fig. 14. Reflectance spectra from the surface of C-group asteroid 704 Interamnia. Data were digitized from the source listed in the figure.

mary crust and from exogenous sources, among other possibilities. Laboratory data are sparse, and UV reflectance spectra of other potential surface materials should be collected, e.g., hydroxides, chlorides, NH4 -materials, carbonates, serpentines, refractory organic materials, etc. There is some indications that carbon from the dark regions on Vesta could contribute of a trigonal carbon Fresnel peak, since the global average I/F show a strong increase in reflectance with decreasing wavelength down to ∼260 nm (supplementary material), but telescopic observations have not shown this, and it is therefore likely a detector artifact. Reflectance spectra from the surface of asteroid 704 Interamnia, which is classified variably in the C-group, have been measured on numerous occasions with suggestions that the spectral properties are similar to carbonaceous chondrites (Hiroi et al., 1993) and a mixture of carbonaceous chondrites, serpentine, and Fe3+ oxides and/or hydroxides (Busarev et al., 2015). The spectral properties of Interamnia have been shown to be very similar to those of asteroid 1 Ceres (Takir and Emery, 2012; Rivkin et al., 2014). Previous UV observations show some indication for a peak in reflectance situated near ∼255 nm in wavelength before and after division by lunar surface spectra (Fig. 14; Roettger and Buratti, 1994). A trigonal carbon peak in reflectance for a C-group asteroid may be expected, particularly if the carbon content of the surface is high, and the surface is highly weathered potentially leaving a high concentration of sp2 -bearing carbon. However, the carbon of carbonaceous chondrites and kerogens appear to be dominated by amorphous tetrahedral sp3 (Figs. 1d and 1e). A number of scenarios may be possible that could explain sp2 -rich carbon, including weathering of igneous graphite such as on a ureilite parent body or space weathering mechanisms converting sp3 to sp2 , or sublimation of more volatile carbon over solar system timescales that leave a lag of highly stable phases such as diamond, graphite, and sp2 -rich

amorphous carbon. Future UV observations of C-group asteroids should provide details on their carbon fractions and how they relate to carbonaceous chondrites. Visible and near-infrared reflectance spectra from the surface of 624 Hektor have been acquired on multiple occasions. The reflectance properties are often indicated to be consistent with organic materials (Cruikshank et al., 2001 and references therein). Perna et al. (2017) show the reflectance to be similar to cometary nuclei and compare models that include amorphous carbon and kerogen suggesting the body can be considers as dead or dormant comets. There is some indication that the UV spectra from the surface of the asteroid show evidence of the trigonal carbon Fresnel peak (Fig. 15), with increasing reflectance with decreasing wavelength to below 350 nm. This change in spectral slope is within error bars, but may be expected for a carbon-rich surface. Additionally, there is some indication that the newly discovered interstellar asteroid 1I/2017 U1 (‘Oumuamua) exhibits similar spectral behavior (Ye et al., 2017; Fitzsimmons et al., 2018), which is consistent with the expectation that it is carbon-rich. The reflectance spectra from the nucleus of comet 67P/Churyumov–Gerasimenko presented by Stern et al. (2015), show clear carbon features near λ∼90 nm (peak) and λ∼170 nm (minimum), and were identified as such. The very low albedo through the UV and VNIR is also evidence for a significant fine-grained carbon fraction. However, it has been shown that submicroscopic sulfides are also sufficient darkening agents, albeit at much higher concentrations (Rousseau et al., 2017). Although the spectra may be too noisy for interpretation, the position and shape of the λ∼90 nm peak appear similar to the coals measured by Papoular et al. (1995; Fig. 1). This may give some indication to the origin and evolution of the carbon. Previous compiled measurements of UV reflectance spectra from the surface of Iapetus have been interpreted as potentially

76

D.M. Applin et al. / Icarus 307 (2018) 40–82

Fig. 15. Reflectance spectra from the surface of the Jupiter Trojan 624 Hektor, from multiple sources. All data are digitized from Cruikshank et al. (2001), and include data from Zellner et al., 1985; Vilas et al., 1993.

Fig. 16a. Reflectance spectra from the nucleus of comet 67P/Churyumov–Gerasimenko collected with two spectrometers onboard the Rosetta spacecraft. Data are digitized from the sources listed in the figure.

D.M. Applin et al. / Icarus 307 (2018) 40–82

77

Fig. 16b. Same as Fig. 16a, but showing only the FUV spectral range. Data are digitized from the source listed in the figure.

Fig. 17a. Reflectance spectra of the leading hemisphere of Iapetus and dark material from Iapetus compared to laboratory reflectance spectra of hematite from Wagner et al. (1987). Observational spectra are digitized from Hendrix et al. (2016a) and Clark et al. (2012), and include data from Vilas et al. (1996), Hendrix and Hansen (2008), and Noll et al. (1997).

78

D.M. Applin et al. / Icarus 307 (2018) 40–82

Fig. 17b. Same as Fig. 17a, but over a longer wavelength range. Observational spectra are digitized from Hendrix et al. (2016a) and Clark et al. (2012).

showing a PAH Fresnel peak near 200 nm (Hendrix et al., 2016b). We find that the ultraviolet reflectance spectra from the surface of the leading hemisphere of Iapetus (Figs. 17a and 17b) appear to be consistent with hematite, which has already been suggested to be present in significant quantities from multiple other studies (e.g., Jarvis et al., 20 0 0; Clark et al., 2012). Evidence comes from the dark visible albedo, the decreasing in reflectance with decreasing wavelength from the visible into the UV, the negative spectral slope from 250 nm to ∼200 nm with a peak in reflectance near 200 nm, followed by a decrease in reflectance to ∼150 nm, and another increase in reflectance to shorter wavelengths. These spectral characteristics appear very much like that of previous measurements of a hematite powder. As such, the UV spectral region may be used to differentiate between darkening by oxides, carbon, and metals in remote sensing.

•

•

•

5. Summary and conclusions Due to strong absorption mechanisms causing high refractive indices, reflectance spectroscopy in the UV (λ = ∼10–400 nm) can be interpreted similar to reflectance in the IR spectral region, with the occurrence of Fresnel peaks, Christiansen-like features, and transition minima. We have compiled refractive indices and reflectance measurements of carbon samples from the literature in addition to collecting high S/N diffuse reflectance spectra from 20 0–50 0 nm of a wide variety of carbon powders. We find that the UV spectral range of carbon materials contains information regarding the structure and bonding (sp3/sp2), crystallinity, and grain size of powders, and have applied this to reflectance spectra from several airless bodies: •

Tetrahedral sites have sp3 bonds that cause σ –σ ∗ transitions primarily near 100 nm in diamond, and slightly longer wavelengths for amorphous tetrahedral sp3 carbon. Amorphous

•

•

tetrahedral carbon can have multiple other σ –σ ∗ peaks between 100 and 210 nm. All of these features can cause Fresnel peaks in reflectance. Insoluble organic matter of the Murchison CM2 chondrite, kerogen, ice Tholins, and Triton Tholins exhibit UV reflectance spectra that appear similar to amorphous tetrahedral carbon, with multiple σ –σ ∗ peaks. Trigonal sites have sp2 bonds that cause σ –σ ∗ and π –π ∗ transitions primarily near ∼85 nm and 250 nm, respectively. Graphite, which is pure sp2 , exhibits these features near 85 nm and 241 nm, and the Fresnel peaks are strong and narrow. With amorphization, these features broaden, weaken, and move to longer wavelengths due to a shortening of the bonding distances and an increase in the overall number of bonding distances. Minima in reflectance that are Christiansen-like occur below 50 nm and in the 160–200 nm range. For tetrahedral sp3 carbon, the feature in the 160–200 nm range is entirely a transition minimum as the spectra move from the “weak surface scattering” to the “strong surface scattering” domains of reflectance with decreasing wavelength. For trigonal sp2 carbon, the 160– 200 nm feature is a combination of a transition minimum from the σ –σ ∗ feature and the Christiansen-like feature of the π –π ∗ transition. In our measurements, LCA101 lampblack with grain sizes <0.021 μm exhibits the longest wavelength π –π ∗ peak in reflectance near 272 nm, but the actual Fresnel peak extends well into the NIR spectrum. Larger macroscopic grained powders of amorphous carbons appear to show the peak in the 250– 260 nm range. Gas phase alkanes appear to show the σ –σ ∗ transition in photoabsorption spectra at ∼92 nm for methane, with peak wavelength decreasing with increasing carbon number. This could be important for future observations in the outer solar system.

D.M. Applin et al. / Icarus 307 (2018) 40–82 •

•

•

•

•

•

•

•

•

•

Coals, with the exception of anthracite, exhibit only the σ –σ ∗ peak and not the π–π∗ peak above background. With decreasing rank, the σ –σ ∗ peak weakens, broadens, and moves to slightly longer wavelengths. Despite not exhibiting a π –π ∗ peak above background, coals do generally decrease with decreasing wavelength from the NUV to the FUV, due to a decrease in n. They exhibit a minimum in reflectance that is the transition minimum before the σ –σ ∗ peak begins to rise in reflectance with decreasing wavelength. Diamond exhibits an absorption edge with decreasing wavelength beginning ∼230 nm, following by a transition minimum near 170 nm before the σ –σ ∗ peak begins to rise in reflectance with decreasing wavelength. Diamond is not opaque until wavelengths shorter than 170 nm. The grain size of carbon powders has a significant effect on overall reflectance and on Fresnel peak positions. The effect is particularly strong for graphite because it is structurally in the form of sheets, so specular reflection to the observer increases with increasing grain size even in diffuse geometry as long as the powder does not exhibit preferred orientation of the particles to mimic a mirror or pellet. As the mean particle size of graphite extends to below roughly 10 μm, the π –π ∗ Fresnel peak moves to longer wavelengths in GRP106, ∼241 nm for all >20 μm powders measured to 248 for the <20 μm powder. We cannot rule out effects from soluble impurities, because the <20 μm fractions contained all of the water and ethanol used to wet-sieve the other fractions during evaporation. The contrast of the π –π ∗ Fresnel peak decreases measurably with decreasing grain size, decreasing strongly only for the <20 μm powder. The overall reflectance of shungite, SHU102, did not change strongly with grain size, presumably due to the particles not forming sheets like those of graphite. The spectral contrast of the π –π ∗ Fresnel peak decreases measurably with decreasing grain size, but much less than that of graphite. As with graphite, the peak position only moved to longer wavelengths with the smallest grain size fraction 251 nm to 265 nm for the >25 μm and <25 μm fractions, respectively. The effects of grain size on the σ –σ ∗ peak are not clear and need to be studied. Our measurements do not extend to these wavelengths, and grain size series of powders at these wavelengths have not been published. Progressive pulsed laser weathering of graphite powder causes amorphization, which is observable by a decrease in contrast, an increase in peak width, and an increase in wavelength position of the π –π ∗ Fresnel peak. These effects progress with the amount of total irradiation energy, which may indicate that graphite on the surface of airless bodies should weather into amorphous carbon, and that amorphous carbon peak position could be developed as a technique to study graphite exposure to space weathering. Progressive thermal alteration of coals and the Murchison CM2 chondrite show similar results. With increasing heating time of coals and increasing temperature of Murchison, both show a progressive increase in overall reflectance and an increase in reflectance with decreasing wavelength, followed by a reversal of the spectral slope at the highest temperatures and heating times. Both are consistent with progressive loss of contribution of carbon to the spectral properties. The 201 nm/250 nm ratio of coals correlates with both the O/C and H/C ratio. This is due to a loss in absorption towards the FUV Christiansen-like feature with decreasing overall carbon content in coals. This is also shown throughout the reflectance literature of coal vitrains. Reflectance spectra of carbonaceous chondrites that have been weathered, dry-heated, or saw-cut appear to show evidence of

•

•

•

•

•

•

•

•

•

•

79

hematite in the UV range, with reflectance increasing with decreasing wavelength towards the hematite Fe-O Fresnel peak near 200 nm Carbonaceous chondrite slabs may show evidence of trigonal and tetrahedral carbon Fresnel peaks due to the very large grain size, but further investigation is needed. CV and CO chondrites can be distinguished from CI and CM chondrites in the NUV due to higher albedo, but spectral slopes do not appear very different. CI and CM chondrites show significant evidence of tetrahedral amorphous carbon peaks, or kerogen-like, in the FUV. When divided by lunar spectra samples, all carbonaceous chondrites appear to show some evidence for this. CV and CO chondrites appear to show a significant peak near 160 nm, which is probably due to olivine. The peak is subdued in CI and CM chondrites, likely due to masking by the carbon. The Goalpara ureilite shows strong evidence for graphite, with a strong increase in reflectance with decreasing wavelength to below 95 nm in wavelength. There may also be evidence for the diamond σ –σ ∗ peak. Kerogen like materials we measured show evidence of tetrahedral amorphous, or kerogen-like peaks near 207, 243, and 276 nm. There may also be evidence of PAHs, but further work is needed. Other concentrated carbon materials, like gilsonite and albertite, appear to show similar spectra peaks near 200 nm. The lunar surface is an excellent reflectance standard for the detection of carbon on other airless bodies due to the extremely low carbon concentration found in samples returned from the Apollo missions. We find that the graphite π –π ∗ peak is easily detectable at a concentration of ∼2.5 wt.% in a lunar mare simulant powder. Graphite is likely detectable at much lower concentrations given the relative strength above background at ∼2.5 wt.%. In probably all carbon powders, the σ –σ ∗ peak will be stronger, and therefore likely detectable at even lower concentrations. We find evidence of this in Goalpara ureilite powder. A comparison of the Mercury spectral data to those presented here and in the literature indicates that the ratioed spectra look significantly similar to the UV–Vis-NIR spectrum of our lampblack LCA101 sample, with a negative spectral slope from 1400 to ∼320 nm, with an abrupt downturn in reflectance towards the FUV. If these data are indicative of carbon, the source can still not be easily discerned, since igneous graphite would presumably space-weather into an amorphous state. Previous UV observations of Phobos and Deimos show a distinct peak in reflectance situated near ∼280 nm in wavelength. By comparison with our laboratory data and those from the literature, this likely indicates that the surface of these satellites contains a significant concentration of submicroscopic amorphous carbon, which is also consistent with their extremely low albedos. An extremely high concentration of carbon has previously been suggested to explain the existing UV spectra from the surface of Ceres. We find that current reflectance databases cannot sufficiently ascribe the 160 nm reflectance peak observed in reflectance from Ceres to spectra of carbon powders, particularly because data from the Framing Camera and VIR spectrometer on board Dawn appear to show that the visible and near-IR reflectance cannot be explained if the Cerean surface contains 85% carbon. We find that the peak near 160 nm can be explained by olivine, and potentially carboxylic acids, but much laboratory work is required in order to accurately assign this feature. In particular, the spectra from the surface of Ceres resemble spectra of CV3 and CO3 chondrite powders measured by Wagner et al. (1987) with the addition to increases darkening.

80 •

•

•

•

D.M. Applin et al. / Icarus 307 (2018) 40–82

Reflectance spectra from the surface of C-group asteroid 704 Interamnia show a distinct peak in reflectance near 255 nm before and after ratioing against spectra from the lunar surface. The peak in reflectance may be indicative of carbon, possibly nanophase graphite or amorphous carbon. Visible and near-infrared spectra of 704 Interamnia have previous been described as similar to carbonaceous chondrites with a mixture of serpentine and Fe3+ -oxides. The UV spectra appear anomalous compare to the rest of the survey by Roettger and Buratti (1994), and therefore may be indicative of error or an emission peak. There is some indication that existing reflectance spectra from the surface of the largest Jupiter trojan, 624 Hektor, exhibit a negative spectral slope with decreasing wavelength into the UV from the visible spectrum. This negative slope, however, does plot within error bars. If real, it may be indicative of an increase in reflectance towards the sp2 π –π ∗ transition. Asteroid 624 Hektor is among the “reddest” bodies observed, and has reflectance spectra similar to D-type asteroids and cometary nuclei, which are expected to be carbon-rich based on this spectral slope. Measurements using the UV range from the nucleus of comet 67P-Churyomov–Gerasimenko appear to show a significant σ –σ ∗ peak near ∼90 nm and minima in reflectance near 170 nm indicative of a transition minimum or Christiansen like feature. The S/N of and artifacts in the measurements make the phase identification difficult, but the carbon appears to be significantly non-graphitic, similar to high sp3 /sp2 carbon, or similar to that found in coals measured by Papoular et al. (1995). Ultraviolet reflectance spectra from the surface of the leading hemisphere of Iapetus appear to be consistent with hematite, which has already been suggested to be present in significant quantities from multiple other studies. Evidence comes from the low visible albedo, the decreasing in reflectance with decreasing wavelength from the visible into the UV, the negative spectral slope from 250 nm to ∼200 nm with a peak in reflectance near 200 nm, followed by a decrease in reflectance to ∼150 nm, and another increase in reflectance to shorter wavelengths. These spectral characteristics appear very much like that of previous measurements of a hematite powder. As such, the UV spectral region may be used to differentiate between darkening by oxides, carbon, and metals in remote sensing.

To fully exploit the promise of this wavelength region for remote geochemical analysis of carbon, advances in a number of areas are required: •

•

•

•

•

Continued acquisition of observational UV spectra of sufficient spectral resolution, wavelength coverage (∼70–350 nm), and S/N of airless bodies that would allow the presence of characteristic carbon-related spectral features to be studied. Additional laboratory studies of carbonaceous powders that extend to wavelengths of ∼70 nm, so that the σ –σ ∗ features can be measured and characterized. Extension of the sample suite measured here and available in the literature to a larger variety of carbonaceous powders with a high range of sp3 /sp2 , crystallinity, structure, and grain sizes extending from the nanophase to slabs/pressed pellets. Reflectance detection limit studies and more space weathering experiments on relevant natural and synthetic materials need to be performed. Laboratory reflectance measurements at temperatures and pressures that are relevant to airless bodies in the solar system to ascertain the extent to which non-ambient conditions affect spectral reflectance in the UV.

Acknowledgements The University of Winnipeg’s HOSERLab was established with funding from the Canada Foundation for Innovation, the Manitoba Research Innovations Fund and the Canadian Space Agency, whose support is gratefully acknowledged. This study was supported by research grants from NSERC, the Canadian Space Agency and the University of Winnipeg. We thank Mr. Old Gregg, with whom there was significant discussions. We wish to thank Dr. Kim Tait of the Royal Ontario Museum for providing samples of Murchison and NWA 801 meteorites, and Dr. Andreas Nathues of the Max Planck Institute in Göttingen, Germany for providing samples of the Jbilet CM2 chondrite. We thank two anonymous referees for constructive reviews. Supplementary materials Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.icarus.2018.02.012. References Arakawa, E.T., Williams, M.W., 1968. Optical properties of aluminum oxide in the vacuum ultraviolet. J. Phys. Chem. Solids 29 (5), 735–744. Barnes, E.E., Simpson, W.T., 1963. Correlations among electronic transitions for carbonyl and for carboxyl in the vacuum ultraviolet. The Journal of Chemical Physics 39 (3), 670–675. Benner, S.A., et al., 20 0 0. The missing organic molecules on Mars. Proc. Natl. Acad. Sci. 97 (6), 2425–2430. Bertaux, J.L., et al., 2011. Combined albedo spectrum of Phobos in UV and visible with SPICAM and OMEGA on Mars express. In: Proceedings of the EPSC-DPS Joint Meeting. Bertaux, J.-L., Montmessin, F., Gondet, B., Bibring, J.-P., Rebérac, A., Baggio, L., 2016. UV–visible reflectance of Phobos from SPICAM and OMEGA and comparison with Deimos. Lunar Planet. Sci. Conf., 47 abstract #2177. Brearley, A.J., Jones, R.H., 1998. Chondritic meteorites. Rev. Mineral. Geochem. 36 (1) 3-1. Busarev, V.V., et al., 2015. Spectrophotometry of (32) Pomona, (145) Adeona, (704) Interamnia, (779) Nina, (330825) 2008 XE3, and 2012 QG42 and laboratory study of possible analog samples. Icarus 262, 44–57. Carlon, H.R., 1979. Christiansen effect in IR spectra of soil-derived atmospheric dusts. Appl. Opt. 18 (21), 3610–3614. Chandrasekharan, V., Damany, H., 1968. Birefringence of sapphire, magnesium fluoride, and quartz in the vacuum ultraviolet, and retardation plates. Appl. Opt. 7 (5), 939–941. Chang, H., Charalampopoulos, T.T., 1990. Determination of the wavelength dependence of refractive indices of flame soot. Proc. R. Soc. Lond. A. 430 (1880) The Royal Society. Christiansen, C., 1884. Untersuchungenüber die optischen Eigenschaften von fein vertheilten Korpern. Ann. Phys. Chem. 23, 298. Clark, RogerN., et al., 2012. The surface composition of Iapetus: mapping results from Cassini VIMS. Icarus 218 (2), 831–860. Clarke, R.H., 1968. A theory for the Christiansen filter. Appl. Opt. 7 (5), 861–868. Cloupeau, M., Klarsfeld, S., 1973. Visualization of thermal fields in saturated porous media by the Christiansen effect. Appl. Opt. 12 (2), 198–204. Cloutis, E.A., Mann, P., Izawa, M.R.M., Applin, D.M., Samson, C., Kruzelecky, R., Glotch, T.D., Mertzman, S.A., Mertzman, K.R., Haltigin, T.W., Fry, C., 2015. The Canadian Space Agency planetary analogue materials suite. Planet. Space Sci. 119, 155–172. Cloutis, E.A., McCormack, K.A., Bell III, J.F., Hendrix, A.R., Bailey, D.T., Craig, M.A., Mertzman, S.A., Robinson, M.S., Riner, M.A., 2008. Ultraviolet spectral reflectance properties of common planetary minerals. Icarus 197, 321–347. Cloutis, E.A., et al., 2011a. Spectral reflectance properties of carbonaceous chondrites: 1. CI chondrites. Icarus 212 (1), 180–209. Cloutis, E.A., et al., 2011b. Spectral reflectance properties of carbonaceous chondrites: 2. CM chondrites. Icarus 216 (1), 309–346. Cloutis, E.A., et al., 2010. Spectral reflectance properties of ureilites. Meteoritics & Planetary Science 45 (10–11), 1668–1694. Cloutis, E.A., et al., 2012. Spectral reflectance properties of carbonaceous chondrites: 8. “Other” carbonaceous chondrites: CH, ungrouped, polymict, xenolithic inclusions, and R chondrites. Icarus 221 (2), 984–1001. Cruikshank, D.P., et al., 2014. Aromatic and aliphatic organic materials on Iapetus: analysis of Cassini VIMS data. Icarus 233, 306–315 Cruikshank et al., 2005. Cruikshank, D.P., et al., 2008. Hydrocarbons on Saturn’s satellites Iapetus and Phoebe. Icarus 193 (2), 334–343. Cruikshank, DaleP., et al., 2001. Constraints on the composition of Trojan asteroid 624 Hektor. Icarus 153 (2), 348–360. DeMeo, F.E., Carry, Benoit, 2013. The taxonomic distribution of asteroids from multi-filter all-sky photometric surveys. Icarus 226 (1), 723–741.

D.M. Applin et al. / Icarus 307 (2018) 40–82 De Sanctis, M.C., et al., 2017. Ac-H-11 Sintana and Ac-H-12 Toharu quadrangles: assessing the large and small scale heterogeneities of Ceres’ surface. Icarus. ´ A.B., Li, E.H., 1999. Optical properties of graphite. J. Appl. Phys. 85, Djurišic, 7404–7410. Duley, W.W., 1984. Refractive indices for amorphous carbon. Astrophys. J. 287, 694–696. Duley, W.W., Lazarev, S., 2004. Ultraviolet absorption in amorphous carbons: polycyclic aromatic hydrocarbons and the 2175 A˚ extinction feature. Astrophys. J. Lett. 612 (1), L33. Ergun, S., 1968. Structure of carbon. Carbon 6 (2), 141–159. Ergun, S., et al., 1967. Transverse and longitudinal optical properties of graphite. Carbon 5, 403–408. Ergun, S., 1976. Analysis of coherence, strain, thermal vibration and preferred orientation in carbons by X-ray diffraction. Carbon 14 (3), 139–150. Ferrari, A.C., Robertson, J.F., 20 0 0. Interpretation of Raman spectra of disordered and amorphous carbon. Physical review B 61 (20), 14095. Fitzsimmons, A., et al., 2018. Spectroscopy and thermal modelling of the first interstellar object 1I/2017 U1 ‘Oumuamua. Nature Astronomy 2 (2), 133. Flynn, G.J., et al., 2003. The origin of organic matter in the solar system: evidence from the interplanetary dust particles. Geochim. Cosmochim. Acta 67 (24), 4791–4806. Foster, P.J., Howarth, C.R., 1968. Optical constants of carbons and coals in the infrared. Carbon 6 (5) 719IN23IN24725-724729. Fraeman, A.A., et al., 2012. Analysis of disk-resolved OMEGA and CRISM spectral observations of Phobos and Deimos. J. Geophys. Res.: Planets 117, E11. Freissinet, C., et al., 2015. Organic molecules in the sheepbed mudstone, gale crater, mars. J. Geophys. Res.: Planets 120 (3), 495–514. French, R.H., Müllejans, H., Jones, D.J., 1998. Optical properties of aluminum oxide: determined from vacuum ultraviolet and electron energy-loss spectroscopies. J. Am. Ceram. Soc. 81 (10), 2549–2557. Friedel, R.A., Breger, I.A., 1959. Free-radical concentrations and other properties of pile-irradiated coals. Science 130 (3391), 1762–1763. Fries, M., et al., 2017. Dust infall onto Phobos and Deimos can explain their carbonaceous reflectance signature, perhaps overlying a Mars-impact-origin core: a hypothesis. LPSC, 2017 abstract # 2570. Gilbert, L.A., 1960. The reflectivity spectra of coal vitrains in the visible and the ultra-violet. Fuel 39, 393–400. Gillis-Davis, J.J., et al., 2015. Laser space weathering of Allende (Cv2) and Murchison (Cm2) carbonaceous chondrites. LPSC, 2015 abstract # 1607. Greenaway, D.L., et al., 1969. Anisotropy of the optical constants and the band structure of graphite. Phys. Rev. 178 (3), 1340. Haensel, R., et al., 1969. Optical absorption of solid krypton and xenon in the far ultraviolet. Phys. Rev. 188 (3), 1375. Hagemann, H-J., Gudat, W., Kunz, C., 1975. Optical constants from the far infrared to the x-ray region: Mg, Al, Cu, Ag, Au, Bi, C, and Al2 O3 . JOSA 65 (6), 742–744. Hapke, B., 2012. Theory of Reflectance and Emittance Spectroscopy. Cambridge University Press, Cambridge, UK. Hartman, P.L., Nelson, J.R., 1957a. Hydrogen lamp of good intensity and reliability for the vacuum ultraviolet. JOSA 47 (7), 646–648. Hartman, P.L., Nelson, J.R., Siegfried, J.G., 1957b. Reflection spectrum and structure in the exciton absorption peak of NaCl and KCl. Phys. Rev. 105 (1), 123. Heavens, O.S., 1992. Handbook of optical constants of solids II. J. Mod. Opt. 39 (1) 189-189. Hendrix, A.R., Hansen, C.J., 2008. The albedo dichotomy of Iapetus measured at UV wavelengths. Icarus 193 (2), 344–351. Hendrix, A.R., Vilas, F., Li., J.-Y., 2016a. Ceres: sulfur deposits and graphitized carbon. Geophys. Res. Lett. 43 (17), 8920–8927. Hendrix, A.R., Vilas, F., Li, J.-Y., 2016b. The UV signature of carbon in the solar system. Meteorit. Planet. Sci. 51, 105–115. Hennig, G.R., 1965. Organic constituents of meteorites – a review. Geochim. Cosmochim. Acta 31, 1395–1440 Holland et al., 1972. Henning, Th., et al., 1999. WWW database of optical constants for astronomy. Astron. Astrophys. Suppl. Ser. 136 (2), 405–406. Hiroi, Takahiro, et al., 1993. Evidence of thermal metamorphism on the C, G, B, and F asteroids. Science 261 (5124), 1016–1018. Holland, P.T., et al., 1972. Study of carbon compounds in Apollo 12 and 14 lunar samples. Space Life Sci. 3 (4), 551–561. Hosterman, J.W., Meyer, R.F., Palmer, C.A., Doughten, M.W., Anders, D.E., 1990. Chemistry and Mineralogy of Natural Bitumens and Heavy Oils and Their Reservoir Rocks from the United States, Canada, Trinidad and Tobago, and Venezuela. U.S. Geological Survey Circular 1047, 19. Huffman, D.R., 1975. Optical properties of particulates. Astrophys. Space Sci. 34 (1), 175–184. Humphreys-Owen, S.P.F., Gilbert, L.A., 1958. Industrial Carbon and Graphite. Society of Chemical Industry, London, p. 37. Jarvis, K.S., et al., 20 0 0. Are Hyperion and Phoebe linked to Iapetus? Icarus 146 (1), 125–132. Kalkreuth, Wolfgang, et al., 1991. Petrographic and chemical characterization of Canadian and German coals in relation to utilization potential: 1. Petrographic and chemical characterization of feedcoals. Fuel 70 (6), 683–694. Kameta, K., et al., 2002. Photoabsorption, photoionization, and neutral-dissociation cross sections of simple hydrocarbons in the vacuum ultraviolet range. J. Electron Spectrosc. Relat. Phenom. 123 (2-3), 225–238. Kaneko, Y., Morimoto, K., Koda, T., 1983. Optical properties of alkaline-Earth Chalcogenides. II. Vacuum ultraviolet reflection spectra in the synchrotron radiation region of 4–40 eV. J. Phys. Soc. Jpn. 52 (12), 4385–4396.

81

Khare, B.N., Thompson, W.R., Sagan, C., Arakawa, E.T., Meisse, C., Gilmour, I., 1990. Optical constants of kerogen from 0.15 to 40 micrometers: comparison with meteoritic organics. In: Proc. of the First International Conference on Laboratory Research for Planetary Atmospheres, CP-3077. NASA, pp. 340–356. Khare, B.N., Thompson, W.R., Cheng, L., Chyba, C., Sagan, C., Arakawa, E.T., Meisse, C., Tuminello, P.S., 1993. Production and optical constants of ice tholin from charged particle irradiation of (1:6) C2 H6 /H2 O at 77 K. Icarus 103, 290–300. Khare, B.N., et al., 1984. Optical constants of organic tholins produced in a simulated titanian atmosphere: from soft X-ray to microwave frequencies. Icarus 60 (1), 127–137. Khare, B.N., et al., 1994. Optical properties of poly-HCN and their astronomical applications. Can. J. Chem. 72 (3), 678–694. Koch, E.-E., Skibowski, D.M., 1971. Optical absorption of gaseous methane, ethane, propane and butane and reflection of solid methane and ethane in the vacuum ultraviolet. Chem. Phys. Lett. 9 (5), 429–432. Krasnopolsky, V.A., Maillard, J.P., Owen, T.C., 2004. Detection of methane in the martian atmosphere: evidence for life? Icarus 172 (2), 537–547. Larruquert, J.I., et al., 2013. High reflectance ta-C coatings in the extreme ultraviolet. Opt. Exp. 21 (23), 27537–27549. Lawrence, SamuelJ., Lucey, PaulG., 2007. Radiative transfer mixing models of meteoritic assemblages. J. Geophys. Res.: Planets 112, E7. Li, J.-Y., et al., 2009. LPSC, XXXX #2101. Limao-Vieira, P., et al., 2006. Acetic acid electronic state spectroscopy by high-resolution vacuum ultraviolet photo-absorption, electron impact, He (I) photoelectron spectroscopy and ab initio calculations. Chem. Phys. 324 (2-3), 339–349. Liu, Y., et al., 2016. The SWRI ultraviolet reflectance chamber (SWURC): progress toward a far ultraviolet surface reflectance library. LPSC, 2016 abstract # 2496. Loh, Eugene, 1964. Ultraviolet reflectance of Al2O3, SiO2 and BeO. Solid State Communications 2 (9), 269–272. Lochmüller, C.H., Röhl, R., Marshall, D.B., 1981. Use of carbon black as a reference material in photoacoustic Spectroscopy. Anal. Lett. Part A: Chem. Anal. 14 (1), 41–46. Lyon, R.J.P., 1965. Analysis of rocks by spectral infrared emission (8 to 25 micron). Econ. Geol. 60, 715–736. Malloci, G., Mulas, G., Joblin, C., 2004. Electronic absorption spectra of PAHs up to vacuum UV—towards a detailed model of interstellar PAH photophysics. Astron. Astrophys. 426 (1), 105–117. Mann, P., et al., 2016. The stability of Spectralon, a potential calibration reference for Mars2020. LPSC abstract# 2362. McClintock, W.E., et al., 2008. Spectroscopic observations of Mercury’s surface reflectance during MESSENGER’s first Mercury flyby. Science 321 (5885), 62–65. McSween, H.Y., et al., 2017. Carbonaceous chondrites as analogs for the composition and alteration of Ceres. Meteorit. Planet. Sci.. McCartney, J.T., Ergun, S., 1967. Optical properties of coals and graphite. Bulletin 641. U. S. Department of the Interior, Bureau of Mines, Washington, D.C. Meech, K.J., et al., 2017. A brief visit from a red and extremely elongated interstellar asteroid. Nature 552 (7685), 378. Minkoff, G.J., Gaydon, A.G., 1946. A Christiansen filter for the ultraviolet. Nature 158 (4022) 788-788. Mittlefehldt, D.W., Lindstrom, M.M., 1998. Petrology and geochemistry of lodranite GRA 95209. Meteorit. Planet. Sci. Suppl. 33, A111. Moore, C.B., Lewis, C.F., Gibson Jr, E.K., 1973. Total carbon contents of Apollo 15 and 16 lunar samples. In: Lunar and Planetary Science Conference Proceedings, 4. Moroz, L.V., et al., 1998. Natural solid bitumens as possible analogs for cometary and asteroid organics: 1. reflectance spectroscopy of pure bitumens. Icarus 134 (2), 253–268. Nagakura, S., Kaya, K., Tsubomura, H., 1964. Vacuum ultraviolet absorption spectra and electronic structures of formic acid, acetic acid and ethyl acetate. Journal of Molecular Spectroscopy 13 (1-4), 1–8. Noll, K.S., et al., 1997. Detection of ozone on Saturn’s satellites Rhea and Dione. Nature 388 (6637), 45–47. Pajola, M., et al., 2013. Phobos as a d-type captured asteroid, spectral modeling from 0.25 to 4.0 μm. Astrophys. J. 777 (2), 127. Papoular, R., Breton, J., Gensterblum, G., Nenner, I., Papoular, R.J., Pireaux, J.-J., 1993. The vis/UV spectrum of coals and the interstellar extinction curve. Astron. Astrophys 270, L5–L8. Papoular, R., Guillois, O., Nenner, I., Perrin, J.-MK., Reynaud, C., Sivan, J.-P., 1995. On the nature of interstellar carbonaceous dust. Planet. Space Sci. 43, 1287–1291. Papoular, R., et al., 1996. A comparison of solid-state carbonaceous models of cosmic dust. Astron. Astrophys. 315, 222–236. Parker, J.W., et al., 2002. Analysis of the first disk-resolved images of Ceres from ultraviolet observations with the Hubble Space Telescope. Astron. J. 123 (1), 549. Perna, D., et al., 2017. Rotationally resolved spectroscopy of Jupiter Trojans (624) Hektor and (911) Agamemnon. Monthly Notices of the Royal Astronomical Society 475 (1), 974–980. Peplowski, P.N., et al., 2011. Radioactive elements on Mercury’s surface from MESSENGER: Implications for the planet’s formation and evolution. science 333 (6051), 1850–1852. Peplowski, P.N., et al., 2016. Remote sensing evidence for an ancient carbon-bearing crust on Mercury. Nature Geoscience 9 (4), 273. Philipp, H.R., Taft, E.A., 1962. Optical properties of diamond in the vacuum ultraviolet. Phys. Rev. 127 (1), 159. Phillip, H.R., Taft, E.A., 1964. Kramers-Kronig analysis of reflectance data for diamond. Physical Review 136 (5A), A1445.

82

D.M. Applin et al. / Icarus 307 (2018) 40–82

Pradeep, T., Rao, C.N.R., 1989. Electronic structures of electron donor-acceptor complexes: results from ultraviolet photoelectron spectroscopy and molecular orbital calculations. J. Mol. Struct.: THEOCHEM 200, 339–352. Prost, R., 1973. The influence of the Christiansen effect on IR spectra of powders. Clays Clay Miner. 21 (5), 363–368. Quirico, E., et al., 2016. Refractory and semi-volatile organics at the surface of comet 67P/Churyumov-Gerasimenko: insights from the VIRTIS/Rosetta imaging spectrometer. Icarus 272, 32–47. Reddy, V., et al., 2012. Delivery of dark material to Vesta via carbonaceous chondritic impacts. Icarus 221 (2), 544–559. Rietmeijer, F.J.M., 1998. Interplanetary dust particles. Planet. Mater. 36 2-1. Rivkin, A.S., Asphaug, E., Bottke, W.F., 2014. The case of the missing Ceres family. Icarus 243, 429–439. Robertson, J., 1986. Amorphous carbon. Advances in Physics 35 (4), 317–374. Robinson, M.S., Lucey, P.G., 1997. Recalibrated Mariner 10 color mosaics: implications for mercurian volcanism. Science 275 (5297), 197–200. Roettger, E.E., Buratti, B.J., 1994. Ultraviolet spectra and geometric albedos of 45 asteroids. Icarus 112 (2), 496–512. Rouleau, F., Martin, P.G., 1991. Shape and clustering effects on the optical properties of amorphous carbon. The Astrophys. J. 377, 526–540. Rousseau, Batiste, et al., 2017. Laboratory simulations of the Vis-NIR spectra of comet 67P using sub-μm sized cosmochemical analogues. Icarus. Rubloff, G.W., 1972. Far-ultraviolet reflectance spectra and the electronic structure of ionic crystals. Phys. Rev. B 5 (2), 662. Salisbury, J.W., Wald, A., 1992. The role of volume scattering in reducing spectral contrast of reststrahlen bands in spectra of powdered minerals. Icarus 96 (1), 121–128. Salisbury, J.W., Walter, L.S., 1989. Thermal infrared (2.5–13.5 μm) spectroscopic remote sensing of igneous rock types on particulate planetary surfaces. J. Geophys. Res.: Solid Earth 94 (B7), 9192–9202. Sandford, ScottA., et al., 2006. Organics captured from comet 81P/Wild 2 by the Stardust spacecraft. Science 314 (5806), 1720–1724. Schäfer, T., et al., 2016. The appearance of carbonaceous chondrites on (1) Ceres from observations by the Dawn Framing Camera. EGU General Assembly Conference Abstracts, 18. Sigel, G.H., 1971. Vacuum ultraviolet absorption in alkali doped fused silica and silicate glasses. J. Phys. Chem. Solids 32 (10), 2373–2383. Sinsheimer, R.L., Loofbourow, J.R., 1947. Christiansen filters for the ultra-violet. Nature 160 (4072) 674-674. Steele, A., et al., 2010. Graphite in an Apollo 17 impact melt breccia. Science 329 (5987) 51-51. Steele, A., et al., 2012. A reduced organic carbon component in martian basalts. Science 337 (6091), 212–215. Steglich, M., et al., 2010. Electronic spectroscopy of medium-sized polycyclic aromatic hydrocarbons: implications for the carriers of the 2175 A˚ UV bump. Astrophys. J. Lett. 712 (1), L16. Stern, S.A., et al., 2015. First extreme and far ultraviolet spectrum of a Comet Nucleus: results from 67P/Churyumov-Gerasimenko. Icarus 256, 117–119.

Stiegman, A.E., Bruegge, C.J., Springsteen, A.W., 1993. Ultraviolet stability and contamination analysis of Spectralon diffuse reflectance material. Optical Engineering 32 (4), 799–805. Syal, M.B., Schultz, P.H., Riner, M.A., 2015. Darkening of Mercury’s surface by cometary carbon. Nat. Geosci. 8 (5), 352–356. Taft, E.A., Philipp, H.R., 1965. Optical properties of graphite. Phys. Rev. 138 (1A), A197. Takir, D., Emery, J.P., 2012. Outer main belt asteroids: identification and distribution of four 3-μm spectral groups. Icarus 219 (2), 641–654. Tan, G.L., et al., 2005. Optical properties and London dispersion interaction of amorphous and crystalline SiO2 determined by vacuum ultraviolet spectroscopy and spectroscopic ellipsometry. Phys. Rev. B 72 (20), 205117. Thangjam, G., et al. Spectral properties and geology of bright and dark materials on dwarf planet Ceres. (2018). arXiv:1712.05203. Thomas-Keprta, K.L., et al., 2014. Organic matter on the Earth’s Moon. Geochim. Cosmochim. Acta 134, 1–15. Tuminello, P.S., Clark, M., 1994b. Optical constants of Triton tholin: preliminary results. Bull. Am. Astron. Soc. 26, 1176–1177. Trang, D., Lucey, P.G., Izenberg, N.R., 2017. Radiative transfer modeling of MESSENGER VIRS spectra: detection and mapping of submicroscopic iron and carbon. Icarus 293, 206–217. Vilas, Faith, et al., 1993. CCD reflectance spectra of selected asteroids. II. Low-albedo asteroid spectra and data extraction techniques. Icarus 105 (1), 67–78. Vorres, K.S., 1993. Users handbook for the argonne premium coal sample program. Argonne National Lab., IL (United States) No. ANL/PCSP–93/1. Vilas, Faith, et al., 1996. Unraveling the zebra: clues to the Iapetus dark material composition. Icarus 124 (1), 262–267. Wagner, J.K., Hapke, B.W., Wells, E.N., 1987. Atlas of reflectance spectra of terrestrial, lunar, and meteoritic powders and frosts from 92 to 1800 nm. Icarus 69, 14–28. Weider, S.Z., et al., 2015. Evidence for geochemical terranes on Mercury: Global mapping of major elements with MESSENGER’s X-Ray Spectrometer. Earth and Planetary Science Letters 416, 109–120. Williams, M.W., Arakawa, E.T., 1972. Optical properties of glassy carbon from 0 to 82 eV. J. Appl. Phys. 43 (8), 3460–3463. Wojak, U., Czarnetzki, U., Döbele, H.F., 1987. Christiansen filters for the far ultraviolet: an old spectral device in a new light. Appl. Opt. 26 (22), 4788–4790. Yasinsky, J.B., Ergun, S., 1965. Transmittance of single crystals of graphite in the infrared spectrum. Carbon 2 (4), 355–358. Ye, Q.-Z., et al., 2017. 1I/2017 U1 (‘Oumuamua) is Hot: Imaging, Spectroscopy, and Search of Meteor Activity. The Astrophysical Journal Letters 851 (1), L5. Yokoyama, H., et al., 1990. Optical and mechanical properties of hard hydrogenated amorphous carbon films deposited by plasma CVD. Jpn. J. Appl. Phys. 29 (12R), 2815. Zellner, B., Tholen, D.J., Tedesco, E.F., 1985. The eight-color asteroid survey: results for 589 minor planets. Icarus 61 (3), 355–416. Zubko, V.G., et al., 1996. Optical constants of cosmic carbon analogue grains—I. Simulation of clustering by a modified continuous distribution of ellipsoids. Mon. Not. R. Astron. Soc. 282 (4), 1321–1329.