Icarus 240 (2014) 73–85

Contents lists available at ScienceDirect

Icarus

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

Spectral analysis of the bright materials on the asteroid Vesta

http://dx.doi.org/10.1016/j.icarus.2014.04.0370019-1035/� 2014 Elsevier Inc. All rights reserved.

⇑ Corresponding author. Fax: +39 06 45488188.E-mail address: francesca.zambon@iaps.inaf.it (F. Zambon).

F. Zambon a,⇑, Maria Cristina De Sanctis a, Stefan Schröder b, Federico Tosi a, Andrea Longobardo a,Eleonora Ammannito a, David T. Blewett c, David W. Mittlefehldt d, Jian-Yang Li e, E. Palomba a,Fabrizio Capaccioni a, Alessandro Frigeri a, Maria Teresa Capria a, Sergio Fonte a, Andreas Nathues f,Carle M. Pieters g, Christofer T. Russell h, Carol A. Raymond i

a INAF-IAPS Istituto di Astrofisica e Planetologia Spaziali, Via del Fosso del Cavaliere, 100, 00133 Rome, Italyb Institute of Planetary Research, German Aerospace Center (DLR), Rutherfordstrasse 2, D-12489 Berlin, Germanyc The Johns Hopkins University, Applied Physics Laboratory, Laurel, MD 20723, USAd NASA Johnson Space Center, 2101 NASA Parkway, Houston, TX 77058, USAe Planetary Science Institute, 1700 East Fort Lowell, Tucson, AZ 85719-2395, USAf Max Planck Institute for Solar System Research, Max-Planck-Strasse 2, D-37191 Katlenburg-Lindau, Germanyg Brown University, 324 Brook Street, Providence, RI 02912, USAh Institute of Geophysics and Planetary Physics, University of California at Los Angeles, 3845 Slichter Hall, 603 Charles E. Young Drive, East, Los Angeles, CA 90095-1567, USAi NASA/Jet Propulsion Laboratory and California Institute of Technology, 4800 Oak Grove Drive, Pasadena, CA 91109, USA

a r t i c l e i n f o

Article history:Received 25 September 2013Revised 11 April 2014Accepted 20 April 2014Available online 14 May 2014

Keywords:Asteroid VestaSpectroscopyMineralogy

a b s t r a c t

Vesta spectra have prominent near-infrared absorption bands characteristic of pyroxenes, indicating adirect link to the howardite, eucrite and diogenite meteorites. Many localized dark and bright materialsare present on Vesta’s surface. Here we focus on the bright material (BM) units to determine their spectralproperties, their origin, the presence of mineralogical phases different from pyroxenes, and whether dif-ferent bright units share a common lithology. VIR, the Visible and Infrared spectrometer onboard Dawn,allows us to first do a detailed analysis of the spectral properties of a large number of bright materialunits on Vesta including examples of the different morphological classes. The spectral parameters usedare band centers, band depths, and Band Area Ratio (BAR) for the pyroxene bands at �0.9 and�1.9 lm. The mineralogies of most bright regions are consistent with those of the howardite, eucriteand diogenite meteorites typical of Vesta’s surface. We find that bright material units exhibit the fullrange of HED pyroxene composition, from eucrites to diogenites. Large part of the bright materials areeucrite-rich, according with the Vesta’s mineralogy. In most cases, the bright materials have the samemineralogy of the surrounding terrain, but have larger band depth values. The band depths can be relatedto the abundance of the absorbing minerals, the abundance of Fe2+, grain size, and/or to the abundance ofopaque materials. We found a positive correlation between albedo and band depth, which suggests thatthe grain size is not the main factor responsible for the higher albedo. The analysis of the band parame-ters indicates that most of the bright materials, excluding the few olivine-rich units, represent freshuncontaminated Vestan pyroxenes from a variety of lithologies exposed from beneath the surface byimpacts.

� 2014 Elsevier Inc. All rights reserved.

1. Introduction Vesta has been the subject of numerous studies, observational

Dawn is a NASA Discovery mission devoted to the exploration ofthe two largest main belt asteroids: Vesta and Ceres (Russell et al.,2007, 2012, 2013). Dawn was launched in September 2007 andentered orbit around Vesta in July 2011. The mission at Vestaended in September 2012. After that, Dawn started its journey toCeres (Russell and Raymond, 2011).

and theoretical. Interest in Vesta increased when McCord et al.(1970) proposed a link with certain basaltic achondrites.

Vesta is the second-most massive object in the main asteroidbelt, with an average diameter of 525 km (Thomas et al., 1997;Zuber et al., 2011), and its reflectance spectrum is characterizedby pyroxene absorption bands (McCord et al., 1970; Gaffey,1997; Zuber et al., 2011; Pieters et al., 2011; De Sanctis et al.,2012a, 2013a). Previous ground-based and Hubble Space Telescope(HST) observations allowed for a global color and spectroscopicanalysis (Thomas et al., 1997; Binzel et al., 1997; Zellner et al.,2005; Li et al., 2010). Vesta’s spectra are characterized by two

74 F. Zambon et al. / Icarus 240 (2014) 73–85

strong absorption bands at 0.9 lm (band I) and 1.9 lm (band II)typical of Fe-bearing pyroxenes, and are very similar to reflectancespectra of howardite, eucrite and diogenite (hereafter referred to asHED) meteorites (Drake, 1979; Feierberg and Drake, 1980; DeSanctis et al., 2012a, 2013b). HED meteorites harbor a largevariety of igneous rocks, including basalts, cumulate gabbros,orthopyroxenites and igneous brecciated mixtures (Mittlefehldtet al., 1998). The diogenites are coarse-grained cumulatesoriginated from a plutonic layer deep in the crust (Mittlefehldtet al., 1998; Beck and McSween, 2010; McSween et al., 2011,2013b). The mineralogy of diogenites is dominated by orthopyrox-enes (from �87 to 99%); all diogenites contain <5% chromite andsome contain olivine usually with a content <10% (McSweenet al., 2011). Eucrites occur as basaltic or cumulate rocks. Theyare dominated by pyroxenes and plagioclase, with minor amountsof metal troilite, chromite, ilmenite, and silica (Mayne et al., 2010;McSween et al., 2011). Eucrites are believed to have crystallized aslavas on the surface or within relatively shallow dikes and plutons(McSween et al., 2011). Basaltic eucrites contain Fe-rich pyroxenes.Cumulate eucrites are predominantly unbrecciated and theirmineralogy is similar to that of the basaltic eucrites, but richer inMg (Mittlefehldt et al., 1998; McSween et al., 2011). Impact mixingof eucrite and diogenite has produced the polymict breccias andhowardites. Howardites are a mixture of eucrite and diogeniteclasts (Mittlefehldt et al., 1998; McSween et al., 2011).

Dawn carries a Visible and Infrared imaging spectrometer (VIR),designed to map the mineralogy of Vesta and Ceres by collectinghyperspectral images in the wavelength range �0.25 to 5.0 lm(De Sanctis et al., 2011; Russell and Raymond, 2011). All VIRspectra of Vesta’s surface show both the 0.9- and 1.9-lm absorp-tion bands, confirming the widespread occurrence of iron-bearinglow-calcium pyroxenes (De Sanctis et al., 2012a). These bandsare caused by absorption of photons, primarily by Fe2+. The exactband position and shape are determined by the relative proportionof Fe2+ and Mg2+ in the M1 and M2 sites of the pyroxene crystalstructure (Burns, 1993), as well as by the calcium content.Laboratory studies indicate that the band centers for the band Iand band II pyroxene absorptions are systematically differentbetween diogenites and eucrites, because a larger abundance ofFe2+ in the eucrites is responsible for a shift of the band centerstowards longer wavelengths (Gaffey, 1976; Klima et al., 2007,2011). Plotting band I versus band II centers is a powerful methodfor determining the composition of the pyroxenes (Gaffey, 1997)and also for discriminating among the different lithologies onVesta. VIR data have been used to build detailed mineralogicalmaps at low and high spatial resolutions, allowing local study ofthe mineralogy of its surface (De Sanctis et al., 2012a;Ammannito et al., 2013a; McSween et al., 2013a). VIR spectra indi-cate that all the different HED lithologies are present on Vesta’ssurface: howardite, eucrite and diogenite; their distributions pro-vide the geologic context for these meteorites (De Sanctis et al.,2012a; Ammannito et al., 2013a). The dominant lithology is aeucrite-rich howardite (De Sanctis et al., 2013b). Pyroxene isubiquitous on Vesta at all the investigated spatial scales, and thereis no convincing spectral evidence for lithologies not representedby HED meteorites. However, some spectra, especially those ofthe lowest-reflectance regions, show an OH feature, thought tobe caused by the presence of hydrated carbonaceous chondritematerial (McCord et al., 2012; De Sanctis et al., 2012b; Reddyet al., 2012a). The spectral characteristics of the dark material aredescribed in McCord et al. (2012), Reddy et al. (2012a), andPalomba et al. (2014). Stephan et al. (2014) studied the propertiesof small, fresh craters containing bright and dark materials. Thethermal behavior of both the dark and bright materials on Vestais described in Tosi et al. (2014).

This paper presents the first detailed study of the spectral prop-erties of the bright materials on Vesta. Bright units are character-ized by high albedo, up to 40% larger than the Vesta’s average (Liet al., 2010; Schröder et al., 2013). Bright areas are non-uniformlydistributed and are not spatially correlated with the dark material,suggesting a different origin (McCord et al., 2012), even if some-times bright and dark materials are mixed together (McCordet al., 2012; Palomba et al., 2014). McCord et al. (2012) employeda multiple-endmember linear spectral unmixing model to showthat VIR spectra of Vesta’s surface can be modeled as the linearcombination of two endmembers, one representative of the brightregions and one of the dark regions.

Vesta has a high reflectance compared to other airless bodies ofthe Solar System. The lack of spectral reddening and the presenceof prominent mineralogical absorption bands suggest that nano-phase iron (npFe0) is not being produced in substantial quantitiesby space weathering on Vesta, in contrast to other airless bodies,particularly the Moon (Pieters et al., 2012). The strength of theabsorption bands is correlated with the albedo (Pieters et al.,2012; McCord et al., 2012). Low-reflectance regions have shallowerbands than bright regions, due to the presence of carbon-richmaterial delivered by low-velocity impactors (McCord et al.,2012; Palomba et al., 2014). Mittlefehldt et al. (2012) produced aglobal map of the bright regions on Vesta, with a classificationbased on their morphology, and Schröder et al. (2013) provided aglobal map of the normal visual albedo at latitudes up to 30�, fromwhich we derived the albedo of the bright regions.

The most diagnostic spectral parameter for mineralogical char-acterization of Vesta material are band I and band II centers(Gaffey, 1997; De Sanctis et al., 2012a; McSween et al., 2013b).Band depths give information on the abundance of the absorbingminerals, and the grain size, while the Band Area Ratio (BAR) isuseful for evaluating the presence of olivine in an olivine–pyroxenemixture (Cloutis et al., 1986). Band centers allow for a computationof the abundance of the wollastonite (Wo) and ferrosilite (Fs)pyroxene endmembers. Burbine et al. (2007, 2009) and Gaffeyet al. (2002) derived formulas to retrieve the molar content ofWo and Fs in two different ways by using their band center posi-tions. Here, we apply the same methods to VIR data to retrievethe wollastonite and ferrosilite contents for the BM units.

2. Classification and description of the bright material units

BM units are defined as areas with higher reflectance than theirsurroundings. Detection and morphological classification of BMunits was performed by Mittlefehldt et al. (2012). We studied mostof the BM units mapped by Mittlefehldt et al. (2012). The geo-graphic distribution and classification were improved, and we con-sidered the latest version of the map (Mittlefehldt and Li, personalcommunication). The distribution of BM units is non-uniform.Large areas in the northern hemisphere appear to lack BM com-pared with the southern hemisphere. However, this may be dueto the presence of shadows at northern latitudes during the timeof Dawn’s orbital mission (Mittlefehldt et al., 2012). Four morpho-logical classes of BM have been defined. Type 1 contains two sub-groups: type 1a and type 1b. Type 1a represents Crater WallMaterial (CWM), discontinuous outcrops or large blocks of rock;type 1b, called Slope Material (SM), generally has a wedge/tongueshape and often occurs as debris found downslope from the CWM(Mittlefehldt et al., 2012; Li et al., 2012). Most type 1a and 1bmaterials appear together on the walls of the craters, suggestinga common origin (Li et al., 2012). Type 2 is Radial Material (RM),always associated with the ejecta of large craters. Type 3 is SpotMaterial (SpM), frequently associated with the ejecta of the small

F. Zambon et al. / Icarus 240 (2014) 73–85 75

craters (diameter <100 m). Type 4, diffuse material, is present at sixlocations and is not correlated with impact craters. Here we ana-lyze the spectral properties of types 1 and 2, which are the mostabundant on Vesta and are easily recognized in the VIR data, whiletype 3 is typically too small for analysis at the spatial resolution ofVIR data.

In Fig. 1, some examples of type 1 and 2 BM units are shown. SMis particularly evident in Cornelia (green arrow), Tuccia and Myiacraters are a clear example of RM (blue arrows), while Cornelia,Aelia and Myia provide an example of the coexistence of brightand dark material (Palomba et al., 2014; Stephan et al., 2014).

We carried out a classification based on the normal visualalbedo (Table 1). The normal visual albedo of each candidate BMoccurrence was compared with the Vesta geometric albedo of0.38 (Tedesco et al., 1989; Li et al., 2010; Schröder et al., 2013).We recognize five classes according to albedo: BM units with analbedo that exceeds the geometric albedo Vesta by up to 10%,between 10% and 20%, between 20% and 30%, more than 30%.A fifth class consists of bright regions that have an albedo lowerthan the Vesta average (Table 1). It is important to note that insome cases there is a coexistence of bright and dark material. Thuswe do not define BM units on Vesta on the basis of an absolute cri-terion, but adopting a relative criterion: if a region is brighter thanthe surroundings, then it is tagged as a BM unit. For this reason, wehave some BM units whose albedo is lower than the geometricalbedo of Vesta. Values of normal visual albedo were derived from

Fig. 1. Equirectangular projection of the Framing Camera clear filter images relative toarrows the SM, and the blue arrows the RM. Images credit: NASA/JPL-Caltech/UCLA/MPS/DLreferred to the web version of this article.)

the Vesta albedo map (Schröder et al., 2013) and are available inthe latitude range from �90� to 30�. Fig. 2 shows the locations ofthe BM occurrences studied in this paper (supplementary materialTables 7–9).

We assign names to the BM sites according to the name of thecrater in which they are located. Bright units in unnamed cratersare assigned a ‘‘BU’’ followed by a progressive number.

3. Dataset description

Dawn’s payload is made up of three instruments: the FramingCamera (FC), the Visible and InfraRed Mapping Spectrometer(VIR), and the Gamma Ray and Neutron Detector (GRaND) (Sierkset al., 2011; De Sanctis et al., 2011; Prettyman et al., 2011). Themission at Vesta was divided into four principal phases based onthe altitude of the spacecraft: Survey (2735 km), HAMO (High Alti-tude Mapping Orbit) (695 km), LAMO (Low Altitude MappingOrbit) (210 km) and HAMO-2, an extension of the mission similarto the HAMO phase (Russell and Raymond, 2011). The VIR instru-ment is made up of two distinct detectors, or ‘‘channels’’: the vis-ible, covering the wavelengths range of 0.25–1.07 lm, and theinfrared, with a sensitivity from 1.02 lm to 5.10 lm (De Sanctiset al., 2011). Each channel has 432 spectral bands, so the averagespectral sampling is 1.8 nm/band for the visible channel and9.8 nm/band for the infrared channel (De Sanctis et al., 2011).The nominal spatial resolution changes according to the mission

the bright material units of types 1 and 2. Violet arrows represent the CWM, greenR/IDA. (For interpretation of the references to color in this figure legend, the reader is

Table 1Classification of the BM units according to the normal visual albedo. The bright regions are classified in five categories. The first column indicates how much BM units brightness,with respect to Vesta geometric albedo.

Normal visual albedo classification

% CWM SM RM

<0% BU14, Calpurnia, Eumchia, Oppia Eumachia, Oppia, Aelia Aelia0–10% Licinia, BU12, Marcia-1, Marcia-2, BU14, Licinia, Lepida, BU12, BU5

BU2-2, BU9, D10-1, BU10-2, Sextilia, BU8 Marcia-1, Marcia-2, Cornelia-1,Drusilla, BU9, BU11-1, BU11-2, BU11-3,BU5, Sextilia, BU8

10–20% Drusilla, Pinaria, Canuleia BU13, Cornelia-2, Cornelia-3, BU7, Canuleia-1, Canuleia-2, BU1Justina-1, Justina-2, BU15 BU2-1, BU2-2, Pinaria, BU10-1, BU10-2, Justina-1, Justina-2, BU6

Canuleia, Justina-1, Justina-2, BU620–30% Cornelia-1, Cornelia-2, BU4 BU3, Myia>30% BU15 Tuccia

Fig. 2. Location of the BM on Vesta global albedo map. As in Fig. 1, we represent inviolet the CWM, in green the SM, and in blue the RM. BM names are indicated withthe first three letters of the crater name in which they are found. BM in unnamedcraters are indicated with the letters ‘‘BU’’ followed by a progressive number. Weemploy the Claudia coordinates system. (For interpretation of the references tocolor in this figure legend, the reader is referred to the web version of this article.)

76 F. Zambon et al. / Icarus 240 (2014) 73–85

phase and was roughly 700 m/px for Survey phase, 180 m/px theHAMO and HAMO-2, and 60 m/px for LAMO. In this work we usedVIR data in units of calibrated reflectance factor (I/F) from 0.4 lmto 2.5 lm, focusing our analysis on the two pyroxene bands. Thebridging between the two VIR channels is performed in the post-processing. A break in the spectra near 1.1 lm (seen in Figs. 3and 4) is due to the junction between the visible and infrared chan-nels. To reduce the noise, we removed recurrent spikes andsmoothed the spectra before deriving the spectral parameters. Adetailed description of the band parameter computation is givenin Ammannito et al. (2013a) and De Sanctis et al. (2012a) supple-mentary material.

For this study, we chose VIR data from HAMO and HAMO-2. Thisdataset provides good coverage of most of the Vesta’s surface at highspatial resolution, except for latitudes northward of 70�. The HAMOand HAMO-2 VIR coverage obtained by VIR is shown in Ammannitoet al. (2013a). VIR data with the highest spatial resolution (LAMO)cover only a very small portion of the surface near the south pole.For the BU15, only LAMO data are available. We selected 75 brightmaterial locations for analysis: 29 observations of CWM, 32 observa-tions of SM and 14 observations RM. The choice of the BM sites wasmade according to the quality and availability of the VIR data. Duringthe different phases of the mission, some areas were mapped morethan once. In such cases, we refer to the acquisitions as follows:BU10-1, BU10-2, Marcia-1, Marcia-2, etc. The identification of thebright units in the VIR data was done by using the coordinates

reported by Mittlefehldt et al. (2012). Subsequently, we selectedas bright pixels those with an albedo 20% larger than their surround-ings. The final spectrum for each bright area is the average of thosepixels (Figs. 3 and 4). The locations of the BM units are given in thecoordinate system known as ‘‘Claudia’’ (supplemental materialof Russell et al. (2012), Roatsch et al. (2012), http://sbn.psi.edu/archive/dawn/grand/DWNVGRD_2/DOCUMENT/VESTA_COORDI-NATES/VESTA _COORDINATES _121214. PDF).

4. Description of the spectral parameters

The aim of this work is to gather information about the spectralcharacteristics of the BM units and to highlight possible differenceswith respect to the other Vesta terrains. We chose three spectralparameters that are diagnostic for mineralogical composition andphysical properties of the material: band center (BC), band depth(BD), and Band Area Ratio (BAR). Band centers and BAR do notstrongly depend on the viewing and illumination geometry(Cloutis et al., 1986), but they can show small shifts in case of rel-evant temperature changes (Hinrichs et al., 1999; Moroz et al.,2000; Hinrichs and Lucey, 2002; Burbine et al., 2009; Moskovitzet al., 2010). As we will show later, in our case the temperaturevariations are small, and therefore do not affect the evaluation ofthe composition of the BM units.

Band depth is the only spectral parameter considered that canvary depending on the illumination conditions. In our case, weselect observations with phase angles between 28� and 52� andincidence angles less than 60�. Longobardo et al. (2014) this issueshow that in this range, band depths are only weakly dependenton the phase angle, hence a photometric correction is notmandatory.

To calculate the spectral parameters, we used the methoddescribed in Ammannito et al. (2013a), to be consistent withpreviously published work (De Sanctis et al., 2012a, 2013b;Ammannito et al., 2013a). In this approach, the band I continuumis fitted by a straight line between the reflectances at 0.7 lm andat 1.238 lm; while the band II continuum is fitted between1.513 lm and 2.487 lm. We have verified that by computing thespectral continuum in alternative ways (small deviations in thewavelength location of the two fit points) does not alter the bandcenters substantially; changes being within the uncertainties asso-ciated with the band center computation. All the spectral parame-ters discussed in the present work were estimated after theremoval of the spectral continuum. The band center is taken asthe minimum of the second-order polynomial that provides thebest fit with to the absorption band (Ammannito et al., 2013a).The band depth is calculated following the definition by Clarkand Roush (1984):

1� r=rc

Fig. 3. Spectra for the Arruntia, Bellicia, and BU5 BM locations. CWM are plotted in violet, SM are green, RM are in blue, the surrounding region (SR) are in red. Left column.Reflectance spectra normalized at 0.55 lm. Right column. Continuum-removed spectra. Arruntia and Bellicia are olivine-bearing locations, while BU5 is a more typicalexample of BM unit. The gap at 1.1 lm is the connection between the VIR visible and infrared channels. (For interpretation of the references to color in this figure legend, thereader is referred to the web version of this article.)

F. Zambon et al. / Icarus 240 (2014) 73–85 77

where r is the reflectance at the minimum and rc is the reflectanceof the continuum at the wavelength of the band minimum (see thesupplementary material of De Sanctis et al. (2012a) and Ammannitoet al. (2013a)).

The Band Area Ratio (BAR) is defined as the ratio of band areas IIand I (Cloutis et al., 1986). We calculate the band area as the areaenclosed by the band and the continuum.

In the supplementary online material we provide values of thespectral parameters computed for all the BM units.

4.1. Error estimation and temperature correction

For each location, we consider the average values of the spectralparameters for the selected VIR pixels, and we take the standarddeviation as an estimate of the uncertainty. Furthermore, theuncertainty associated with the spectral parameters, is consistentwith the standard deviation; for example the error associated withthe band centers is 0.004 lm for band I and 0.02 lm for the band II

(Ammannito et al., 2013a). Since we consider average values, itmay happen that the BM units are mixed with dark material atvery small spatial scales or that bright regions are represented byonly a few pixels. The largest standard deviation values are foundfor those BM that are mixed with dark material at a small localscale because of the difficulty in selecting ‘‘pure’’ BM-only pixels,as for Aelia, Arruntia and Sextilia. High standard deviation valuesare also found for the CWM, which are usually very small and moredifficult to isolate in VIR data.

Laboratory experiments show that band center positionsdepend on temperature (e.g., Roush and Singer, 1986; Hinrichset al., 1999; Moroz et al., 2000; Hinrichs and Lucey, 2002;Burbine et al., 2009). A thorough analysis of the temperatures ofBM units as retrieved from VIR infrared data was carried out byTosi et al. (2014). Limiting the analysis only to the period of max-imum daily insolation, the average temperatures of BM units lie inthe range between 252 and 265 K with maximum values between255 and 266 K (Tosi et al., 2014). Due to Dawn’s observational

Fig. 4. The mean spectra of Tuccia, Aelia and BU15 BM units. Tuccia and BU15 are the brightest BM units in our study set. BU15 contains a very bright streak (region boundedby the black circle in Fig. 10) that has band depths shallower than the surroundings, unlike the other bright units. The magenta spectra in BU15 plot are an example of spectraof the very bright streak. In Aelia, the reduction of the bands in the RM is due to the presence of admixed dark material. The colors of the spectra are indicative of the BM types,as described in Fig. 3. Left column. Reflectance spectra normalized at 0.55 lm. Right column. Continuum-removed spectra. The gap at 1.1 lm is the connection between theVIR visible and infrared channels. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

78 F. Zambon et al. / Icarus 240 (2014) 73–85

planning, aimed at optimizing the signal-to-noise, most BM unitswere observed by VIR in a relatively narrow range of local solartimes, during which only small thermal excursions were recorded.Applying the empirical relation of Burbine et al. (2009) for the bandcenter shifts and the Longobardo et al. (2014) correction, we pre-dict only very small shifts of the band centers. The Burbine methodleads a shift on the order of �0.001 lm for the band I center and�0.01 lm for the band II center. The Longobardo et al. (2014) cor-rection provides the same results for the band II center, while theband I center shift is slightly larger (up to 0.004 lm). The possiblethermal shifts of the band centers are comparable to the standarddeviation of the BM data, therefore confirming that, for the purposeof our mineralogical analysis, temperature variations are notrelevant.

5. Bright material spectral analysis

In this section we describe the results of the spectral analysisperformed on BM units. We use these parameters to derive infor-mation on the origin and nature of the BM units on Vesta.

5.1. Band centers

Band centers of the BM units, computed as explained above, canbe compared to laboratory analogue band center values.

A classical way to infer the mineralogy of a sample is to plot theband I and band II center positions (Fig. 5). The positions for the 75observations of BM are shown in Tables 1–3 of the supplementarymaterial. Most of the band I and band II centers lie on a linear

F. Zambon et al. / Icarus 240 (2014) 73–85 79

trend, as expected if variations are caused by changes in pyroxenechemistry (Adams, 1974; Gaffey et al., 2002). The only clear devia-tion from the trend is shown by the Bellicia site, described in latersections.

Fig. 6. Summary diagram of the bright material composition. Each color representsthe percentage of diogenitic, howarditic and eucritic BM unit. (For interpretation ofthe references to color in this figure legend, the reader is referred to the web versionof this article.)

5.1.1. Bright material mineralogy and relation with HEDFrom laboratory data it is known that the band centers of

pyroxenes increase with increasing iron content (Klima et al.,2007, 2011). HED meteorites show a similar trend as shown inthe plot of band I center versus band II center (Gaffey, 1997; DeSanctis et al., 2012a; McSween et al., 2013b). The band I and II cen-ters occur at shorter wavelengths in diogenites than in eucrites,due to a higher abundance of Mg-rich pyroxenes with lower Caconcentrations in diogenites (Gaffey, 1976). Howardites overlapthe band-center fields of diogenites and eucrites. The areas occu-pied by each subgroup are outlined in Fig. 5 with different colors:red for diogenites, green for howardites, and yellow for eucrites.The definition of the HED regions is given by Ammannito et al.(2012, 2013a), De Sanctis et al. (2012a), and De Sanctis et al.(2013a). The values of the band centers derived from the BM unitsplot in different regions, spanning from diogenites to eucrites, sug-gesting a large range of mineralogies present among the differentBM units. However, the values of the band centers of the BM unitsshow that most of them are compatible with eucrite-rich howar-dites, while only a few bright areas are diogenite-rich howardites(Figs. 5 and 6). In particular, among the 75 observations analyzed,42 can be associated to eucrite-rich howardites, 24 to eucrites, 2 tohowardites and 5 to diogenite-rich howardites while only 2 arediogenitic (Fig. 6). The proportion of the different mineralogiesamong the BM units is similar to the proportion found on Vesta,where VIR data show a preponderance of eucrite-rich howardite,indicating that eucritic lithologies dominate, as they do among

Fig. 5. Plot of Band I vs. Band II centers of the BM units. Symbols and colorsdesignate the BM types. The colored circles delimit the HED regions. (Forinterpretation of the references to color in this figure legend, the reader is referredto the web version of this article.)

the HEDs (De Sanctis et al., 2013b). Diogenites are the least abun-dant class among HEDs, and VIR spectral data confirm that they arethe least common lithology on Vesta (De Sanctis et al., 2013a) aswell as among the BM units.

5.1.2. Band centers and albedoWe searched for a relationships between the band II center

position and the normal visual albedo for each BM units, but wedid not find any (Fig. 7A). Similarly, no correlation is found forthe band I center and the normal visual albedo. This means thatVesta has both eucritic and diogenitic bright units, and also thatthe two lithologies do not exhibit major differences in terms ofreflectance (Fig. 7A).

5.2. Band depth analysis

Band depths are indicative of content of the pyroxene, opaqueminerals or other mineralogical phases, the iron, and the grain sizedistribution (Adams, 1974; Clark, 1999; Mayne et al., 2010;Serventi et al., 2013). The band depths of the BM units are reportedin Tables 7–9 of the supplementary material.

A key characteristic of BM units is a larger band depth com-pared with their surroundings (Pieters et al., 2012; Reddy et al.,2012b).

We calculated band I and band II depths for the different typesof BM units to determine if there are differences between the BMclasses and whether all the BM units show this increased banddepth (Fig. 8A). We found that only a few units, namely BU14,Arruntia, Bellicia and Aelia-RM, depart from the general trendfor their BM type (see Section 7). Those cases have shallowerband depths with respect to the surrounding areas. Band I depthand band II depth are correlated (Burns, 1993). We calculated thebest-fit line for each BM type (Fig. 8A), to emphasize the differ-ences among the BM types. We found that the best-fit line forRM has a lower slope than the other BM types, suggestive ofgreater degrees of mixing between RM and the surroundingregions. In Fig. 8B, we show band depths of the different BM unitsand band depths of HED meteorites for different grain sizes(Table 11 of the supplementary material). Pyroxene absorptionbands have been studied by many authors (e.g., Crown and

Fig. 7. (A) The plot of normal visual albedo vs. band II center does not show arelationship between albedo and band centers. (B) Plot of normal visual albedo vs.band II depth. Solid line represents the best fit.

Fig. 8. (A) Band I vs. Band II depth for the BM units, with the corresponding best fitlines. (B) Band depths of the BM regions compared with those of the HED RELABsamples at different grain size.

80 F. Zambon et al. / Icarus 240 (2014) 73–85

Pieters, 1987; Craig et al., 2007, 2008; Pompilio et al., 2009), andit is well known that both the band I and band II depths increasewith increasing grain size. Thus band depths can be used to inferinformation on the grain size of the observed units. The banddepths of the BM units are similar to band depths of HEDs withsmaller grain sizes (0–25 lm). Only a few BM examples are con-sistent with slightly larger grains (25–45 lm) (Fig. 8B and Tables7–9 and 11 of the supplementary material). This indicates thatthe BM units are composed of small particles. On the other hand,BM units generally have band depths larger than those of the sur-rounding terrains, suggesting that the regolith average grain sizenearby is even smaller. Only a very few BM cases show a reduc-tion of one or both the bands compared with the surroundings.

5.2.1. Agents that modify band depthIn the analyzed dataset, we found some bright regions with

bands shallower than average. Below we discuss possible causesfor the band contrast changes.

� Presence of other mineralogical phases. A possible reason for thereduction of the bands could be related to a larger abundancesof minerals with weak or no absorption bands, like plagioclase.Alternatively, the presence of a mineral like olivine could mod-ify the band I shape and reduce the band II depth (see for exam-ple Bellicia spectrum in Fig. 3).

� Grain size. Band depth is also a function of the grain size. If mul-tiple scattering dominates, as is usually the case in the visibleand near-infrared, then the reflectance decreases as the grainsize increases, while band depth increase with increasing grainsize, up to a maximum before to diminish (Clark, 1999; Harloffand Arnold, 2001; Craig et al., 2007, 2008; Cloutis et al., 2013).Hiroi et al. (1994) reported an average grain size <25 lm forVesta’s regolith based on analysis of telescopic spectra. Fig. 8Bshows the variation of the band depth of HED samples providedby the Reflectance Experiment Laboratory (RELAB) for differentgrain sizes (the values are reported in Table 11 of the supple-mentary material). The plot shows the good correspondencebetween the band depths of the finer HED grain (<45 lm) andthose of the BM units.� Opaque material. The presence of low-reflectance carbon grains

or other opaque phases reduces the brightness and the banddepth in a mixture. Studies made on intimate mixtures with dif-ferent amounts of dark carbon grains, demonstrate the relation-ship between the band depth and the presence of dark material(e.g., Clark, 1983; Le Corre et al., 2011; Cloutis et al., 2013). Therelationship between the band depth and the amount of opaquematerials is generally not linear and thus small quantities ofopaques can drastically reduce the band depth (Clark, 1983).Carbonaceous material on Vesta is suspected to be present indark regions which display very shallow bands (McCord et al.,2012; Reddy et al., 2012b; Palomba et al., 2014). In our BMcases, we can exclude a reduction of the band depth due to a

Fig. 9. Plot of Band I center vs. BAR for the different types of BM units along withthose for the HED RELAB samples of grain size <25 lm.

F. Zambon et al. / Icarus 240 (2014) 73–85 81

high abundance of opaques because of the high reflectance ofthese areas, except for a few cases such as Aelia-RM, wherebright and dark materials are mixed at a small scale.

5.2.2. Normal visual albedo and band depth of the bright materialBM units generally have an albedo greater than the average

value for Vesta. We found a linear relationship between normalvisual albedo and band depths for the BM units (Fig. 7B), asopposed to what is expected if stronger bands are due to largergrain sizes. In fact, Fig. 7B shows that the albedo increases forgreater band II depths. For the band I depth the results are similar.Albedo, band depth and grain size are correlated. A small particlesize of the BM units should lead to high albedo values withdecreasing band depth, contrary to what we observe. Hence grainsize, is not the most important cause affecting the band depths.

Thus, excluding the few cases in which BM units show smallerband depths, we conclude that the band depths of BM regions prin-cipally depend on the abundance of pyroxenes and the amount ofFe2+.

5.2.3. Space weathering effectsSpace weathering could also reduce the band depth. On the

Moon this process is linked to the presence of nanophase metalliciron (npFe0), which causes darkening of the soil, and a reduction ofthe band depth. ‘‘Classical’’ space weathering, like the processesaffecting the lunar soils, introduces a strong red slope in thenear-infrared continuum between �0.7 and 1.5 lm and reducethe pyroxenes bands and the reflectance (Pieters et al., 2012;McSween et al., 2013b). As mentioned earlier, Vesta is one of thebrightest airless silicate bodies in the Solar System, and Vesta’sspectra have prominent absorption bands, especially for brightregions. Vesta does not show the characteristics typical of lunar-style space weathering (Pieters et al., 2012). However, the space-weathering environment is different on Vesta compared to theMoon. The evolution of Vesta’s surface appears to be dominatedby large- and small-scale impacts, which produce and redistributeparticulate material (Bottke et al., 1994; Pieters et al., 2012). More-over, the low average velocity (<5 km/s) of impacts in the mainasteroid belt lead to a dominance of mechanical brecciation overmelting and vaporization (Pieters et al., 2012) while micrometeor-oid bombardment causes a continuous redistribution of the rego-lith (Pieters et al., 2012; McSween et al., 2013b). This leads to adifferent style of space weathering on Vesta, as mainly constitutingmixing of different endogenic and exogenic materials.

5.3. Band Area Ratio analysis

The Band Area Ratio (BAR) is a linear function of the abundanceof pyroxenes, and is a good parameter with which the presence ofolivine can be highlighted (Cloutis et al., 1986). Since olivine lacksthe absorption band near 2 lm, an increase in the olivine contentof a mixture with pyroxenes causes the BAR value to decrease.Thus, a very small BAR could represent a high amount of olivinein a pyroxene–olivine mixture (Cloutis et al., 1986). We evaluatedthe BAR of the BM units and the BAR of RELAB samples of HEDmeteorites with a grain size smaller than 25 lm. The plot inFig. 9 shows the distribution of the BAR vs. band I center of theBM compared with those of the HEDs. Most of the BM units havea BAR comparable to those of the diogenites, howardites and someeucrites. Moreover, one can see that the bright regions do not haveBAR value exceeding 1.7, while some eucrites show values largerthan 1.7. However, the eucrites with a larger BAR also have theband I center at longer wavelengths, which we do not find forthe BM units on Vesta. It is interesting to notice that some brightunits, such as Arruntia, BU14 and especially Bellicia have small val-ues of BAR and relatively long band I centers. This behavior is

indicative of the presence of olivine in the pyroxene mixture. BandI shapes of these units are atypical for pyroxenes (Fig. 3). Moreover,the spectra of these regions have a smaller band II depth, whichcauses the small BAR values. These characteristics are compatiblewith the presence of an amount of olivine >50%, as reported inAmmannito et al. (2013b). Tables 7–9 and 12 of the supplementarymaterial show the BAR values of the BM units and those of the HEDused in the analysis.

5.4. Molar content of Wo and Fs in the bright material

Indirect methods based on the band center position allow us toderive the average molar content of wollastonite (Wo) (CaSiO3),ferrosilite (Fs) (FeSiO3), and Enstatite (En) (MgSiO3) in pyroxenes,thereby revealing relative variations of Ca, Mg and Fe.

Gaffey et al. (2002) and Burbine et al. (2007) addressed thisissue in two different ways. Gaffey et al. (2002) predicted the abun-dance of Wo and Fs iterative method using pyroxenes data; eachformula is valid for a given interval of Wo and Fs. Burbine et al.(2007) considered a restricted number of HEDs, and found simplerrelations valid for all abundances of Wo and Fs (Burbine et al.,2007; Moskovitz et al., 2010).

We derive the molar content of Wo and Fs with both methodsfounding similar results. In particular, we observe that the meandifferences between the values obtained with the two methodsare �1% for Wo and �5% for Fs. In Fig. 10, we show the resultsobtained by using the approach of Burbine et al. (2007) (see alsoTables 4–6 in the supplementary material). We plot results onpyroxene quadrilateral (Fig. 10). Such pyroxene diagram is gener-ally used for single mineralogical phases, while the Vesta spectraare characterized by mixtures of several minerals, dominated bypyroxenes of various compositions. We note that Bellicia-CWMhas a very different behavior compared with the rest of the BMunits, a result of its high olivine content, and thus its presence inthis diagram is inappropriate. Nonetheless, the location of Belli-cia-CWM in the pyroxene diagram highlights its unusual spectralproperties (and composition).

6. Peculiar cases of bright material units

Some bright areas show peculiar characteristics compared withother regions. Below we describe these units in detail.

6.1. Diogenitic bright material units

A few BM units (BU5 – CWM and RM, BU8 – CWM, Aelia – SM,)exhibit spectral properties consistent with a diogenite-rich howar-dite composition: their band centers occur at the shortest wave-length in the entire dataset. Except for one case of radial material

Fig. 10. Pyroxene diagram for Vesta’s BM units and the surrounding regions. Tuccia and Eumachia, the most eucritic BM units exibit high amounts of Wo and Fs, while BU8,BU5 and Aelia are those with the lowest content of Wo and Fs. Bellicia’s position in the plot is not representative of its pyroxenes because of the presence of large amounts ofolivine. We include it in the diagram only to show the strong contrast between Bellicia and the other BM.

82 F. Zambon et al. / Icarus 240 (2014) 73–85

(BU5-RM), all diogenitic BM regions belong to the Crater WallMaterial (CWM) and Slope Material (SM) morphological classes.These BM are located in the longitude range of 0� to 100�E and lat-itude range �50� to +50� (Fig. 2) corresponding to a region possiblycovered with Rheasilvia ejecta (De Sanctis et al., 2012a;Ammannito et al., 2013a; McSween et al., 2013a). BU5-RM andBU8-CWM are the most diogenitic BM with the lowest amount ofFs and Wo. The low abundance of Fs and Wo results in a highamount of En (up to 60%) indicating a more Mg-rich lithology.Their spectra (Figs. 3 and 4) have very prominent bands that areamong the deepest of all the BM sample analyzed (Figs. 4 and8A). Nevertheless, BU5-CWM and BU8-CWM have an albedo thatis only 10% higher than the average of Vesta.

Another diogenite-rich site is Aelia, a small crater located atlongitude 140.6�E and latitude �14.3�. The ejecta of Aelia containboth dark and bright materials (Stephan et al., 2014; Palombaet al., 2014). Band depths for this diogenitic BM region are similarto those of other BM, except for the Aelia-RM, which has banddepths similar to the surroundings (Fig. 8A). However, in interpret-ing the results obtained for the Aelia region, we must consider thedifficulty in selecting its BM areas, due to the small dimensions ofthe crater and to the presence of dark material. For the same rea-sons, the normal visual albedo appears to be lower than the Vesta’saverage.

Diogenitic BM units are very rare on Vesta’s surface. Diogeniteis generally less abundant than howardite and eucrite, beingmainly concentrated in the Rheasilvia basin (De Sanctis et al.,

2012a) and in a few other locations in the northern hemisphere(Ammannito et al., 2013a). According to the classical magma oceanmodels (Bowman et al., 1997; Righter and Drake, 1997; Ruzickaet al., 1997; McSween et al., 2011; Beck et al., 2012), diogenite isformed in the lower crust/upper mantle and not on the surface.In this scheme, diogenite is exposed only by large impacts, suchas the event that formed the Rheasilvia basin. The deep excavationpermitted diogenite to be incorporated in the ejecta deposited nearthe basin rim, with some ejecta reaching higher latitudes. Nor-mally, areas surrounding the bright units do not show diogenite.An exception is the region that surrounds BU8, implying that atthis location, the crater interior exposes a higher quantity of diog-enite that is less reworked.

6.2. Low band depth bright material units

BM units commonly have prominent absorption bands, gener-ally higher than Vesta’s average band depths. However, three ofthe BM units analyzed here, (Arruntia, Bellicia and BU14) haveshallow bands, similar to those of their surrounding terrain (Figs. 3and 8A). These BM units were identified by Ammannito et al.(2013b) as howarditic regions rich in olivine. The Bellicia spectrumis very different from the average of other BM spectra. The band Icenter (0.944 ± 0.009), band I shape, and the low BAR value(0.677 ± 0.077) are indicative for the presence of olivine, whilethe band II center value is typical for howardite (1.964 ± 0.009).Arruntia and BU14 have characteristics similar to those of Bellicia,

F. Zambon et al. / Icarus 240 (2014) 73–85 83

but less extreme (Fig. 3). For Arruntia and BU14, band I is moresimilar to those of other BM units, and the band center positionis in the howarditic region.

The Bellicia bright units are small CWM and small fresh cratersin the Bellicia ejecta field. The Arruntia crater has extended ejecta,while BU14 is a small unit near Arruntia. Arruntia, Bellicia andBU14 are located in the same region in the northern hemisphere,and the similarity of their spectral parameters suggests a commonorigin. Their spectra have been interpreted as a mixture of howar-dites and olivine in different proportions (Ammannito et al.,2013b). A possible explanation for the presence of olivine in thisarea is as follows: an ancient large impact, which excavated andincorporated large blocks of diogenite-rich and olivine-rich mate-rial into the eucritic crust. Subsequent impacts exposed the oliv-ine-rich material in this area, producing olivine-bearing terrainsin a howarditic background (Ammannito et al., 2013b).

6.3. High albedo bright material units

Tuccia and BU15, the highest-reflectance features on Vesta, arelocated in the southern hemisphere. The albedo of these two loca-tions is about 40% larger than the geometric albedo of Vesta. Inaddition to its very high albedo, Tuccia is one of the most eucriticof the BM units. Band I center is similar to the average of the otherBM (BCI 0.932 ± 0.002 lm), while its band II center(1.993 ± 0.006 lm) is at longer wavelength than the other BM,indicating a higher content of iron and calcium (Wo 8.8 ± 0.8, Fs43.1 ± 1.3) (Fig. 10). The very high albedo and prominent bandssuggest that Tuccia is an example of unweathered Vestan soil, con-sidering that the Vestan-style space-weathering reduces thealbedo and band depths without reddening the spectra (Pieterset al., 2012). BU15-SM (Fig. 11) is the brightest BM unit with a nor-mal visual albedo of 0.531 ± 0.013, and it is located in the southernhemisphere at �65� latitude and 358�E longitude. For this BM unit

Fig. 11. VIR LAMO image of BU15 at 0.55 lm (VIR cube 385962328) overlappedwith the corresponding Framing Camera image (FC21B0015548 11361091538F1D).The images are represented in stereographic projection. The black circle indicatesthe very bright streak inside the BU15-SM.

we analyzed only LAMO data, because of the lack of HAMO andHAMO-2 data. The derived composition of the BU15-SM iseucrite-rich howardites (Table 10 in supplementary material),and the band depths of the BU15-CWM are among the deepest inthe entire BM dataset (Fig. 8A). The SM band depths are similarto those of the surroundings. A very bright streak of �1 km longextends inside the SM (Fig. 11); the band depth of this steak islower than that of the surrounding terrain (Fig. 4, Table 10 of thesupplementary material) (Schröder et al., 2013). This bright streakis peculiar because it has the highest albedo of any feature on Vesta(Schröder et al., 2013), but bands shallower than its surroundings.This anomalous behavior can be due to the presence of a high-albedo mineral with a featureless spectrum (e.g. plagioclase)enriched in this mineral (Crown and Pieters, 1987; Harloff andArnold, 2001; Mayne et al., 2010; Serventi et al., 2013).

7. Discussion and conclusions

Vesta has a high average albedo and shows very large albedocontrasts with many bright and dark units. In this work we reportthe first detailed analysis of a large sample of BM units of Vesta,using VIR data with the highest available spatial resolution in orderto understand the origin and the nature of these materials. Wesought to determine whether BM units are all composed of thesame material or whether they have different mineralogies. Fur-ther, we classified the units as eucritic, diogenitic or howarditicin order to determine any dominant lithology among them. Wealso compared the BM regions to their immediate surroundingsand, searched for the presence of other non-pyroxene minerals.

In general, BM units have the same mineralogy as the surround-ing terrain, but they often show greater band depths. Band depth islinked to factors such as the abundance of the absorbing minerals,the grain size, the presence of opaque material, and the extent ofspace weathering (Clark, 1983; Pieters et al., 2012; Reddy et al.,2012a).

Band depth analysis for the BM units indicates that grain-sizeeffects are not responsible for the enhanced band depths. Albedodecreases when grain size increases, while band depths increasewith increasing grain size. By comparing the band depths of BMunits with those of HEDs for different grain sizes, we found thatthe bright regions are similar to those of the HEDs with grain size<45 lm. In the bright regions we observe that band depthsincrease linearly with albedo, thus we can exclude small grain sizeto be responsible for the high reflectance of the BM units.

Small differences in band depths between the BM types havebeen observed. Crater wall material (CWM) band depths are simi-lar to those of the slope material (SM); this is due to their locationand common origin. CWM may represent discontinuous outcrops,or large megablocks of ejecta from previous, large impacts (Li et al.,2012; Mittlefehldt et al., 2012), while SM may originate as debrismass-wasted from the CWM (Li et al., 2012; Mittlefehldt et al.,2012). Radial material (RM) represents the crater ejecta. In mostcases, RM have shallower bands than CWM and SM present inthe same region. Ejecta are generally more contaminated andmixed with Vesta’s regolith, leading to a reduction of the banddepths.

Space-weathering processes could also affect the bands. Vestaexperience space-weathering processes operate differently thanthose found on, for example, the Moon. These changes are alsocharacteristic of lunar space weathering, but on Vesta there is noassociated increase (‘‘reddening’’) of the continuum slope. This isattributed to the absence of npFe0 on Vesta (Pieters et al., 2012).However, fresh ejecta from Vesta craters show relatively strongband depths, indicating that the physical mixing process on Vestais sufficiently robust to alter their optical properties, reducing both

84 F. Zambon et al. / Icarus 240 (2014) 73–85

the albedo and the band depth (Pieters et al., 2012). Thus, the deepbands and the high albedo of the bright regions are indicative ofpure, less mixed, material.

Using the BI/BII diagram we can associate BM spectra with spe-cific HED lithologies. It is also possible to compute the percentageof BM that shows a specific lithology. All the HED lithologies, fromdiogenite to eucrite, are represented in the BM units, and themajority are eucrite-rich howardite. The lithological distributionof BM units is similar to that found on Vesta as a whole, wheremost of the surface is compatible with eucrite-rich howardite(De Sanctis et al., 2013a). Eucrites are the predominant type amongHEDs, while diogenites are the less abundant class. Pure diogenitesare rare on Vesta, and constitute only 0.2% of the surface (DeSanctis et al., 2013a). Our analysis confirms that diogenites areuncommon on Vesta, but they are a larger fraction (2%) withrespect the other BM regions. The few examples of diogenitic BMunits are located in regions of Rheasilvia ejecta, in agreement withthe global mineralogy of Vesta (De Sanctis et al., 2012a;Ammannito et al., 2013a; McSween et al., 2013a). The spectralparameters used in this study reveal that some BM units containminerals other than pyroxenes. Olivine-rich material was foundin Arruntia, BU14 and in the Bellicia bright unit. Unlike other BMregions, these three units have low BAR values (<0.9), the band IIis significantly shallower, and the shape of band I is consistent withthe presence of olivine. Moreover, the band I center is at longerwavelengths with respect to the other BM units. The effects pro-duced by olivine on band I can be easily detected only for olivinecontents of >50 vol.% (Singer, 1981; Cloutis et al., 1986).Ammannito et al. (2013b) found that the composition in thesebright regions is consistent with a mixture of pyroxenes and oliv-ine, with an olivine concentration between 50 vol.% and 80 vol.%.

The bright streak in BU15, the site with the highest albedo onVesta (Schröder et al., 2013), shows band depths shallower thanits surroundings. This reduction in band depth could be due to adifferent grain size or to a mixture with another mineral. Adecrease in grain size should produce an increase in albedo, butis expected to also result in weaker absorption bands. Thereforewe favor an alternative explanation for the very high albedo.BU15 could contain plagioclase, a high-albedo mineral with a fea-tureless spectrum. When mixed with pyroxene, plagioclaseincreases the albedo and decreases the band depth (Crown andPieters, 1987; Harloff and Arnold, 2001; Mayne et al., 2010;Serventi et al., 2013).

The results of our spectral analysis of BM units indicate that thebright regions have an endogenous origin. Bright regions representfresh, less contaminated material derived from below the surfaceof Vesta.

Acknowledgments

This work was supported by the Italian Space Agency (ASI) andNASA’s Dawn at Vesta Participating Scientists Program. The VIRinstrument was developed under the leadership of INAF, Italy’sNational Institute for Astrophysics, Rome. The instrument wasbuilt by SELEX-Galileo, Florence, Italy. The authors acknowledgethe Dawn Science, Operation and Instrument Teams. A specialacknowledgment to Dr. Cristian Carli for his helpful comments.

Appendix A. Supplementary material

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.icarus.2014.04.037.

References

Adams, J.B., 1974. Visible and near-infrared diffuse reflectance spectra of pyroxenesas applied to remote sensing of solid objects in the Solar System. J. Geophys.Res. 79, 4829–4836. http://dx.doi.org/10.1029/JB079i032p04829.

Ammannito, E., et al., 2012. Vesta and the HED meteorites: comparison of spectralproperties. ACM.

Ammannito, E. et al., 2013a. Vestan lithologies mapped by the visual and infraredspectrometer on Dawn. Maps 48, 2185–2198. http://dx.doi.org/10.1111/maps.12192.

Ammannito, E. et al., 2013b. Olivine from Vesta’s mantle exposed on the surface.Nature 504, 112–125. http://dx.doi.org/10.1038/nature12665.

Beck, A.W., McSween, H.Y., 2010. Diogenites as polymict breccias composed oforthopyroxenite and harzburgite. Meteoritics Planet. Sci. 45, 850–872. http://dx.doi.org/10.1111/j.1945–5100.2010.01061.x.

Beck, A.W., Welten, K.C., McSween, H.Y., Viviano, C.E., Caffee, M.W., 2012. Petrologicand textural diversity among the PCA 02 howardite group, one of the largestpieces of the Vestan surface. Meteoritics Planet. Sci. 47 (6), 947–969. http://dx.doi.org/10.1111/j.1945-5100.2012.01360.x.

Binzel, R.P., Gaffey, M.J., Thomas, P.C., Zellner, B.H., Storrs, W.E.N., Wells, A.D., 1997.Geologic mapping of Vesta from 1994 Hubble Space Telescope Images. Icarus128, 95–103. http://dx.doi.org/10.1006/icar.1997.5734.

Bottke, W.F., Nolan, M.C., Greenberg, R., Kolvoord, R.A., 1994. Velocity distributionsamong colliding asteroids. Icarus 107, 255–268. http://dx.doi.org/10.1006/icar.1994.1021.

Bowman, L.E., Spilde, M.N., Papike, J.J., 1997. Automated EDS modal analysis appliedto the diogenites. Meteoritics Planet. Sci. 32, 869–875. http://dx.doi.org/10.1111/j.1945-5100.1997.tb01577.x.

Burbine, T.H., Buchanan, Binzel, R.P., 2007. Deriving formulas from HED spectra fordetermining the pyroxenes mineralogy of Vesta and asteroids. Lunar Planet.Sci., 38.

Burbine, T.H., Buchanan, P.C., Dolkar, T., Binzel, R.P., 2009. Pyroxene mineralogies ofnear-Earth vestoids. Meteoritics Planet. Sci. 44, 1331–1341. http://dx.doi.org/10.1111/j.1945-5100.2009.tb01225.x.

Burns, R., 1993. Mineralogical Applications of Crystal Field Theory, second ed.Cambridge University Press.

Clark, R.N., 1983. Spectral properties of mixtures of montmorillonite and darkgrains – Implications for remote sensing minerals containing chemically andphysically adsorbed water. J. Geophys. Res. 88, 10635–10644. http://dx.doi.org/10.1029/JB088iB12p10635.

Clark, R.N., 1999. Remote Sensing for the Earth Sciences – Manual of RemoteSensing, vol. 3. John Wiley and Sons, New York.

Clark, R.N., Roush, T.L., 1984. Reflectance spectroscopy: Quantitative analysistechniques for remote sensing applications. J. Geophys. Res. 89, 6329–6340.http://dx.doi.org/10.1029/JB089iB07p06329.

Cloutis, E.A., Gaffey, M.J., Jackowski, T.L., Reed, K.L., 1986. Calibrations of phaseabundance, composition, and particle size distribution for olivine–orthopyroxene mixtures from reflectance spectra. J. Geophys. Res. 91, 641–653. http://dx.doi.org/10.1029/JB091iB11p11641.

Cloutis, E.A., Izawa, M.R.M., Pompilio, L., Reddy, V., Hiesinger, H., Nathues, A., Mann,P.J., Le Corre, L., Palomba, E., Bell, J.F., 2013. Spectral reflectance properties ofHED meteorites + CM2 carbonaceous chondrites: Comparison to HED grain sizeand compositional variations and implications for the nature of low-albedofeatures on Asteroid 4 Vesta. Icarus 223, 850–877. http://dx.doi.org/10.1016/j.icarus.2013.02.003.

Craig, M.A., Cloutis, E.A., Bailey, D.T., 2007. The Effects of Grain Size, <45–1000 lm,on the Reflectance Spectrum of Planetary Analogs from 0.35–2.5 lm. LunarPlanet. Sci., 38.

Craig, M.A., Cloutis, E.A., Reddy, V., Bailey, D.T., Gaffey, M.J., 2008. The Effects ofGrain Size, <10 lm - 4.75 mm, on the Reflectance Spectrum of PlanetaryAnalogs from 0.35-2.5 lm. 39th Lunar and Planetary Science Conference.

Crown, D.A., Pieters, C.M., 1987. Spectral properties of plagioclase and pyroxenemixtures and the interpretation of lunar soil spectra. Icarus 72, 492–506. http://dx.doi.org/10.1016/0019-1035(87)90047-9.

De Sanctis, M. et al., 2011. The VIR spectrometer. Space Sci. Rev. 163, 329–369.http://dx.doi.org/10.1007/s11214-010-9668-5.

De Sanctis, M. et al., 2012a. Spectroscopic characterization of mineralogy and itsdiversity Across Vesta. Science 336, 697–700. http://dx.doi.org/10.1126/science.1219270.

De Sanctis, M.C. et al., 2012b. Detection of widespread hydrated materials on Vestaby the VIR imaging spectrometer on board the DAWN mission. Astrophys. J.Lett. 758, L36. http://dx.doi.org/10.1088/2041-8205/758/2/L36.

De Sanctis, M.C. et al., 2013a. Vesta’s mineralogical composition as revealed by thevisible and infrared spectrometer on Dawn. Meteoritics Planet. Sci.. http://dx.doi.org/10.1111/maps.12138.

De Sanctis, M. et al., 2013b. Possible detection of Olivine on Vesta. Lunar Planet. Sci., 44.Drake, M.J., 1979. Geochemical Evolution of the Eucrite Parent Body-Possible Nature

and Evolution of Asteroid 4 Vesta. University of Arizona Press.Feierberg, M.A., Drake, M.J., 1980. The meteorite-asteroid connection – The infrared

spectra of eucrites, shergottites, and Vesta. Science 209, 805–807. http://dx.doi.org/10.1126/science.209.4458.805.

Gaffey, M.J., 1976. Spectral reflectance characteristics of the meteorite classes. J.Geophys. Res. 81, 905–920. http://dx.doi.org/10.1029/JB081i005p00905.

Gaffey, M.J., 1997. Surface lithologic heterogeneity of Asteroid 4 Vesta. Icarus 127,130–157. http://dx.doi.org/10.1006/icar.1997.5680.

F. Zambon et al. / Icarus 240 (2014) 73–85 85

Gaffey, M.J., Cloutis, E.A., Kelley, M.S., Reed, K.L., 2002. Mineralogy of asteroids.Asteroids III-University of Arizona Press, Tucson.

Harloff, J., Arnold, G., 2001. Near-infrared reflectance spectroscopy of bulk analogmaterials for planetary crust. Planet. Space Sci. 49, 191–211. http://dx.doi.org/10.1016/S0032-0633(00)00132-X.

Hinrichs, J.L., Lucey, P.G., 2002. Temperature-dependent near-infrared spectralproperties of minerals, meteorites, and lunar soil. Icarus 155, 169–180. http://dx.doi.org/10.1006/icar.2001.6754.

Hinrichs, J.L., Lucey, P.G., Robinson, M.S., Meibom, A., Krot, A.N., 1999. Implicationsof temperature-dependent near-IR spectral properties of common minerals andmeteorites for remote sensing of asteroids. Geophys. Res. Lett. 26, 1661–1664.http://dx.doi.org/10.1029/1999GL900334.

Hiroi, T., Pieters, C.M., Takeda, H., 1994. Grain size of the surface regolith of asteroid4 Vesta estimated from its reflectance spectrum in comparison with HEDmeteorites. Meteoritics 29, 394–396.

Klima, R.L., Pieters, C.M., Dyar, M.D., 2007. Spectroscopy of synthetic Mg–Fepyroxenes I: Spin-allowed and spin-forbidden crystal field bands in the visibleand near-infrared. Meteoritics Planet. Sci. 42, 235–253. http://dx.doi.org/10.1111/j.1945-5100.2007.tb00230.x.

Klima, R.L., Dyar, M.D., Pieters, C.M., 2011. Near-infrared spectra of clinopyroxenes:Effects of calcium content and crystal structure. Meteoritics Planet. Sci. 46, 379–395. http://dx.doi.org/10.1111/j.1945–5100.2010.01158.x.

Le Corre, L., Reddy, V., Nathues, A., Cloutis, E.A., 2011. How to characterize terrainson 4 Vesta using Dawn Framing Camera color bands? Icarus 216, 376–386.http://dx.doi.org/10.1016/j.icarus.2011.09.014.

Li, J.Y., McFadden, L.A., Thomas, P.C., Mutchler, M.J., Parker, J.W., Young, E.F., Russell,C., Sykes, M.V., Schmidt, B.E., 2010. Photometric mapping of Asteroid (4) Vesta’ssouthern hemisphere with Hubble Space Telescope. Icarus 208, 238–251.http://dx.doi.org/10.1016/j.icarus.2010.02.008.

Li, J.-Y. et al., 2012. Investigating the origin of bright materials on Vesta: Synthesis,conclusions, and implications. Lunar Planet. Sci..

Longobardo, A., Palomba, E., Capaccioni, F., De Sanctis, M.C., Tosi, F., Ammannito, E.,Schröder, S.E., Raymond, C.A., Russell, C., 2014. Photometric behavior of spectralparameters in Vesta dark and bright region as inferred by the Dawn VIRspectrometer. Icarus 240, 20–35.

Mayne, R.G., Sunshine, J.M., McSween, H.Y., McCoy, T.J., Corrigan, C.M., Gale, A.,2010. Petrologic insights from the spectra of the unbrecciated eucrites:Implications for Vesta and basaltic asteroids. Meteoritics Planet. Sci. 45,1074–1092. http://dx.doi.org/10.1111/j.1945–5100.2010.01090.x.

McCord, T.B., Adams, J.B., Johnson, T.V., 1970. Asteroid Vesta: Spectral reflectivityand compositional implications. Science 168, 1445–1447. http://dx.doi.org/10.1126/science.168.3938.1445.

McCord, T.B. et al., 2012. Dark material on Vesta from the infall of carbonaceousvolatile-rich material. Nature 461, 83–86. http://dx.doi.org/10.1038/nature11561.

McSween, H.Y., Mittlefehldt, D.W., Beck, A.W., Mayne, R.G., McCoy, T.J., 2011.Meteorites and their relationship to the geology of Vesta and the dawn mission.Space Sci. Rev. 163, 141–174. http://dx.doi.org/10.1007/s11214-010-9637-z.

McSween, H.Y. et al., 2013a. Composition of the Rheasilvia basin, a window intoVesta’s interior. J. Geophys. Res. 118, 335–346. http://dx.doi.org/10.1002/jgre.20057.

McSween, H.Y. et al., 2013b. Dawn; the Vesta–HED connection; and the geologiccontext for eucrites, diogenites, and howardites. Meteoritics Planet. Sci., 1–15.http://dx.doi.org/10.1111/maps.12108.

Mittlefehldt, D.W., McCoy, T.J., Goodrich, C.A., Kracher, A., 1998. Non-ChondriticMeteorites from Asteroids Bodies. Planetary Materials, vol. 36(Chapter 4).

Mittlefehldt, D.W. et al., 2012. Types and distribution of bright materials on 4 Vesta.Lunar Planet. Sci..

Moroz, L., Schade, U., Wasch, R., 2000. Reflectance spectra of olivine–orthopyroxene-bearing assemblages at decreased temperatures: Implicationsfor remote sensing of asteroids. Icarus 147, 79–93.

Moskovitz, N.A., Willman, M., Burbine, T.H., Binzel, R.P., Bus, S.J., 2010. A spectroscopiccomparison of HED meteorites and V-type asteroids in the inner Main Belt. Icarus208, 773–788. http://dx.doi.org/10.1016/j.icarus.2010. 03.002.

Palomba, E., Longobardo, A., De Sanctis, M.C., Zambon, F., Tosi, F., Ammannito, E.,Capaccioni, F., Frigeri, A., Capria, M.T., Cloutis, E.A., Jaumann, R., Combe, J.-Ph.,Raymond, C., A., Russell, C.T., 2014. Composition and mineralogy of darkmaterial deposits on Vesta. Icarus 240, 58–72.

Pieters, C.M. et al., 2011. Surface composition of Vesta: Issues and integratedapproach. Space Sci. Rev. 163, 117–139. http://dx.doi.org/10.1007/s11214-011-9809-5.

Pieters, C.M. et al., 2012. Distinctive space weathering on Vesta from regolithmixing processes. Nature 491, 79–82. http://dx.doi.org/10.1038/nature11534.

Pompilio, L., Pedrazzi, G., Sgavetti, M., Cloutis, E.A., Craig, M.A., Roush, T.L., 2009.Exponential Gaussian approach for spectral modeling: The EGO algorithm I.Band saturation. Icarus 201, 781–794. http://dx.doi.org/10.1016/j.icarus.2009.01.022.

Prettyman, T.H. et al., 2011. Dawn’s gamma ray and neutron detector. Space Sci.Rev. 163, 371–459. http://dx.doi.org/10.1007/s11214-011-9862-0.

Reddy, V. et al., 2012b. Color and Albedo heterogeneity of Vesta from Dawn. Science336, 700–704. http://dx.doi.org/10.1126/science.1219088.

Reddy, V., Le Corre, L., O’Brien, D.P., Nathues, E.A., Cloutis, D., Durda, B.M.U., Bottke,W.F., Nervorny, D., Buczkowski, D., Scully, J.E.C., Palmer, E.M., Sierks, H., Mann,P.J., Becker, K.J., Beck, A.W., Mittlefehldt, D.W., Li, J.-Y., Gaskell, R., Russell, C.T.,Gaffey, M.J., McSween, H.Y., McCord, T.B., Combe, J.-P., Blewett, D.T., 2012a.Delivery of dark material to Vesta via carbonaceous chondritic impacts. Icarus223, 544–559. http://dx.doi.org/10.1016/j.icarus.2012.10.006.

Righter, K., Drake, M.J., 1997. A magma ocean on Vesta: Core formation andpetrogenesis of eucrites and diogenites. Meteoritics Planet. Sci. 32, 929–944.http://dx.doi.org/10.1111/j.1945-5100.1997.tb01582.x.

Roatsch, T. et al., 2012. High resolution Vesta High Altitude Mapping Orbit (HAMO)Atlas derived from Dawn framing camera images. Planet. Space Sci. 73, 283–286. http://dx.doi.org/10.1016/j.pss.2012.08.021.

Roush, T.L., Singer, R.B., 1986. Gaussian analysis of temperature effects on thereflectance spectra of mafic minerals in the 1-lm region. J. Geophys. Res. 91,10301–10308. http://dx.doi.org/10.1029/JB091iB10p10301.

Russell, C.T. et al., 2007. Dawn mission to Vesta and Ceres. Symbiosis betweenterrestrial observations and robotic exploration. Earth Moon Planets 101, 65–91. http://dx.doi.org/10.1007/s11038–007–9151–9.

Russell, C.T. et al., 2012. Dawn at Vesta: Testing the protoplanetary paradigm.Science 336, 684–686. http://dx.doi.org/10.1126/science. 1219381.

Russell, C.T., Raymond, C.A., 2011. The dawn mission to Vesta and Ceres. Space Sci.Rev. 163, 3–23, 10.1007/s11214–011–9836–2.

Russell, C. et al., 2013. Vesta in the light of Dawn. Lunar Planet. Sci., 44.Ruzicka, A., Snyder, G.A., Taylor, L.A., 1997. Vesta as the HED parent body:

Implications for the size of a core and for large-scale differentiation. MeteoriticsPlanet. Sci. 32, 825–840. http://dx.doi.org/10.1111/j.1945–5100.1997.tb01573.x.

Schröder, S.E., Mottola, S., Keller, H., 2013. Resolved photometry of Vesta revealsphysical properties of crater regolith. Planet. Space Sci.. http://dx.doi.org/10.1016/j.pss.2013.06.009.

Serventi, G., Carli, C., Sgavetti, M., Ciarniello, M., Capaccioni, F., Pedrazzi, G., 2013.Spectral variability of plagioclase–mafic mixtures (1): Effects of chemistry andmodal abundance in reflectance spectra of rocks and mineral mixtures. Icarus226, 282–298. http://dx.doi.org/10.1016/j.icarus.2013.05.041.

Sierks, H. et al., 2011. The dawn framing camera. Space Sci. Rev. 163, 263–327.Singer, R.B., 1981. Near-Infrared spectral reflectance of mineral mixtures:

Systematic combinations of pyroxenes, olivine, and iron oxides. J. Geophys.Res. 86, 7967–7982. http://dx.doi.org/10.1029/JB086iB09p07967.

Stephan, K. et al., 2014. Small fresh impact craters on asteroid 4 Vesta: Acompositional and geological fingerprint. J. Geophys. Res. 119, 771–797.http://dx.doi.org/10.1002/2013JE004388.

Tedesco, E.F., Williams, J.G., Matson, D.L., Weeder, G.J., Gradie, J.C., Lebofsky, L.A.,1989. A three-parameter asteroid taxonomy. Astron. J. 97, 580–606.

Thomas, P.C., Binzel, R.P., Gaffey, M.J., Storrs, A.D., Wells, E.N., Zellner, B.H., 1997.Impact excavation on Asteroid 4 Vesta: Hubble space telescope results. Science227, 1492–1495. http://dx.doi.org/10.1126/science.277.5331.1492.

Tosi, F., Capria, M.T., De Sanctis, M.C., Combe, J.-P., Zambon, F., Nathues, A., Schröder,S.E., Li, J.-Y., Palomba, E., Longobardo, A., Blewett, D.T., Denevi, B.W., Palmer, E.,Capaccioni, F., Sunshine, J.M., Ammannito, E., Titus, T., Mittlefehldt, D.W.,Blewett, D.T., Russell, C.T., Raymond, C.A., The Dawn/VIR Team, 2014. Thermalmeasurements of dark and bright surface features on Vesta as derived fromDawn/VIR. Icarus 240, 36–57.

Zellner, B.H., Gibbard, S., de Pater, I., Marchis, F., Gaffey, M.J., 2005. Near-IR imagingof Asteroid 4 Vesta. Icarus 177, 190–195. http://dx.doi.org/10.1016/j.icarus.2005.03.024.

Zuber, M.T. et al., 2011. Origin, internal structure and evolution of 4 Vesta. Space Sci.Rev. 163, 77–93. http://dx.doi.org/10.1007/s11214-011-9806-8.