Icarus 208 (2010) 945–954

Contents lists available at ScienceDirect

Icarus

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

Surface composition and physical properties of several trans-neptunian objectsfrom the Hapke scattering theory and Shkuratov model q

F. Merlin a,b,*, M.A. Barucci a, C. de Bergh a, S. Fornasier a, A. Doressoundiram a, D. Perna a,c, S. Protopapa d

a LESIA/Observatoire de Paris, 5 Place Jules Janssen, 92195 Meudon Cedex, Franceb Department of Astronomy, University of Maryland, College Park, MD 20742, USAc INAF, Osservatorio Astronomico di Roma, via Frascati 33, 00040 Monteporzio Catone (Roma), Italyd Max-Planck Institute for Solar System Research, Max-Planck-Str. 2, 37191 Katlenburg-Lindau, Germany

a r t i c l e i n f o

Article history:Received 19 March 2009Revised 3 March 2010Accepted 7 March 2010Available online 16 March 2010

Keywords:Kuiper BeltIcesSpectroscopyTrans-neptunian objectsCentaurs

0019-1035/$ - see front matter � 2010 Elsevier Inc. Adoi:10.1016/j.icarus.2010.03.014

q Based on observations made with ESO Telescopesunder programme ID 178.C-0036.

* Corresponding author at: Department of AstronoCollege Park, MD 20742, USA.

E-mail addresses: merlin@astro.umd.edu, frederic.m

a b s t r a c t

Several different trans-neptunian objects have been studied in order to investigate their physical andchemical properties. New observations in the 1.1–1.4 lm range, obtained with the ISAAC instrument,are presented in order to complete previous observations carried out with FORS1 in the visible and SINF-ONI in the near infrared. All of the observations have been performed at the ESO/Very Large Telescope.We analyze the spectra of six different objects (2003 AZ83, Echeclus, Ixion, 2002 AW197, 1999 DE9 and2003 FY128) in the 0.45–2.3 lm range with the model of Hapke (Hapke, B. [1981]. J. Geophys. Res. 86,4571–4586) and the method of Shkuratov et al. (Shkuratov, Y., Starukhina, L., Hoffmann, H., Arnold, G.[1999]. Icarus 137, 235–246). Water ice is found on two objects, and in particular it is confirmed in itsamorphous and crystalline states on 2003 AZ84 surface. Upper limits on the water ice content are givenfor the other four TNOs investigated, confirming previous results (Barkume, K.M., Brown, M.E., Schaller,E.L. [2008]. Astron. J. 135, 55–67; Guilbert, A., Alvarez-Candal, A., Merlin, F., Barucci, M.A., Dumas, C., deBergh, C., Delsanti, A. [2009]. Icarus 201, 272–283). Whatever the absorption features in the near infrared,all objects but one exhibit a moderate red slope in the visible, as most TNOs and Centaurs. We discuss theimplications of the presence of water ice and the probable sources of the red slope.

� 2010 Elsevier Inc. All rights reserved.

1. Introduction

Trans-neptunian objects (TNOs) are known as primordial andremnant bodies of the Solar System. Located beyond Neptune, theyare separated into four dynamical classes, that are the ScatteringDisk Objects (SDOs), the classical Kuiper Belt Objects (KBOs), theDetached Objects (DO) and the Resonant Objects with Neptune(ROs). Centaurs, which orbit completely inside the orbit of Neptuneare considered as SDOs; see Davies et al. (2008) and Gladman et al.(2008) for a complete review on the dynamical classes. AlthoughTNOs and Centaurs seem to have a common origin, the observationof TNOs has revealed a high diversity of colors (in the visible andnear infrared: Doressoundiram et al., 2008; Tegler et al., 2008,etc.), of absorption features (e.g.: Barucci et al., 2008a; Guilbertet al., 2009) and multiple surface roughnesses (from polarimetryand photometry; e.g., Belskaya et al., 2008) that imply a varietyof surface composition. A few trends are drawn from these recent

ll rights reserved.

at the Paranal Observatories

my, University of Maryland,

erlin@obspm.fr (F. Merlin).

observations. Objects with a diameter close or greater than1000 km show clear absorption bands in the near infrared (Barucciet al., 2008a), neutral or moderate red slope from visible spectros-copy or photometry (e.g.: Eris, Pluto or Haumea observed byDumas et al. (2007), Merlin et al. (2009, 2007), Licandro et al.(2006), Brown et al. (2005), Trujillo et al. (2007), Pinilla-Alonso etal. (2009) for instance). The spectra of smaller objects are mainlyfeatureless in the near infrared and exhibit different slopes in thevisible (for a sample see, e.g.: Dotto et al., 2003a). It is possible toidentify the physical state or/and dilution state of the different icespresent on the surface of the biggest and brightest objects, but ourknowledge on the surface of the smallest and darkest ones is stilllimited due to low signal-to-noise level. However, it is very impor-tant to compare the chemical composition of the smallest oneswith that of the biggest ones to confirm or refute recent theorieson the physical processes that govern their surface evolution.

These theories imply that the surface of any atmospherelessbody in the Solar System is directly exposed to cosmic rays, solarwind or high energy particles coming from the Sun or from theinterstellar medium. These phenomena are very important in thetransformation of the species located in the first layers of the sur-face, those that are accessible from visible and near-infrared spec-troscopy. These physical processes, called space weathering, tend

946 F. Merlin et al. / Icarus 208 (2010) 945–954

to modify the initial rich icy layers into organic or carbon richmaterials that are usually dark and red (Strazzulla et al., 2003) withpossible physical state changes. Crystalline water ice, for example,requires high temperature (more than 100 K) to be formed andseems to be sensitive to space weathering that transforms it intoits amorphous state at low temperature (Mastrapa et al., 2008). Abalance can exist between crystalline and amorphous water iceabove 40 K, from thermal recrystallization and irradiation pro-cesses (Zheng et al., 2009). However, the authors consider onlyion irradiation and other irradiation sources need to be investi-gated too. Crystalline water ice, that displays a clear absorptionband at 1.65 lm (see Grundy and Schmitt, 1998; Mastrapa et al.,2008), is usually detected on the surface of big objects which arecovered by water ice. Recent studies suggest that big objects aremore easily submitted to major rejuvenation events than the smallones, as condensation of volatile species (Schaller and Brown,2007a) or cryo-volcanism processes (Jewitt and Luu, 2004). Thesescenarios also explain the presence of large amounts of very vola-tile species, such as nitrogen or methane, on the surface of thethree biggest objects (Pluto, Eris, 2005FY9 or Makemake). Contraryto the big objects, small bodies would be dominated by spaceweathering (Strazzulla et al., 2003; Brunetto et al., 2006) which ex-plains why the small TNOs are usually the darkest ones (Lykawkaand Mukai, 2005; Stansberry et al., 2008). However, cometary likeactivity, non-disruptive collisions, or disruptive ones, could renewthe surface of any object, small or big, with fresh material (seeGil-Hutton (2002) for instance). It seems to be the case, for in-stance, for the crystalline water ice rich objects of the Haumeafamily (Brown et al., 2007; Pinilla-Alonso et al., 2007; Barkumeet al., 2008). Therefore, it seems that collisional history and sizegive stronger constraints on the surface evolution of these bodiesthan the objects’ dynamics (no clear trends have been obtainedyet from photometry, Doressoundiram et al., 2008). For example,Gil-Hutton et al. (2009) find that the presence of crystalline andamorphous water ice on the surface of the Haumea family mem-bers could perhaps be explained as a balance between irradiationand collisions. Observation of small icy bodies and recurrent spec-tral analyses are required to better constrain these scenarios andbetter investigate the surface properties of TNOs and Centaurs.

In this paper, the spectra of six objects are presented as well as theresults of spectral models (obtained with the models of Hapke (1981,1993), Shkuratov et al. (1999)). These six objects are representativeof most TNOs observed until now, they are relatively dark, have dif-ferent spectral slopes in the visible, and have diameters between 300and 850 km (except Echeclus that is smaller, as the other Centaurs,see Table 1 for details on the physical properties). The use of the com-plete spectrum from 0.4 to 2.35 lm and of the albedo determined inthe V band by Stansberry et al. (2008) is very helpful to minimize theerrors on the models’ free parameters, which are the particle sizesand the relative amounts of the chemical compounds. In fact, thealbedo and the absorption features depend on the quantity and theparticle size of bright and dark compounds, respectively. The goal

Table 1Physical properties of the six studied objects. Dynamical class comes from Gladmanet al. (2008) whereas visual albedo and diameter come from Stansberry et al. (2008).

Name Dynamical class Albedo Diameter (km)

2003 AZ84 Resonant 0:123þ4:31�2:91 585:8�95:5

þ98:8

Echeclus Centaur 0:038þ1:89�1:08 83:6�15:2

þ15:0

Ixion Resonant 0:156þ12:0�5:53 573:1�141:9

þ139:7

55565 (2002 AW197) Classic 0:118þ4:42�3:00 734:6�108:3

þ116:4

26375 (1999 DE9) Resonant 0:069þ1:58�1:19 461�45:3

þ46:1

120132 (2003 FY128) Scattered 0:07�þ0:�0: 307�0:0

0:0

*Arbitrary value.

of this work is to better constrain the chemical composition, espe-cially the icy species as water ice.

2. Observations and data reduction

In this paper, we analyze the spectra of six objects observed inthe framework of a ESO Large Program (P.I.: M.A. Barucci). Thespectroscopic and photometric data were obtained from almostsimultaneous visible and near-infrared observations carried outat UT1, UT2 and UT4 VLT–ESO telescopes (Cerro Paranal, Chile).The visible spectroscopic data were obtained using FORS1 andwere already presented in Alvarez-Candal et al. (2008). The near-infrared H and K spectroscopy was obtained with SINFONI andthe results are presented in Guilbert et al. (2009). The V, J, H andK photometry, used to calibrate and align the different spectro-scopic ranges, is reported and discussed in DeMeo et al. (2009).The near-infrared spectroscopy in J, presented in this paper, wascarried out using the ISAAC instrument in its SW mode (1.1–1.4lm spectral range and equipped with a Rockwell Hawaii1024 � 1024 pixel Hg:Cd:Te array). The spectral resolution is about500 with a 100 slit. The observations (see Table 1) were done bynodding the object along the slit by 1000 between two positions Aand B. The A and B images were combined using the ESO softwarepackages Eclipse and MIDAS following the procedure described byBarucci et al. (2000). Several solar analogs were observed duringeach night at similar airmasses as the objects, and the TNOs reflec-tivity was obtained by dividing their spectra by that of the solaranalog star closest in time and airmass, as reported in Table 2. Eachspectrum has been smoothed with a median filter technique (seeMerlin et al., 2009) to increase the S/N ratio. The final spectral res-olution is 250. The obtained J spectra together with V and H + Kspectra, calibrated with the simultaneous photometry, are re-ported in Fig. 1.

These six complete spectra from 0.4 to 2.3 lm have a spectralresolution between 250 and 1500 (spectral resolution is oversam-pled in the V band). In order to increase the signal-to-noise level,we reduced the spectral resolution to 250. This spectral resolutionis accurate enough to detect the broad absorption bands we arelooking for (water ice, methane ice, methanol ice, etc. see, de Berghet al. (2008) for a review of probable chemical compounds) and getsome information on their physical states (for instance, water icein amorphous or crystalline phase). Fig. 1 shows the results ofour ‘‘degraded resolution” spectra for all the objects. The signal-to-noise level is improved by a factor 3–10, depending on thewavelength range. Alvarez-Candal et al. (2008) reported theabsorption features detected in the visible range, whereas Guilbertet al. (2009) discuss about those observed in the H and K bands.The main absorption features are detected in the near infrared,close to 1.5 and 2.0 lm, which indicates the presence of waterice. Except for 2003 AZ84, the visible spectra appear featurelessand show similar mean red slopes (around 20 ± 5%/100 nm, seeAlvarez-Candal et al., 2008). With the available 0.4–2.3 lm rangespectra, we aim to constrain the surface composition using differ-ent models to compare our data with the existing ones, and to de-rive new information on these objects that have different dynamicproperties. Usually, the signal-to-noise level is good in the V band,relatively good in the H band and is lower in the K and J bands,especially in the cases of 1999 DE9, Echeclus and 2003 FY128.

3. Spectral analyses tools

3.1. The models

In order to investigate the surface properties of these objects,we use the spectral models developed by Hapke (1981, 1993)

Table 2Observational conditions of objects spectroscopically investigated during the first and second semester of the Large Program on TNOs. For each object we report the observationaldate and universal time (UT of the beginning of the exposure), the total exposure time (second), the airmass (mean of the airmass value at the beginning and at the end ofobservation), and the observed solar analog stars (Landolt catalog) with their airmass used to remove the solar contribution.

Object Date UTstart Texp Airmass Analog (airmass)

2003 AZ84 24/01/2007 04:53 2000 1.31 Land. 98–978 (1.39)Echeclus 14/05/2007 01:24 1200 1.13 Land. 107–684 (1.40)Ixion 15/07/2007 04:37 1800 1.30 Land. 147–935 (1.32)55565 (2002 AW197) 23/01/2007 05:57 1800 1.18 Land. 102–1081 (1.13)26375 (1999 DE9) 22/01/2007 06:22 1800 1.18 Land. 102–1081 (1.11)120132 (2003 FY128) 22/01/2007 07:59 2000 1.06 Land. 102–1081 (1.11)

Fig. 1. Reflectance spectra of the six objects. Each spectrum is normalized at 1 in Vband (at 0.55 lm). Raw (gray scale) and smoothed spectra (black line) are adjustedwith photometry (see text). Photometric data (big black dots) give relativereflectance in R, J, H and K bands from left to right. The photometric errors areincluded in the dots. Spectra of Echeclus, Ixion, 2002 AW197, 1999 DE9 and 2003FY128 are shifted by +2, +4, +6, +8 and +10 units in reflectance, respectively, forclarity.

F. Merlin et al. / Icarus 208 (2010) 945–954 947

and Shkuratov et al. (1999). These models allow us to determinethe reflectance spectra or the albedo of a medium from individualphysical properties of the different chemical components (withoptical constants). We first briefly present the basis of the Hapkemodel to investigate the surface properties of the TNOs and theCentaur. All the developments of the model are presented in Hapke(1981, 1993). In our case, we neglect interferences and approxi-mate the geometric albedo at zero phase angle from Eq. (44) ofHapke (1981):

Alb ¼ r0ð0:5þ r0=6Þ þ ðw=8Þðð1þ B0ÞPð0Þ � 1Þ ð1Þ

where w is the single-scattering albedo and r0 ¼ 1�ffiffiffiffiffiffiffi

1�wp

1þffiffiffiffiffiffiffi

1�wp is the

bihemispherical reflectance. w depends on the optical constants of

the particles and is described in Hapke (1981). B0 is the ratio ofthe near-surface contribution to the total particle scattering at zerophase (Hapke, 1986). PðaÞ is the phase function, which depends onthe phase angle a and the asymmetry parameter n (see below). Thephase function describes the angular distribution of light reflectedfrom a body. We follow the formalism of Emery and Brown(2004) to compute w from different compounds, assuming a ‘‘saltand pepper” or an ‘‘intimate” mixture. This choice is mainly sup-ported by literature for 2003 FY128 (Sheppard, 2007; Dotto et al.,2008), 2002 AW197 (Sheppard, 2007; Guilbert et al., 2009) and1999 DE9 (Alvarez-Candal et al., 2007), that do not exhibit clear het-erogeneous surfaces. Even if Echeclus and Ixion could have hetero-geneous surfaces (this paper) we conserve the same models forconsistency between all objects and all models (the Shkuratovmethod is performed for intimate mixtures).

When the particles are opaque, then all the scattered lightcomes from the surface of the particle, B0 is assumed to take its ex-treme value of 1 (0 if the particles are completely transparent). Inour models, B0 is not wavelength dependent, just as Pð0Þ, the phasefunction chosen at 0 phase angle (the phase angle of our objectsbeing very small). There are different ways to take into accountthe phase function, which can be described by a first-orderLegendre polynomial, by a real phase function or by an alternativeexpression such as the Henyey–Greenstein function (Henyey andGreenstein, 1941). In our case, we compute the phase functionwith a single (Eq. (2)) and double-lobe (Eq. (3)) Henyey–Greenstein(HG) functions (Hillier, 1993), which depend on the asymmetryparameter n:

Pð0; nÞ ¼ ð1� n2Þ=ð1þ 2nþ n2Þ3=2 ð2ÞPð0Þ ¼ ð1� f ÞPð0; n1Þ þ fPð0; n2Þ ð3Þ

The asymmetry parameter determines whether the particle isbackscattering ðn � 0Þ, isotropically scattering ðn ¼ 0Þ or forward-scattering ðn � 0Þ. The last function depends on three parametersinstead of 1, with n1 and n2 that are the angular width of the for-ward and backward scattering lobes, respectively, and f that givesthe ratio between the two contributions. In this work, we assumethat all the particles of the medium have a similar phase function.Here, the asymmetry parameters and B0 are free parameters andthe valid ranges are fixed to [�1.0:1.0] and [0:1.0], respectively.Hapke also developed refinements of his model, as the correctionsfor macroscopic surface roughness (Hapke, 1984) or the porosity(Hapke, 2008). They are not taken into account due to the largenumber of parameters already included in the models.

The use of the Shkuratov et al. (1999) method is more trivial, inthe sense that the number of free parameters, except for the opticalconstants, is reduced to 1; the porosity or the volume fraction filledby particles (a free parameter in the model in the [0.0:1.0] range). Aprevious work on the comparison on Hapke and Shkuratov modelshas been done by Poulet et al. (2002) on two Centaurs. The authorsreport that the phase function can be described in the Shkuratovet al. model from the individual contributions of light scatteredby a particle into the backward and forward hemisphere. In our

948 F. Merlin et al. / Icarus 208 (2010) 945–954

models based on Hapke’s theory, we make use of a constant valueof the phase function along the wavelength. The phase function iswavelength dependent in the Shkuratov method, and changes ap-pear with the absorption coefficient variations. Thus, differencesbetween the two models are expected, in terms of width and depthof the absorption bands for a given particle size.

3.2. Reduced v2

Aiming to search for the best free parameters that fit well theobjects’ spectra, we derive abundances and particle size from Mar-quardt–Levenberg minimization. This algorithm allows us to ob-tain a minimization of the reduced v2 between synthetic spectraand the reflectance spectra of the objects. However, these resultsare not unique because similar reduced v2 can exist with differentparameters combinations. On one hand, different combinationscould give similar absorption features. The same depths andwidths of the water ice bands could be obtained from a combina-tion of small amounts of big size particles or large amounts ofsmall size particles, for example. In the same time, the use of moreor less dark material can reduce or increase the albedo of the syn-thetic spectrum and compensate the use of larger or smalleramounts of bright compounds. However, we use different initialconditions to take into account the best result in terms of reducedv2. On the other hand, the reduced v2 is a mathematical result thatcan favor large scale features (spectrum continuum for instance)rather than the small ones (absorption bands). Moreover, our spec-tra have better signal-to-noise in the visible range and the resultsare mainly constrained in this part because the standard deviationof the data is used in the algorithm to determine the best v2. Evenif we do not expect to have a unique solution, especially if we takeinto account the other parameters of the models that are usuallyunknown (porosity, asymmetry parameter, etc.), these modelshave the advantage to better constrain the surface composition(such as water ice, showing clear absorption features in the 0.4–2.35 lm range).

3.3. Optical constants and limits of the models

The first limit is the lack of optical constant data (usually notedn and k). It is especially the case for irradiated products and icymixtures. An additional problem is that the conditions underwhich the laboratory measurements are performed (different tem-perature, deposition processes, etc.) are not always appropriate forthe study of TNOs. From the small data set available, and consider-ing previous results on analogous chemical compounds (see, for in-stance, de Bergh et al., 2008), we have chosen our best candidates.The first one is water ice, which is detected here and is verycommon in the Solar System (for example Barkume et al., 2008;Guilbert et al., 2009). As our objects are mainly at 35–40 AU fromthe Sun and have albedos from 0.04 to 0.15, we can derive a surfacetemperature close to 40 K from Stefan–Boltzmann law. As previ-ously reported, a balance of crystalline and amorphous water icecould occur near 40 K. Therefore, we use optical constants of waterice in both states obtained close to 40 K (Grundy and Schmitt,1998). As the objects do not show any evidence of other icy brightcomponents and since it is not expected that they have them(Schaller and Brown, 2007a), we do not use other ices in the mod-els. In order to take into account the possible presence of mineralsor irradiated products on the surface of these bodies, we use opti-cal constants of olivine and ortho-pyroxene from the Jena database (available on line at: http://www.astro.spbu.ru/JPDOC/f-dba-se.html). We use optical constants of amorphous carbon (Zubkoet al., 1996) as dark compound, and those of Triton, Titan and IceTholins (Khare et al., 1986, 1993) as organic compounds, all ofthem supposed to be produced by space weathering. Titan Tholin,

for instance, is supposed to be formed in Titan haze or Titan upperatmosphere where lightning discharge, electron bombardment, UVphotolysis and high energetic particles act on nitrogen and meth-ane gases. Even if these volatile species are not detectable on thesurface of the small objects, as opposed to the biggest ones, we as-sume that this kind of irradiated products could be present on thesurface of TNOs where nitrogen and methane ice should be, orshould have been, present in icy or gas state (see for more detailson volatile losses, Schaller and Brown, 2007a). Moreover, nitrogenand methane are frequently present in the short and long periodcomets that strongly suffer from solar heating during their multi-ple perihelion passages. Imanaka et al. (2004) formed Titan Tholinfrom a low pressure (a few pascal) mixture of gaseous N2 and CH4

in a cold plasma that reproduces different irradiation sources (elec-tron, UV light and high energy particles). This indicates that spaceweathering could be efficient enough to form products similar toTitan Tholin in ancient or recent transient thin atmospheres ofTNOs.

The approximations on the B0 and P(0) functions, as well as theones made on the Chandrasekhar H functions (Hapke, 1981) thatlead to the Eq. (44) of Hapke (1981) are the first limitations ofthe model. The first two functions are simplified to account forthe lack of knowledge on the surface properties (multi-angleobservations are required to better constrain them). This may givesome variations on the relative amount and size of the compoundscompared to wavelength dependent functions (Mishchenko et al.,2006; Poulet et al., 2002). Hapke (1981) claims that the approxima-tion on the H functions agrees with exact values to better than 3%and that the agreement is better for low albedo surfaces, whichjustifies this last approximation. Another limitation is the particlesize range used in our models. Indeed, the models are based ongeometric optic, and they are valid when the particle size is largerthan the wavelength. The particle size should therefore be largerthan several micron in the 0.4–2.3 lm wavelength range. In severalcases, the only way to obtain a good enough fit of the objects’ spec-tra, is to use smaller particle sizes compared to the wavelength.However, Piatek et al. (2004) show that the Hapke model is still va-lid for grains having size smaller or similar grain sizes compared tothe wavelength.

4. Results

In this section, we will present our best fit synthetic spectra forthe objects. Our results are obtained with albedos in the V bandadopted from Stansberry et al. (2008) and reported in Table 1.For 2003 FY128, we consider an arbitrary albedo of 0.07.

4.1. 2003 AZ84

The signal-to-noise (S/N) level of the spectrum is high enoughto see clear absorption of water ice at 1.5 and 2.0 lm (see Fig. 2).It is possible to report the signature of the crystalline water ice,with the absorption band located at 1.65 lm (already mentionedby Guilbert et al., 2009), but it is not possible to detect otherabsorption bands in the near infrared, within the error bars. Inthe visible part, the spectrum shows a broad absorption feature be-tween 0.5 and 0.7 lm. Fornasier et al. (2004) suggest that this bandis relative to the presence of aqueously altered materials(confirmed by Alvarez-Candal et al. (2008) on the spectrum ac-quired in January 2007 and used in this paper for the spectral mod-eling). For a detailed discussion on the aqueous alteration of TNOs,see the work of de Bergh et al. (2004). Our best fit models are ob-tained with large quantities of water ice; 59%, 66% or 48% from thedifferent models (see Table 3). Small amounts of organic com-pounds and large amounts of amorphous carbon (particle size

Fig. 2. Spectrum of 2003 AZ84 (in gray) with a resolution of 250 and error bars (1r).Continuous and dotted lines represent the synthetic spectra obtained with Hapkemodel, with single HG function and double HG function, respectively. The dashedline represents the synthetic spectrum obtained with Shkuratov model.

F. Merlin et al. / Icarus 208 (2010) 945–954 949

close to 10 lm) are always required with the Hapke models, wherethe fits are better and obtained with a good confidence level (betterthan 99% with reduced v2 close to 0.81). Main differences occur inthe visible part between the models, see Fig. 2, due to variations onthe Titan and Ice Tholins amounts) and around 1.7 lm (due to thedifferent contributions of the water ice). The visible part is not fit-ted very well with the Shkuratov method. Optical constants of aqu-eously altered materials, and more refined model assumptions,would be required to fit the visible range, but the identificationof this kind of compound is beyond the scope of this paper. Irradi-ated products are probably different than the Tholin used, but ourmodels suggest that Titan and Ice Tholins are the best ones to fitthe spectrum among our different Tholins and minerals. These re-sults correspond to more water ice on the surface than what is ex-pected by Guilbert et al. (2009) but are similar to those of Barkumeet al. (2008), in terms of total amount of water ice. Our resultsagree on the presence of large amounts of amorphous water iceand presence of some crystalline water ice, unlike the results ofGuilbert et al. (2009) and Barkume et al. (2008) that suggest largeramounts of crystalline water ice. The reduced v2 is slightly largerwhen only amorphous water ice is used (0.82 for Hapke models)

Table 3Results (% and size in lm) on the composition and physical parameters obtained with sobtained with Shkuratov method (S.). The particles size are given in parentheses. The num

Cr. H2O Olivine Ice T. Titan

1-SL 7 (7) – 5 (50) 6 (20)1-DL 9 (5) – 4 (50) 5 (20)1-S. 21 (4) – – –2-SL – 11 (15) 8 (4) 4 (1)2-DL – 13 (1400) 4 (13) 5 (1.42-S. – 18 (2000) 3 (6) –3-SL 5 (140) – 62 (1240) 16 (1)3-DL 8 (64) 10 (445) 49 (744) 20 (1)3-S. 8 (243) 20 (41) 34 (938) 18 (1)4-SL 6 (38) 33 (37) 24 (360) 37 (1)4-DL – 26 (51) 9 (190) 36 (1)4-S. – 18 (65) 6 (63) 26 (1)5-SL 4 (2000) – 64 (473) 17 (1)5-DL 4 (2000) – 64 (473) 17 (1)5-S. 6 (280) – 66 (299) 12 (1)

a The results have been obtained with the following parameters: 1 – 2003 AZ84: n = 0n = 0.52, B0 = 0.38 or n1 = �0.05, n2 = 0.28, B0 = 0.35, f = 1.0 or Porosity = 0.01. 3 – Ixion4 – 2002 AW197: n = �0.12, B0 = 0 or n1 = �0.46, n2 = 0.21, B0 = 0.9, f = 0.94 or Porosity = 0Porosity = 0.86.

and even more when only crystalline water ice is used (0.95 forHapke models). The models with no water ice have reduced v2 en-hanced (2.32 in the best case with the Hapke model with doubleHG functions). It seems that both crystalline and amorphous statesof water ice are possible (both possibilities are within 99% of con-fidence level). However, new near infrared spectra with higher sig-nal-to-noise level are required to constrain the relative amount ofeach form (see Fig. 3 on the difference of the two models). In thiswork, we do not need any dark and blue components, as used inGuilbert et al. (2009). Abundant water ice, amorphous carbonand small quantities of Tholins appear sufficient to reproduce thegeneral behavior of the spectrum of 2003 AZ84 with Hapke models.Detection of large quantities of water ice, partially in the crystal-line state, can involve heating processes and support the hypothe-sis of Fornasier et al. (2004) on the presence of aqueously alteredmaterials. Nevertheless, the very last data obtained on 2003 AZ84

do not show any absorption feature (Fornasier et al., 2009). So itis possible that 2003 AZ84 has an heterogeneous surfacecomposition.

4.2. Echeclus

The spectrum of Echeclus is noisy in the 0.7–1.4 lm range andappears featureless (see Fig. 4). Our best synthetic spectra are ob-tained with 11% of olivine (15 lm), 8% of Ice Tholin (�4 lm), 4% ofTitan Tholin (1 lm), 13% of Triton Tholin (�6 lm) and 64% ofamorphous carbon (10 lm) with Hapke models, and 18% of olivine(2 mm), 3% of Ice Tholin (6 lm particle size), 3% of amorphouswater ice (2 mm) and 76% of amorphous carbon (10 lm) with Shk-uratov method. Results are slightly different in terms of spectralbehavior (see Fig. 4). The main difference appears on the particlesize required by the models and the possible presence of a smallamounts of water ice with the Shkuratov method. The photometryis not completely in agreement with the reflectance spectrum ofEcheclus. Indeed, the K value from photometry is clearly abovethe reflectance derived from spectroscopy (see Fig. 1). In the pres-ent work, the H and K spectra that are obtained at the same timeare arbitrarily adjusted with the photometry performed in the Hband. Usually, photometry and spectroscopy give similar results(it is true for most objects discussed here), but here, an adjustmentof the H + K spectrum with the K photometry instead of the H pho-tometry would provide a better agreement with the syntheticspectrum obtained with the Hapke model. In this case, the water

ingle HG function (SL) and double HG functions (DL) in the Hapke model and thoseber of degrees of freedom is 252, 250 or 253 for SL, DL and S. models, respectively.a

T. Triton T. Am. H2O Am. C Red. v2

– 52 (14) 30 (10) 0.8138– 57 (11) 25 (10) 0.8142– 27 (24) 52 (10) 1.950813 (6) - 64 (10) 0.7006

) 12 (4.5) – 66 (10) 0.7096– 3 (2000) 76 (10) 0.766617 (2.4) – – 1.058813 (2) – – 1.299620 (2.2) – – 1.3975– – – 1.8396– 5 (144) 24 (10) 1.8261– 7 (138) 43 (10) 1.741915 (2.5) – – 1.627515 (2.3) – – 1.634116 (1.8) – – 2.3443

.25, B0 = 0 or n1 = �0.08, n2 = 0.52, B0 = 0.15, f = 1.0 or Porosity = 0.00. 2 – Echeclus:: n = �0.33, B0 = 0.51 or n1 = �0.41, n2 = 0.28, B0 = 0.38, f = 0.37 or Porosity = 1.00..86. 5 – 1999 DE9: n = �0.09, B0 = 0.51 or n1 = �0.27, n2 = 0.31, B0 = 0.50, f = 0.54 or

Fig. 3. Spectrum of 2003 AZ84 (in gray) with a resolution of 250 and error bars (1r).Continuous and dotted lines represent the synthetic spectra obtained with Hapkemodel with crystalline and amorphous water ice, respectively. The main differenceappears in the near infrared range where the model performed with crystallinewater ice is better in the 1.45–1.8 lm range but worse in the 1.9–2.3 lm range.

Fig. 4. Spectrum of Echeclus (in gray) with a resolution of 250 and error bars at 1r.See previous figure for details of the synthetic spectra (in dashed, continuous anddotted lines).

Fig. 5. Spectrum of Ixion (in gray) with a resolution of 250 and error bars at 1r.Continuous and dotted lines represent the synthetic spectra obtained with theHapke model (single HG function), with single HG function and double HG function,respectively. The dashed line represents the synthetic spectrum obtained with theShkuratov model.

Fig. 6. Spectrum of Ixion (in gray) with a resolution of 250 and error bars at 1r.Continuous and dotted lines represent the synthetic spectra obtained with theHapke model (single HG function), with and without water ice, respectively.

950 F. Merlin et al. / Icarus 208 (2010) 945–954

ice amount could be larger than that predicted in this paper. Differ-ent red slope have been measured in the visible range and reportedby Alvarez-Candal et al. (2008). These variations reinforce the ideathat the object is probably heterogeneous and explain the differenttrends observed between photometry and spectroscopy. Echeclushad experienced a comet activity event on December 30, 2005(Choi and Weissman, 2006). However, our observations do not re-veal any evidence of strong absorption bands of volatiles speciesthat should be redeposited (the diameter of this object being largeenough to accumulate them). The lack of volatiles species (as thelow assumed abundance of the water ice compounds) is in agree-ment with the observations of Rousselot (2008) who measures alow gas-to-dust ratio. However, new observations are required inorder to cover the whole surface of this object and conclude on thisissue.

4.3. Ixion

Except for the 0.7–1.4 lm range of the spectrum which is verynoisy, the noise level is low enough to distinguish the main absorp-tion feature of water ice at 2.0 lm (see Fig. 5). Our best modelswith the Hapke models are obtained with amounts of crystalline

water ice (5–8%), between 34% and 49% of Ice Tholin (with grainslarger than 700 lm), small size particles of Triton Tholin and of Ti-tan Tholin (1–2.5 lm) and possible presence of olivine; see Table 3for details. Both models are very similar to that obtained with theShkuratov method, that suggests 8% of crystalline water ice (parti-cle size of 243 lm), 20% of olivine (37 lm), 24% of Ice Tholin(360 lm), 20% of Triton Tholin (2.2 lm) and 18% of Titan Tholin(1 lm). Models with no water ice can be excluded within 99% con-fidence level with the Hapke model (single HG function). Fig. 6illustrates the need of few amount of water ice to fit the spectrumof Ixion. Models with amorphous water ice are quite goodðv2 ¼ 1:42Þ but outside the 99% confidence level. From these re-sults, water ice, in its crystalline form, is quite probable on the sur-face of Ixion, but new data are required to clearly state on thisissue. We note that bad fitting on the entire wavelength rangecould be attributed to the heterogeneous nature of Ixion. Indeed,the R and K photometric points do not adjust well the reflectancespectrum (see Fig. 1), whereas the K and H spectroscopic rangeshave been obtained simultaneously, and photometry in the R bandis done quasi-simultaneously with photometry in V band. The redslope in the visible may correspond to a different part of the objectthan that observed in the J band on the one hand and in the H and K

Fig. 8. Spectrum of 2002 AW197 (in gray) with a resolution of 250 and error bars at1r. The sharp feature around 2.05 lm is due to bad telluric subtraction. Syntheticspectra obtained with and without water ice are presented in continuous anddotted lines, respectively.

Fig. 9. Spectrum of 1999 DE9 (in gray) with a resolution of 250 and error bars at 1r.See Fig. 5 for a detailed caption.

F. Merlin et al. / Icarus 208 (2010) 945–954 951

bands on the other hand. This could also explain why Barkumeet al. (2008) suspect large amounts of water ice contrary toLicandro et al. (2002) and Boehnhardt et al. (2004).

4.4. 2002 AW197

The spectrum of the object has a low noise level; unfortunately,sky residuals are quite important in the J and K parts (see absorp-tion features at 1.15, 1.4 and 2.07 lm in Fig. 7). Our models lead toa surface composed of large amounts of Tholins (32–61%), olivine(18–33%) and, depending on the models, amorphous carbon. Asmall amount of water ice is included in the different models(see Table 3). The models cannot exclude any of the crystalline oramorphous water ice phase because the reduced v2 are very closefor each kind of model (1.84/1.88 with Hapke model and single HGfunction, 1.83/1.85 with Hapke model and double HG functionsand 1.74/1.87 with Shkuratov method). We note very small differ-ences between the models (see Fig. 7). Synthetic spectra have thesame shape, and the main features, except in the H spectroscopicpart, are mainly due to the use of crystalline or amorphous waterice. The reduced v2 is larger when the models are performed withno water ice (2.21 in the best case), showing that water ice is pos-sible on the surface of this object (see Fig. 8 to note the differenceswith each model) but not certain because all models are outsidethe 99% confidence level. The statistics are limited by the badagreement of our models in the visible range, especially around0.7 lm. This implies that other red analogs are mandatory in thiscase. This result on the possible presence of water ice on the sur-face of this object agrees with previous ones (Doressoundiramet al., 2005; Barkume et al., 2008; Guilbert et al., 2009) where�20% of this ice is also used.

4.5. 1999 DE9

The signal-to-noise level is low in the J band, but much higher inthe rest of the spectrum (see Fig. 9). However, no strong absorptionband is detected. Our best fit models are obtained with similarchemical composition from the different models. They suggestaround 5% of crystalline water ice (large particle size), largeamounts of Ice Tholin (64–66%, �300–470 lm), moderate amountsof small particles of Triton Tholin and Titan Tholin (a total of28–32%). The results are better with Hapke models than with the

Fig. 7. Spectrum of 2002 AW197 (in gray) with a resolution of 250 and error bars at1r. The sharp feature around 2.05 lm is due to bad telluric subtraction. Syntheticspectra obtained with the three methods are very similar (see Fig. 5 for details onthe lines properties).

Shkuratov method, in terms of reduced v2 (see Table 3) and differ-ences between synthetic spectra (see Fig. 9). The reduced v2 in-crease with models with no water ice (1.89, 1.98 and 2.81,respectively for Hapke models and Shkuratov method). From thesenew models, we cannot fully confirm the presence of water ice,suggested by Jewitt and Luu (2001), due to the small amounts de-rived from our models and the lack of significance of the results.Accurate rotational analyses are required to know if the surfaceof 1999 DE9 is ice depleted, as suspected by Alvarez-Candal et al.(2007) or Barkume et al. (2008) or covered by small patches ofwater ice.

4.6. 2003 FY128

The spectrum of 2003 FY128 is the noisiest among the six spectrapresented in this paper. It seems very difficult to detect anyabsorption feature, and models are mandatory if we want to getsome constraints on its surface properties, and particularly to seepossible differences or similarities with other objects. The Hapkemodels propose a surface composed of around 15% of Triton Tholin,around 20% of olivine, 17–18% of Titan Tholin, 7–20% of Ice Tholinand 36–42% of amorphous carbon. Results obtained with the Shk-uratov method confirm this trend, with no more than 5% of crystal-line water ice, 16–19% of Ice Tholin, 7–14% of olivine, 5–15% of

Fig. 10. Spectrum of 2003 FY128 (in gray) with a resolution of 250 and error bars at1r. See Fig. 5 for a detailed caption.

952 F. Merlin et al. / Icarus 208 (2010) 945–954

Triton Tholin, 9–14% of Titan Tholin and 34–61% of amorphous car-bon. In all cases, models need lots of different organics or mineralcompounds to fit the spectrum of 2003 FY128. We note also that alarge variety of grain sizes is required in both cases (from the mi-cron to almost the millimeter, see Table 4). The need for so manyred components with a large difference in particle sizes suggeststhat the object is covered by a complex organic crust. Thus, evenif the visible part is correctly fitted (see Fig. 10), other red com-pound analogs are probably present on the surface of this object.These results show evidence for the need of other chemical com-pounds. If we run the models with the Shkuratov method, the re-duced v2 are slightly increased when no water ice is usedcompared to the case with water ice. Results obtained with Hapkemodels are within 99% confidence level and do not require waterice. In other words, water ice is probably absent from the surfaceof 2003 FY128. Our results obtained with the Shkuratov and Hapkemodels are not so far from those obtained from the near infraredrange by Barkume et al. (2008) or Guilbert et al. (2009), that sug-gest small amounts of water ice on the surface of this object, ifpresent. However, in order to better constrain the surface of2003 FY128, we need to know its albedo, which is not yet con-strained. In this paper, we mainly use a possible albedo of 0.07, de-rived from the work of Lykawka and Mukai (2005) on the albedoand size measurements and the absolute magnitude of this object(5.0). This value is also the typical geometric albedo for all of theTNOs and Centaurs (6.9–8.0%, see Stansberry et al., 2008). In orderto derive some limits on the water ice amount, we also report inTable 4 the results using an albedo of 0.14, twice the initial as-sumed albedo, that seems a realistic upper limit, but water ice isstill not required with Hapke models, within 99% confidence level.

5. Main results and discussion

Hapke and Shkuratov models usually give, at the first order,similar results on the amounts and particle sizes of the differentspecies (see Table 3). However, a few disagreements come from,among other things, the treatment of the phase function (see pre-vious section), and the fact that the volume absorption coefficient,which depends on the particle size S and the wavelength l, isslightly different (equal to ð8=3ÞnkpS=l in the Hapke (1981) modeland 4kpS=l in the Shkuratov et al. (1999) model, which gives somedifferences when the real part of the refractive index n is very dif-ferent than 3/2).

Usually, the v2 obtained with Hapke models are better thanthose obtained with Shkuratov method, except in the case of theobject 2002 AW197. This result is mainly due to the use of more freeparameters in the models of Hapke. New optical constants of darkand red organic material are required to improve the accuracy ofthe models, because particles too tiny compared to the geometricoptic application, are often used (especially for Titan Tholin parti-cles). Echeclus and 2003 FY128 seem depleted of water ice (less

Table 4Results (% and size in micron) on the surface composition of 2003 FY128 obtained with twparticles size are given in parentheses.

Red. v2 Cr. H2O Olivine Triton Tholin

Hapke, single HG function, n = 0.15 (�0.11*) and B0 = 1.01.1869 0 21 (2000) 14 (3.4)1.1644* 0 18 (1800) 16 (5.4)

Hapke, double HG function, n1 = �0.22 (�0.40*), n2 = 0.27 (0.26*), B0 � 0.1 and f = 0.51.1917 0 20 (2000) 14 (3.4)1.1684* 0 20 (2000) 16 (5.6)

Shkuratov, porosity = 0.91.4366 4 (46) 7 (236) 5 (3.5)1.3045* 5 (1700) 14 (152) 15 (5.8)1.4584 None 7 (198) 5 (3.5)

than a few percent), while 1999 DE9 could be covered by smallamounts of water ice (with particle sizes close to the millimeter).It is probable that water ice is present on Ixion and possible on2002 AW197. For these objects, the models cannot suggest any gen-eral trend on the state of the water ice. New observations withhigher signal-to-noise levels are required to firmly identify theform of the water ice on these objects. However, we confirm thepresence of large amounts of water ice (probably in both states)on the surface of 2003 AZ84 (Barkume et al., 2008; Guilbert et al.,2009). A balance between the amorphous and crystalline statesof water ice could agree with the work of Zheng et al. (2009) basedon amorphization and recrystallization processes, or that ofGil-Hutton et al. (2009) on irradiation and collisions, for instance.However, other phenomena and parameters can play a role onthe armorphization and/or recrystallization rates. For instance,the work of Zheng et al. (2009) does not take into account otherradiation sources, such as the effects of solar wind protons, inter-stellar neutrals, cosmic rays, etc. that also play an important role(Strazzulla et al., 2003). Cryo-volcanism phenomenona (Jewittand Luu, 2001; Cook et al., 2006) cannot explain the crystallinewater ice because this object is too small (diameter lower than685 km, which is below the limit for cryo-volcanism, seeStansberry et al., 2008; McKinnon et al., 2008).

Our results also suggest the presence of particles with very smallsizes on the surface of some of the objects presented in this work.From laboratory measurements, Hapke et al. (1997) showed thatnumerous small grain sizes of the order of the wavelength range, clo-sely spaced together, seem to be acting like larger grain sizes in thevisible and in the near infrared range. This could mean that if highred slopes come from small particles sizes, according to the models,these tiny particles are probably not so closely spaced. Grundy

o different albedos; 0.07 and 0.14 (marked with �), and for each kind of model. The

Titan Tholin Ice Tholin Amorphous carbon

17 (1) 7 (8.7) 41 (10)18 (1) 10 (6.7) 38 (10)

4 (0.37*)17 (1) 7 (8.7) 42 (10)18 (1) 10 (6.5) 36 (10)

9 (1) 16 (3) 59 (10)14 (1) 18 (2.3) 34 (10)8 (1) 19 (2.7) 61 (10)

F. Merlin et al. / Icarus 208 (2010) 945–954 953

(2009) shows that small organic compounds, scattered in a water icematrix, give reflectance spectra with redder slopes and shallowerfeatures in the near infrared (which is the case here) than expectedwhen particles are homogeneously mixed with the water ice. Thehypothesis that the red slope comes from small particle sizes is alsosupported by other laboratory experiments. For instance, in Fig. 4 ofde Bergh et al. (2008), huge red slopes are obtained with organiccompounds with the smallest grain size. From thin moderately irra-diated samples (of the order of the micrometer), Brunetto et al.(2006) also obtained huge red slopes in the visible range.

Irradiation products depend on the irradiated sources andamounts (Strazzulla et al., 2003). Even if the strongest variationson the irradiation products are expected beyond 80 AU, Stansberryet al. (2008) note that the red slope of most objects is correlatedwith the albedo, which is itself correlated with the perihelion dis-tance. On the other hand, irradiated surfaces can be refreshed par-tially or completely by random non-disruptive collisions or icecondensation. All of these phenomena contribute to the colordiversity of this population, along with possibly different originalcompositions. Our results propose that the reddest TNOs couldbe covered with a thin layer of organic compounds or with largeramounts of very small grain, confirming previous results obtainedon Centaurs (Cruikshank et al., 1998; Poulet et al., 2002). This sug-gests that the surface of these objects has experienced refreshmentprocesses or/and is moderately irradiated. Indeed, the dark objectswith neutral or small red slopes, are mainly covered by larger par-ticles or neutral carbon material, assumed to be a result of heavilyirradiated material (see Strazzulla et al. (2003) on the carbonenrichment of irradiated products). Our observations propose thatthe spectral behavior is sensitive to the size and arrangement ofthe particles in the bulk, even if the importance of the red slopecan be due to the chemical composition itself. Indeed, Titan andTriton Tholin spectra have different behaviors in the visible, evenif the chemical composition is quite similar (while the first one isformed from a gaseous mixture made of 90% of nitrogen and 10%of methane, the second one is formed from a mixture composedof 99% of nitrogen and 1% of methane). Moreover, physical con-straints, like the initial atmospheric pressure in the case of the Tho-lin formation, are important. Imanaka et al. (2004) note thatTholins formed at low pressure are red and brown, while they tendto be more yellow when they are synthesized at high pressure.

2003 AZ84 is a gray object and has a relative high albedo. Simi-larly to Orcus (Barucci et al., 2008b), it is one of the TNOs that defythe color albedo trend (correlation between red color and higheralbedo, found by Stansberry et al. (2008)). Both exhibit strongwater ice absorption features, and these objects have perhapsexperienced a collisional event, even if cryo-volcanism processesare not excluded for the large TNO Orcus. These objects whichare partly covered by water ice, are binary objects (Noll et al.,2008). Other binary objects, such as 47171 and Quaoar (Stansberryet al., 2005; Noll et al., 2008) are also covered by water ice (detec-tion of water ice performed by for instance, Dotto et al., 2003b;Merlin et al., 2005; Protopapa et al., 2009; Schaller and Brown,2007b). It is possible that binaries have more water ice, whichseems to be better preserved or renewed on the surface of theseobjects. This could partly explain why binaries seem to be brighterthan the single TNOs (see, Lykawka and Mukai (2005), who men-tion that all discovered binaries have albedos greater than 0.04),but new observations are required to deeply investigate this ideaand search for the related mechanisms.

6. Conclusion

In this paper, we presented surface properties of six objects,that are representative of relatively dark or moderately bright

TNOs and Centaurs, whatever their dynamical properties. We pro-vided information on the chemical composition of their surfaceand on the physical processes that may govern their surface evolu-tion. We mainly confirm previous results (Barkume et al., 2008;Guilbert et al., 2009) performed on more restricted wavelengthrange and without albedo constraints. The results obtained on2003 AZ84 could agree with the work of Zheng et al. (2009) orGil-Hutton et al. (2009) to explain the observation of crystallinewater ice, but new laboratory measurements are needed to dealwith other radiation sources. New observations must be performedat higher signal-to-noise level to better constrain the ratio betweenamorphous and crystalline water ice, especially for 2003 AZ84 andIxion where the presence of water is probable. Our work showsthat Shkuratov et al. (1999) and Hapke (1981) models give similarresults and can therefore be used without preference. Finally weshow that phase functions, when unknown, can be simplified bysingle HG functions for dark objects instead of a double HG func-tion that depends on three parameters, but models refinementsare required.

Acknowledgments

We are very grateful to the two anonymous referees for theirvaluable comments that contributed to the improvement of thispaper.

References

Alvarez-Candal, A., Barucci, M.A., Merlin, F., Guilbert, A., de Bergh, C., 2007. A searchfor rotational variations on trans-neptunian objects. Astron. Astrophys. 475,369–374.

Alvarez-Candal, A., Fornasier, S., Barucci, M.A., de Bergh, C., Merlin, F., 2008. Visiblespectroscopy of the new ESO Large Program on trans-neptunian objects andCentaurs. Part 1. Astron. Astrophys. 487, 741–748.

Barkume, K.M., Brown, M.E., Schaller, E.L., 2008. Near-infrared spectra of Centaursand Kuiper Belt Objects. Astron. J. 135, 55–67.

Barucci, M.A., de Bergh, C., Cuby, J.-G., Le Bras, A., Schmitt, B., Romon, J., 2000.Infrared spectroscopy of the Centaur 8405 Asbolus: First observations at ESO–VLT. Astron. Astrophys. 357, L53–L56.

Barucci, M.A., Brown, M.E., Emery, J.P., Merlin, F., 2008a. Composition and SurfaceProperties of Transneptunian Objects and Centaurs. The Solar System beyondNeptune. University of Arizona Press, Tucson. pp. 143–160.

Barucci, M.A., Merlin, F., Guilbert, A., de Bergh, C., Alvarez-Candal, A., Hainaut, O.,Doressoundiram, A., Dumas, C., Owen, T., Coradini, A., 2008b. Surfacecomposition and temperature of the TNO Orcus. Astron. Astrophys. 479, L13–L16.

Belskaya, I.N., Levasseur-Regourd, A.-C., Shkuratov, Y.G., Muinonen, K., 2008.Surface properties of Kuiper Belt Objects and Centaurs from photometry andpolarimetry. The Solar System beyond Neptune. University of Arizona Press,Tucson. pp. 115–127.

Boehnhardt, H., Bagnulo, S., Muinonen, K., Barucci, M.A., Kolokolova, L., Dotto, E.,Tozzi, G.P., 2004. Surface characterization of 28978 Ixion (2001 KX76). Astron.Astrophys. 415, L21–L25.

Brown, M.E., Trujillo, C.A., Rabinowitz, D.L., 2005. Discovery of a planetary-sizedobject in the scattered Kuiper Belt. Astrophys. J. 635, L97–L100.

Brown, M.E., Barkume, K.M., Ragozzine, D., Schaller, E.L., 2007. A collisional family oficy objects in the Kuiper Belt. Nature 446, 294–296.

Brunetto, R., Barucci, M.A., Dotto, E., Strazzulla, G., 2006. Ion irradiation of frozenmethanol, methane, and benzene: Linking to the colors of Centaurs and trans-neptunian objects. Astrophys. J. 644, 646–650.

Choi, Y.J., Weissman, P., 2006. Discovery of cometary activity for Centaur 174P/Echeclus (60558). Bull. Am. Astron. Soc. 38, 551.

Cook, J.C., Desch, S.J., Roush, T.L., Trujillo, C.A., Geballe, T.R., 2006. Near-infraredspectroscopy of charon: Possible evidence for cryovolcanism on Kuiper BeltObjects. Astrophys. J. 663, 1406–1419.

Cruikshank, D.P., and 12 colleagues, 1998. The composition of Centaur 5145 Pholus.Icarus 135, 389–407.

Davies, J.K., McFarland, J., Bailey, M.E., Marsden, B.G., Ip, W.-H., 2008. The EarlyDevelopment of Ideas Concerning the Transneptunian Region. The Solar Systembeyond Neptune. University of Arizona Press, Tucson. pp. 11–23.

de Bergh, C., Boehnhardt, H., Barucci, M.A., Lazzarin, M., Fornasier, S., Romon-Martin, J., Tozzi, G.P., Doressoundiram, A., Dotto, E., 2004. Aqueous alteredsilicates at the surface of two Plutinos? Astron. Astrophys. 416, 791–798.

de Bergh, C., Schmitt, B., Moroz, L.V., Quirico, E., Cruikshank, D.P., 2008. LaboratoryData on Ices, Refractory Carbonaceous Materials, and Minerals Relevant toTransneptunian Objects and Centaurs, The Solar System beyond Neptune.University of Arizona Press, Tucson. pp. 483–506.

954 F. Merlin et al. / Icarus 208 (2010) 945–954

DeMeo, F.E., Fornasier, S., Barucci, M.A., Perna, D., Protopapa, S., Alvarez-Candal, A.,Delsanti, A., Doressoundiram, A., Merlin, F., de Bergh, C., 2009. Visible and near-infrared colors of transneptunian objects and Centaurs from the second ESOLarge Program. Astron. Astrophys. 493, 283–290.

Doressoundiram, A., Barucci, M.A., Tozzi, G.P., Poulet, F., Boehnhardt, H., de Bergh, C.,Peixinho, N., 2005. Spectral characteristics and modeling of the trans-neptunianobject (55565) 2002 AW197 and the Centaurs (55576) 2002 GB10 and (83982)2002 GO9: ESO Large Program on TNOs and Centaurs. Planet. Space Sci. 53,1501–1509.

Doressoundiram, A., Boehnhardt, H., Tegler, S.C., Trujillo, C., 2008. Color Propertiesand Trends of the Transneptunian Objects. The Solar System beyond Neptune.University of Arizona Press, Tucson. pp. 91–104.

Dotto, E., Barucci, M.A., de Bergh, C., 2003a. Surface composition of TNOs andCentaurs: Visible and near-infrared spectroscopy. Comptes rendus del’academie des sciences 4, 775–782.

Dotto, E., Barucci, M.A., Boehnhardt, H., Romon, J., Doressoundiram, A., Peixinho, N.,de Bergh, C., Lazzarin, M., 2003b. Searching for water ice on 47171 1999 TC36,1998 SG35, and 2000 QC243: ESO Large Program on TNOs and Centaurs. Icarus162, 408–414.

Dotto, E., Perna, D., Barucci, M.A., Rossi, A., de Bergh, C., Doressoundiram, A.,Fornasier, S., 2008. Rotational properties of Centaurs and trans-neptunianobjects. Lightcurves and densities. Astron. Astrophys. 490, 829–833.

Dumas, C., Merlin, F., Barucci, M.A., de Bergh, C., Hainault, O., Guilbert, A., Vernazza,P., Doressoundiram, A., 2007. Surface composition of the largest dwarf planet136199 Eris (2003 UB313). Astron. Astrophys. 471, 331–336.

Emery, J.P., Brown, R.H., 2004. The surface composition of Trojan asteroids:Constraints set by scattering theory. Icarus 170, 131–152.

Fornasier, S., Doressoundiram, A., Tozzi, G.P., Barucci, M.A., Boehnhardt, H., deBergh, C., Delsanti, A., Davies, J., Dotto, E., 2004. ESO Large Program on physicalstudies of trans-neptunian objects and Centaurs: Final results of the visiblespectrophotometric observations. Astron. Astrophys. 421, 353–363.

Fornasier, S., and 10 colleagues, 2009. Visible spectroscopy of the new ESO LargeProgram on trans-neptunian objects and Centaurs: Final results. Astron.Astrophys. 508, 457–465.

Gil-Hutton, R., 2002. Color diversity among Kuiper Belt Objects: The collisionalresurfacing model revisited. Planet. Space Sci. 50, 57–62.

Gil-Hutton, R., Licandro, J., Pinilla-Alonso, N., Brunetto, R., 2009. The trans-neptunianobject size distribution at small sizes. Astron. Astrophys. 500, 909–916.

Gladman, B., Marsden, B.G., Vanlaerhoven, C., 2008. Nomenclature in the OuterSolar System. The Solar System beyond Neptune. University of Arizona Press,Tucson. pp. 43–57.

Grundy, W.M., 2009. Is the missing ultra-red material colorless ice? Icarus 199,560–563.

Grundy, W.M., Schmitt, B., 1998. The temperature-dependent near-infraredabsorption spectrum of hexagonal H2O ice. J. Geophys. Res. 103, 25809–25822.

Guilbert, A., Alvarez-Candal, A., Merlin, F., Barucci, M.A., Dumas, C., de Bergh, C.,Delsanti, A., 2009. ESO-Large Program on TNOs: Near-infrared spectroscopywith SINFONI. Icarus 201, 272–283.

Hapke, B., 1981. Bidirectional reflectance spectroscopy. 1. Theory. J. Geophys. Res.86, 4571–4586.

Hapke, B., 1984. Bidirectional reflectance spectroscopy. 3. Correction formacroscopic surface roughness. Icarus 59, 41–59.

Hapke, B., 1986. Bidirectional reflectance spectroscopy. IV – The extinctioncoefficient and the opposition effect. Icarus 67, 264–280.

Hapke, B., 1993. Theory of Reflectance and Emittance Spectroscopy. Topics inRemote Sensing. Cambridge University Press, c1993, Cambridge, UK.

Hapke, B., 2008. Bidirectional reflectance spectroscopy. 6. Effects of porosity. Icarus195, 918–926.

Hapke, B., Dimucci, D., Nelson, R., Smythe, W., 1997. Lunar Planet. Inst. Conf. 28,513–514 (Abstracts).

Henyey, L.G., Greenstein, J.L., 1941. Diffuse radiation in the Galaxy. Astrophys. J. 93,70–83.

Hillier, J.K., 1993. Voyager photometry of Triton. Ph.D. Thesis.Imanaka, H., Khare, B.N., Elsila, J.E., Bakes, E.L.O., McKay, C.P., Cruikshank, D.P.,

Sugita, S., Matsui, T., Zare, R.N., 2004. Laboratory experiments of Titan Tholinformed in cold plasma at various pressures: Implications for nitrogen-containing polycyclic aromatic compounds in Titan haze. Icarus 168, 344–366.

Jewitt, D.C., Luu, J.X., 2001. Colors and spectra of Kuiper Belt Objects. Astron. J. 122,2099–2114.

Jewitt, D.C., Luu, J., 2004. Crystalline water ice on the Kuiper Belt Object (50000)Quaoar. Nature 432, 731–733.

Khare, B.N., Sagan, C., Ogino, H., Nagy, B., Er, C., Schram, K.H., Arakawa, E.T., 1986.Amino acids derived from Titan Tholins. Icarus 67, 176–184.

Khare, B.N., Thompson, W.R., Cheng, L., Chyba, C., Sagan, C., Arakawa, E.T., Meisse, C.,Tuminello, P.S., 1993. Production and optical constraints of Ice Tholin fromcharged particle irradiation of (1:6) C2H6/H2O at 77 K. Icarus 103, 290–300.

Licandro, J., Ghinassi, F., Testi, L., 2002. Infrared spectroscopy of the largest knowntrans-neptunian object 2001 KX75. Astron. Astrophys. 388, L9–L12.

Licandro, J., Pinilla-Alonso, N., Pedani, M., Oliva, E., Tozzi, G.P., Grundy, W.M., 2006.The methane ice rich surface of large TNO 2005 FY9: A Pluto-twin in the trans-neptunian belt? Astron. Astrophys. 445, L35–L38.

Lykawka, P.S., Mukai, T., 2005. Higher albedos and size distribution of largetransneptunian objects. Planet. Space Sci. 53, 1319–1330.

Mastrapa, R.M., Bernstein, M.P., Sandford, S.A., Roush, T.L., Cruikshank, D.P., Ore,C.M.D., 2008. Optical constants of amorphous and crystalline H2O–ice in thenear infrared from 1.1 to 2.6 lm. Icarus 197, 307–320.

McKinnon, W.B., Prialnik, D., Stern, S.A., Coradini, A., 2008. Structure and Evolutionof Kuiper Belt Objects and Dwarf Planets. The Solar System beyond Neptune.University of Arizona Press, Tucson. pp. 213–241.

Merlin, F., Barucci, M.A., Dotto, E., de Bergh, C., Lo Curto, G., 2005. Search for surfacevariations on TNO 47171 and Centaur 32532. Astron. Astrophys. 444, 977–982.

Merlin, F., Guilbert, A., Dumas, C., Barucci, M.A., de Bergh, C., Vernazza, P., 2007.Properties of the icy surface of the TNO 136108 (2003 EL61). Astron. Astrophys.466, 1185–1188.

Merlin, F., Alvarez-Candal, A., Delsanti, A., Fornasier, S., Barucci, M.A., DeMeo, F.E., deBergh, C., Doressoundiram, A., Quirico, E., Schmitt, B., 2009. Stratification ofMethane ice on Eris’ surface. Astron. J. 137, 315–328.

Mishchenko, M.L., Travis, L.D., Lacis, A.A., 2006. Multiple Scattering of Light byParticles. Cambridge University Press.

Noll, K.S., Grundy, W.M., Chiang, E.L., Margot, J-L., Kern, S.D., 2008. Binaries in theKuiper Belt. The Solar System beyond Neptune. University of Arizona Press,Tucson. pp. 345–363.

Piatek, J.L., Hapke, B.W., Nelson, R.M., Smythe, W.D., Snyder-Hale, A., 2004.Scattering properties of planetary regolith analogs. Icarus 171, 531–545.

Pinilla-Alonso, N., Licandro, J., Gil-Hutton, R., Brunetto, R., 2007. The water ice richsurface of (145453) 2005 RR43: A case for a carbon-depleted population ofTNOs? Astron. Astrophys. 468, L25–L28.

Pinilla-Alonso, N., Brunetto, R., Licandro, J., Gil-Hutton, R., Roush, T.L., Strazzulla, G.,2009. The surface of (136108) Haumea (2003 EL61), the largest carbon-depletedobject in the trans-neptunian belt. Astron. Astrophys. 496, 547–556.

Poulet, F., Cuzzi, J.N., Cruikshank, D.P., Roush, T., Dalle Ore, C.M., 2002. Comparisonbetween the Shkuratov and Hapke scattering theories for solid planetarysurfaces: Application to the surface composition of two Centaurs. Icarus 160,313–324.

Protopapa, S., Alvarez-Candal, A., Barucci, M.A., Tozzi, G.P., Fornasier, S., Delsanti, A.,Merlin, F., 2009. ESO Large Program about transneptunian objects: Surfacevariations on (47171) 1999 TC36. Astron. Astrophys. 501, 375–380.

Rousselot, P., 2008. 174P/Echeclus: A strange case of outburst. Astron. Astrophys.480, 543–550.

Schaller, E.L., Brown, M.E., 2007a. Volatile loss and retention on Kuiper Belt Objects.Astrophys. J. 659, L61–L64.

Schaller, E.L., Brown, M.E., 2007b. Detection of methane on Kuiper Belt Object(50000) Quaoar. Astrophys. J. 670, L49–L51.

Sheppard, S.S., 2007. Light curves of dwarf plutonian planets and other large KuiperBelt Objects: Their rotations, phase functions, and absolute magnitudes. Astron.J. 134, 787–798.

Shkuratov, Y., Starukhina, L., Hoffmann, H., Arnold, G., 1999. A model of spectralalbedo of particulate surfaces: Implications for optical properties of the Moon.Icarus 137, 235–246.

Stansberry, J.A., Cruikshank, D.P., Grundy, W.G., Margot, J.L., Emery, J.P., Fernandez,Y.R., Rieke, G.H., 2005. Bull. Am. Astron. Soc. 37, 737.

Stansberry, J., Grundy, W., Brown, M., Cruikshank, D., Spencer, J., Trilling, D., Margot,J.-L., 2008. Physical Properties of Kuiper Belt and Centaur Objects: Constraintsfrom the Spitzer Space Telescope. The Solar System beyond Neptune. Universityof Arizona Press, Tucson. pp. 161–179.

Strazzulla, G., Cooper, J.F., Christian, E.R., Jonhson, R.E., 2003. Ion irradiation ofTNOs: From the fluxes measured in space to the laboratory experiments.Comptes rendus de l’academie des sciences 4, 791–802.

Tegler, S.C., Bauer, J.M., Romanishin, W., Peixinho, N., 2008. Colors of Centaurs. TheSolar System beyond Neptune. University of Arizona Press, Tucson. pp. 105–114.

Trujillo, C.A., Brown, M.E., Barkume, K.M., Schaller, E.L., Rabinowitz, D.L., 2007. TheSurface of 2003 EL61 in the near-infrared. Astrophys. J. 655, 1172–1178.

Zheng, W., Jewitt, D., Kaiser, R.I., 2009. On the state of water ice on Saturn’s moonTitan and implications to icy bodies in the outer Solar System. J. Phys. Chem.113, 11174–11181.

Zubko, V.G., Mennella, V., Colangeli, L., Bussoletti, E., (1996). Optical Constants ofAmorphous Carbon Extracted from Recent Laboratory ExtinctionMeasurements, The Role of Dust in the Formation of Stars, ESO AstrophysicsSymposia, p. 333.