The Rest of the Story
Data Reduction
The spectra were then shifted to place 2003 EL61 in the center of the array, and individual spectra at a similar air mass were summed to maximize the signal from the satellite. For comparison, the spectrum of 2003 EL61 was extracted in the same way as the spectrum of the outer satellite.
Discussion
The poor fit of the water ice model short to 1.2 µm is probably due to the uncertainties in the model in this region (Grundy & Schmitt 1998). The absorption features in the spectrum of the satellite are significantly deeper than water ice features typically observed on KBOs.
Observations & Data Reduction
No classical cold objects were observed due to the lack of bright targets in this population. The signal to noise was low in the individual spectra and therefore multiple spectra were added together to obtain higher signal to noise. Multiple spectra of a single target were summed together, weighted by their noise to produce a final spectrum.
We examined individual spectra to look for substantial variability with time or rotation phase. We did not detect any significant variation in the spectra of any object, although small changes are undetectable due to noise. High signal-to-noise spectra were obtained for most objects with an average total integration time of 6000 seconds.
Analysis
With these problems in mind, we do not attempt to make exact surface models of the KBO in our survey. Much more can be learned by parameterizing the spectra of the entire population and comparing the relative strengths of the spectral signatures of surface water ice using Hapke models. To use the full information available in our spectra, we create spectral models of the surface reflectance between 1.4 and 2.4µm using the equations of radiative transfer developed by Hapke (1993).
A linear reflectance model of water ice and a continuum component was fitted to each spectrum and solved for the fraction of water in the spectrum, and m, the slope of the continuum component. To obtain better estimates of the true errors, we selected objects in our survey whose spectra appeared consistent with no water ice by visual inspection. The magnitudes and albedos are taken from Stansberry et al. 2007) and are reproduced in Table 3.6 together with the orbital elements of the objects in our survey.
Results
The 2003 EL61 Collisional Family
In addition to the absorption features of the deep water, the family members have uniform neutral color in the visible. The family members' model fits revealed that another blue continuum component significantly reduced the minimum χ2 value with the exception of OP32 from 2003. As such, the blue continuum component appears to be a real feature in the spectra for most of the family members.
The presence of crystalline water ice in members of the family seems to rule out thermal processes as the main crystallization mechanism. Larger grain sizes in family members may be consistent with water ice surfaces where crystals have grown due to less impurities (Brown et al., 2007a). The large fraction of water ice detected in the family members may indicate that proto 2003 EL61 was differentiated by material derived primarily from an icy mantle.
KBOs and Centaurs
Specifically, we find that the fraction of water ice in the spectrum can vary greatly among the objects in a particular group, and that the groups have similar distributions with the exception of the BB group. We assume properties for the water ice and the characteristic component based on those observed on icy bodies. We assumed a 60% albedo for the visible part of the water ice reflectance and a grain size of 50 µm was used.
Intimate mixture models were generated by linear mixing of the albedo of individual diffuse water ice and carbon-rich material. However, we do not detect a relationship between size and water ice fraction in our sample of Centaurs. This suggests that the processes controlling the presence of water ice on the Centauri surface could be different from those on the KBO.
Discussion
See Table 3.6 fraction of water ice detected in the spectrum and the other fit results for the spectral models. Once the 2003 EL61 collision family is excluded, no correlation is found between the fraction of water ice detected in the spectrum of KBOs and Centaurs and their visible colors. For the remaining groups, RB (red-blue), IR (intermediate red) and RR (red), we find no correlation with the fraction of water ice detected in the spectrum.
We find no correlation between the fraction of water ice detected in the spectrum and the albedo measured for the object. The thick gray line represents the expected albedo for a given surface fraction of water ice and is determined from our two-component synthetic surface model. We find a correlation between the size and extent of detected water ice for KBOs.
Introduction
Mineral absorptions in the blue-UV wavelength region as well as broad absorptions at 7000 ˚A have been identified in C-type asteroids and a number of its sub-classes. -type asteroids are found in the outer asteroid belt and are thought to contain water ice, although their surfaces appear dehydrated (Lazzarin et al., 1995). Finally, methane ices O2 and O3 have also been detected in the visible spectra of icy solar system bodies.
Both species have been detected on Ganymede while O2 is also detected on Callisto and Europa (Spencer et al., 1995; Spencer and Calvin, 2002). So far, only a handful of visible spectra of KBOs have been published (see Barucci et al. 2008)) giving a limited view of the spectral visible character of KBOs. With few exceptions, they show largely few spectroscopic features and only a small number of these objects are observed at wavelengths shorter than 5000 ˚A. A U-band photometric survey of KBOs conducted by Jewitt et al. 2007) provides information on the spectral properties in the region.
The Sample
Observations & Data Analysis
Observations
Stars dimmer than magnitude 16 were also excluded, as they are too dim to be seen reliably in the slit images. To observe the KBO, the USNO astrometric standard was first placed in the slot, and then the telescope was moved to the KBO position. During the observation, the telescope tracked the KBO as it moved, providing another check that the object in the slit was the intended target for the KBO near the opposition.
In general, the observations were long enough for background field stars to move out of the slit. The star was also used to jiggle the telescope along the slit by 10-20 arcseconds between observations. During observations, the slit was maintained along the parallactic angle, where the parallactic angle is the angle perpendicular to the horizon.
Data Analysis
Corrections to the curvature in the spatial direction were centered on the target so that the area around the target is least changed by the corrective shifts in the spectral direction. In general this leads to a good correction of the background sky, although telluric OH lines become stronger in the red and increase the noise in our data for the region beyond 7000 ˚A. The final spectrum is extracted by summing the signal from the target in the spatial direction, typically over 6 to 18 pixel strips, depending on the viewing conditions.
The spectral images of the lamps were corrected in the same way as the target data and a lamp spectrum was extracted from the same spatial location as the target. The error in the wavelength calibration was found by measuring the wavelength of bright telluric Hg and OH lines and comparing them with reported values. The final spectra have a higher resolution than most of the spectral features we expect to find in the optics, where typical absorptions are hundreds of Angstroms wide.
Results
Methane Giants
This is about half the grain size reported by Brown et al. Shifts in the position of the methane bands have been previously detected for Pluto and Eris (Dout´e et al., 1999; Licandro et al., 2006b). Detecting the differences in shifts at different wavelengths can also provide information about compositional gradients on the surfaces of the KBOs as different bands probe different depths in the ice (Licandro et al., 2006b).
Ethane was also observed in the NIR spectrum in 2005 FY9, and higher-order hydrocarbons are also expected (Brown et al., 2007a), which may be responsible for the reversal. In addition to their spectral characteristics, these objects were also found to have orbits very similar to those of 2003 EL61 (Brown et al., 2007b). A second blue NIR component has been reported to increase the fit for almost all fragments (see Chapter 3, Licandro et al. 2007)), allowing for the possibility that the surfaces of the fragments and 2003 EL61 are not pure water ice.
Discussion
The higher noise at 5890 ˚A and 7600 ˚A is due to poor correction of the Na doublet line and the O2Fraunhofer lines. The gradient of the spectral slope in our data is similar to those reported in photometric studies. A line fit to the red data is used to scale the blue data to match the red data. The gradient of the spectral slope in our data is similar to those reported in photometric studies.
The higher noise at 5890 ˚A and 7600 ˚A is due to weak correction of the Na double line and O2 Fraunhofer lines. A negative value for the blue gradient indicates an upward curve or a concave upward departure in the slope of the spectrum. We also find no significant correlation between the blue and visible gradients, although the noise is high in the blue gradient measurements.
Data represent a smoothed mean of the original data with formal 1-sigma error bars indicated. The ESO Major Program for Physical Studies of Trans-Neptunian Objects and Centaurs: Final Results of Visible Spectrophotometric Observations.
