A close look at Saturn's rings with Cassini VIMS (original) (raw)

Introduction

Saturn's rings provided one of the earliest successful applications of ground-based near-infrared spectroscopy with the discovery that the rings are composed of—or covered by—H2O in the form of ice or frost (Kuiper, 1952, Kuiper, 1957; Moroz, 1967). In his chapter in ‘The Atmospheres of the Earth and Planets’ (second ed., 1957, pp. 364–365) Kuiper writes, “Of special interest is the question of the composition of the rings of Saturn. The writer had expected to find a neutral reflection spectrum such as that from rock, although the high visual albedo of the rings might have led one to suspect otherwise. The reflection spectrum appears to be very similar to that of the polar cap of Mars; the water-cell equivalent of about 2/3 mm. It is provisionally concluded that the rings are covered by frost, if not composed of ice.” Although this conclusion was initially based simply on the depression of the near-infrared spectrum in the region beyond 1.5 μm, by 1970 the two strong water ice bands centered at 1.5 and 2.0 μm had been spectrally resolved and the identification clinched (Kuiper et al., 1970, Pilcher et al., 1970, Lebofsky et al., 1970).

While the work cited above was based on simple comparisons of 1–3 μm telescopic spectra with laboratory spectra of H2O ice at various temperatures, Pollack et al. (1973) published the first models of the reflectance spectra of the rings, using scattering theory and the optical properties of water ice, concluding that typical particle radii were in the 25–125 μm range.

The 1970s and 1980s also saw the first microwave observations of the rings, motivated by their surprisingly strong detection in radar observations at 13 cm wavelength (Goldstein and Morris, 1973, Goldstein et al., 1977, Ostro et al., 1980). Because the rings had never been convincingly detected in microwave emission, the large radar return quickly led to the conclusion that the ring particles had to be on the order of the wavelength in size (Pollack et al., 1973, Pettengill and Hagfors, 1974, Pollack, 1975). Subsequent radiometric observations of finite, but very low, ring brightness temperatures (see Esposito et al., 1984 for a review) could be reconciled with the strong radar returns only if the ring particles were not only wavelength-size, with power-law size distributions, but also made of nearly pure water ice (Cuzzi and Pollack, 1978, Cuzzi et al., 1980, Epstein et al., 1984). Grossman (1990) estimated the fraction of silicate material within the rings to be 1% or less, on the assumption that it is uniformly distributed within the ring particles.

These conclusions were reinforced by the Voyager radio occultation experiment (Tyler et al., 1983) from which Marouf et al. (1983) and Zebker et al. (1985) derived particle size distributions for optically thin parts of the A and C rings and the Cassini Division which spanned the range 1 cm to 10 m. A power-law size distribution with an incremental index, q≃3 and a lower cutoff of 1 cm was found to fit these observations, with relatively few particles exceeding a radius of ∼5 m. Observations of Voyager and ground-based stellar occultations (Showalter and Nicholson, 1990, French and Nicholson, 2000) have confirmed that the optical cross-section of the rings is also dominated by particles in the 1 cm to 10 m size range, while analysis of Voyager photometry for the A ring showed a negligible contribution by micron-sized ‘dust’ to scattered light at optical wavelengths, even at high phase angles (Dones et al., 1993).

Clark and McCord (1980) published new 1.0–2.5 μm spectra of the rings and summarized the existing observational data, including ultraviolet and visual spectra by Lebofsky and Fegley (1976) as well as observations by Puetter and Russell (1977) which extended the spectral coverage of ground-based infrared data to 4.1 μm. This revealed the strong fundamental absorption band of water ice centered at ∼2.9 μm, as well as the Fresnel reflectance peak at 3.1 μm and a broad peak in reflectance at 3.6 μm. Model fits to these data by Clark et al. (1986), using the methods of Clark and Roush (1984), showed that the 3.6 μm peak—which is defined by strong ice absorptions at 2.9 and ∼4.5 μm—is quite sensitive to grain size. They obtained an average grain size of 30 μm, but this was now understood as referring to the frosty regolith which must cover the individual decimeter-sized ring particles. A similar conclusion was reached by Doyle et al. (1989), based on both the 3.1 and 3.6 μm peaks, who estimated that rg=20μm.

However, pure water ice should have a flat visible spectrum down to ∼0.3 μm, whereas the average ring spectrum exhibits a strong red slope shortward of 0.6 μm [see the summary by Esposito et al. (1984) or Karkoschka (1994) for a more recent HST spectrum]. Analyses of the ultraviolet–visual color of the A and B rings have led several investigators, including Lebofsky et al. (1970), Irvine and Lane (1973) and Clark (1980), to conclude that the rings must also contain a well-mixed, non-icy component in order to account for their pronounced reddish color. Initial hypotheses centered on frosts containing either irradiated NH3 or H2S hydrates (Lebofsky and Fegley, 1976), sulfur allotropes (Gradie et al., 1980) or some form of Fe3+-bearing silicate material (Clark, 1980), possibly akin to carbonaceous chondrites. Less than 2% of the latter would be required, given the high albedo of the rings in the 0.7–1.4 μm region (Clark and Lucey, 1984).

Voyager images revealed that the bright A and B rings, which dominate the rings' average spectrum, have much redder visual colors than the more neutral-colored C ring and Cassini Division (Estrada and Cuzzi, 1996). Subsequent modeling by Cuzzi and Estrada (1998) suggested that an initial few percent by mass of reddish organics might provide the reddening agent that dominates the properties of the optically-thick A and B rings, but that eons of meteoritic infall of dark, neutrally-colored material had polluted the less-massive C ring and Cassini Division, leading to their lower albedos and less-red visual colors.

Support for this latter interpretation was provided by Poulet and Cuzzi (2002) and Poulet et al. (2003), who fitted the regolith light-scattering model of Shkuratov (1999) to visible photometry acquired with HST (Cuzzi et al., 2002) and to new high-resolution near-infrared spectra obtained at the IRTF. In their best-fitting models, Poulet et al. (2003) invoke a ring particle regolith with ice grains ranging from 10 μm to 1 mm in size, containing inclusions of less than 1% of tholins1 to provide the red color in the visible and intimately mixed with small but variable amounts of carbon grains to lower the albedo of the ice–tholin mixture. Ice–tholin mixtures have also been invoked to model the near-infrared spectra of the icy saturnian satellites (Cruikshank et al., 2005b).

To date, no spectroscopic evidence for ices other than H2O has been reported for Saturn's rings, and no clearly diagnostic features of the non-icy component(s) have been identified, with the possible exception of a very weak absorption in the 0.8–0.9 μm region (Clark and McCord, 1980). Moreover, as neither Voyager spacecraft carried a near-infrared spectrometer, the spatial resolution of all existing near-IR spectra is limited to ∼6000 km (1 arcsec as seen from Earth). This is sufficient to separate the A, B and C rings (cf. Poulet et al., 2003), but not to look for spectral gradients within these major ring regions or to study the Cassini Division.

We report here on the first observations of Saturn's rings by Cassini's Visual and Infrared Mapping Spectrometer (VIMS), made shortly after the spacecraft's Saturn orbit insertion (SOI) burn on 1 July 2004. Although of modest spectral resolution compared with recent ground-based data sets, VIMS data provide the first near-infrared spectral observations of the rings that are capable of resolving much of their fine-scale structure. Free from the effects of telluric absorption, especially atmospheric water vapor, VIMS data also provide the first complete 1–5 μm spectra of the rings. Preliminary results were presented by Brown et al. (2006).

In this paper we present average visible and near-infrared spectra recorded by VIMS for the major ring regions, and describe both their regional and smaller-scale spectral variations. Simple comparisons of the observed radial brightness profiles with the predictions of single-scattering by a classical, many-particle-thick layer are made, but full radiative-transfer modeling of the observed spectra will be the subject of future work. After a brief overview of the VIMS instrument and the SOI observation sequence in Sections 2 The VIMS instrument, 3 Observations, we describe our data reduction procedures in Section 4. Section 5 presents a summary of the principal features seen in the data, which are then described in more detail in Sections 6 (radial profiles and single-scattering models) and 7 (spectral variations). Our principal results are summarized in Section 8.

Section snippets

The VIMS instrument

The VIMS instrument is designed to obtain spatially-resolved reflectance spectra of planetary targets in the visual (hereafter VIS) and near-infrared (IR) wavelength range (0.35 to 5.10 μm). Details on the instrument design and calibration are provided by Brown et al. (2004); we summarize here the principal features and operational modes used at SOI. Between 0.35 and 1.05 μm, the VIS channel's long-slit spectrometer acquires data using a CCD array detector in ‘pushbroom’ mode, with a spectral

Observations

At approximately 3:07 Spacecraft Event Time (SCET) at Saturn on 1 July 2004, about 10 min after the successful completion of the 96-min main engine burn that placed Cassini in orbit around Saturn, the spacecraft executed a short turn to aim its suite of four optical remote sensing (ORS) instruments down at the rings. At the time, the orbiter was at a distance of ∼20,000 km above Saturn's cloud tops and a similar altitude above the unlit northern side of the C ring. Between 3:11 and 3:55 SCET,

IR data

Standard calibration of VIMS-IR imaging data involves four steps (McCord et al., 2004). First, an internal background spectrum, recorded at the end of each line of data and due primarily to thermal emission from the instrument, is subtracted from each raw spectrum. Next, a flat-field correction is applied to correct for slight variations in response with scanning mirror position and a despiking algorithm is used to identify and flag spikes due to cosmic rays and gamma rays from the radioisotope

Overview of the data

We organize our discussion of the VIMS SOI data as follows. In this section we present average spectra for the classical A, B and C rings, plus the Cassini Division, with qualitative interpretations in terms of ice grain size and possible non-icy contaminants; radial profiles of I/F extracted from the IR data; and an overview of the large-scale spectral variations seen in the data. In Section 6 we examine the radial I/F profiles in some detail, drawing comparisons between the VIMS SOI data and

Ring structure and radial I/F profiles

In this section we examine the VIMS radial I/F profiles of the rings in more detail, drawing comparisons between the VIMS data and optical depth and brightness profiles obtained from Voyager observations. We also present a simple radiative transfer model for the observed I/F profiles based on the assumption of single-scattering in a classical, many-particle-thick layer. The reader interested primarily in the spectral results may wish to skip ahead to Section 7.

Ring spectral variations

In this section we explore the spectral properties of the rings as a function of radial location. We begin by comparing the IR spectra of several broad, structurally-homogeneous regions. Next we quantify several observed spectral variations by plotting radial profiles of band depths and spectral slopes at selected wavelengths. We present these data for scans 1 and 2 separately, and then discuss their similarities and differences.

Conclusions

We summarize here our principal results, referring to the relevant figure(s), and end with some suggestions for future work.

(1) VIMS data obtained at SOI cover the outer C and inner B rings (IR only), the Cassini Division and entire A ring, at a radial resolution in the 1–5 μm range which varies from 15 to 30 km (Fig. 2).

(2) The spectra of all the major regions sampled are dominated by water ice, with typical regolith grain radii of 5–20 μm (Fig. 3, Fig. 4). The presence of relatively weak

Acknowledgements

We thank the VIMS engineering and operational teams, without whom neither the instrument nor any of the data presented here would exist. We also acknowledge the work of Brad Wallis and the SOI target working team, whose lengthy deliberations led to a workable compromise for the design of this unique opportunity in the Cassini orbital tour. Thanks are also due to Ron Draper, Cassini's first project manager, whose early encouragement in the face of engineering skepticism led to the first