First detection of [OI] 630 nm emission in the Enceladus torus
Planetary Plasma and Atmospheric Research Center, Graduate School of Science, Tohoku University, Sendai, Japan
Corresponding author: K. Kodama, Planetary Plasma and Atmospheric Research Center, Graduate School of Science, Tohoku University, Aramaki-aza-aoba, Aoba, Sendai 980-8578, Japan. (firstname.lastname@example.org)
 Observations of [OI] 630 nm emission in the Enceladus torus around Saturn have been made at the summit of Mt. Haleakala in Hawaii using a high-dispersion echelle spectrograph coupled to a 40 cm telescope in the period of 13 May through 19 June 2011. A slit of the spectrograph was aligned perpendicular to the equatorial plane of Saturn and placed at a distance of 4 Saturn's radii (Rs) from the planetary center in the dawn side to put the Enceladus torus within the field of view. As a result, [OI] 630 nm torus emission was detected with S/N ~ 7 for summed exposure of 20 h during the observing period. The observed brightness has a maximum value of 4.1 ± 0.6 Rayleighs (R) near the equator, and it extends to north-south (N-S) direction with a full width at half maximum of 0.8 Rs. We made estimation to explain mechanism of the observed brightness taking into account an excitation of [OI] 630 nm by electron impact and photodissociation of water group molecule (OH and H2O). Densities of electron, O, OH, and H2O and electron temperature derived from data taken by Cassini and Hubble telescope were used for the estimation. The observed brightness is reasonably explained, taking into account an uncertainty of estimation depending on N-S distributions of species and quiet solar activity conditions. The estimation also suggests that [OI] 630 nm emission is excited by photodissociation of OH and H2O and by electron impact of O with their contributions of 50%, 30%, and 20%, respectively, for quiet solar activity. We also note that the intensity due to photodissociation has considerable variability depending on the level of solar activity by a factor of 2.5.
 Neutral particles in Saturn's magnetosphere are predominately composed of H2O and their dissociative products, which mainly come from active water plumes on the icy moon, Enceladus. These particles and their dissociative products (water group particles) distribute around the Enceladus orbit as a donut-shaped ring called the Enceladus torus. Remote sensing observations by spacecraft have been made for molecules and the atoms in the torus. Hubble Space Telescope (HST) observed resonant scattering of OH 310 nm [e.g., Shemansky et al., 1993]. Resonant scattering of O 130.4 nm are also observed by Ultraviolet Imaging Spectrograph (UVIS) on board the Cassini spacecraft [Esposito et al., 2005; Melin et al., 2009]. The result of these observations provides the density distribution of OH and O. These distributions restrict models that performed to solve the evolution process of neutrals and source rate of H2O jetted from a plume on Enceladus [e.g., Cassidy and Johnson, 2010]. Ground-based observation of these emissions is difficult due to absorption by telluric atmosphere. Despite such difficulty, ground-based observations are expected to make it possible to monitor the torus which should have some variations in response to plume activities or magnetospheric environment over long time scales.
 Among water group particles, oxygen atoms have a forbidden emission line that is potentially observable in visible range at 630 nm. [OI] 630 nm emission has been observed near Jupiter's moon Io, in a torus around Jupiter, and in the coma of a comet [e.g., Oliversen et al., 2001; Brown, 1981; Morgenthaler et al., 2001]. Excitation processes of this emission in Enceladus torus are expected to be a combination of direct electron impact on oxygen atom and water vapor molecule, and photodissociation of water group particles followed by emission. Here we present the first ground-based detection of [OI] 630 nm emission in the Enceladus torus.
2 Observation and Data Reductions
 Observations of the [OI] 630 nm emission in the Enceladus torus were made at Haleakala observatory, Hawaii, from 13 May through 19 June 2011 using a high-dispersion echelle spectrograph (λ/δλ ~ 36,000) coupled to a 40 cm Schmidt-Cassegrain telescope. Table 1 shows detailed specifications of the spectrograph. The spectrograph has a slit-viewing camera to make precise pointing and tracking of observing targets relative to an entrance slit of the spectrograph. The left panel of Figure 1 indicates geometry of the slit with respect to the Enceladus torus. The spectrograph slit was aligned perpendicular to the equatorial plane of Saturn and placed at a distance of 4 Saturn's radii (Rs) from its center on the dusk side. Although the length of the slit was 500 arcsecond (") on the sky, a stripe of neutral density mask was put at the center of the slit for other observing targets. Thus, actual slit length (north-south direction) and width (east-west direction) used for this observation was 200" by 5.1", corresponding to ~22 by ~0.57 Rs on the sky, respectively. A typical value of spectral resolution (a full width at half maximum (FWHM) of point spread functions) was 1.5" measured from the slit-viewing camera. The exposure time of each spectral image was 40 min. Short exposures of Saturn's disk continuum were acquired immediately before and after the torus exposure for absolute intensity calibration. Tungsten and neon lamp spectra were also taken before and after the torus exposures for flat-fielding and wavelength calibration.
Table 1. Specification of the Telescope and the Spectrograph
40 cm Schmidt-Cassegrain
Near-Littrow mounting using echelle grating
5.1" on sky
Dispersion (630 nm)
Spectral resolution (FWHM)
36,000 (~8 km/s)
 The raw spectral image data were processed as follows. The first step of the data reduction was general CCD data processing that includes subtraction of the DC offset bias and dark current, and flat fielding using tungsten lamp data. Cosmic ray events and hot pixel were removed by sigma filter. Spectral curvature of the image was corrected using two Ne emission lines at 631.6855 nm and 630.4789 nm. Absolute wavelength calibration was carried out using the telluric airglow [OI] 630.0304 nm and Ne lamp spectrum. Count rates of the torus image and the Saturn's disk image were compared to convert the count value to Rayleighs. The intensity of Saturn's disk near 630 nm was 1.82 × 107 R/nm calculated from the solar flux (=16.6 × 10−5 W/cm2/nm at 1 AU from Livingston ) and Saturn's reflectance near the equator (=0.8 at 635 nm from Ortiz et al. ). The right panel of Figure 1 shows the reduced spectral image acquired on 29 May 2011. As seen in the spectral image, the spectrum is contaminated with stray light from Saturn and its rings scattered by telluric atmosphere and in the instrument.
 After theses procedures, 30 data sets were shifted in the Saturn frame individually to negate the variation of relative velocity between Earth and Saturn. The velocity was ~19 to ~ 28 km/s in our observation period. Finally, the data sets were coadded to increase the signal-to-noise (S/N) ratio, making total exposure time of 20 h.
3 Result and Discussion
 Figure 2 shows the spectral profile of the observed emission averaged over a spatial range of ±0.33 Rs (3 pixel) around Saturn's equatorial plane. The horizontal axis indicates wavelengths in the Saturn rest frame. The protruded line between 630.96 and 630.00 nm is telluric [OI] 630 nm emission, sufficiently Doppler shifted that does not affect the analysis of the torus emission. The spectrum indicated by the red line shows the scattered light spectrum from spatial regions above and below the Saturn equator where the torus emission was not expected. The scattered light spectrum was scaled up to match the continuum of the equatorial spectrum. Two vertical lines indicate the range of plausible Doppler shifts due to Keplerian motion of particles when observing at an orbital location of 4 Rs in the dusk side from the Saturn, 0 to 12.6 km/s. The limitation velocity of particle traveling across the line of sight is 0 km/s. The orbital velocity of 4 Rs is 12.6 km/s. The error bar represents 1 σ of measurement errors considering photon noise of the signal and the background continuum and the dark noise of the CCD sensor. The torus emission is clearly seen within a range indicated by the two red lines. The scattered light spectrum showed that there was no absorption feature around torus emission so we can derive the torus emission by lining the base. Figure 3 shows the brightness distribution along the slit in Rayleighs. This north-south (N-S) spatial distribution was derived by integrating the spectrum between 630.02 and 630.07 nm, the expected range from Doppler shift and instrumental broadening. The obtained peak intensity of the torus emission is 4.1 ± 0.6 R, and an FWHM of N-S distribution is 0.8 ± 0.2 Rs. Note that the brightness is a value averaged over 1 month time span.
 Next we evaluate candidates of emission mechanisms. Below are the expected processes which produce 1D state of oxygen atom, from which [OI] 630 nm emission is emitted. The processes we assessed are electron impact and photodissociation of water group molecule. Electron impact reactions lead to direct excitation of atomic O and dissociation of H2O. Photodissociative reactions are dissociation of H2O and OH. The reactions are as follows.
 Photodissociative reactions are dissociation of H2O and OH. The reactions are as follows.
 The [OI] 630 nm emission is emitted when the energy level drops to the 3P state from the 1D state. If we neglect the quenching effect due to low densities, brightness of [OI] 630 nm emission caused by electron impact reactions, IEI(O) and IEI(H2O), are given by the following equations:
where k1 and k2 are the rate coefficient for electron temperature, T, respectively. For photodissociative reactions, brightness of 630 nm, IPD(OH) and IPD (H2O), are given by the following equations:
where k3 and k4 are the rate coefficient of for photodissociative reactions at a distance of Saturn from the Sun. In equations (5)–(8), r is radial distance from Saturn's center, and ne, nO, nH2O, and nOH are the number densities of electron, O, H2O, and OH, respectively. The brightness of 630 nm looking from an observer is obtained by integrating the volume emission rate along the line of sight (LOS), s. In this estimation, we used O and OH density distributions derived from observations by Cassini/UVIS and HST and values are around 700/cm3 at 4 Rs [Melin et al., 2009]. The H2O density model by Cassidy and Johnson  showing 6 × 103/cm3 at 4 Rs was used for the H2O density distribution. The electron density distribution, derived from Cassini/Radio and Plasma Wave Spectrometer data, is given as a function of radial distance [Persoon et al., 2006]. Because the electron density inside of 5 Rs is highly variable from orbit to orbit [Persoon et al., 2005, Figure 2], values of 100 and 50/cm3 were taken as representative values for inside of 5 Rs. The radial dependence of electron temperature derived from the Cassini Plasma Spectrometer and Electron Spectrometer data was also employed [Schippers et al., 2008]. The function of electron temperature in unit of eV is given as follows:
where r is radial distance from Saturn's center in unit of Rs. We adapted values for a and b given for cold electron by Schippers et al. , a = 6.8 × 10−5, b = 5.9 for r ≤ 9 Rs and a = 4245, b = −2.3 for r > 9 Rs. They reported discontinuity at 9 Rs coinciding with edge of OH cloud. As shown in Wegmann et al. , the rate coefficient for dissociation of H2O has higher value for hotter electron. We need to consider the dissociation by superthermal electron for further accurate estimation; such discussion is well beyond the scope of this paper. For rate coefficients of electron impact reactions, we used the value given by Smyth and Shemansky  for O and that by Wegmann et al.  for H2O. The reaction rates for photodissociation of OH and H2O were given by Huebner et al.  as the value at 1 AU distance from the Sun. For reaction rate of OH, the experimental and theoretical values were given. We applied the experimental value in this estimation. This value can be changeable depending on the solar activity. Therefore, we used the maximum and minimum values against solar activity in our calculation.
 Figure 4 gives the radial distribution of estimated brightness for each reaction looking from the equatorial plane of Saturn. Note that an inclination of LOS against the equatorial plane, 9° during the observation period, can decrease the brightness by 50% assuming N-S distribution of electron (scale height = 0.56 Rs at 4 Rs) [Persoon et al., 2006], O (FWHM = 1 Rs) [Melin et al., 2009], and OH (e.g., scale height = 0.35 Rs) [Richardson et al., 1998]. The LOS effect for the brightness will be dealt with more accurately for the next step of this study. From Figures 4a and 4b, the brightness by electron impact excitation near 4 Rs is ~1 R for O and less than 0.1 R for H2O at 4 Rs. Additionally, the difference in electron density within 5 Rs does not have Figures 4c and 4d show the brightness due to photodissociative excitation range between 2.5 and 6 R for OH and 1 and 2.5 R for H2O at 4 Rs depending on the reaction rate for photodissociation. Therefore, the total estimated brightness is 4.5–9.5 R viewing from the equatorial plane, and it can decrease to about 2.3–4.8 R viewing from the Earth depending on the solar activity. Thus, the observed brightness, 4.1 ± 0.6 R, is consistent with the estimation taking into account the range of solar activity, 2.3–4.8 R. The estimation also implies that the observed [OI] 630 nm emission results from the combination of following three emission processes, primarily photodissociation of OH with 50% contribution, photodissociation of H2O with 30%, and electron impact of O with 20% for quiet solar activity. Additionally, our calculations indicate the possibility of photodissociation as the dominant contributor for this emission when solar activity is high. In such a condition, the [OI] 630 nm emission will directly reflect the density distribution of neutrals since the emission does not depend on electron conditions. Therefore, long-term observations covering a solar activity cycle will be important for identifying the main source for the [OI] 630 nm Enceladus torus emission and detecting variations in the density of the Enceladus torus.
 [OI] 630 nm emission from Enceladus torus were detected by ground-based observation using a high-dispersion echelle spectrograph coupled to a 40 cm telescope. Observed value of [OI] 630 nm brightness averaged during 13 May through 19 June 2011 is 4.1 ± 0.6 R at 4 Rs in the dusk side of the torus with N-S extent of 0.8 Rs in FWHM. The observed brightness is in reasonable agreement, considering an uncertainty of N-S distribution of emitting atoms and that the observation period was in the quiet solar activity condition, with the estimated value which takes direct electron impact and photodissociative processes into account. This estimation also suggests that brightness of [OI] 630 nm emission should have a range depending on solar activity. In active solar condition, the emission will mainly be caused by photodissociation reaction. In such condition, brightness distribution could be directly converted into neutrals distribution.
 In any case, ground-based observation is a strong tool for observing torus environment that have active source like a plume. We need more long-term observation and data to find out the process of [OI] 630 nm emission from Enceladus torus.
 The authors would like to thank the anonymous reviewers for their comments and helpful suggestions in evaluating this letter. We would like to express our appreciation to Jeffrey Kuhn, David Harrington, and Michael Maberry for their kind support for our observations at Haleakala High Altitude Observatory of the Institute for Astronomy, University of Hawaii. We also thank Lester Hieda and Daniel O'gara for their technical support at the summit.
 The Editor thanks two anonymous reviewers for their assistance in evaluating this paper.