The Comet SL9 impact observation run at the Steward Observatory 90-Inch extended
from July 15 through July 24 (UT). Observers were Milagros Ruiz, George and Marcia
Rieke, and various visitors. Telescope operators were Patrick Frawley and Dennis
Means. The instrument was FSPEC, a long slit cryogenic spectrometer with a
NICMOS3 array. FSPEC provides two gratings, giving per--pixel resolutions
respectively of 1600 and 7000. Slits can be selected that are 1, 2, or 4 pixels wide;
the pixels project to 1.2" on the sky. For this event, FSPEC was outfitted by Earl
Montgomery, Kevin Luhman, and Craig Thompson with an infrared guide camera
using a NICMOS2 128x128 array with pixels projecting to 0.5" on the sky. The guide
camera views off the slit to provide accurate registration of the spectra. The infrared
guide camera contributed substantially to the success of our observations, as it and
the excellent support by the telescope operators allowed us to catch events during the
day as well as at night. Also, the impact sites were very obvious through the guide
camera's 2.0 - 2.3um filter, allowing accurate registration and guiding. We used the
guide camera to record images of Jupiter periodically during our observations to
document the appearance of the planet.
The first day of the run we obtained baseline data, mostly low resolution spectra aligned along the belt where the impacts would occur and high resolution spectra aligned pole to pole to document the level of auroral emission. The next night was lost to weather. On July 17, 0 hours UT (5 pm local time), we began the first observations of the A impact site, including images through the guide camera and a low resolution spectrum. With the adjacent disk emission subtracted off the spectrum of the A impact site, the albedo was perfectly neutral from 2.1 to 2.4um, indicating that the materials reflecting the light from the site were above the Jovian atmosphere (to avoid methane and molecular hydrogen absorption) and that they were unlikely to be either water, ammonia, or methane, all of which have spectral structure that would have appeared in our data. Later the same day, we "observed" the impact B fireball, meaning we set the high resolution grating to the wavelengths of molecular hydrogen and H3+ but saw no action (through on and off clouds).
July 18 (UT) was lost to weather. July 19, we opened at 03:10 UT (7:10 pm local time) and obtained a set of low resolution spectra with the slit aligned along the belt where the impacts occurred. We followed these observations with a series of grating settings to obtain a high resolution spectrum of the same belt from 2.06 to 2.36um.
On the next day, we opened at 22:15 (UT) July 19 and successfully observed the fireball from the L impact, using the high resolution grating set at 2.10um to include both molecular hydrogen and H3+. No emission lines were detected in this spectrum. We also obtained a spectrum to look for Brackett gamma after the fireball had partially faded, again with no detection of a line. The remainder of this night was spent repeating the set of high resolution observations with the slit oriented along the belt containing the impact sites.
On July 21 (UT) we used the high resolution grating with the slit oriented pole-to-pole to check for the presence of the molecular hydrogen and H3+ lines at the polar caps from aurorae. The lines were present at levels comparable with our baseline data at a similar longitude prior to the impact, indicating that the auroral zones were not significantly affected by the atmospheric disturbances associated with the comet impacts. Later that night, we observed the fireball from the impact of fragment R, this time with the high resolution grating set to 2.32um. Despite the low altitude of Jupiter and a substantial layer of cloud, we obtained a spectrum that beautifully shows CO emission, with a bandhead followed by approximately 40 emission lines. The development of the fireball was followed with 30 sec time resolution; however, we have not yet subdivided the data to determine the time evolution of the CO emission.
July 22 (UT) was dedicated to an attempt to frequency switch between the first CO overtone (at 2.3um) and the second overtone (at 1.6um) to gain further information on the temperature structure of the fireball from impact V, which was to be well placed above our site. The sky was perfectly clear, but no fireball occurred. The observing strategy we used will result in a low resolution spectrum of the impact sites that were visible that night as a byproduct of the effort to observe the V fireball.
July 23 (UT) was lost to weather. July 24 (UT) we obtained a low resolution spectrum of the impact sites (including the first, A) at 01:30 and a set of high resolution spectra at 04:00. This last set of observations was terminated due to heavy cloud. The low resolution spectrum obtained on this date of the K impact site shows a modest decrease in albedo from 2.1 to 2.3um (after subtracting the disk background), possibly indicating that the reflected light is coming from a sufficiently low level that there is weak methane absorption overlying the reflecting material. This behavior contrasts with the neutral albedo from the A site a few hours after impact, and may be associated with the fact that the K site was observed about 4 days after the impact. Further information on the variation of albedo of the impact sites with age after impact should be available when the remainder of our data have been reduced.