During a program of very distant comet observations at the Keck II Telescope on 4 March 1997 we targeted 55P/Tempel-Tuttle for recovery (in part because the weather conditions were too poor for the main, very faint targets at ). A moving object, most likely the comet, was easily visible in a 900 s integration. As this object did not display any evidence of coma or activity, it was not possible to identify it securely with Tempel-Tuttle on morphological grounds, since previously unknown asteroids in the field could not be ruled out a priori. Moreover, is also not possible to unambiguously identify a comet by position and rate of motion alone without a second night of data to verify the orbit (Marsden, 1981; Marsden, 1994; Marsden, 1996a; Marsden, 1996b; Marsden, 1997; Bowell et al., 1989).
Since we had only obtained one night of data, insufficient for a safe recovery, we put out an urgent request for follow-up observations. Observations were made on 7 March 1997 using the ESO NTT telescope and the comet was subsequently reported as recovered (Hainaut et al., 1997; Martin et al., 1997). Following the recovery observations, we analyzed data obtained prior to 1997, and were able to identify several pre-recovery images of the comet. Table 1 lists the circumstances of each of the observing runs; composite images of the comet are shown in Fig. 1.
2.1. 1994 May 10-11; r = 10.8 AU (NTT)
The images were obtained with the 3.54-m New Technology Telescope (NTT) at La Silla, on the Superb Seeing Imager (SuSI) installed at the f/11 Nasmyth-A focus, equipped with the CCD ESO#25, a Tektronix TEK 1024 Grade 1 (10242 pix) detector. The filter used was a Bessel V (ESO#641, Å, Å FWHM, peak transmission 80%). The 24µm pixels, covering 0 128 on the sky, were read binned . The read-out noise was about per binned pixel. The weather was perfect over the observatory, in spite of some worrisome cirrus over the Western horizon. The pre-recovery orbital elements used for these observations were unfortunately off by about with respect to the position computed using the post-recovery elements, which is more than half the field-of-view of the SuSI camera (2 2). Thanks to the large offsets (up to ) applied to the telescope for the image dithering, the actual position has been imaged on one of the individual frames. However, the limiting magnitude of that frame is only , which does not provide a very constraining limit to the size of a cometary nucleus at 10 AU.
2.2. 1995 February 3-4; r = 9.5 AU (NTT)
The instrumental set-up was the same as for the 1994 run (the CCD was read un-binned; read noise ); the images were taken through the Bessel R filter ESO#642 ( Å, Å FWHM, peak transmission 85%); the weather was perfect. As for the 1994 observations, the ephemerides computed using the pre-recovery elements were off by almost . Fortunately, the images obtained during the second night show the post-recovery-expected position for the comet close to the edge of the composite image. A 2.4 object about from the expected position is visible on the composite 4000 s image. It is too faint to be visible in the single frames, or even in partial composites. While the object is probably the comet, the 2.4 level does not ensure a high enough statistical significance (the probability for the candidate to be a noise feature, or the probability to have a noise feature of about diam. within of the expected position, is about ), therefore, we cannot consider it to be a secure detection. The 3 limiting magnitude is reported as upper limit for the brightness in Table 3.
2.3. 1996 May 12; r = 6.7 AU (ESO/MPI 2.2-m)
The observations were performed on the ESO/MPI 2.2-m telescope at La Silla, using the EFOSC2 spectro-imager equipped with the CCD ESO#19, a front-side illuminated, UV coated Thomson 31156 Grade A detector (1024 1024, 19µm pix), whose readout noise was rms, and gain . The camera was operated at f/4.9, giving a 5.7 field (0 332 pix). The filter used was a Bessel R (ESO#585, Å, Å FWHM, peak transmission 86%). While the comet was not visible on most of the single frames (300 s exposure each, except for one of 270 s), it appeared on some of them, and is clearly visible on a composite made from half of the available single frames. Because of its faintness, the comet was not immediately identified when the images were taken, and it was only after the recovery with the Keck telescope had led to an improved orbit that we were able to identify the comet on these pre-recovery frames.
2.4. 1997 March; r = 4.5 AU
March 4, Keck-II - Images were obtained with the LRIS imaging spectrometer from the remote observing room in Waimea. In the imaging mode, the plate scale is 0 215 pix-1, with a useful area of 1660 x 2048 pixels. The gain of the detector is with a readout noise of . Several hours during the first half of the night were lost because of a coolant leak from the instrument. After repairs, we were just coming on line again when thick clouds appeared. We quickly changed our program to look at brighter objects and targeted 55P/Tempel-Tuttle. We obtained 7 images of the comet - 4 of which were very bad because of clouds (very poor S/N, poor image quality). The 3 remaining ones were fortunately obtained through gaps between clouds and are the deepest of all the Tempel-Tuttle images reported here.
March 7, NTT - While the S/N of the March 4 images did not leave any doubt about the reality of the object, its stellar appearance and the short time span covered by the Keck observations left some doubt about the identity of the object. However, a few days later, a single, 900 s image was obtained under excellent seeing condition at the NTT by P. Martin and the NTT team, showing the object at the expected position (and motion) and thus confirming its identity. The image was obtained with the SuSI camera, equipped with CCD ESO#42, a back illuminated Tektronix TeK1024A (24µm pix, readout noise ), through a Bessel R filter (ESO#642). The frame was processed by the NTT team, and transfered electronically to Hawai`i where the comet was measured (photometry and astrometry).
2.5. 1997 April 15,17; r = 4.08 AU (UH 2.2m)
Images were obtained on 1997 April 15 using the UH 2.2m telescope with the Tektronix 2048 CCD camera at the f/10 Cassegrain focus. The plate scale in this configuration is 0 219 pix-1, and the detector had a gain of and a readout noise of . The night started with some high altitude clouds and high humidity, however by 8 UT the sky appeared clear, and the humidity had dropped enough to open by 10:30 UT. Although the night appeared photometric after opening, and some standards were taken, the 55P/Tempel-Tuttle fields were re-calibrated on the following night. Images were also obtained on 1997 April 17, also during non-photometric conditions.
2.6. 1997 June 18; r = 3.5 AU (UH 2.2m)
Director's discretionary time was obtained on the UH 2.2m from 1997 June 17-19 UT for the purpose of calibrating comet data taken on previous runs under non-photometric conditions. The same Tektronix 2048 CCD was used with the f/31 focus giving a plate scale of per binned pixel and a FOV. On the first night of the run all of the previous 55P/Tempel-Tuttle images were re-calibrated, and on 1997 June 18 additional images of the comet were obtained.
© European Southern Observatory (ESO) 1998
Online publication: April 20, 1998