THE ASTRONOMICAL JOURNAL, 115:1200-1205, 1998 March
© 1998. The American Astronomical Society. All rights reserved. Printed in U.S.A.
OH spectra of comet Hale-Bopp were obtained at 1.667 GHz between 1997 February 19 and June 1 using the 26 m parabolic antenna of the Dominion Radio Astrophysical Observatory. Spectra showed strong (2 Jy) emission in mid-February. The comet's flux density then decreased, until it was barely detectable by March 11. An absorption line dominated the spectrum for several days in March. Flux density remained low for most of the subsequent observations. At the end of April a strong emission line appeared close to the expected cometary velocity. It is argued that this line did not arise in the comet itself, but came from a background source, probably IRAS 04361+2547, whose radiation had undergone maser amplification in the coma of the comet.
Key words:
comets: individual (Hale-Bopp 1995 O1)
Comet Hale-Bopp provided an unprecedented opportunity to study gaseous emission from a very massive comet. Water vapor sublimating from a comet's core is photodissociated to produce OH that can be observed with several transitions in the radio spectrum. Of the four J = 3/2 transitions between the lowest rotational states of the 2
3/2 levels, the line at 1.667 GHz is usually the strongest (Snyder 1986), and was therefore selected for this investigation. The present spectra have higher signal-to-noise ratios than those obtained for comets Halley (Galt 1987a) and Giacobini-Zinner (Galt 1987b) with the same antenna.
Observations were made with the 26 m paraboloid of the Dominion Radio Astrophysical Observatory, low-noise receivers, and a three-level autocorrelation spectrometer that produced 512-channel spectra (Hovey 1997). The antenna was pointed using an ephemeris (JPL No. 45-DASTCOM3) provided by the Jet Propulsion Laboratory (see Yeomans 1997). This ephemeris also controlled a receiver local oscillator to compensate for the topocentric velocity of the comet. Two independent signal paths preserved the two orthogonal polarizations from the feed horn through the receiver and spectrometer to produce separate spectra for each polarization. Spectra were obtained by switching the observing frequency half a bandwidth every 5.625 s. Subsequent software operations removed frequency switching and averaged the two polarizations. Most spectra were observed with a bandwidth of 500 kHz, which, after removal of frequency switching, produced a velocity range of 45 km s-1 in 256 channels. Apodization lowered the resolution to about 0.2 km s-1 per channel. Some spectra were taken with a narrower bandwidth of 250 kHz.
The observing session began with a broadband linearly polarized feed horn and an ambient-temperature receiver. After 2 weeks a circularly polarized horn and cryogenically cooled receiver were installed. The remainder of the session after 1997 March 20 used the broadband receiver again. The broadband receiver had a system temperature of 
50 K, while the cooled receiver's system temperature was 30 K.
Observations began 1997 February 19, when the comet was 1.16 AU from the Sun, and continued intermittently through perihelion until the comet had receded to 1.39 AU on 1997 June 1. Flux calibration was tied to the 9.5 km s-1 absorption line in Orion B (= W12 = NGC 2024), which has a depth of 15.7 Jy (Claussen & Schloerb 1987).
Spectra of the comet are plotted in Figures 1 and 2. Unfortunately, these spectra do not form a homogeneous set. In particular, the observations were not spaced evenly in time, and integration times differ by a factor of 14. As the eight spectra observed at 250 kHz bandwidth show no fine structure, they have been smoothed to the same resolution as the rest of the spectra for Figure 1.
 | FIG. 1. Spectra of comet Hale-Bopp from 1997 February 19 to March 19 |
 | FIG. 2.Spectra of comet Hale-Bopp from 1997 March 27 to June 1 |
At least eight of the spectra in Figure 1 show narrowband man-made interference of unknown origin. In particular, the top five spectra show a negative spike near velocity -12 km s-1. The fact that the velocity (and hence the frequency) of the spike remained constant while the receiving frequency was changed to follow the comet's Doppler velocity is strong evidence that the spike is interference. (Although a spike appears as absorption in this display, it is actually emission observed during the reference half of the switching cycle, and thus its frequency is 250 kHz higher than the scale of Fig. 1 would imply.) Spikes that were obviously larger than the rms noise in the uncontaminated baseline were edited out before analysis of the spectra was undertaken. Figure 1 shows the spectra before they were edited.
Many of the spectra show obvious asymmetry, with edges steeper on the low-velocity side than on the high-velocity side. The asymmetry is in the same direction as that observed at Nançay in 1996 November 17
30 (Biver et al. 1997). The asymmetry probably arises as a geometric effect similar to that studied by Rauer et al. (Rauer et al. 1997) in optical cometary spectra of H2O+. In spite of the asymmetry, each spectrum was analyzed by fitting one or two Gaussian curves; the choice of whether to use one or two was usually, but not always, obvious. Numerical results of the fitting process are shown in Table 1. Areas under the Gaussian curves are plotted in Figure 3.
 | TABLE 1 COMET HALE-BOPP SPECTRA |
 | FIG. 3.Areas of Gaussian curves fitted to spectra of Figs. 1 and 2. The gap on the time axis indicates when the spectrum appeared to be dominated by a background source. UT dates are shown at the top. |
That error bars shown in Figure 3 are dominated by thermal noise can be seen by comparing spectra in Figure 1 taken with the ambient temperature receiver and with the cryogenically cooled receiver. For example, the spectrum for day 517.2 is noisier than that for day 520.5, even though the observation times are nearly equal. Of course, the assymmetry of the lines also contributes to the errors produced by a Gaussian fitting process.
As the telescope was engaged in another project, no comet observations were obtained between 1997 April 13 and May 13. However, observations attempted 1997 April 30 to May 1 showed an anomalously strong line near the expected velocity of the comet. This line was nearly 3 times stronger than any comet line observed in this session. Indeed, it was 30 times stronger than expected from interpolation between previous and subsequent comet observations. The line had a width of 1.28 ± 0.05 km s-1, narrower than most of the comet lines in Table 1, and its velocity appeared to be 2.0 km s-1 lower than the expected comet velocity. The amplitude variation shown in Figure 4 suggests that the line originated in a background source near the comet's trajectory. The comet's path is plotted in Figure 5, with heavy lines indicating when the telescope was observing. Arrows point to the Sun for two of the observations.
 | FIG. 4.Peak flux density of the 1.667 GHz OH spectral line that was recorded while following comet Hale-Bopp. The line probably originates in a background source. A 15 km s-1 wide section of each spectrum is also shown at one-fifth vertical scale. UT dates are shown at the top. |
 | FIG. 5.Path of comet Hale-Bopp in relation to the IRAS source that may be responsible for the emission observed 1997 April 29 to May 1. The five heavy lines along the comet's path indicate times when spectra were being recorded. The plus sign marks the position of the emitting region measured 27 days after the comet had passed. Arrows show the direction of the Sun for the first and last observations. UT dates are shown at the top. |
The position where the line had been seen was reobserved 27 days later when the comet had moved 22°, or 44 beamwidths, away. The line was still present, but its amplitude was only half what it had been when the comet was in the line of sight. It would appear that maser action in the coma of the comet had amplified the background source when it was first observed. It is possible that the background source is variable, but this explanation seems unlikely considering the evidence shown in Figure 4 that the intensity was a maximum when the comet was near the line of sight.
A rough position for this source was determined by pointing the antenna 0.5 beamwidth north, south, east, and west of the place on the comet's path where the line was strongest. These measurements yielded the position 4h 40m, +26°20
(J2000.0), which is marked on Figure 5. The line's radial velocity, measured with respect to the local standard of rest, was 6.11 ± 0.02 km s-1. A search was made using the Canadian Astronomical Data Centre at the Dominion Astrophysical Observatory, Victoria, British Columbia, in an attempt to identify the source responsible. A possible candidate is IRAS 04361+2547, which is a known H2O and CO emitter. Its CO velocity of 5.9 km s-1 (Wouterloot, Brand, & Fiegle 1993) is close to the OH velocity of the background source. The catalog position of this IRAS source is also shown on Figure 5. Although the positional agreement is not perfect, one should remember that there is considerable uncertainty in the measurement of the radio position.
The comet spectrum for 1997 May 19 (JD 2,450,588.3) shows two distinct peaks. It was first suspected that the extra peak was also a background source. However, a spectrum of this region, obtained after the comet had passed, was featureless. It is therefore likely that both lines are associated with the comet. Presumably, the high-velocity peak arises in gas ejected from the comet's surface with a 6.4 km s-1 velocity component in the line of sight. Although not obvious at first glance, an indication of this ejected material can be seen by looking obliquely at adjacent spectra of Figure 2. This would imply that the event lasted 2 or 3 days.
Perhaps the most striking phenomenon demonstrated by spectra of Figure 1 is the gradual weakening of emission followed by the appearance of absorption as the comet approaches the Sun. For about 7 days both emission and absorption are present. Emission and absorption were roughly equal on JD 2,450,521.2, when the heliocentric velocity of the comet was -11 km s-1. A comparison of the circumstances under which comet Halley changed from emission to absorption is appropriate. Comet Halley had a velocity with respect to the sun of -11 km s-1 on 1986 February 1.3. Spectra taken 1986 January 31.9 and February 1.8 show weak emission. The spectrum of 1986 February 2.9 shows strong absorption, as do the spectra for the next 3 days (Galt 1987a). The spectral sequence observed at Nançay also shows a changeover from emission to absorption about February 1 (Gerard et al. 1987).
The change from emission to absorption has been attributed to the radio analogy of the Swings effect and results from a change in the relative populations of the F = 2 + and - parity states of the 23/2 levels of OH. These populations in turn depend on excitation by narrow lines in the solar spectrum, whose frequencies, as seen by the OH molecules in the comet, will rise as the comet approaches the Sun. The beam of the radio telescope, however, covers a large fraction of the coma and therefore accepts radiation from molecules ejected from the comet's surface with a wide range of velocities. The amplitude of the spectrum at any frequency (or velocity) depends on the vector sum of the ejected gas velocity and the component of the comet's velocity in the Sun's direction, integrated over the beam of the antenna.
Although maser amplification of the background source W9 was predicted for comet Halley (Schloerb & Gerard 1985), this phenomenon was not seen in the spectra taken by Galt (Galt 1987a). A point with large error bars in the Nançay results (Gerard et al. 1987) may be evidence of such maser amplification. The evidence presented here, however, would appear to be a better example of the phenomenon.
I wish to thank my colleagues, M. Davies, G. Hovey, K. Tapping, and A. Willis, for help with various aspects of the project.
. 1987b, AJ, 94, 174 First citation in article | CrossRef Link | NASA ADS | Inspec Abstract | Order from Infotrieve