Science with the Hubble Space Telescope -- II
Book Editors: P. Benvenuti, F. D. Macchetto, and E. J. Schreier
Electronic Editor: H. Payne

---

Comets

Paul D. Feldman
Department of Physics and Astronomy, The Johns Hopkins University, Baltimore, MD 21218, USA

Abstract:

Keywords: comets,spectroscopy,ultraviolet

Introduction

Although comets are extended objects to the eye, the source of the cometary activity is the icy nucleus whose typical size is of the order of 10 km. The only resolved images of a nucleus were obtained by the Giotto and Vega missions to comet Halley in 1986. Otherwise, size estimates are obtained from photometry of comets at large heliocentric distances where their comae are insignificant. The unique spatial resolution of HST, which translates to a WFPC2 PC pixel size of 30 km for a comet 1 AU from the Earth, leads to the expectation of being able to partially resolve the nucleus of a comet whose orbit takes it reasonably close to the Earth. The high spatial resolution also makes it possible to separate the sunlight reflected from the nucleus from that reflected by the grains of the coma. This has been done successfully for two periodic comets, Faye and Borrelly, and was attempted for the separate fragments of comet Shoemaker-Levy 9 in order to attempt to estimate the mass, and consequently the kinetic energy input to the upper atmosphere of Jupiter during the 60 impacts in July 1994 (Weaver et al. 1994a,1995).

Spectroscopic observations of volatiles produced near the nucleus, particularly in the vacuum ultraviolet where cosmically abundant atoms and small molecules resonantly fluoresce solar photons, is an extremely powerful tool for determining the relative abundances of the vaporizing species and studying the photochemical and physical processes acting in the densest region of the coma. However, except for the detection of CO Cameron band emission in two moderately active comets in 1991 and 1992 (Weaver et al. 1994b), the full potential of HST for ultraviolet spectroscopy of comets is still to be achieved.

Observations

Comets observed to date by HST are listed in Table 1. The nomenclature used follows the IAU recommendations that went into effect at the beginning of 1995 (Marsden 1995). In this text we will refer to each comet by name only. The very interesting HST search for Halley-sized bodies at the edge of the outer solar system (``Kuiper belt'' objects) is excluded from this listing since discrete objects have not been identified (Cochran et al. 1995). Some of the early results have been presented at the previous HST meeting (Weaver & Feldman 1992), so this review will concentrate on the observations made since mid-1992.

  
Table 1: Hubble Space Telescope Comet Observations

Imaging

HST played a major role in the campaign to study the tidally disrupted comet D/Shoemaker-Levy 9 (SL9) and its subsequent impact into the atmosphere of Jupiter in July 1994 (McGrath 1996). The first observations of SL9 were made in July 1993 and continued, following the HST servicing mission, in 1994 at regular intervals up to a few days before impact. A mosaic from 17 May 1994 is shown in Fig. 1 (Weaver et al. 1995).

  
Figure: Mosaic of SL9 taken with the WFPC2 in wide-field mode on 17 May 1994 (Weaver et al. 1995).

The primary goal of the imaging was to determine the sizes (or significant upper limits) of the largest fragments in order to assess the amount of energy (the fragments entered at a velocity of 60 ) that would be deposited in the Jovian atmosphere and the nature of the effects that would be produced as well as their observability from Earth. A secondary objective was to study the spatial distribution and evolution of the ``dust tails'' associated with each fragment to determine if, in fact, each fragment was an active comet nucleus. Although there was no expectation to resolve a fragment (at the mean distance of SL9, a PC pixel translates to a linear size of 150 km), an unresolved point source is detectable above the background of an optically thin dust coma. Such a point source was not clearly detected in the July 1993 images (the V magnitudes of the brightest fragments, after subtraction of the coma, were 24), leading to upper limits of 4--5 km diameters for the largest fragments (Weaver et al. 1994a). These were based on the assumption of a Halley-like geometric albedo of 0.04. Subsequent to the servicing mission, the unaberrated images confirmed these estimates (Weaver et al. 1995). Analysis of the impact events themselves leads to estimates of slightly under 1 km diameter for the largest fragments and these are consistent with the data. However, this issue is still being debated and the interested reader is encouraged to consult the proceedings of the STScI Workshop on SL9 held in May 1995 (IAU Colloquium 156, The Collision of Comet P/Shoemaker-Levy 9 and Jupiter, Noll et al. 1996) for the various points of view.

The difficulty in deriving nucleus diameters lies in the extrapolation of the coma to the pixel of peak brightness. This is illustrated in Fig. 2, which shows the azimuthally averaged brightness extrapolated to the central pixel together with a model PSF for a point nucleus.

  
Figure: Comparison of azimuthally average spatial brightness profiles for fragment G1 of SL9 (left) (Weaver et al. 1995, revised) and comet Borrelly (right)(adapted from Lamy et al. 1996). The nucleus of comet Borrelly is clearly detected photometrically.

The uncertainty in the derived point source upper limit is largely determined by our lack of understanding of the inner coma. For pure radial outflow, the coma brightness should vary as , where is the projected cometocentric distance on the sky. This condition appears to hold on the tailward side for distances greater than , but not closer to the nucleus. If the cometary activity was very low, the slope might be influenced by other factors including Jovian gravity and solar radiation pressure. However, the validity of this approach can be seen in the PC images of comet Borrelly obtained by Lamy et al. (1996) in November 1994 when the comet was 0.62 AU from the Earth (one PC pixel 21 km). A typical brightness profile is also shown in Fig. 2, clearly demonstrating the detection of sunlight reflected from the nucleus. Moreover, the azimuthally averaged brightness of this moderately active comet is found to follow the same ``law'' near the nucleus as was found for SL9. From observations spaced over nearly one day, Lamy et al. derive a synodic period of P/Borrelly of 24.7 hours and model the lightcurve with a prolate spheroid, smaller but similar in shape to P/Halley. A less robust but similar detection of a cometary nucleus was obtained by Lamy et al. (1995) for comet Faye from 1991 WFPC-1 images.

PC images of some of the brighter fragments of SL9 obtained a few days before impact show a pronounced elongation along the line connecting the fragments (Weaver et al. 1995). Rettig et al. (1996) attribute this to tidal dispersal and use the relative elongation to derive a mean dust outflow velocity of m s. They thus conclude that the large fragments of SL9 were indeed ``active''. However, this outflow velocity is three orders of magnitude smaller than typically found for micron-sized grains leaving an active comet near 1 AU (e.g., Weaver et al. (1992) found a projected velocity of 160 m s from WFPC-1 images of comet Levy), implying that the dust in SL9 may have included grains as large as 0.4 mm. Alternately, Sekanina et al. (1994) have argued that such a low outflow velocity implies that the observed dust coma is a remnant of large dust particles produced when the original comet split in July 1992 and that continuous dust production is not required.

A unique opportunity arose in July 1995 with the discovery of comet Hale-Bopp with V10 at a heliocentric distance of 7 AU. Conservative extrapolations suggest that this comet will be a bright naked-eye object during early 1997 when it approaches perihelion on April 1 at a heliocentric distance of 0.9 AU, possibly rivalling the brightness of the ``Great Comet'' of 1811. Comet Hale-Bopp is not dynamically new (its orbital period is 4000 years) and so is not likely to fade as it approaches perihelion. WFPC2 images were first obtained in September and October 1995 (Fig. 3) providing detailed views of the pinwheel-shaped coma and its temporal variability and allowing a size estimate of a 40 km diameter nucleus (H. A. Weaver, private communication).

  
Figure: WFPC2 images of comet Hale-Bopp taken in September and October 1995.

A striking advantage of HST relative to ground-based imaging is the ability to discern the details of the comet structure on the sub arc-second scale even in a crowded region of sky near the galactic plane due to the small PSFs of the trailed star images. Further observations are planned for the period between March and October 1996 as the comet moves from 5.0 to 2.5 AU from the sun, but the geometry at the time of perihelion in 1997 is very unfavorable for HST observations (the maximum solar elongation angle is 47 degrees) and there is the programmatic constraint of the second servicing mission scheduled for February 1997.

Spectroscopy

Weaver & Feldman (1992) presented a composite spectrum of Hartley 2 that spanned the wavelength range 1200--7000Å utilizing spectra taken with all five of the FOS gratings. This approach has the advantage of providing a complete inventory of the volatile species in the inner coma, including those regularly monitored from the ground, with a single aperture and over a short period of time. The primary use of HST, however, is for ultraviolet spectroscopy (1200--3100Å ), where the dominant atomic and molecular constituents fluoresce. This spectral region has proven to be very productive with the International Ultraviolet Explorer (IUE) (over 50 comets observed in 17 years) and both FOS and GHRS have significant advantages over the IUE spectrographs in terms of sensitivity and spatial resolution. Aside from SL9 and Hale-Bopp, both observed at heliocentric distances greater than 5 AU, only three comets have been studied spectroscopically in detail and none of these was highly active. Nevertheless, the initial results have been of interest.

As an example, spectra of comet Shoemaker-Levy (note: this was a dynamically new eighth magnitude comet discovered by the same team that discovered SL9) taken with both the FOS and the GHRS in July 1992 are shown in Figs. 4 and 5 and illustrate the improvement over comparable IUE long wavelength spectra. The FOS spectrum of Fig. 4 was taken with the aperture which partially resolves the OH bands and clearly separates the various CS bands, as indicated on the figure. The right panel, which shows the same spectrum after subtraction of dust-scattered solar radiation, clearly resolves the two components of the CO band at 2884 and 2896Å .

  
Figure: FOS spectrum of comet Shoemaker-Levy obtained in July 1992 using the square aperture. The panel on the right shows the same spectrum after subtraction of dust scattered solar radiation.

One expectation of FOS is that the small field-of-view will allow for the detection of short-lived species vaporized directly from the nucleus. Of particular interest is S, whose photodissociation lifetime at 1 AU is 300 seconds, and which has been detected to date in one comet, IRAS-Araki-Alcock, in 1983. FOS observations to date provide limits slightly more sensitive than those derived by Budzien & Feldman (1992) from IUE observations of more active comets. Fig. 5 shows a medium-resolution GHRS spectrum of the region including the OH (1,0) band in which the rotational structure is completely resolved. The relative intensities are in good agreement with the fluorescence equilibrium calculations of Schleicher and A'Hearn (1988).

  
Figure: GHRS spectrum of comet Shoemaker-Levy in the region of the OH (1,0) band.

The first detection of CO Cameron band () emission, made in comet Hartley 2 (Weaver et al. 1994b), has already been noted above. The emission was subsequently also detected in comet Shoemaker-Levy. The primary source of CO in the metastable state is photodissociation of CO, so the measurement provides a means of determining the relative CO abundance in the inner coma, and for both of these comets this abundance, relative to that of water, was found to be comparable to that derived for P/Halley from in situ mass spectrometer measurements made during the Giotto encounter in March 1986. However, CO Fourth Positive () fluorescence, a measure of the CO abundance in the coma, was not seen, and the derived 3- upper limits, 1%, indicated that CO was significantly depleted relative to Halley as well as to several other comets previously observed spectroscopically by IUE and sounding rocket experiments. This analysis has been extended by Feldman et al. (1996a) by the subsequent discovery that the (1,0) Cameron band at 1993Å is detected in IUE spectra of four comets, located at the edge of the SWP camera image using the new IUE processing software developed for the IUE final archive. Earlier processing had truncated these spectra at 1980Å . Fig. 6 illustrates the difference between spectra of P/Halley, which shows strong Fourth Positive emission, and Hartley 2, which does not. Despite the large uncertainties in the derived relative CO abundances, the comets (including Halley) in which this feature is detected in IUE spectra have values in the range 3--6%, comparable to that derived for Hartley 2 and Shoemaker-Levy, while for CO, their relative abundance varies between 3 and 8%.

  
Figure: Left: IUE low dispersion spectrum of P/Halley (SWP 27884) from 9 March 1986. The exposure time was 45 minutes. Right: Composite FOS spectrum using G130H (40 minutes) and G190H (32 minutes) of 103P/Hartley 2 from 18--19 September 1991. Dust scattered continuum (longward of 2000Å ) and grating scattering background have been subtracted.

Two possible explanations for the low derived CO abundance in the comets observed by HST have been advanced. The first suggests that CO is not produced directly from the nucleus (whereas CO is) but rather from an extended source as in the case of Halley (Eberhardt et al. 1987). The smaller field-of-view of the FOS, relative to that of the IUE spectrographs, would then exclude the region of the coma in which most of the CO is present. This hypothesis requires detailed modeling to verify, as the line-of-sight to the comet would pass through the CO. Moreover, Eberhardt et al. claim that one-half of the CO (i.e., 7% relative to water) in Halley did come from a region within a radius of 1000 km from the nucleus, so if the same distribution occurred in Hartley 2 the CO would have been detected. The other possibility is that there is a correlation between the CO abundance of comets and their intrinsic activity level. For both Hartley 2 and Shoemaker-Levy, the water production rates were found to be molecules s, while the comets in which CO was detected with IUE had production rates 2 to 10 times larger. IUE observed many comets with lower activity levels but did not have the sensitivity to CO required. Further comet investigations with HST, particularly after the next servicing mission when the long-slit capability of STIS becomes available, should shed more light on this issue.

Since the presence of water ice is a primary physical characteristic of comets, the FOS was used to search for the hydroxyl radical (OH) at 3085Å in spectra of SL9 taken at three times during the HST campaign. In all of the spectra obtained, OH was not detected. The derived upper limit to the water production rate is compatible with what is expected from a comet with a small outgassing surface area near 5 AU. However, on one occasion, four days before impact (on 14 July 1994), FOS spectra of the G1 fragment show a brief outburst of magnesium ions (Feldman et al. 1996b). The Mg emission decays back to its quiescent (undetectable) level with a time constant of 1 minute, consistent with the ions being swept out of the field of view by the motional electric field induced by Jupiter's rotating magnetic field. Although this event occurred when the G1 fragment was inside the Jovian magnetosphere, the role of energetic trapped electrons in producing the observed event is not clear and the exact mechanism responsible for the transient population of Mg remains unidentified.

  
Figure: Comparison of the observed Mg II doublet (left) with synthetic spectra (right). The solid line is a synthetic spectrum for a circularly symmetric brightness distribution that varies as the inverse cometocentric distance, while the dashed line is that for a source uniformly filling the aperture (from Feldman et al. 1996b).

The robustness of this detection is shown in Fig. 7, where the Mg II 2800 doublet is shown together with synthetic spectra for both a uniform source filling the aperture and a source whose spatial distribution follows that of the dust coma. It is clear that terrestrial emission, which would fill the aperture, cannot account for the observed spectral shape, although the serendipitous nature of this event is always likely to raise some degree of skepticism.

Conclusion

There remains much to anticipate from future cometary programs, particularly with the new capabilities of STIS and NICMOS that will be available after the February 1997 servicing mission. The goal of a detailed ultraviolet spectroscopic investigation of a naked-eye comet is ever present, together with the disappointment that comet Hale-Bopp will not satisfy that goal due to its orbital geometry. The results presented here only serve to whet the appetite.

Acknowledgments:

Sincere thanks go to Hal Weaver for a stimulating and fruitful collaboration in HST comet programs and for his assistance in the preparation of this paper. We thank Philippe Lamy and Terry Rettig for permission to quote their results prior to publication. This work is based on observations with the National Aeronautics and Space Administration--European Space Agency HST obtained at the Space Telescope Science Institute, which is operated by the AURA, Inc., under NASA contract NAS5-26555. Support for this work was provided by NASA through grant numbers GO-2442.01-87A, GO-5624.05-93A and GO-5844.02-94A from the STScI.

References:

Budzien, S. A. & Feldman, P. D. 1992, Icarus, 99, 143

Cochran, A. L., Levison, H. F., Stern, S. A., & Duncan, M. J. 1995, ApJ, 455, 342

Eberhardt, P. et al. 1987, A&A, 187, 481

Feldman, P. D., Festou, M. C., Tozzi, G. P., & Weaver, H. A. 1996a, ApJ, submitted

Feldman, P. D., Weaver, H. A., Boice, D. C., & Stern, S. A. 1996b, Icarus, submitted

Lamy, P. L. & Toth, I. 1995, A&A, 293, L43

Lamy, P. L., Toth, I., & Weaver, H. A. 1996, A&A, submitted

Marsden, B. G. 1995, Internat. Comet Q., 17, 3

McGrath, M. A. 1996, this volume, p.

Noll, K. S., Feldman, P. D., & Weaver, H. A. (eds.) 1996, The Collision of Comet P/Shoemaker-Levy 9 and Jupiter, Cambridge: Cambridge University Press, in press

Rettig, T. W., Sobczak, G., & Hahn, J. M. 1996, Icarus, submitted

Schleicher, D. G. & A'Hearn, M. F. 1988, ApJ, 331, 1058

Sekanina, Z., Chodas, P. W., & Yeomans, D. K. 1994, A&A, 289, 607

Weaver, H. A. & Feldman, P. D. 1992, in Science with the Hubble Space Telescope, P. Benvenuti & E. Schreier, Garching: ESO, 475

Weaver, H. A. et al. 1992, Icarus, 97, 85

Weaver, H. A. et al. 1994a, Science, 263, 787

Weaver, H. A., Feldman, P. D., McPhate, J. B., A'Hearn, M. F., Arpigny, C., & Smith, T. E. 1994b, ApJ, 422, 374 et al. 1995, Science, 267, 1282