STScI Newsletter
2021 / Volume 38 / Issue 02

About this Article

P. D. Feldman (c/o hal.weaver[at]

Department of Physics and Astronomy, The Johns Hopkins University, 3400 N. Charles Street, Baltimore, Maryland 21218, USA


Editor’s Note: Part of the article below is an abridged and updated version of a paper originally published in a Conference Proceedings: From Giotto to Rosetta: 30 Years of Cometary Science from Space and Ground, eds. Cesare Barbieri and Carlo Giacomo Someda, Galilean Academy of the Sciences, Humanities, and Arts and CISAS University of Padua, held 27–29 October 2016 in Padua, Italy.

1. Introduction

Early discussions about the Space Telescope (the name Hubble was added later), originally scheduled for launch in 1983, featured detailed plans for the observation of comet 1P/Halley in 1986 (Morrison 1979; Brandt 1982), including the spacecraft’s maximum moving target tracking rate that was based on the motion of Halley on the sky at its closest approach to Earth in April 1986. Regrettably, various problems led to multiple postponements and HST was finally launched in April 1990, albeit with a primary mirror defect of spherical aberration, which was later corrected. Nevertheless, the nearly 30 years of Hubble’s far ultraviolet (FUV) spectroscopy of comets provides an enduring legacy of probing cometary composition, which we summarize here.

The original complement of instruments on board Hubble included two with FUV capability, the Goddard High Resolution Spectrograph (GHRS) and the Faint Object Spectrograph (FOS). The latter suffered from the use of visible light sensitive detectors that made it less sensitive in the far-ultraviolet than the GHRS. In 1997, these were replaced by the Space Telescope Imaging Spectrograph (STIS), and in 2009 the Cosmic Origins Spectrograph was installed in HST. We will limit our discussion to the wavelength range 1150–3100 Å covered by the Hubble instruments. Comets observed spectroscopically by HST are listed in Table 1. In the following sections, we will highlight some of the main results by instrument.

Table 1: Comets observed in the far-ultraviolet by Hubble




103P/Hartley 2

September 1991


C/1991 T2 (Shoemaker-Levy)

July 1992



March 1993


D/Shoemaker-Levy 9

July 1993

January–July 1994



C/1995 O1 (Hale-Bopp)

September 1995–October 1996


29P/Schwassmann-Wachmann 1

March 1996


C/1996 B2 (Hyakutake)

March/June 1996



June–July 1996



July 1996/January 1997


C/1995 O1 (Hale-Bopp)

August 1997/February 1998


103P/Hartley 2

January 1998


C/1999 H1 (Lee)

June/August 1999


C/1999 S4 (LINEAR)

July 2000


C/2000 WM1 (LINEAR)

December 2001



April 2002


C/2001 Q4 (NEAT)

May 2004


C/2002 T7 (LINEAR)

June 2004


103P/Hartley 2

September/November 2010


C/2009 P1 (Garradd)

January 2012


C/2012 S1 (ISON)

May, October–November 2013


C/2014 Q2 (Lovejoy)

February 2015



December 2019–January 2020




2. Faint Object Spectrograph (FOS)

The first comet to be observed by the FOS, in September 1991, was 103P/Hartley, a Jupiter family comet that was to be observed on successive apparitions by later instruments and was the target of NASA’s EPOXI fly-by mission in 2010. The first detection of CO Cameron band (a3Π−X1Σ+) emission was made in this comet (Weaver et al. 1994), and also subsequently observed by FOS in comets C/1991 T2 (Shoemaker-Levy) and C/1996 B2 (Hyakutake), the latter at higher spectral resolution (McPhate et al. 1996). The primary source of CO in the metastable a3Π state is photodissociation of CO2, so this detection in 103P allowed a means of determining the relative CO2 coma abundance, which relative to that of water, was found to be comparable to that derived for 1P/Halley from in situ mass spectrometer measurements made during the Giotto encounter in March 1986. However, CO Fourth Positive Group (A1Π−X1Σ+, 4PG) fluorescence, a measure of the CO abundance in the coma, was not seen, and the derived 3σ upper limit, ∼1%, indicated that CO was significantly depleted relative to Halley. This conclusion was later borne out by more sensitive COS observations of 103P in 2010 (Weaver et al. 2011), which are discussed below.

The close passage of C/1996 B2 (Hyakutake) to Earth in April 1996 (0.10 AU) allowed the FOS detection of S2, which had only been seen once before, in comet C/1983 H1 (IRAS-Araki-Alcock) with IUE under similar observing geometry (A'Hearn et al. 1983). The detection of this species in 67P/Churyumov-Gerasimenko by the mass spectrometer ROSINA on Rosetta has stimulated renewed interest in its formation process in cometary ice (Mousis et al. 2017). S2 was subsequently also observed by STIS in comets C/1999 H1 (Lee) and 153P/Ikeya-Zhang, taking advantage of the very high spatial resolution of this instrument.

3. Goddard High Resolution Spectrograph (GHRS)

The close approach of C/Hyakutake also allowed for significant measurements by the GHRS. The H Ⅰ Lyman-α line profile was used to probe the dissociation dynamics of H2O in the coma (Combi et al. 1998). The D Ⅰ Lyman-α line, shifted 0.3 Å shortward of the H Ⅰ line, one of the initial objectives of the GHRS (Brandt 1982), was not detected, and it remained for STIS, using an echelle mode with a resolving power of ∼110,000, to make this detection in comet C/2001 Q4 (NEAT) in 2004 (Weaver et al. 2004). CO and H2 fluorescence induced by solar Lyman-α and O Ⅰλ1304 were also detected in C/Hyakutake, primarily because of the high CO abundance in this comet (Wolven & Feldman 1998; Lupu et al. 2007).

4. Space Telescope Imaging Spectrograph (STIS)

STIS is an almost ideal instrument for long-slit spectroscopy of extended sources such as comets. In addition to visible channels using CCD detectors, STIS includes two UV channels, FUV and NUV, together spanning 1150–3100 Å, with photon counting detectors and multiple gratings for a range of spectral resolving powers. With detector pixels of ∼0.025" square, the spatial resolution along the 25" long slit translates to 35 km for a comet at a geocentric distance of 1 AU. The long wavelength end of the NUV spectral range includes the strong (0,0) band of the OH A2Σ+−X2Π system centered near 3090 Å, and which is used for a surrogate for the H2O abundance in the comet (Schleicher & A'Hearn 1988). Models used to determine the H2O production rate, Q(H2O) (Festou 1981), can be compared to the observed spatial profiles, often showing asymmetrical outflow of the OH dissociation product. Nevertheless, the determination of the nearly simultaneous H2O production rate allows for the abundances of the other detected species to be given in terms of abundance relative to H2O.

The NUV range also allows for the detection of S2, which because of its short photodissociation lifetime (∼300 s at 1 AU) and relatively low abundance requires the high spatial resolution of STIS to be seen. As noted above, S2 has been detected in only two of the recent comets observed by STIS, moderately active ones that can be defined as having Q(H2O) ≥ 1029 molecules s–1. The same holds for the detection of CO Cameron bands, which despite the usually large relative abundance of CO2 remain weak features of the coma spectrum because of their high rotational temperature that distributes the photons of a single band over multiple detector pixels. We note also that most comets observed by Hubble are roughly 1 AU from the Sun, due in part to the HST solar exclusion zone of 50º that limits the ability to observe closer to the Sun where the solar fluorescence is higher.

An FUV spectral image of comet 153P/Ikeya-Zhang, obtained on 2002 April 20, is shown in Figure 1 (from Lupu et al. 2007). The geocentric distance of the comet was 0.43 AU so that the image, top to bottom, spans 2,800 km on either side of the nucleus. The brightest spectral features are C Ⅰ λ1561 and C Ⅰλ1657, which are dissociation products and extend fairly uniformly far into the coma. The others are primarily bands of the CO 4PG that peak near the nucleus (the zero-offset position) and then decrease due to the radial outflow of CO from the nucleus. Analysis of the individual bands by Lupu et al. showed that the strongest bands were optically thick near the nucleus due to saturation of the solar pumping radiation, which was then modeled to give a good fit to the observed spectra. These models have been widely used in the interpretation of subsequent observations of this band system, both by COS, and by the Alice far-ultraviolet spectrograph on Rosetta.

Spectra of comet showing bands of bright pixels
Figure 1: STIS spectral image of comet 153P/Ikeya-Zhang, showing the brightness variations along the slit. The spectral region containing the strong geocoronal H I λ1216 and O I λ1302 has been excluded. The carbon multiplets at 1561 Å and 1657 Å are the strongest features and are relatively constant as they are dissociation products extending far into the coma. The zero point in the vertical direction marks the location of the nucleus. From Lupu et al. (2007).

5. Cosmic Origins Spectrograph (COS)

COS, installed in Hubble in 2009, is exclusively an ultraviolet spectrograph comprising both FUV and NUV channels. The FUV channel is complementary to that of STIS in that its sensitivity and spectral resolution (in a standard observing mode) are significantly better, but without any spatial resolution that makes STIS valuable for extended source observations. With a 2.5" diameter entrance aperture, spectral resolution of ∼1.0 Å is achieved for an extended source, making it possible to resolve multiplet structure and molecular band structure. The small aperture also favors species originating on or near the nucleus.

As noted above, the first comet observations using COS were made of the Jupiter family comet 103P/Hartley before, during and after the fly-by of this comet by the EPOXI spacecraft in November 2010. The high FUV sensitivity and small field of view permitted several bands of the CO 4PG to be detected, leading to a derived relative CO/H2O abundance of 0.15%–0.45% (Weaver et al. 2011), consistent with the previous upper limit from 1991 (Weaver et al. 2004), even though the H2O production rate had decreased by a factor of ∼5 from 1991. The strongest features in the spectrum, however, were partially resolved multiplets of S Ⅰ, centered at 1429 Å and 1479 Å, and multiplets of C Ⅰ at 1561 Å and 1657 Å. These are characteristic of all of the comets observed by COS, and are discussed further below.

Comets C/2009 P1 (Garradd) and C/2014 Q2 (Lovejoy) were significantly more active and more abundant in CO than 103P so that COS was able to detect nineteen individual bands of the CO 4PG system in these comets. As an example, the spectrum of C/Garradd is shown in Figure 2, together with a synthetic CO 4PG spectrum using the model of Lupu et al. (2007). Nearly simultaneous STIS spectra were used to obtain the H2O production rate from the observed OH emission so that the relative abundance of CO to H2O, often found to be temporally variable, could be determined. C/Garradd was particularly interesting in this respect as due to favorable viewing geometry, it was observed over a long period of time and was found to exhibit very asymmetric outgassing around perihelion. Following perihelion, while the visual magnitude remained relatively constant, the H2O production rate declined rapidly with time, while that of CO increased rapidly (Feaga et al. 2014).

Line chart of brightness and wavelength of comet
Figure 2: COS G160M spectrum of comet C/2009 P1 (Garradd) taken on 2012 January 19, exposure time 999 s. Top panel: FUV A channel; bottom panel: FUV B channel. A CO Fourth Positive Group (4PG) synthetic spectrum from the model of Lupu et al. (2007) is overplotted in red. Atomic multiplets of C and S are indicated. Also identified (in the top panel) are two He I lines (green) due to terrestrial emission, seen in third order.

Long-slit spectroscopy has demonstrated that all the S I emissions observed in comets are very extended, confirming that they were produced by resonance scattering of S atoms following photodissociation of short-lived sulfur-bearing molecules and not by direct excitation of these same molecules. The only sulfur-bearing molecules detected in the UV are CS (a dissociation product of CS2) and S2, and these are not sufficient to produce the amount of atomic sulfur inferred from the S Ⅰ emissions (Meier & A'Hearn 1997). Several other sulfur-bearing molecules were detected through millimeter wave spectroscopy (Bockelée-Morvan et al. 2004), but the full complement of sulfur-bearing molecules was not revealed until mass spectroscopic measurements by the ROSINA instrument on Rosetta (Calmonte et al. 2016). They confirmed earlier results that H2S was the most abundant of these species in 67P, and it is probable that this is the case for most comets, including long-period comets. The short photodissociation lifetime of H2S can account for the relatively high atomic sulfur column densities derived from the COS spectra.

Remarkably, two objects on clearly interstellar orbits (i.e., with hyperbolic heliocentric orbital elements) have been discovered during the last several years: 1I/'Oumuamua in 2017 and 2I/Borisov in 2019. Although 'Oumuamua exhibited non-gravitational accelerations usually associated with mass ejection and its reaction force on the nucleus, neither dust nor gas emissions were ever detected (ISSI Team 2019), which suggests that 'Oumuamua was unlike typical comets. In contrast, 2I/Borisov displayed both dust and gas emissions and had other comet-like properties. Hubble COS observations of 2I/Borisov detected CO 4PG emissions during four different sets of observations taken in December 2019 and January 2020 (Bodewits et al. 2020). The estimated CO/H2O ratio of ~1.7 was unusually large, more than three times larger than the largest value measured at a similar heliocentric distance (2 AU) in solar system comets, which suggests that 2I/Borisov formed outside the CO snow line in an extrasolar planetary system.

6. JWST Investigations of Cometary Composition

HST's FUV spectroscopic investigations are sensitive to atoms and small molecules. The only "parent" species (i.e., those sublimating directly from the nucleus) probed by Hubble are CO and S2. Moderate resolution ( l/dl ~ 2700) infrared (IR) spectroscopy of comets by Webb can provide direct measurements of the key parent molecules H2O, CO, and CO2, which usually are responsible for driving cometary activity. The JWST spectral range also provides access to the fundamental vibrational transitions of many organic species, including those that don’t have allowed radio/mm/sub-mm transitions (e.g., CO2, CH4, C2H6, C2H2,…).

The high sensitivity of Webb can extend IR compositional measurements to larger heliocentric distances than previously possible. For example, JWST monitoring of a single comet starting at ~10 AU pre-perihelion (where both CO and CO2 are active) and extending to ~10 AU post-perihelion could provide a unique and uniform set of data that could be used to investigate in detail the various processes and conditions that affect cometary outgassing behavior (e.g., sublimation temperature, exposed active area, seasonal effects, etc.). Further details on potential Webb spectroscopic investigations of comets can be found in Kelley et al. (2016) and Milam et al. (2016), with the latter also providing information on the various operational issues associated with JWST observations of moving objects (e.g., targets that require non-sidereal tracking).

With the successful launch of JWST on Christmas Day 2021, we eagerly anticipate the initial spectroscopic observations of a comet and trust that the compositional diversity of comets will be revealed by subsequent observations of multiple comets over Webb's lifetime.


The author wishes to thank many students and collaborators, and in particular Hal Weaver, over the past 30~years. Thanks are also due to the operations teams at STScI for the superb planning and execution of difficult observations of moving targets. This work is based on observations made with the NASA/ESA Hubble Space Telescope, which is operated by the Association of Universities for Research in Astronomy, Inc., under NASA contract NAS 5-26555. Financial support was provided by NASA through grants from the Space Telescope Science Institute.


A'Hearn, M. F., Schleicher, D. G., & Feldman, P. D. 1983, ApJL 274, L99

Bodewits, D., Noonan, J. W., Feldman, P. D., et al. 2020, Nature Astronomy 4, 867

Bockelée-Morvan, D., Crovisier, J., Mumma, M. J., & Weaver, H. A. 2004, in Comets II, ed. M. C. Festou, H. A. Weaver, & H. U. Keller (Tucson: Univ. of Arizona), pp. 391–423

Brandt, J. C. 1982, in NASA Conference Publication, ed. D. N. B. Hall, Vol. 2244

Calmonte, U., Altwegg, K., Balsiger, H., et al. 2016, MNRAS 462, S253

Combi, M. R., Brown, M. E., Feldman, P. D., et al. 1998, ApJ 494, 816

Feaga, L. M., A'Hearn, M. F., Farnham, T. L., et al. 2014, AJ 147, 24

Feldman, P. D., A'Hearn, M. F., Bertaux, J.-L., et al. 2015, A&A 583, A8

Festou, M. C. 1981, A&A 95, 69

'Oumuamua ISSI Consortium, 2019, Nature Astronomy 3, 594

Kelley, M. S. P., et al. 2016, PASP, 128:018009

Lupu, R. E., Feldman, P. D., Weaver, H. A., & Tozzi, G.-P. 2007, ApJ 670, 1473

McPhate, J. B., Feldman, P. D., Weaver, H. A., et al. 1996, BAAS 28, 1093–1094

Meier, R., & A'Hearn, M. F. 1997, Icarus 125, 164

Milam, S. N., Stansberry, J. A., Sonneborn, G., & Thomas, C. 2016, PASP, 128:018001

Morrison, D. 1979, in NASA Conference Publication, Vol. 2111, pp. 77–97

Mousis, O., Ozgurel, O., Lunine, J. I., et al. 2017, ApJ 835, 134

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

Stern, S. A., Slater, D. C., Scherrer, J., et al. 2007, Space Sci. Rev. 128, 507

Weaver, H. A., A'Hearn, M. F., Arpigny, C., et al. 2004, in AAS/Division for Planetary Sciences Meeting Abstracts #36, Vol. 36, p. 1120

Weaver, H. A., Feldman, P. D., A'Hearn, M. F., Dello Russo, N., & Stern, S. A. 2011, ApJL 734, L5

Weaver, H. A., Feldman, P. D., McPhate, J. B., et al. 1994, ApJ 422, 374

Wolven, B. C., & Feldman, P. D. 1998, in The Scientific Impact of the Goddard High Resolution Spectrograph, ed. J. C. Brandt, T. B. Ake, & C. C. Petersen, Vol. 143, p. 373



This site is protected by reCAPTCHA and the Google

Contact our News Team 

Contact our Outreach Office