Exocomets from a Solar System Perspective

For comets, "activity" usually refers to the sublimation of ices stored in the nucleus. The expanding gas drags small dust (and sometimes ice) particles along. The main ices in most comet nuclei are H₂O, CO₂, and CO; their relative abundances appear to vary greatly among comets and may be a primordial property (A'Hearn et al. 2012; Ootsubo et al. 2012; Bockelée-Morvan & Biver 2017).

These ices have very different sublimation temperatures and can therefore start to effectively sublimate at different distances from the Sun (Meech & Svoren 2004; Burke & Brown 2010). The onset of significant effective sublimation of H₂O ice from the comet nucleus occurs at 180 K or 2.5 au from the Sun; CO₂ at 80 K or out to 13 au; and CO at 25 K or 120 au. It is of note that sublimation does not stop outside these distances but rates drop off exponentially. Consequently, comets can be active at great distances from the Sun (Jewitt et al. 2019).

Typical water sublimation rates of comets observed from Earth are between 10²⁷ and 10³⁰ water molecules s⁻¹, equivalent to 30–30,000 kg s⁻¹. Jupiter Family comets typically reach water production rates of order 10²⁸ molecules s⁻¹; whereas Oort Cloud comets frequently reach 10²⁹ molecules s⁻¹. Sustained comet water production rates above a level of ∼5 × 10²⁹ molecules s⁻¹ are exceptional (A'Hearn et al. 1995). The largest reported are for comets C/1975 V1 (West) and C/1995 (Hale–Bopp), which both reached peak water production rates above 10³¹ molecules s⁻¹ (A'Hearn et al. 1995; Biver et al. 1997), though higher production rates likely briefly occur in the case of comets that get extremely close to the Sun. For example, C/2011 W3 (Lovejoy) had production rates ∼10³¹ molecules s⁻¹ despite having a nucleus that was at least an order of magnitude smaller than that of Hale–Bopp (Raymond et al. 2018a).

The Rosetta spacecraft traveled alongside comet 67P/Churyumov–Gerasimenko for two years and found that the percentage of surface that can be considered as "active" varies greatly along the orbit, and is influenced not only by the illumination geometry but also by local structural and compositional properties (Combi et al. 2020). For most comets, only a small fraction of the surface, of order 5%, contributes to the sublimation activity (A'Hearn et al. 1995), but there are comets for which significantly more of the surface appears to be active (Bodewits et al. 2014). Other comets eject large amounts of icy grains that seemingly exaggerate the nucleus's active surface area (such as 103P/Hartley 2; A'Hearn et al. 2011).

Comet-like, sublimation-driven activity can occur even if an object contains no water ice. Silicates will start to sublimate at 1000–1500 K, depending on their Mg/Fe content (see Figure 17 in Jones et al. 2018). Such temperatures occur only when objects get within ∼0.1 au of the Sun. Asteroid (3200) Phaethon has displayed faint comet-like activity near its perihelion at 0.14 au, with a variety of mechanisms suggested for this behavior (Jewitt & Li 2010).

Very little is known about comet interiors and how ices are stored and mixed within them (A'Hearn 2017). The dust-to-ice ratio of comets is an area of active debate; the Rosetta mission to comet 67P/Churyumov–Gerasimenko suggested that its nucleus may contain far less volatile material than was previously assumed for comets, with a refractory to ice ratio of 6 ± 3 in this object (Fulle et al. 2017, 2019; Choukroun et al. 2020). In addition, it is a long-standing question whether comets (and other small bodies such as Centaurs) contain amorphous water ice, whose conversion to crystalline ice could drive activity at large distances from the Sun (Prialnik & Bar-Nun 1987; Meech & Svoren 2004; Guilbert-Lepoutre 2012; Agarwal et al. 2017).

Once ejected material has left the vicinity of a solar system comet's nucleus, it forms the comet's dust and ion tails. Neutral cometary gas is ionized by photoionization, charge transfer with solar wind protons, and electron impact ionization (Cravens 1991). Solar system comets are immersed in the solar wind, a continuous anti-sunward stream of ionized material flowing at several hundred km per second. Once ionized, this plasma is subsumed by the solar wind, forming an ion tail which can be observed remotely. Ion tails can span several astronomical units in length and often appear blue due to the resonant fluorescent emission of CO⁺ (next section). Cometary dust is accelerated anti-sunward by radiation pressure, generally following the classic dust tail formation process first formulated by Finson & Probstein (1968). There are indications that dust can also be affected by the solar wind, i.e., that it is electrically charged, and is affected by the plasma flowing from the Sun, e.g., Kramer et al. (2014), Price et al. (2019).

Most of our knowledge about the composition of comets comes from observations of the gas and dust surrounding them. The elemental composition of comet dust has been studied in situ by instruments on board spacecraft like Vega 1 and 2, Giotto, and Rosetta (Cochran et al. 2015; Bardyn et al. 2017), in laboratories on Earth by collecting particles in Earth's upper atmosphere (Sykes et al. 2004), and directly from a comet's atmosphere by the Stardust mission (Brownlee 2014). The Giotto, Stardust, and Rosetta–COSIMA results were consistent and showed that approximately 50% of the mass of the cometary dust is solid organic matter (the other half consisting of minerals) and that the average elemental composition is close to that of the Sun (Bardyn et al. 2017) with H and He being notable exceptions (they are more abundant in the Sun). Remote observations of dust are most diagnostic in the mid-infrared, where there are absorption features of minerals, but this has only been possible for relatively few bright comets (Wooden et al. 2017). In optical and near-IR wavelengths, the spectrum of comet dust is mostly featureless and slightly reddened with respect to sunlight. There is a very strong forward scattering phase effect (e.g., Marcus 2007; Bertini et al. 2017) that can enhance the dust's brightness by a factor 1000 or more at phase angles above 175° (Hui 2013).

Most remote gas-phase chemical abundances are measured with respect to H₂O. At distances further than 2.5 au from the Sun, the sublimation rates of water decline quickly. The relative abundance at larger heliocentric distances thus does not represent the ice composition of the nucleus (Ootsubo et al. 2012). After H₂O, the main components of gas comae are usually CO and CO₂ (both 0%–30% with respect to H₂O; A'Hearn et al. 2012). Additionally, smaller amounts of other molecules are routinely observed in active comets, including CH₄, HCN, CH₃OH, H₂S, etc. Inventories of molecules are given in dedicated review papers (Mumma & Charnley 2011; Bockelée-Morvan & Biver 2017; Altwegg & ROSINA Team 2018; Altwegg et al. 2019). The ROSINA instrument on board Rosetta detected more than 55 different species surrounding 67P/Churyumov–Gerasimenko (Altwegg & ROSINA Team 2018), including the amino acid glycine (Altwegg et al. 2016), molecular oxygen (Bieler et al. 2015), and the noble gases Ar and Xe (Balsiger et al. 2015; Marty et al. 2017).

Finally, there are many ions, radicals, and fragment species that are formed by physical reactions in the coma or as products of photodissociation, such as CO⁺, H₂O⁺, CN, CS, OH, C₃, C₂ etc. (A'Hearn et al. 1995; Fink 2009; Cochran et al. 2012). Many of these species have been observed from the ground in numerous comets because they have relatively large fluorescence efficiencies, are often longer lived than their parents, and they are accessible at multiple wavelengths and with different techniques. However, the origin of many fragment species is unknown or ambiguous (Feldman et al. 2004). An often encountered question in cometary volatiles is therefore whether they are native to the comet nucleus or formed from other processes on the surface or in the coma.

Cometary contents have been studied and detected in almost every part of the electromagnetic spectrum (see Table A1 for a summary of features). In this section, we aim to identify the brightest and most useful features, the processes that drive them, and some of the diagnostics they provide. We recall here that this paper is intended as a broad overview which applies to remote observations of comets, typically at a distance of 1 au from Earth and the Sun. On a one-by-one case, comets can behave very differently, caused by, for example, different observing circumstances (extremely close approaches to Earth, space missions), or intrinsic physical properties (very low or high production rates, close proximity to the Sun, unusual chemical composition). Given the depth and long history of comet science, this work is also not intended as an exhaustive review, but merely to provide some useful starting points and examples. There are a number of reviews that provide a more comprehensive overview of the field (including Feldman et al. 2004; Mumma & Charnley 2011; Cochran et al. 2015; Bockelée-Morvan & Biver 2017).

Gases in the coma of solar system comets emit light throughout the electromagnetic spectrum through different excitation mechanisms. At most wavelengths, the dominant process observed is solar pumped fluorescence where small molecules re-emit upon excitation with high efficiency. A second process is emissive photodissociation, also known as prompt emission, where photodissociation produces excited fragments ([O i], [N i], [C i]; see McKay et al. 2013; Opitom et al. 2019b). Third, in the inner coma, electron impact dissociation reactions may produce excited atomic and molecular fragments such as H, [O i], CO, OH, etc. (Feldman et al. 2015; Bodewits et al. 2016). Given the right conditions this emission mechanism can reveal small traces of volatiles. In a different setting it also led to the discovery of H₂O plumes above the surface of Europa (Roth et al. 2014).

In 1996, Chandra and the Extreme UltraViolet Explorer unexpectedly discovered that comets can be bright (>1 GW) extreme ultraviolet and X-ray sources (Lisse et al. 1996; Krasnopolsky et al. 1997). When highly charged solar wind ions (O⁸⁺, O⁷⁺, C⁶⁺, ...) collide with the neutral gas surrounding comets, the ions capture one or more electrons into an excited state. As they cascade to the ground state, they emit X-rays (Cravens 1997). This charge exchange emission has distinct spectral features different from thermally excited plasmas and can be used to help characterize the solar wind plasma and structures such as comets' solar wind bow shocks (Wegmann et al. 2004), the amount of neutral gas present, and possibly even its composition (Bodewits et al. 2007; Mullen et al. 2017).

In the far- and mid-ultraviolet (120–300 nm), several bright emission lines from atoms such as H, C, and O can be detected. The fluorescent Lyα emission of atomic hydrogen is routinely used to determine comet water production with the Solar Wind Anisotropies/Solar and Heliospheric Observatory (SWAN/SOHO) survey instrument (see Combi et al. 2011). Notable other bright emission features are the CO fourth positive system (130–190 nm; Lupu et al. 2007), the forbidden Cameron bands of CO (190–280 nm; Weaver et al. 1994), and the CO₂⁺ feature around 289 nm (Festou 1981).

At near-UV and optical wavelengths (300–700 nm), the most commonly observed lines are from fragment species (OH, CN, C₂, etc.), which emit at easily accessible optical wavelengths and have very high fluorescence efficiencies. Several high-quality line atlases are available for optical spectra (including Cochran & Cochran 2002; Cremonese et al. 2002, and Brown et al. 1996). Comet-specific narrowband filters (Farnham et al. 2000; Schleicher & Farnham 2004) are commonly used to image the distribution of fragment species in the coma and to survey production and mixing rates (A'Hearn et al. 1995).

The strongest feature in this region is the OH A ²Σ⁺–X ²Π_i (0-0) band around 308.5 nm but due to significant atmospheric extinction in the near-UV and limited telescope facilities that are sufficiently blue sensitive, it is often challenging to study. Instead, the strong CN B ²Σ⁺–X ²Σ⁺ violet band near 388.3 nm is generally the most accessible bright cometary feature (see Feldman et al. 2004) that was already identified 139 yr ago (Huggins 1881). Although the chemical source of CN is debated (several comets have more CN than its presumed parent HCN), its emission has been detected at large distances (>5 au) from the Sun, e.g., Cochran & Cochran (1991), Schleicher et al. (1997). Another feature that can be very bright is the CO⁺ A ²Π–X ²Σ "comet tail" emission between 300 and 500 nm (see Opitom et al. 2019b), although the brightness of this feature tends to vary substantially between comets, presumably due to the wide range of involved CO abundances (Dello Russo et al. 2016). Also small carbon chain radicals have been observed. In the 19th century, Huggins (1881) observed a blue band at 405.2 nm in the spectrum of comet Tebbutt, that later was identified as arising from C₃ emission (Douglas 1951).

Emission from neutral Na may also be seen in solar system cometary spectra, particularly those of near-Sun comets (Jones et al. 2018 and references therein). The origin of this species in these comets is undetermined. Occasionally, spectra have Na features at two distinct velocities (Cremonese 1999; Leblanc et al. 2008), possibly suggesting multiple sources. The Rosetta spacecraft did detect Na in dust grains (Schulz et al. 2015), but a consensus on a dust and/or nuclear ice source in different comets has not yet been reached (e.g., see Cremonese 1999 and Cremonese et al. 2002 for reviews). Ellinger et al. (2015) have suggested an ice source of Na. Sodium observations are discussed further in Section 4.3. Despite being key species in exocomet studies, the Ca⁺ ion, and the neutral Ca species, have been observed in a few solar system comets, with the bright sungrazer C/1965 (Ikeya–Seki) being the first where Ca i and Ca ii were detected remotely in significant abundances (Slaughter 1969). Their sources are undetermined.

Parent molecules are generally observed at infrared, submillimeter, and millimeter wavelengths (Bockelée-Morvan et al. 2004; Mumma & Charnley 2011; Cordiner et al. 2014). Symmetric species lacking permanent dipoles and therefore pure rotational spectra (CO₂, CH₄, C₂H₂, C₂H₆, C₂H₄,...) can only be observed through rovibrational transitions in the IR upon asymmetric excitations, while other polar complex volatiles excited by solar radiation and/or collision processes, also exhibit pure rotational transitions at radio wavelengths. Measurements at these wavelengths offer insight into the molecular complexity of cometary comae (Biver et al. 2015), distribution of species throughout the comae (Cordiner et al. 2014), as well as other physical mechanisms such as jet activity, rotation, gas temperature profile, distributed sources, etc. (see Drahus et al. 2010; Bonev et al. 2013; Cordiner et al. 2017). These techniques are entering a new era in sensitivity, resolution, and bandwidth (allowing the coverage of multiple molecules simultaneously) that will not only significantly advance comet studies within the solar system, but likely also be employed for future studies of exocomets.

As is the case with other astronomical spectroscopic fields, the complete and accurate interpretation of cometary composition is limited by the availability of high-quality spectral data, such as comprehensive line lists and accurate broadening parameters. Currently, complete spectral data only exists for a fraction of the molecular species that could be present in comets (Sousa-Silva et al. 2019). As such, cometary spectra are vulnerable to misinterpretation, and molecular detections susceptible to misassignments. Table A1 highlights the complexity of molecular composition of comet spectra, with dozens of molecular species already discovered. It is therefore plausible that many more molecules are present in comets that cannot yet be correctly identified due to lack of spectral data; it is crucial that these fundamental molecular spectra are obtained, either through experimental measurements (e.g., Gordon et al. 2017) or theoretical calculations (e.g., Tennyson et al. 2016; Fortenberry et al. 2019; Sousa-Silva et al. 2019).

The properties of comets and other small bodies in the solar system are not strictly defined by the International Astronomical Union (IAU). Resolution 5A for its General Assembly XXVI in 2006 defined planets and dwarf planets. By exclusion, all other bodies, except satellites, orbiting the Sun were collectively defined as small solar system bodies. This population includes a broad variety of objects, including asteroids, Trojans, moons, Centaurs and comets. These classes are mostly defined by their dynamical properties such as their location in the solar system and their relation with Jupiter (Levison 1996). In practice, there is a strong phenomenological component to the qualification of small bodies. Objects are identified as comets when they show activity consisting of a cloud of gas and dust (coma) and/or tails of ions and/or dust. In that simple view, comets are active due to the sublimation of volatile ices while asteroids are rocky and do not show such activity. This phenomenological approach is not perfect; a small number of asteroids have been observed to demonstrate activity (i.e., mass loss) through a variety of non-volatile sublimation processes, including impacts, rotational spin up, and in some cases repeated comet-like activity driven by sublimation of refractory material (Jewitt et al. 2015). Comets can cease to show activity far away from the Sun or as an evolutionary end-state when volatiles near their surface are depleted (Jewitt 2004). Finally, different small body populations in the solar system may be connected as objects evolve dynamically and can thus technically change classification over time—for example objects that are now in the so called "scattered disk" may become Centaurs and eventually Jupiter Family Comets (Bernstein et al. 2004; Sarid et al. 2019).

The identification of an object as a comet, asteroid, Centaur, or even Kuiper Belt object depends on its location in the solar system architecture and these terms raise expectations about the nature of the small body considered. In this paper, we use the word "exocomet" to describe comets which orbit other stars than the Sun and which exhibit some form of observable activity such as the release of gas or dust, e.g., through a coma or tails of ions or dust. The term "FEBs" has been used extensively as a synonym for exocomets (e.g., Lagrange-Henri et al. 1992; Beust & Lissauer 1994), though this term is misleading: the activity in such objects is likely driven by sublimation rather than evaporation, and "falling" implies objects seen only prior to periastron, while objects are now also detected after periastron (e.g., Kiefer et al. 2014b) when they could be described as "rising." The variable gas absorption signatures described in Section 3.2.1 clearly indicate the presence of exocometary gas and are thus characterized as bonafide exocomets. The asymmetrical transit signatures observed photometrically thought to be caused by transiting dust tails (as described in Section 3.2.2). The presence of trace amounts of cold gas within debris disks surrounding nearby main-sequence stars is the product of the release of gas from exocometary ices (as described in Section 3.2.3). Observational evidence has opened up the possibility that exocomets may be accreting onto WDs (e.g., Xu et al. 2017) as observed indirectly through the analysis of WD atmospheres (see Section 3.2.4). This suggests the accreting bodies might in some cases show compositional properties akin to solar system comets which indicates the bodies might indeed represent their extra solar equivalents, although the accretion is more likely dominated by asteroids.

For cases where no cometary activity signatures are observed, we suggest the more inclusive term "Extrasolar Small Bodies" (ESBs). The properties of ESBs are not well constrained and we thus caution that what an object is identified as might later change as new information becomes available—a situation not uncommon for small bodies in our solar system.

3.2.1. Spectroscopic Observations of Variable Absorption Lines

Exocomets can be discovered spectrocopically by the variable absorption they cause in addition to the stationary stellar Ca ii H & K lines (see Figure 1). The absorption occurs when the exocometary gas passes in front of the star, absorbing part of the stellar light as it transits. The interstellar medium also causes absorption in addition to the exocomets. However, this absorption contribution remains constant when compared to the exocomet features which can vary from every few hours to a few times per month. Due to the small equivalent width (EW) of these variable exocomet features (few mÅ), only high-resolution instruments (R > 60,000; Δv_r < 5 km s⁻¹) are suitable for these observations.

Zoom In Zoom Out Reset image size

β Pic is a young (∼23 Myr) and bright (V_mag = 4) A-star which exhibits the largest amount of exocometary activity of any star observed to date. Due to the edge-on inclination of the system as seen from Earth, β Pic is well positioned to detect exocomets in the same orbital plane, but much closer to the star than the debris disk observed at tens of au (e.g., Matrà et al. 2019b). The orbits of the exocomets around β Pic have been estimated indirectly using radial velocity observations combined with a physical sublimation model in an effort to calculate the stellocentric distance at transit (Beust et al. 1990; Beust & Tagger 1993). Rapidly varying absorption features (on the order of hours) are observed in sequential radial velocity spectra and are interpreted as being due to accelerating exocomets. The measured acceleration constrains the distance at which the exocomets pass in front of the star which is found to be within a few tens of stellar radii (Kennedy 2018). These observations are consistent with analytical estimates by Beust et al. (1996) who estimated that low-velocity Ca ii absorption features correspond to distances of ∼15–30 stellar radii (R_*) whereas high-velocity features correspond to distances of 5–8 R_*. At distances ≲3R_* the radiation pressure becomes too high for a large and thus detectable cloud to form. This is not the case for other lines such as Mg ii and Al iii where the radiation pressure is at least 10 times smaller (Beust et al. 1996). There is now a large set of β Pic spectra, routinely showing deep short term variations in several spectral lines including Ca ii H & K in the visible and, for instance, Mg ii and Fe ii in the UV. There is a wide variety in the variable features' EW and velocities. They spread between −200 and +200 km s⁻¹ in radial velocity (see e.g., Kiefer et al. 2014b), and are in most cases redshifted compared to the star's Doppler shift, i.e., moving toward the central star. The redshift is due to the projection of the velocity of the exocomet onto the line of sight. These properties led the first discoverers of the β Pic phenomenon, to call them "FEBs," as discussed earlier. With the exception of the most conspicuous events in β Pic, they have EWs of a few mÅ. No system has been found yet (with the possible exception of ϕ Leo; Eiroa et al. 2016) with levels of variability comparable to the canonical case of β Pic. Their high frequency (hundreds of events per year) implies a large number of objects.

There are three exocomet hosting systems showing variable absorption features attributed to exocomets at both optical and UV wavelengths (see Table 1). The low number of objects is mainly due to the limited number of UV-facilities (recently only the Hubble Space Telescope (HST) is capable of UV-exocomet observations). Apart from β Pic, HD 172555 (Kiefer et al. 2014a; Grady et al. 2018) and 49 Cet (Miles et al. 2016), have shown exocometary-like absorption at both optical and UV wavelengths. Examples include the UV lines Fe ii (e.g., Grady et al. 1996, 2018) and carbon or oxygen (Roberge et al. 2006, 2014; Grady et al. 2018) lines.

A few other systems have showed similar variable spectroscopic features (see e.g., Eiroa et al. 2016; Iglesias et al. 2018; Welsh & Montgomery 2018, and references therein) at optical wavelengths. The absorption features have predominantly been found when observing late B and A-type stars. The systems are all young with ages typically ranging from a few tens to hundred of million years old (see Wyatt et al. 2007b; Welsh & Montgomery 2018 and references therein). There is currently no clear explanation whether this is an observational bias or a physical effect. A possible exception, both in spectral type and age, is the 1.4 Gyr, F2V star η Crv where a tentative absorption signature (2.9σ detection) was detected by Welsh & Montgomery (2019). The majority of targets show possible exocomet induced variability in the Ca ii H and/or K lines. With a limited number of observational epochs and lower signal to noise, their identification as exocometary, or at least circumstellar in origin is less certain than those of β Pic, 49 Cet and HD 172555. These candidate exocomet hosting systems which exhibit variability in one of the Ca ii H or K lines or weak photometric signatures are listed in Table 2. All stars in this table require further follow-up to discard periodicity and ensure the variability is not caused by some other process (e.g., stellar pulsations).

3.2.2. Photometric Transit Observations

The milli-magnitude precision of space-based, wide field imagers such as Kepler/K2 and TESS have allowed the first detections of exocomets transiting other stars via photometry (Ansdell et al. 2018; Rappaport et al. 2018; Zieba et al. 2019). These transit events often have a distinct light curve shape reminiscent of a saw tooth whose shape depends on the angle of the trajectory. The shape was first predicted by Lecavelier Des Etangs et al. (1999), see Figure 2. The sharp decrease in flux is caused by a steep increase in absorption from the leading edge of the comet's coma followed by an exponential decay back to the full flux level of the star which is caused by the decreasing absorption from the cometary tail. This technique has proven the presence of comet-like bodies around stars with later spectral types (other than A-type), confirming the possible bias in the spectroscopic method. Estimating the size of exocomets is particularly hard due to the large number of degeneracies involved. As long as the exocometary orbital parameters remain unknown there is a large degeneracy between the transverse velocity, the length of the exocomet tail and the impact parameter (see Zieba et al. 2019).

Zoom In Zoom Out Reset image size

3.2.3. Exocomet Populations within Debris Disks

3.2.4. WD Pollution

Gas is released by exocomets transiting their host star at a few stellar radii as well as by the population of exocomets further out at tens of au within cold debris disks, giving us access to their composition. Several gas species attributed to the presence of exocomets have been detected to date (see Table A2).

At a few stellar radii, gas is thought to originate from sublimation as exocomets move away from or toward the star, producing red or blueshifted gas absorption lines. The most readily detected species is Ca⁺, where absorption is seen in the H&K lines at 3933.7 Å and 3968.5 Å, respectively. The absorption signatures typically vary on timescales of hr to days relative to the absorption features caused by circumstellar and interstellar gas which vary on much longer timescales (Kiefer et al. 2019). Time-variable ultraviolet lines such as Al iii, C ii, C iv, Mg ii, Fe i and Fe ii have been detected in observations of β Pic (Deleuil et al. 1993; Vidal-Madjar et al. 1994; Miles et al. 2016; Grady et al. 2018), see Table A2 for a complete list of species. Although the star itself is unable to photoionize some of these species, it is thought that exocomets sufficiently close to the star (a few stellar radii) form a shock surface where the heat generated by compression and collisions is high enough to produce highly ionized atoms (Beust & Tagger 1993).

The connection between these red or blueshifted exocomets at a few stellar radii from their stars and the exocomets that are much more distant orbiting in exo-Kuiper belts (with lines at the stellar velocity, i.e., no shifts) is not clear. The most intuitive interpretation is that we are seeing the same sort of exocomets but at different locations, i.e., the shifted signatures of exocomets appear when the objects are pushed onto small pericentre orbits from large distances. There are dynamical processes that look promising at producing this inward flow of material (Beust & Morbidelli 1996, 2000; Bonsor et al. 2012, 2013, 2014; Faramaz et al. 2017; Marino et al. 2018; Sezestre et al. 2019). If this is correct, observations of gas released within debris disks at tens of au can be used to constrain the volatile content, while observations of transits close to the star probe the refractories within exocomets. Observations of exocometary gas within debris disks makes it possible to study the composition of the bulk exocomet population. Assuming the collisional cascade that is producing observable dust is also releasing gas at a steady-state, and that all the ice is lost to the gas phase by the time solids are ground down to the small grains (which are then ejected through radiation pressure), allows derivation of exocometary mass fractions of ice species within the debris disk (Zuckerman & Song 2012; Matrà et al. 2015; Kral et al. 2017b). In the systems where CO has been measured by ALMA, the CO (and/or CO₂) mass fraction in exocomets is consistent with solar system cometary compositions, within about an order of magnitude (Matrà et al. 2017). Searches for gas molecules other than CO are underway with ALMA, though these are harder to detect due to their significantly shorter photodissociation timescales (Matrà et al. 2018). Using the excitation of the observed O i line, Kral et al. (2016) quantified the maximum H₂O-to-CO ratio of exocomets within the β Pic debris disk and found that little water is released together with CO; this is consistent with a direct upper limit on H₂O gas emission from Herschel (Cavallius et al. 2019).

Detectable CO release rates observed in debris disks vary by orders of magnitude from 10⁻¹ to less than 10⁻⁴ M_⊕ Myr⁻¹. Taking a typical production rate of 10⁻² M_⊕ Myr⁻¹, we find a production rate of ∼10³⁴ CO molecules s⁻¹, which is much higher than what is observed in the solar system for a given comet (see Section 2.1). This is however not surprising, as the rate in debris disks arises from the collective release from a large amount of exocomets (as first proposed by Lecavelier Des Etangs et al. 1996 for the β Pic debris disk), rather than the sublimation-driven release by a single object typically observed in the solar system.

A major difference between the observations of solar system comets and exocomets is that the former are studied individually, whereas the latter generally cannot be resolved. Persistent and robotic observing campaigns have accrued chemical abundances for hundreds of comets for fragment species (A'Hearn et al. 1995; Fink 2009; Cochran et al. 2012), and for dozens of comets for molecules that are thought to be released directly from the nucleus (A'Hearn et al. 2012; Dello Russo et al. 2016). These surveys indicate that comets have a broad range of properties that are likely related to their origin in the protoplanetary disk (Davidsson et al. 2016; Eistrup et al. 2019), but that are also altered by solar heating and interstellar processing. The extent of these effects on comets, and the exact connection between volatiles stored in the nucleus and the observable coma are among the major questions in comet research (A'Hearn 2017; Keller & Kührt 2020). Similar to the comets of the solar system, exocomet likely also vary in chemical composition depending on their formation environments which is governed by their location in the debris disk and type of star in which they orbit.

For exocomets the determination of the properties of individual objects is a challenge as we can never be sure we are only observing a single object or collection of fragments originating from a single object. Some photometric and spectroscopic transits of exocomets indeed suggest the multiplicity of transiting objects (Beust et al. 1996; Kiefer et al. 2017; Neslušan & Budaj 2017). The exception to this are interstellar visitors (see Section 5) which can be studied individually. Photometric observations which show transit light curves consistent with what is expected from a single exocomet transit may increase the likelihood that a single exocomet is being characterized, although there is currently no method in place for verifying the single nature of photometrically detected exocomets. Observations of the gas in planetesimal belts likely measure the combined composition of large numbers of exocomets. Similarly, the elemental composition derived from WD pollution and spectroscopic observations is likely the product of multiple exocomets and therefore reflects the chemical properties of an entire population of objects.

Exocomets can be detected through the gas and dust they release, producing absorption of starlight and/or emission. First, exocomets at a few stellar radii can be detected with high resolution spectrographs through their gas coma, which covers a significant fraction of the stellar surface and hence makes them detectable in absorption against the star (Lagrange-Henri et al. 1992). High resolving power is needed, as delivered by the current generation of optical and NIR echelle spectrographs have spectral resolutions of 120,000 (Mayor et al. 2003) to 190,000 (Pepe 2017). This allows measurement of the amount of absorption as a function of radial velocity at resolutions of ∼2–5 km s⁻¹. This is more than suitable for detecting exocomets which typically display red and blueshifted absorption signatures with a radial velocity in the range from 0 to ±200 km s⁻¹.

Second, with the introduction of sensitive space based photometers such as Kepler/K2 and TESS, exocomets close to the star can also be detected by the light blocked out as the dust released from their surface transits the host star. Similar to exoplanets, which display a unique light curve as they transit, transiting exocomets produce their own unique light curve as shown in Section 3.2.2. This provides information about the dust density in the tail. They could also be potentially a source of false positives for single transiting exoplanets when their trajectory causes a symmetric lightcurve (see Lecavelier Des Etangs et al. 1999).

Third, exocomets further away from their host stars, orbiting within debris belts, can also be observed through the gas they release as part of the collisional cascade. This gas can be seen in absorption in the UV/optical for edge-on systems (e.g., Brandeker et al. 2004; Roberge et al. 2006) and in emission in the far-IR (e.g., Cii, Oi) or in the sub-mm (e.g., CO, C i).

Solar system comet nuclei are obscured when active, but their nuclei can be directly imaged (optical/IR/radar) when they are far from the Sun, weakly active, and/or very close to Earth. Similar observations of exocomet nuclei will not be possible for the foreseeable future. Even detections of transits by solid bodies will only be possible for bodies much larger than any minor bodies in the solar system. Currently, the detection of exocomet transits requires much higher optical depth than is seen for typical comets in the solar system. This suggests either large exocomets (∼10–100 km), larger than most solar system comets (Bauer et al. 2017) but comparable to many Kuiper Belt objects (Schlichting et al. 2013) with extended dust tails transiting the star and/or a system containing a group of exocomets, possibly disintegrating.

Comet comae in the solar system are seen in emission. The only solar system comet successfully observed while transiting the Sun's disk was the sungrazer C/2011 N3. This was only observed for a few minutes in both absorption and emission at EUV wavelengths, before the comet was apparently destroyed (Schrijver et al. 2012). To date, direct spectroscopic exocomet detections have been entirely in absorption. Lines that are commonly bright in solar system comets (Na i in sungrazers and a few other comets further from the Sun with very high production rates, CN in regular comets; see Table A1) have been searched for in emission in exocomets without success. Spectroscopic observations of exocomets have not yet shown the presence of CN. The exception to this is 2I/Borisov which clearly showed CN emission (e.g., Fitzsimmons et al. 2019) while passing through our solar system. An overview of the cometary environments and how they are observed is presented in Figure 3.

Zoom In Zoom Out Reset image size

The Ca ii lines commonly seen in the spectra of β Pic and polluted WDs (see Section 3.2) have been detected in the extreme case of the large sungrazing comet C/1965 S1 Ikeya–Seki (Preston 1967; Slaughter 1969). Interestingly, the typical optical cometary emission lines (CN, C₂, C₃, etc.) were faint or undetectable in Ikeya–Seki close to its perihelion; this behavior may provide insight into what is seen (or not seen) in exocomets close to their stars. Nitrogen-bearing molecules such as N₂, NH₃ and CN, although detectable in solar system comets, are considered minor constituents (e.g., Krankowsky et al. 1986; Eberhardt et al. 1987; Wyckoff et al. 1991). Observations of the Ni line in β Pic showed no short term variable absorption signatures of Ni which is consistent with N only being a minor exocometary consituent (Wilson et al. 2019).

The main volatile of most solar system comets is water ice. Observations of the Lyα emission line in β Pic showed a strong asymmetric line profile caused by additional redshifted absorption. The asymmetric line shape has been interpreted as hydrogen gas falling toward the star which may have arisen from the dissociation of water originating from sublimating exocomets (Wilson et al. 2017) or from the gas disk accreting onto the star (Kral et al. 2017b).

The Na D lines at 589.0 and 589.6 nm have been detected in over a dozen comets (Cremonese et al. 2002), primarily at heliocentric distances less than 1 au due to their very large fluorescence efficiencies for heliocentric velocities exceeding a few 10 s of km s⁻¹ (Watanabe et al. 2003). Perhaps counterintuitively, sodium has not conclusively been detected in close-in transiting exocomets. Observations of the Na D lines are challenging from the ground due to telluric contamination. Future observations to look for variable Na D lines in exocomets clearly warrants further investigation.

No subsequent sungrazing comets have approached Ikeya–Seki's size or brightness, so we have yet to study such a comet with modern instrumentation for a more direct comparison to exocomets. Comet C/2012 S1 (ISON) held great promise for such observations, as it was discovered a year before perihelion and was the subject of a worldwide observing campaign. However, it evidently began disintegrating before perihelion (Knight & Battams 2014), and was undetected in the EUV around its perihelion passage (Bryans & Pesnell 2016). More than 4000 sungrazing comets have been detected to date (Battams & Knight 2017), but the vast majority are smaller than 100 m in diameter and only observed via broadband imaging (e.g., Knight et al. 2010). Limited spectroscopy of a dozen or so sungrazing comets has been obtained at UV and EUV wavelengths by solar observatories over the last two decades (e.g., Bemporad et al. 2007; Bryans & Pesnell 2012; Schrijver et al. 2012). These observations were optimized for solar observing so the specific comet lines detected were serendipitous and not necessarily the most prominent or diagnostic. The detections of highly ionized O, C, and Fe by Solar Dynamics Observatory's AIA instrument (McCauley et al. 2013; Pesnell & Bryans 2014) hold promise for direct comparisons to exocomet systems since conditions are akin to star-grazing exocomet systems, but will likely require a future generation of space-based X-ray or UV facilities to be detected. The most highly ionized species attributed to exocomets are Al iii, C iv and Si iv. Even higher ionized species may exist, but have not yet been detected. A model of cometary debris in the solar corona by Pesnell & Bryans (2014) opens up the possibility of detecting higher ionized species in exocomets (they model the detection of C iv, Fe viii through Fe x, and O iii through O vi. A review of observations of Near-Sun Comets in our solar system has been provided by Jones et al. (2018).

Solar system comet observations indicate that some intermediate charge state ions are created through photoionization, e.g., C iii detected in C/2002 X5 (Kudo-Fujikawa) when it was at ∼0.2 au from the Sun can be explained by the double photoionization of C originating from cometary dust (Povich et al. 2003). To explain the presence of highly ionized species such as Al iii and C iv in exocomet spectra, Beust & Tagger (1993) invoked the process of ionization at a collisional shock occurring around exocomets. This is because the central star (in this case β Pic) is unable to photoionize such highly ionized species. The heat generated by compression and collisions within the shock surface of exocomets sufficiently close to the star were thought to be high enough to explain the formation of the highly ionized ions.

Shocks have been detected in situ at numerous solar system comets during comet encounter missions (e.g., Gringauz et al. 1986; Neubauer et al. 1993; Coates et al. 1997; Gunell et al. 2018). These arise due to the slow-down and deflection of the supersonic solar wind when it reaches the cometary obstacle, where significant mass, in the form of freshly ionized cometary gas, is added to the wind. It is important to note that all solar wind shocks detected to date are collisionless rather than collisional; this is possible due to the presence of the heliospheric magnetic field that is carried by the solar wind.

Emission from higher ionization state species (Hydrogen and Helium-like ions) has been observed in solar system comets, too (Bodewits et al. 2007). Rather than being neutral species originating at the comet that are subsequently ionized, these ions' parent species originate at the Sun as multiply ionized heavy ions, e.g., O^7,8+, and are carried to the comet by the solar wind (Cravens 1997). Instead of moving to higher ionization states, these are partially neutralized at the comet through charge exchange processes with neutral species in the cometary coma, as described in Section 2.3. O iv was detected in situ in the ion tail of C/2006 P1 McNaught by the Ulysses spacecraft, but it is unclear whether that ion resulted from the ionization of cometary oxygen atoms, or the multistage neutralization of highly charged oxygen ions in the solar wind (Neugebauer et al. 2007).

Exocomets around stars with stellar winds could exhibit similar charge exchange emission at EUV and X-ray energies. This would be most apparent in charge states not expected in cold cometary environments. It is clear that stellar winds could be present at several of the systems where exocomets' presence has been inferred. Stellar winds are, however, difficult to detect; as noted by Suess & Tsurutani (2015), the Sun's solar wind would be invisible at stellar distances. The presence and nature of stellar winds in many of these systems will possibly remain undetermined. The conceivable presence of a stellar wind carrying a magnetic field, as is the case in our solar system, should be considered when interpreting observations of the ionized components of exocomets. A decoupling in the line-of-sight velocities of neutral and ionized components of an exocomet could indicate that the ions are carried by a magnetized stellar wind, i.e., as an ion tail.

Finally, as discussed in Sections 3.2.3 and 3.3, far-IR and millimeter observations of molecular emission lines within debris disks can be used to probe the composition of the population of exocomets within debris disks. The derived fractions of CO(+CO₂) ice by mass are so far found to be largely consistent with the compositions observed in solar system comets (Matrà et al. 2017), which may indicate similar formation conditions in the starting protoplanetary disks (A'Hearn et al. 2012; Eistrup et al. 2019). In addition to the CO(+CO₂) ice mass fractions, mm observations have started setting upper limits on CN emission, which (assuming CN is produced by HCN photodissociation alone) set tight constraints on the HCN/(CO+CO₂) outgassing rate ratio. This ratio is at the low end of what is expected from typical solar system HCN/(CO+CO₂) compositions in a few systems (Matrà et al. 2018; Kral et al. 2020). As mentioned in Section 3.3, upper limits on the presence of exocometary water vapor have also been set, directly (Cavallius et al. 2019) and indirectly (Kral et al. 2016), within the β Pic disk. These measurements suggest that the H₂O outgassing rate, when compared to that of CO(+CO₂), is, like HCN, also at the low end of solar system comet range. If depletions of water and HCN are widely confirmed compared to CO(+CO₂), this could imply either a true depletion of H₂O and HCN compared to CO(+CO₂) ice or, given the low temperatures of tens of K at these distances, decreased outgassing for the less volatile molecules of H₂O and HCN compared to CO (Matrà et al. 2018).