In [ ]:
from IPython.display import HTML
HTML('../style/course.css')
HTML('../style/code_toggle.html')
The Magellanic system consists of two medium-sized satellite galaxies of the Milky Way, the Large Magellanic Cloud (LMC) and the Small Magellanic Cloud (SMC). At a distance of about 50 and 60 kpc, respectively, these two galaxies are among the nearest satellites of the Milky Way and are easily visible to the naked eye from the southern hemisphere of the earth. Both galaxies are comparatively HI-rich and additionally are accompanied by a vast, filamentary network of HI filaments, the so-called Magellanic Bridge and Magellanic Stream, that are thought to be the result of different interaction processes between the two Magellanic Clouds as well as the Milky Way. In this section, the HI properties of the Magellanic system are presented and discussed.
At an angular diameter of about 8° on the sky, the LMC is a very extended source in HI emission that requires large-scale mapping (or mosaicking) to cover in its entirety. Major single-dish observations of the LMC were presented by McGee & Milton (1966)⤴ McGee & Milton (1966), Rohlfs et al. (1984) and Luks & Rohlfs (1992), all made with the 64-m Parkes radio telescope in Australia. In addition, an HI absorption-line study towards bright radio continuum sources in the direction of the LMC was carried out by Dickey et al. (1994) with the Australia Telescope Compact Array (ATCA). High-resolution HI imaging data of the LMC, based on combined observations with the ATCA and Parkes, were presented by Kim et al. (2003).
The total HI mass of the LMC was estimated by Luks & Rohlfs (1992) to be (3.1 ± 0.6) × 108 M☉ (d / 50 kpc)2. Brüns et al. (2005) derived a slightly higher mass of (4.4 ± 0.1) × 108 M☉ (d / 50 kpc)2, and an even higher value of (4.8 ± 0.2) × 108 M☉ (d / 50 kpc)2 was measured by Stavely-Smith et al. (2003). These differences are the result of different fields of view used in the mass measurements. Furthermore, all masses were derived under the assumption that the gas is optically thin and therefore constitute lower limits.
The high-resolution HI imaging data of the LMC by Kim et al. (2003) reveal a flocculent structure of the gas disc with numerous HI clumps and shells. Based on the ATCA data alone, Kim et al. (1999) identified 23 supergiant shells and 101 giant shells across the disc of the LMC. These shells are believed to be the result of gas being pushed back by a combination of kinetic pressure and radiation pressure from high-mass stars and supernovae (e.g., Weaver et al. 1977). Hence, the shells found in the LMC are often aligned with regions of active star formation. Furthermore, expanding shells appear to induce further star formation along their edges, as indicated by the presence of HII regions and OB associations (Kim et al. 1999).
At an angular diameter of about 5° on the sky, the SMC appears slightly smaller than its companion. Major single-dish HI imaging surveys of the SMC were presented by Hindman (1967) based on observations with the Parkes radio telecope in Australia and by Bajaja & Loiseau (1982) based on data taken with the 30-m antenna of the Instituto Argentino de Radioastronomía in Argentina. High-resolution images of the SMC based on combined data from the Australia Telescope Compact Array (ATCA) and the Parkes telescope were published by Stanimirović et al. (1999). They determine the spatial power spectrum, P(k) ∝ kγ, of the SMC and derive an exponent of γ = −3.04 ± 0.02, similar to the values found for the Milky Way.
The total HI mass of the SMC was determined by Loiseau & Bajaja (1981) to be (5.5 ± 0.4) × 108 M☉ (d / 60 kpc)2 based on the SMC survey presented by Bajaja & Loiseau (1982). Stanimirović et al. (1999) derived a slightly smaller value of (3.8 ± 0.5) × 108 M☉ (d / 60 kpc)2, presumably because their data cover a smaller field of view. Both mass estimates were derived without opacity corrections and should therefore be considered as lower limits. After applying basic opacity corrections, Stanimirović et al. (1999) obtain a slightly higher mass estimate of 4.2 × 108 M☉ (d / 60 kpc)2, corresponding to a 10% increase compared to their uncorrected value. A comparison with the HI mass of the LMC above demonstrates that both galaxies have approximately the same total HI mass, even though the LMC’s stellar mass, optical luminosity and diameter are larger. Hence, the HI mass-to-light ratio of the SMC of MHI / LV = 0.85 M☉ / L☉ is somewhat higher that that of the LMC of MHI / LV = 0.32 M☉ / L☉ (see Westmeier et al. 2015).
The Magellanic Stream is a large HI structure associated with the Magellanic Clouds and extending across a significant fraction of the sky. It consists of a bifurcated trailing arm (the Magellanic Stream proper), the fainter Leading Arm with its three major filaments, the generally denser Magellanic Bridge that connects the LMC and SMC, and a broad, fanned-out structure near the SMC called the Interface Region.
First detected by early HI observations in the 1960s and early 1970s (Dieter 1965; Wannier & Wrixon 1972; Wannier, Wrixon & Wilson 1972; van Kuilenburg 1972), the full extent of the stream, including the three major filaments of the Leading Arm, and its association with the Magellanic Clouds were first recognised by Mathewson (1974). More detailed HI emission-line studies of individual parts of the stream followed (Mirabel, Cohen & Davies 1979; Haynes 1979; Cohen 1982; Morras 1983), revealing the complex structure of the stream and its distinct bifurcation into two seemingly interwoven filaments. The full spatial extent and global structure of the Magellanic Stream and Leading Arm were studied in several large-scale HI surveys with the Parkes 64-m radio telescope (Putman et al. 1998; Putman et al. 2003; Brüns et al. 2005). The total HI mass of the system, consisting of the Magellanic Stream, Leading Arm, Magellanic Bridge and Interface Region, is approximately 5 × 108 M☉ (d / 55 kpc)2 (Brüns et al. 2005), although there could potentially be significant distance variation along the extent of the stream, making HI mass estimates highly uncertain.
In recent years, detailed HI studies of individual regions near the Magellanic Stream have revealed that the stream is much more complex and extended than previously thought. Braun & Thilker (2004) used the Westerbork Synthesis Radio Telescope (WSRT) in total-power mode to map a large area around the galaxies M31 and M33 in HI emission down to an unprecedented 3σ column density sensitivity of 1.5 × 1017 cm−2 across a line width of 30 km s−1, but with a fairly large beam size of 48 arcmin. Their observations uncovered that the two parallel filaments of the Magellanic Stream extend to a declination of δ ≈ +50°, tens of degrees further north than seen in previous data (e.g. Putman et al. 2003).
A high-resolution HI study of the region near the tip of the “classical” stream was carried out by Stanimirović et al. (2008) using the Arecibo telescope. They discovered that the two main filaments of the Magellanic Stream break up into four separate, coherent filaments, each of which can be traced over a length of 20° on the sky. At the same time, Westmeier & Koribalski (2008) found coherent filaments of HI clouds to the east of the Magellanic Stream out to a maximum separation of about 20°, using data from the HI Parkes All-Sky Survey (HIPASS; Barnes et al. 2001). These filaments can be traced over tens of degrees on the sky, and their morphology and kinematics suggest that they are part of the stream.
All these new observational results were combined by Nidever et al. (2010) and augmented with additional HI data from the Green Bank Telescope (GBT) to create an overall HI map tracing the Magellanic Stream and Leading Arm across an angle of more than 200° on the sky in the Magellanic longitude range of about −150° to +70°. More recently, Venzmer, Kerp & Kalberla (2012) and For, Staveley-Smith & McClure-Griffiths (2013) independently reported the discovery of an additional Leading Arm filament named LA IV. In addition, Fox et al. (2014) presented the results of an extensive UV absorption-line study of the Magellanic system suggesting that the ionised component of the stream covers about one quarter of the entire sky and contains approximately three quarters of the total gas mass of the stream of about 2 × 109 M☉ (d / 55 kpc)2.
Several of the HI clouds in the Magellanic Stream and Leading Arm shown signs of ram-pressure interaction in the form of the presence of a multi-phase medium as well as a distinct cometary appearence of some clouds (Putman et al. 2003; Ben Bekhti et al. 2006; Venzmer, Kerp & Kalberla 2012; For, Staveley-Smith & McClure-Griffiths 2013; For et al. 2014). Brüns et al. (2005) and Putman et al. (2011) found such head–tail clouds to be particularly common in the Leading Arm near the Galactic plane, suggesting that the gas of the Leading Arm may be interacting with the outer gas disc of the Milky Way. Direct morphological and kinematical evidence for a collision of the Leading Arm with the Galactic disc was reported by McClure-Griffiths et al. (2008) based on HI data from the Parkes telescope. From their data they derive a heliocentric distance to the Leading Arm of 21 kpc, suggesting that the Leading Arm crosses the Galactic plane at a Galactocentric radius of only 17 kpc.
The presence of a two-component HI line profile in parts of the Leading Arm, as observed by Brüns et al. (2005), suggests that a core of cold HI gas is present in some of the clouds. HI gas can cool down if sufficiently shielded from the ambient radiation field such that cooling through different atomic line transitions (but not the HI line itself) becomes more efficient than radiative heating. One way to achieve this is through compression of the gas by ram pressure, resulting in higher gas densities and thus better shielding (Quilis & Moore 2001; Konz, Brüns & Kerp 2002; Heitsch & Putman 2009). This scenario is consistent with the head–tail morphologies found by Brüns et al. (2005), Venzmer, Kerp & Kalberla (2012) and For, Staveley-Smith & McClure-Griffiths (2013) in some of the clouds of the Leading Arm, with the cold HI component usually residing within the head of the cloud. A significant number of head–tail clouds is also found along the Magellanic Stream proper (For et al. 2014), suggesting that the Stream is interacting with the hot Galactic halo. From a Gaussian decomposition of HI profiles, Kalberla & Haud (2006) find that about a quarter of all sightlines towards the near end of the Stream (at positive radial velocities) show evidence of a core–envelope structure, while this fractions drops to about 10% at the far end of the Stream (at negative velocities), suggesting a possible change in environmental conditions (e.g. in Galactic halo density) along the extent of the Stream. The scenario of interaction of the Magellanic Stream with the Galactic halo is also consistent with enhanced Hα fluxes observed along the Stream (Weiner & Williams 1996; Putman et al. 2003).
The presence of a large amount of neutral hydrogen in the Magellanic Stream raises the question whether the formation of molecular gas and possibly even star formation may have occurred in some parts of the Stream. Carbon monoxide (CO) emission, which is generally used as a proxy for molecular hydrogen (H2), has so far only been detected towards the Magellanic Bridge (Muller, Staveley-Smith & Zealey 2003; Mizuno et al. 2006; Muller et al. 2014), but not in the Magellanic Stream or Leading Arm. In addition, Murray et al. (2015) report the first detection of HCO+ (in absorption) towards one sightline in the Magellanic Bridge. Mizuno et al. (2006) conclude that the molecular gas detected in the Magellanic Bridge was formed in situ after the formation of the Stream. This would be consistent with the observational evidence of recent star formation in the Bridge (Meaburn 1986; Nishiyama et al. 2007; Skowron et al. 2014).
While no CO emission has so far been detected towards the remainder of the Magellanic Stream and Leading Arm, Sembach et al. 2001 report the direct detection of H2 in far-UV absorption towards a sightline in the Leading Arm. Given the long formation timescale of the H2 molecule, they conclude that the molecular hydrogen originally may have been formed in the SMC and survived the tidal stripping of the Leading Arm. Searches for stars associated with the Magellanic Stream have long been unsuccessful (Tanaka & Hamajima 1982) or inconclusive (Brück & Hawkins 1983) until recently, when Casetti-Dinescu et al. (2014) reported the discovery of a small population of young O and B stars potentially associated with the Leading Arm, suggesting that recent star formation may have occured along the Magellanic Stream.
The debate about the origin and age of the Magellanic Stream is still ongoing. While there is agreement that the gas in the Stream originates from the Magellanic Clouds and that gravity must have played a role in its formation (because there are leading and trailing arms), there is still no consensus about whether the gas is supplied by the LMC, the SMC or both, whether gravity or ram pressure is the main stripping mechanism and what role the Milky Way plays in the interaction. Mathewson (1974) speculates that the gas could either have been tidally stripped from the SMC by the Milky Way (although this would have required a recent close encounter at 20 kpc), that primordial gas from the LMC–SMC system could have been stripped by ram-pressure forces from the intergalactic medium, or that “explosions” within the SMC could have expelled the gas.
Proponents of the tidal scenario include Lin & Lynden-Bell (1982), Gardiner & Noguchi (1996), Yoshizawa & Noguchi (2003), Connors, Kawata & Gibson (2006) and Diaz & Bekki (2012). The tidal scenario is able to explain the presence of both a leading and trailing arm, but fails to reproduce some of the details in the morphological and kinematical structure of the Magellanic Stream. The ram-pressure scenario, on the other hand, can explain the observed asymmetries between the leading and trailing arms of the Stream and was investigated by Mathewson, Schwarz & Murray (1977), Meurer, Bicknell & Gingold (1985), Liu (1992), Sofue (1994) and Moore & Davis (1994). Ram pressure alone, however, would not be able to generate the Leading Arm, and both tidal forces and ram pressure combined are therefore thought to have shaped the Magellanic Stream. Simulations that include both gravity and ram pressure were presented by Mastropietro et al. (2005) and Besla et al. (2010).
Based on a morphological and kinematical analysis of the Stream, Nidever, Majewski & Burton (2008) instead propose that the gas may have originated from star formation-related outflows in the LMC; they even argue that the twisted appearance of the Stream could be an imprint of the LMC’s rotation velocity. Once expelled, the gas would then have been exposed to tidal and ram-pressure forces to create the Magellanic Stream and Leading Arm. However, a similar kind of analysis carried out by For et al. (2014) finds that the Magellanic Stream can be entirely traced back to the SMC rather than the LMC.
Evidence for the origin of the Stream comes from metallicity measurements based on optical or ultra-violet absorption spectroscopy of background sources. Gibson et al. (2000) use data from the Hubble Space Telescope to measure metal abundances towards three background sources. Their measurement of &lsqbS / H&rsqb = −0.55 towards Fairall 9 is consistent with an origin of the gas in either the LMC or SMC. Lu et al. (1998) measure &lsqbS / H&rsqb = −0.60 and &lsqbFe / H&rsqb = −1.48 towards NGC 3783 in the direction of the Leading Arm, again consisten with either an LMC or SMC origin of the gas. The measurement of &lsqbO / H&rsqb = −1.00 by Fox et al. (2010) suggests that the Stream originates from the SMC rather than the LMC. This scenario is consistent with the measurements of &lsqbS / H&rsqb = −1.13 and &lsqbO / H&rsqb = −1.24 made by Fox et al. (2013) in two separate directions along the Stream. On the other hand, Richter et al. (2013) revisit the Fairall 9 sightline and derive a much higher sulphur abundance of &lsqbS / H&rsqb = −0.30, supporting a scenario in which at least some of the gas in the Stream originated from the LMC and was chemically enriched prior to fuelling the Stream. The chemical evolution of the entire Magellanic system has also been studied with the help of simulations by Yozin & Bekki (2014) who are able to reproduce the one tenth solar metallicity observed along most the Stream, but struggle to reproduce the hypothesised, more metal-right LMC filament postulated by Nidever, Majewski & Burton (2008) and Richter et al. (2013).
To complicate matters further, recent proper-motion measurements of the Magellanic Clouds (Kallivayalil et al. 2006; Kallivayalil, van der Marel & Alcock 2006; Besla et al. 2007; Piatek, Pryor & Olszewski 2008) suggest that the LMC and SMC could be on their first infall towards the Milky Way and may not have been bound to either the Milky Way or each other in the past. This would pose serious problems for most of the formation models proposed for the Magellanic Stream, as they require the influence of strong tidal forces over a long period of time. Improved proper motion measurements published by Kallivayalil et al. (2013) have slightly eased the situation, but still rule out orbital periods of the Magellanic Clouds of ≲ 4 Ga. Astonishingly, the LMC’s proper motion of μW = −1.910 ± 0.020 mas a−1 and μN = 0.229 ± 0.047 mas a−1 as measured by Kallivayalil et al. (2013) is almost identical to the prediction of 2.0 mas a−1 due east made by Lin & Lynden-Bell (1982) based on their N-body simulations of the Magellanic Clouds interacting with the Milky Way’s gravitational potential.
In summary, the origin of the Magellanic Stream and Leading Arm has not been satisfactorily constrained yet, but a combination of tidal forces (from the Milky Way as well as between the two Magellanic Clouds) and ram pressure (from the hot, gaseous halo of the Milky Way) is expected to have shaped the Stream. While most of the gas is thought to have been provided by the SMC, there is some evidence for LMC gas in the Stream as well from morphological studies and metallicity measurements.
The Andromeda Galaxy (M31, NGC 224) is the second large spiral galaxy in the Local Group next to the Milky Way. Together with the smaller Triangulum Galaxy (M33, NGC 598) and dozens of small satellite galaxies it forms a distinct, gravitationally bound sub-group within the Local Group.
At a distance of just under 800 kpc, the Andromeda Galaxy is the nearest large spiral galaxy in the local universe and readily visible to the naked eye. Notable high-resolution synthesis mosaics of the HI emission in M31 were published by Unwin (1980) based on observations with the Cambridge Half-Mile Telescope in the United Kingdom, Braun et al. (2009) based on data taken with the Westerbork Synthesis Radio Telescope (WSRT) in the Netherlands, and Chemin, Carignan & Foster (2009) based on observations with the Synthesis Telescope and the 26-metre Telescope at the Dominion Radio Astrophysical Observatory in Canada. Recent measurements of the total HI mass of M31, based on a distance of 785 kpc (McConnachie et al. 2005), range from 4.2 × 109 M☉ without any opacity corrections (Chemin, Carignan & Foster 2009) to 7.3 × 109 M☉ with basic opacity correction (Braun et al. 2009), although even the latter value should still be considered as a lower limit.
Among the most prominent features of M31’s HI disc are the deficiency of neutral hydrogen in the centre of the galaxy (van de Hulst, Raimond & van Woerden 1957) and the presence of a conspicuous, dense HI ring with a radius of about 10 kpc (Roberts 1966). The outer HI disc of M31 shows significant warping (e.g. Roberts & Whitehurst 1975; Sawa & Sofue 1982), which is particularly prominent on the north-eastern side of the disc and generally believed to be the result of tidal interaction with a satellite galaxy. The warp is also present in the stellar disc of M31 (Innanen et al. 1982), although the stellar warp begins at smaller radii and deviates more strongly from the major axis of the inner disc than the HI warp (Ferguson et al. 2002).
Another conspicuous HI feature in close proximity to M31 on the sky is Davies’ Cloud (Davies 1975). It is located about 1.5 degrees to the north-west of the centre of M31 at a radial velocity of about −450 km s−1 in the LSR frame. Davies (1975) discusses two possible origins of the cloud: it could either be a satellite of M31 or part of the Magellanic Stream. High-resolution HI observations of Davies’ Cloud with the WSRT are presented by de Heij, Braun & Burton (2002). In their discussion they consider the object to be either associated with M31 or part of a population of primordial dark-matter haloes in the Local Group. However, the recent discovery of faint, northern extensions of the Magellanic Stream (Braun & Thilker 2004; Nidever et al. 2010) seems to suggest that Davies’ Cloud is instead part of the Magellanic Stream which runs across the same part of the sky at similar radial velocities.
Detailed rotation curves of M31 from high-resolution HI synthesis observations are presented by Chemin, Carignan & Foster (2009) and Corbelli et al. (2010) based on tilted-ring fitting of the radial velocity field. Chemin, Carignan & Foster (2009) derived the rotation curve within a radius of 38 kpc. They find a strong variation of the rotation velocity from about 230 to 264 km s−1 across the central disc of M31 (at an average inclination angle of about 74°), with a sharp rise to 275 km s−1 near the outer edge of the disc. While the shape of the rotation curve derived by Corbelli et al. (2010) is similar to that of Chemin, Carignan & Foster (2009), their overall rotation velocities are on average about 10 km s−1 lower and in the range of about 219 to 254 km s−1 across the central disc out to a radius of 37 kpc. This difference is due to a generally slightly higher inclination angle of 77° found by Corbelli et al. (2010) as the result of a slightly different approach in their tilted-ring modelling. In particular, they use spherical geometry in their modelling to take into account that the near side of M31’s disc is considerably closer to the Milky Way than the far side due to the small distance of M31.
Both studies also derive a dynamical mass model of M31 from the measured rotation curve. Chemin, Carignan & Foster (2009) derive a total dynamical mass of (4.7 ± 0.5) × 1011 M☉ within a radius of 38 kpc from different ΛCDM models and a total mass of about 1.0 × 1012 M☉ with M31’s virial radius. This is consistent with the dynamical mass of (4.9+12−14) × 1011 M☉ derived by Corbelli et al. (2010) within 37 kpc radius, although they estimate a slightly higher total mass of about 1.3 × 1012 M☉.
M33, the smaller companion of M31, was observed in HI with the WSRT in the Netherlands by Deul & van der Hulst (1987), with the Very Large Array (VLA) in New Mexico by Thilker, Braun & Walterbos (2002) and with the 300-m Arecibo telescope in Puerto Rico by Putman et al. (2009). The total HI mass of M33 is 1.4 × 109 M☉ (d / 730 kpc)2 without opacity corrections (Putman et al. 2009).
The HI disc of M33 shows strong signs of distortion in the form of an extended, warped outer disc as well as numerous clumps and filaments surrounding the entire disc, including a large HI cloud near the southern edge of the disc that appears connected to M33 through a gas bridge. A detailed orbital analysis carried out by Putman et al. (2009) suggests that the strong distortions of M33’s disc could have been the result of a close encounter with M31 that occurred about 1 to 3 Ga ago.
The rotation curve of M33 was determined by Corbelli et al. (2014) from high-resolution HI observations with the VLA. They find a generally rising rotation curve, with the rotation velocity increasing to about 120 to 130 km s−1 at a radius of 23 kpc. They also fit a ΛCDM-based mass model with NFW halo to the rotation curve and derive a total mass of (4.3 ± 1.0) × 1011 M☉ for M33.
A population of about 20 isolated HI clouds within a projected distance of about 50 kpc of M31 was discovered by Thilker et al. (2004) using the Green Bank Telescope. These clouds are thought to be the analogues to the high-velocity clouds (HVCs) found around the Milky Way. The clouds were studied in greater detail with the Effelsberg telescope by Westmeier, Brüns & Kerp (2008) and in high-resolution interferometric observations with the WSRT presented by Westmeier, Braun & Thilker (2005).
The HVCs of M31 have typical diameters of the order of 1 kpc and HI masses of typically a few times 105 M☉ (Westmeier, Braun & Thilker 2005). Many of the clouds show distinct velocity gradients across their extent; if these are assumed to indicate rotation, the resulting dynamical masses would be about 100 times larger than the observed HI masses. This result could either indicate the presence of a significant amount of dark matter or imply that the velocity gradients are not the result of rotation, but of turbulent motions within the clouds.
The origin of the HVCs around M31 is still uncertain, and hypotheses range from primoridal dark-matter haloes and condensations of accreted intergalactic gas to tidal debris from recent interactions between M31 and satellite galaxies. The extended complex of HVCs located near the south-eastern edge of M31’s disc both spatially and kinematically overlaps with the giant stellar stream of M31 discovered by Ibata et al. (2001) and presumably originates from the disruption of the satellite galaxy that formed the stellar stream. Another cloud is close to one of M31’s satellite galaxies, NGC 205, both in position and velocity, and could have been stripped from that galaxy. This would be consistent with optical (McConnachie et al. 2004) and far-infrared (De Looze et al. 2012) observations that uncovered evidence for tidal stripping of stars and gas from NGC 205. Other HVCs appear isolated and not associated with any known satellite galaxy or stellar structure of M31’s disc, and their origin is unknown.
Several HI clouds have also been found in the vicinity of M33, most notably Wright’s Cloud (Wright 1974), which, at an LSR velocity of about −390 km s−1, might be associated with the nearby Magellanic Stream rather than M33. If located at the distance of M33, the HI mass of Wright’s Cloud would be fairly significant at 4.5 × 107 M☉ (Keenan et al. 2016). More recently, sensitive HI observations of a large area around M33 with the Arecibo telescope (Grossi et al. 2008; Keenan et al. 2016) have revealed the presence of about 20 isolated HI clouds surrounding the galaxy. Their HI masses are in the range of about 105 to 107 M☉ with sizes of typically several kpc. While Grossi et al. (2008) treat these clouds as being associated with M33, Keenan et al. (2016) point out that the positions and velocities of most of them are fully consistent with the Magellanic Stream that runs across the same part of the sky; therefore, most of the clouds might not be located at the distance of M33 at all.
Braun & Thilker (2004) carried out a very deep HI survey of a large area around M31 and M33 using the WSRT in total-power mode. The resulting image has a poor angular resolution of 48 arcmin (FWHM), but an unprecendented column density sensitivity of 1.5 × 1017 cm−2. Overall, they detect 95 discrete HVCs in the field. While many of the clouds are consistent with being part of the Magellanic Stream that runs across the same part of the sky, they uncovered a faint HI filament that appears to connect M31 with M33, not just spatially, but also in velocity space. This filament could potentially be the neutral component of the warm–hot intergalactic medium (WHIM) expected to form along filaments of the cosmic web (Braun & Thilker 2004), or the result of tidal stripping during a recent, close encounter between M31 and M33 (Bekki 2008).
Recent follow-up observations of the HI bridge with the GBT (Lockman, Free & Shields 2012; Wolfe et al. 2013; Wolfe, Lockman & Pisano 2016) have confirmed the existence of the filament and revealed that virtually all of the gas is located in discrete clumps with peak column densities in the range of about 1 to 5 × 1018 cm−2 and typical HI masses of the order of 105 M☉ without any evidence for a diffuse component (Wolfe, Lockman & Pisano 2016). The clumps are kinematically distinct from the populations of gas clouds and HVCs surrounding M31 and M33, suggesting that their origin is a different one. The results presented by Wolfe et al. (2013) appear to support the cosmic web hypothesis and suggest that the gas could provide fuel for future star formation in M31 and M33. Wolfe, Lockman & Pisano (2016) rule out the possibility that these gas clouds constitute satellite galaxies of M31 due to the lack of optical counterparts. However, they speculate that the clouds could also be condensations within the extended circumgalactic medium claimed to have been detected around M31 by Lehner, Howk & Wakker (2015).
A comparison between the extragalactic HI structures in the M31 system with faint stellar substructures found in deep optical imaging by the PAndAS project (Lewis et al. 2013) revealed that there is very little spatial overlap between the gas and the stars. The only notable exception is M31’s giant stellar stream which appears to have some HI gas associated with it (also see Westmeier, Braun & Thilker 2005). Lewis et al. (2013) conclude that, while stellar and gaseous structures throughout the M31 system may well have been formed in similar accretion events of satellite galaxies in the past, they have since been shaped by different physical processes, thus explaining the current lack of spatial correlation between the two components. The gas in particular is subject to ram-pressure forces and shock-heating and may therefore not be collocated with the stars.
HI emission has been unambiguously detected in about two dozen dwarf galaxies across the Local Group, including the LMC and SMC, WLM (Kepley et al. 2007), IC 10 (Wilcots & Miller 1998), Pisces I / LGS 3 (Young & Lo 1997), IC 1613 (Silich et al. 2006), Phoenix (Young & Lo 1997; Irwin & Tolstoy 2002; Young et al. 2007), UGCA 92 (Begum et al. 2008), Leo III / Leo A (Young & Lo 1996), Sextans A (Barnes & de Blok 2004), Sextans B (Haynes & Giovanelli 1984), NGC 3109 (Barnes & de Blok 2001), Antlia (Barnes & de Blok 2001), UGC 8091 (Fisher & Tully 1975), Sagittarius DIG (Young & Lo 1997), Aquarius dIrr / DDO 210 (Lo, Sargent & Young 1993), IC 5152 (Whiteoak & Gardner 1977), UGCA 438 (Longmore et al. 1982), NGC 6822 (de Blok & Walter 2000), Leo T (Ryan-Weber et al. 2008), Pegasus DIG (Lo, Sargent & Young 1993) and UGC 4879 (Bellazzini et al. 2011). Also see Grcevich & Putman 2009 for a comprehensive list plus references. Among the known satellites of the Andromeda Galaxy, HI emission has only been unambiguously detected in NGC 185 and NGC 205 (Johnson & Gottesman 1983). With few exceptions, most of the galaxies with HI detections are located far away from the two major galaxies in the Local Group, the Milky Way and the Andromeda Galaxy.
The recent surge in discoveries of new dwarf galaxies across the Local Group and particularly in the vicinity of the Magellanic Clouds (Koposov et al. 2015; Martin et al. 2015; Kim et al. 2015; Bechtol et al. 2015; Kim & Jerjen 2015; Sand et al. 2015; Laevens et al. 2015; Drlica-Wagner et al. 2015; Tollerud et al. 2015; Luque et al. 2016; Torrealba et al. 2016a; Luque et al. 2016; Torrealba et al. 2016b) has led to several unsuccessful follow-up searches for HI emission (Westmeier et al. 2015; Janzen et al. 2015; Crnojević et al. 2016), although HI has been successfully detected in the two more distant and isolated galaxies Pisces A and B with KAT–7 (Carignan et al. 2016). The newly discovered population of ultra-faint dwarf galaxies near the Magellanic Clouds was recently simulated by Yozin & Bekki (2015) who conclude that the population is consistent with being part of a distinct Magellanic group of galaxies. The ultra-faint dwarfs must have lost their gas during the infall of the Magellanic Clouds onto the Milky Way through ram-pressure stripping by the Galactic halo. Yozin & Bekki (2015) speculate that the stripped gas might still be detectable in HI emission, although no longer collocated with its progenitor galaxies on the sky, and should be distinguishable from Galactic gas through its lower metallicities.
The HI morphology of several of the more isolated, HI-rich dwarf galaxies of the Local Group has been studied in great detail using radio interferometers. Many were found to be gas-rich and show signs of ongoing star formation activity. A well-studied example is the dIrr galaxy NGC 6822. Located at a distance of roughly 500 kpc (Gallart, Aparicio & Vílchez 1996), NGC 6822 is not thought to be associated with either the Milky Way or the Andromeda Galaxy. Its HI emission was studied with the ATCA by de Blok & Walter (2000) who discovered numerous HI shells in the galaxy’s disc, including one of the largest supergiant shells every found with a size of 2.0 × 1.4 kpc. They estimate that creation of the supergiant shell would require the energy of about 100 supernovae and speculate that the shell could be the result of a burst of star formation during an interaction with a smaller companion galaxy about 100 Ma ago. Signs of active star formation in the form of HI shells have also been found in a few other Local Group dwarfs, in particular IC 10 (Wilcots & Miller 1998) and IC 1613 (Silich et al. 2006).
A curious case is that of the Phoenix dwarf galaxy in which the HI component is spatially shifted with respect to the stellar component. Young et al. (2007) conclude that the HI gas has most likely been expelled from the galaxy by supernova-driven winds resulting from strong star formation activity rather than tidal or ram-pressure interaction, although the gas is still gravitationally bound to Phoenix and could fuel further star formation in the future. The Pegasus dIrr galaxy, on the other hand, shows signs of ram-pressure stripping of the outer parts of its HI disc (McConnachie et al. 2007) despite its relatively large distance from M31 of about 500 kpc, suggesting the presence of a low-density intergalactic medium in the Local Group with a density of at least nIGM ≈ 10−6 to 10−5 cm−3.
Due to their small distances and hence large angular sizes on the sky, dwarf galaxies in the Local Group are particularly suitable to study the kinematics of the HI gas and the total fraction and spatial distribution of dark matter, although many of the Local Group dwarf galaxies do not show any sign of rotation, but rather appear to be supported by random motions of their gas and stars. Examples include Sagittarius DIG and Pisces I / LGS 3 (Young & Lo 1997). Several of the Local Group dwarfs with previous HI detections were included in the LITTLE THINGS survey with the VLA (Hunter et al. 2012). This sample was used by Oh et al. (2015) to derive the rotation curves and mass composition of 26 dwarf galaxies across the Local Group and the local volume. They conclude that the relatively shallow inner slopes of the dark-matter density profiles disagree with the generally steeper, cusp-like profiles predicted by dark-matter simulations in the ΛCDM paradigm. The discrepancy can potentially be explained by baryonic feedback processes within the galaxies, such as gas outflows driven by supernovae.
Observations of neutral hydrogen in dwarf galaxies throughout the Local Group have revealed that satellite galaxies of the Milky Way and the Andromeda Galaxy are generally HI-deficient (e.g. Young 1999; Young 2000; Grcevich & Putman 2009); despite sensitive and extensive searches, no HI emission has been detected in most of the known Milky Way and Andromeda satellites. A systematic study of the HI content of dwarf spheroidal satellite galaxies around the Milky Way by Spekkens et al. (2014) revealed that there is a sharp transition between gas-rich and gas-poor galaxies, with satellites located within the Milky Way’s virial radius of Rvir ≈ 300 kpc having HI mass-to-light ratios (MHI / LV) and HI-to-dynamical mass ratios (MHI / Mdyn) that are at least a factor of 100 to 1000 lower than in galaxies located outside the virial radius of the Milky Way. This significant HI deficiency has been attributed to a number of different physical processes, including
It is likely that a combination of these processes is responsible for the observed HI deficiency of Local Group dwarf galaxies in the vicinity of both the Milky Way (Spekkens et al. 2014) and the Andromeda Galaxy (Grcevich & Putman 2009). Indeed, evidence of strong tidal and ram-pressure forces is seen in satellites that are currently close to their host galaxy, including the LMC and SMC near the Milky Way (Mathewson 1974) as well as M32 and NGC 205 near the Andromeda Galaxy (Choi et al. 2002; Westmeier, Braun & Thilker 2005).
From their numerical simulations of a Milky Way-sized galaxy, Nickerson et al. (2011) conclude that ionisation by UV radiation and ram-pressure stripping are the dominant factors responsible for removing most of the gas from low-mass satellites. Additional gas loss is caused by tidal stripping from satellites of all masses near the pericentre of their orbit. The results of Nickerson et al. (2011) could help to explain the so-called “missing satellites” problem, whereby numerical simulations of structure formation in a ΛCDM universe produce many more dark-matter satellites around galaxies like the Milky Way that there are known satellite galaxies (e.g. Klypin et al. 1999; Moore et al. 1999).
The idea of gas loss during passages through the halo of either the Milky Way or M31 was also investigated through simulations by Teyssier, Johnston & Kuhlen (2012). They conclude that several of the Local Group dwarf galaxies currently located outside the Milky Way’s virial radius are likely to have passed within the virial radius of the Milky Way in the past, including objects such as NGC 3109, NGC 6822 and Leo T. While still containing HI gas today, most of the potential Milky Way escapees have lower HI fractions and show signatures of past interactions in their stellar and gaseous components, supporting the idea that processing of galaxies occurs as the result of interactions with the major spiral galaxies in the Local Group.
The gas loss of satellite galaxies in the Local Group has significant consequences for the star formation history and quenching of Milky Way satellites (Slater & Bell 2014; Bahé & McCarthy 2015; Wetzel, Tollerud & Weisz 2015). A study of star formation in Local Group galaxies by Slater & Bell (2014) revealed that almost all of the satellite galaxies with stellar masses of M★ ≲ 107 M☉ are quenched, while at higher mass the quenching fraction decreases to about 50%. Their results suggest that star formation quenching is more efficient in low-mass galaxies, e.g. due to ram pressure from the host galaxy’s hot, ionised halo being more effective in removing gas from the satellite’s gravitational potential near the pericenter of the orbit.
In [ ]:
In [ ]: