Local Group galaxies, including the Milky Way and M31, can be dissected star by star, and thus provide a laboratory for understanding the details of hierarchical galaxy formation. However, it is unclear whether the Local Group galaxies represent typical evolution histories. To understand the place of the Milky Way and M31 in the larger context requires studying tidal substructures in galaxies beyond the Local Group, where resolving individual stars is difficult. In this section, we detail some of the methods used to discover tidal debris features, including resolved and unresolved stellar populations, and discuss the unique insights gleaned from some examples of these structures.
2.1. Resolved stellar structures in the Local Group
The era of massive, deep photometric surveys has enabled the detection of tidal debris structures within the Milky Way as stellar overdensities of carefully selected tracers. The famous “Field of Streams” image from SDSS (Belokurov et al. 2006; reproduced in Chapter 1) mapped substructure using main sequence turnoff (MSTO) starcounts to surface brightness limits of fainter than ∼ 32 mag arcsec−2. The identification of individual members of stellar streams enables extremely low surface brightness features to be detected. This can be achieved by kinematical selection of stream members (for example, via spectroscopic velocities), allowing features in the Galactic halo containing fewer than ∼ 1 red giant branch (RGB) star per square degree to be identified.
Likewise, in M31, photometric selection of metal-poor RGB candidates removes much of the contaminating background, and has allowed detections of features as faint as ∼ 30 mag arcsec−2, including the “giant stream” (Ibata et al. 2001) and other low surface brightness features (e.g., Ferguson et al. 2002; see Chapter 8 for more about M31). When individual stars can be resolved and spectroscopically vetted in M31, the measurement of surface brightnesses as faint as ∼ 32 mag arcsec−2 is possible (see, e.g., Gilbert et al. 2012 and other results from the SPLASH survey). Thus, we are able to probe both the Milky Way and M31 to depths of ∼ 32 mag arcsec−2, which is deep enough to sample most of the simulated halo substructures seen in simulations (discussed further in Section 4).
2.2. Detection methods in and beyond the Local Group
Current ground-based telescopes are unable to resolve stellar tidal streams around most galaxies beyond the Local Group into individual stars. Most of the streams that have been found in external galaxies thus appear as elongated, diffuse-light regions extending over several arcminutes on the sky. To map such tidal streams requires deep imaging that is also sensitive to extremely faint surface brightness features; the typical surface brightness of known stellar tidal streams is 27 mag arcsec−2 or fainter, depending both on the luminosity of the progenitor and the time they were accreted (Johnston et al. 2008).
Faint tidal features can be identified on sky-limited archival photographic plates using a process called photographic amplification; the surface brightness limit is even fainter if photographically amplified derivatives of several plates are combined together (Malin 1981). Using these techniques (photographically or digitally) it is possible to detect extended features to 28 mag arcsec−2 from existing photographic surveys (Malin & Hadley 1997). This depth is comparable to that achievable with SDSS in Figure 1).
 |
Figure 1. Left panels: images from the Sloan Digital Sky Survey of the nearby galaxies NGC 4013 (top) and M63 (bottom). These images do not show any obvious signs of tidal streams in the halos of these galaxies at the surface brightness limit of SDSS. Deep images of these same galaxies reveal a low-latitude stellar stream around NGC 4013 (upper right panel; Martínez-Delgado et al. 2009) and a giant tidal stream around the spiral galaxy M63 (lower-right; Chonis et al. (2011)). For reference, a color inset of each galaxy's central regions has been inserted atop the deeper images. These images illustrate the value of deep, sensitive imaging (i.e., beyond that of SDSS) for detecting the faint debris structures predicted by theoretical models in external galaxies. |
Small, fast (i.e., low focal-ratio) telescopes (e.g., Martínez-Delgado et al. 2010, van Dokkum et al. 2014) and modern CCD cameras are capable of imaging unresolved structures in external galaxies to ΣR ∼ 29 mag arcsec−2. Detecting these faint features requires very dark sky conditions and images taken with exquisite flat-field quality over a relatively large angular scale. More specifically, stellar streams are typical found at large galactocentric distances (15 kpc < R < 100 kpc, or farther) and could be found out to a significant portion of the virial radius of the parent galaxy (for the Milky Way or M31, Rvirial ∼ 300 kpc). Thus, surveys for stellar debris must produce images over large angular scales (from > 30′ for systems < 10 Mpc to ∼ 15′ for systems ∼ 50 Mpc away). As an example, the survey strategy employed by Martínez-Delgado et al. (2010) uses stacks of multiple deep exposures of each target taken with high throughput clear “luminance” filters (transmitting 4000 Å < λ < 7000 Å with a near-IR cut-off) with typical exposure times of 7-8 hours.
Recent deep, wide-field imaging surveys have focused on nearby spiral galaxies that were suspected (based on existing data from, e.g., surveys such as POSS-II or SDSS, or from amateur astronomical imaging) to contain diffuse-light over-densities. To date, the combined observational efforts have revealed more than 50 previously undetected stellar structures in galaxies as distant as 80 Mpc. Figure 2 shows eight such galaxies from the survey of Martínez-Delgado et al. (2010), illustrating the variety of tidal debris features – both in their morphologies and in their projected radii. The central disks of the galaxies shown in Figure 2 are of similar overall physical size, but the debris features can span large galactocentric distances. The morphologies of the features include great circle streams resembling the Sagittarius stream around our Galaxy (upper right panel of Figure 2; Sagittarius is discussed in Chapter 2 of this volume), isolated shells, giant clouds of debris floating within galactic halos, jet-like features emerging from galactic disks, and large-scale diffuse structures that may be related to the remnants of ancient, already thoroughly disrupted satellites. The diversity in observed substructure morphology parallels that seen in simulations (e.g., Johnston et al. 2008, Cooper et al. 2010). In addition to the remains of satellites that are likely completely destroyed, there are a few examples (e.g., Martínez-Delgado et al. 2012, Martínez-Delgado et al. 2014, Amorisco et al. 2015) of surviving satellites caught in the act of tidal disruption, displaying long tails departing from the progenitor satellite (i.e., similar in spirit to observations in the Local Group). The extraordinary variety of morphological specimens provides strong evidence to support the hierarchical galaxy formation scenarios predicted by cosmological models (e.g., Cooper et al. 2010).
 |
Figure 2. Luminance filter images of nearby galaxies from the pilot survey of Martínez-Delgado et al. (2008, 2010) showing large, diffuse light structures in their outskirts. These include tidal streams similar to Sagittarius (upper right panel), giant plumes (middle panels in the top row), partially disrupted satellites (top row, third panel from left), umbrella-shaped tidal debris structures (middle two panels in the bottom row), enormous stellar clouds, prominent spikes, and large scale, complex inner halos sprinkled with several debris features. A color inset of the disk of each galaxy has been overplotted for reference. An illustrative comparison of these features to the surviving structures visible in cosmological simulations is given in Martinez-Delgado et al. 2010, their Fig 2. |
2.3. Unresolved features beyond the Local Group
Prior to recent dedicated CCD searches, only a few cases of extragalactic stellar tidal streams have been reported in nearby spiral galaxies. Malin & Hadley (1997) found two possible tidal streams surrounding the galaxies M83 and M 104 by using special contrast enhancement techniques on plates obtained for wide-field photographic surveys. Shortly thereafter, a study of the nearby, edge-on galaxy NGC 5907 by Shang et al. (1998) employed deep CCD images to reveal an elliptical loop in the halo of this galaxy, which they believed to be the remains of a tidally disrupted galaxy similar in size to the Sagittarius dwarf galaxy (Sagittarius was at that time a recent discovery in the Milky Way halo). Shang et al. also identified another Sagittarius-like dwarf galaxy that they suggested might be interacting with the disk of NGC 5907, causing the observed warp in HI. Their photometry reached surface brightnesses of 28.6 and 26.9 mag arcsec−2 in R and I-bands, respectively.
More recently, NGC 5907 was imaged by Martínez-Delgado et al. (2008) (Figure 3), revealing even fainter features of the stream. As shown in Figure 3, this stream is prominently visible as an interwoven, rosette-like structure traversing nearly 720∘ around NGC 5907. Detailed N-body modeling suggests that all of the features can be reproduced from the accretion of a single, low-mass satellite galaxy, with the stream tracing two full orbits of its progenitor. The presence of such a long stream confirms that a stellar substructure can survive several gigayears, which though predicted by N-body simulations of tidally disrupted stellar systems around the Milky Way (e.g., Law et al. 2005, Peñarrubia et al. 2005) lacked direct confirmation. Interestingly, the N-body model of Figure 3 was created using a progenitor with orbital parameters similar those found for Sagittarius in the Milky Way (Chapter 2). The model suggests that the fainter, outer loop material (blue points in Figure 3) became unbound from its progenitor at least 3.6 Gyr ago. The substructure in NGC 5097 is one of the most striking examples of an external great-circle tidal stream to date.
 |
Figure 3. Left: deep image of the stellar tidal stream around NGC 5907 obtained with the 0.5-meter BBO telescope (Martínez-Delgado et al. 2008). The great-circle morphology of this system is likely very similar to that of the Sagittarius stream in the Milky Way. Right: N-body model of the NGC 5907 stellar stream. The satellite is realized as a King model with an initial mass, King core and tidal radii of M = 2 × 108 M⊙, rc = 0.39 kpc and rt = 2.7 kpc, respectively. Different colors denote particles that became unbound after different pericentric passages, whereas black particles are those that remain bound. The fainter, outer loop material (blue points) became unbound at least 3.6 Gyrs ago. For this particular model the orbital period is Tr = 0.9 Gyr. |
While some galaxies have great-circle tidal streams that resemble the Sagittarius stream surrounding our Galaxy (e.g., in NGC 5907: Figure 3, or NGC 4013 and M63: Figure 1), others have enormous structures that resemble open umbrellas, and that extend over tens of kiloparsecs (e.g., the middle two panels in the bottom row of Figure 2). These structures are often located on both sides of the host galaxy, and display long narrow shafts that terminate in a giant shell of debris (e.g., NGC 4651; Foster et al. 2014). Another umbrella-like feature, dubbed the “dog leg stream” (Amorisco et al. 2015), has a long narrow spoke (with an embedded progenitor) that stretches to a radius of ∼ 150 kpc beyond the center of NGC 1097, terminating in a “dog-leg” that appears like an umbrella feature with one half of the shell missing (note that this system has other narrow plumes visible as well). With such examples, we are beginning to see real streams around galaxies in the local universe that resemble the menagerie of morphological features predicted by Λ-CDM hierarchical structure formation models (e.g., Johnston et al. 2008, Cooper et al. 2010; see also Chapter 6 of this volume).
While there have been numerous isolated discoveries of debris features around external galaxies, there have been few large-scale systematic surveys to build a comprehensive census of halo substructures. Only such a survey can inform simulations by providing estimates of the prevalence of streams of different morphologies, and thus different progenitor masses, orbits, and infall times. One systematic search by Miskolczi et al. (2011) analyzed 474 galaxies in SDSS and found clear tidal features around 6% of the galaxies, with 19% exhibiting some features above the surface brightness limit of ∼ 28 mag arcsec−2. From imaging data in the Canada-France-Hawaii Telescope Legacy Survey, Atkinson et al. (2013) find tidal features (including both minor and major merger events) around 12% of the galaxies imaged. Given that the Λ-CDM paradigm predicts that we should see accretion relics around all Milky Way-sized galaxies, this ∼ 10% fraction from the SDSS/CFHT studies would seem to suggest a significant deficit of detected accretion events relative to predictions. However, as illustrated in Figure 1 (see also Figure 6), the surface brightness limits for most large scale surveys are simply too shallow to reveal the complex webs of substructure both predicted in simulations (e.g., Bullock & Johnston 2005) and observed locally in the Milky Way and M31 (i.e., where µ > 28 mag arcsec−2 can be attained using resolved stellar populations). However, it is puzzling that no tidal stream currently known in the Milky Way or M31 is remotely as bright as the “faint limits” of these large scale searches (Sagittarius – by far the brightest stream in the Milky Way – is only ∼ 30 mag arcsec−2 at about 30-40 deg. from the core, according to Mateo et al. 1998).
2.4. Resolved structures beyond the Local Group
With large aperture facilities equipped with wide-field detectors, it is possible to resolve individual stars in some tidal debris structures beyond the Milky Way and M31. Perhaps the most spectacular example of this from ground-based observations is the Milky Way analog NGC 891, at a distance of ≈ 10 Mpc, which was surveyed by Mouhcine et al. (2010) with Suprime-Cam on the 8.2m Subaru telescope. With very long (> 11-hr) exposures, this study resolved stars to ∼ 2 magnitudes below the RGB tip in NGC 891, covering a ∼ 90 × 90 kpc area around the galaxy down to i-band magnitudes fainter than 28th mag (see Figure 4). Surface density maps of RGB stars show a complex of features looping throughout the halo of NGC 891, suggesting that the halo of this galaxy contains numerous accretion remnants. The disk of NGC 891 also appears to be “super-thick” (Mouhcine et al. 2010), providing further evidence of recent accretion. Another recent example of deep, ground-based observations is the study by Greggio et al. (2014), who mapped the density of RGB stars in the halo of the spiral galaxy NGC 253 at a distance of 3 Mpc using deep Z- and J-band imaging from the VISTA telescope. As a whole, Greggio et al. found that the halo of NGC 253 is fairly homogeneous out to ∼ 50 kpc, with the exception of a ∼ 20 kpc wide shell roughly 28 kpc from the plane that is interpreted to be the result of a recent tidal interaction. While these ground-based studies are spectacular, the extremely deep, large-area observations required to resolve individual RGB stars in the accretion relics highlight the difficulty (or perhaps impossibility) of doing similar work for large numbers of Milky Way analogs.
 |
Figure 4. Stellar density map of RGB stars stars with magnitudes 25.8 ≤ i0 ≤ 27.0 over a ∼ 90 × 90 kpc region surrounding the Milky Way analog galaxy NGC 891. Multiple interlocking loops and arc-like (“great circle”) features are visible over vast regions of the NGC 891 halo. [Figure reproduced by permission from Mouhcine et al. (2010).] |
Alternatively, space based telescopes can provide the necessary spatial resolution and depth to trace extra-galactic substructures with individual stars, albeit over significantly smaller angular scales than those on the ground (the HST+ACS field of view is ∼ 4′). The HST+ACS GHOSTS survey (Radburn-Smith et al. 2011) resolved stars at the RGB tip and used them to map low surface brightness features (to ΣV ∼ 30 mag arcsec−2) in the outer regions of 14 disk galaxies out to distances of ∼ 17 Mpc. Bailin et al. (2011) began with GHOSTS images from HST for the spiral galaxy NGC 253, and supplemented these with imaging over a much wider field of view with Magellan/IMACS, reaching well below the RGB tip. They estimated the total stellar luminosity of the NGC 253 halo to be roughly twice that of the Milky Way or M31, and fit profiles to stellar densities out to ∼ 30 kpc from the galaxy center. The shelf-like feature to the south that had been seen by Malin & Hadley (1997) in photographic plates is clearly visible, as well as other substructure at the ∼ kpc level. Thus, targeted follow-up, ground-based or space-based, for stellar streams discovered in integrated light has the potential to build datasets complementary to those generated en masse for the Milky Way and M31, reaching appropriate physical spatial scales and surface brightnesses.