Tidal tails, and more generally the fine structures that surround galaxies (stellar streams, rings, bridges, shells) are among the least ambiguous signposts of galaxy evolution. Indeed, whereas other galactic properties such as the presence of spiral structures, bars, warps or even starbursts, may be accounted for by secular and internal evolution, the formation of stellar filaments can only be explained by a past collision between galaxies. Numerical cosmological simulations predict the formation of many such structures (see among many others Johnston et al. 2008, Peirani et al. 2010). However, their census and interpretation face a number of issues.
Such issues might be addressed by combining predictions from numerical simulations and observations. We present below examples of the use of tidal structures as probes of galaxy evolution and the mass assembly of galaxies.
6.1. Determining the merger rate evolution with tidal tails
Early deep observations with the Hubble Space Telescope revealed that distant galaxies (z > 1) seemed to be much more morphologically perturbed than local galaxies (Griffiths et al. 1994, Glazebrook et al. 1995, Abraham et al. 1996), supporting the idea that a smaller, denser and younger Universe favored galaxy-galaxy collisions. Since then, many studies based on deep surveys, such as the one illustrated in Figure 13, have tried to quantify the evolution of the merger rate as a function of time, without in fact reaching a consensual value. A variety of methods have been used, based on:
However a few remarks need to be made at this stage: first, the most massive component of tidal tails formed in major mergers is by far the atomic hydrogen. As mentioned before, HI surveys might disclose collisional debris that are hardly visible in the optical. Unfortunately, the current technology and antennas sensitivity limit the detection of the 21 cm emission line to redshifts less than 0.3. In the more distant Universe, tidal tails may only be observed through the emission of their stars. Intrinsic dimming with redshift as well as band shifting make them less and less visible and bias surveys in favor of UV emitting, star forming structures. Other difficulties arise at high redshift. The gas fraction of galaxies was higher and their gaseous disks more unstable. Prominent star forming condensations formed in the disk may be mistaken with either multiple nuclei of merging galaxies or even condensations within tidal tails (Elmegreen et al. 2009). Among these "clumpy" galaxies, only a fraction of them (for instance the so-called "tadpoles" systems) may be genuine interacting systems (Elmegreen et al. 2007). One usual hypothesis when counting the number of tidally perturbed systems is that disk-disk collisions at low and high redshift produce similar external structures. However if the colliding progenitors are the gas-rich clumpy disks mentioned above, the mutual interaction between their clumps (which have masses comparable to that of dwarf galaxies) might prevent the formation of tidal tails (Bournaud et al. 2011). Thus, when trying to measure the evolution of the past merger rate by looking at the level of tidal perturbations, one should keep in mind that distant tidally interacting galaxies might differ from those observed in the Local Universe.
A last word of caution: when comparing the merger rate at low and high redshift, it is assumed that the fraction of galaxies involved in a tidal interaction is well known in the nearby Universe (and considered not to exceed a few percent, see Miskolczi et al. 2011). However even tidal tails from past major mergers might have been missed at z = 0 because of their low surface brightness. Indeed, the extremely deep mapping of the Andromeda region, in the Local Group, has revealed an extremely faint stellar bridge between M31 and M33 (McConnachie et al. 2009), suggesting that the two spirals are involved in a tidal collision. Prominent tidal tails of very low-surface brightness were also recently discovered around apparently relaxed massive ellipticals (Duc et al. 2011). An example is shown on the top panel of Figure 8. Such observations indicate that, in the local Universe, the fraction of tidally interacting galaxies is likely underestimated: serious issues plague the determination of the merger rate even at low redshift.
Figure 13. The Hubble Ultra Deep Field, showing several tidally perturbed distant galaxies. The evolution of the fraction of collisions as a function of redshift is the subject of strong debates. Credit: NASA/ESA.
6.2. Determining the mass assembly history of galaxies with tidal tails
The existence of a tidal tail unambiguously establishes the occurrence of merger event in the past history of the host galaxy. Therefore, the census of the collisional debris might, in principle, constrain the recent mass assembly of nearby galaxies. Now that surface brightness limits of unprecedented depth may be reached with the current generation of optical cameras with large-field of view, this method of galactic archeology might be very powerful. However, it faces a number of issues:
As a consequence, it might be difficult to probe collisions older than a few Gyrs.
6.3. Constraining the distribution of dark matter with tidal tails
Tidal tails might not only tell us about the baryonic content of their parent galaxies and how it reacted to the environment; they are as well insightful to constrain the structure and distribution of the most massive component of galaxies: dark matter (DM). Rotation curves of galaxies reveal how much gravitational matter is located within the radius at which velocities are measured but do not constrain the extent and 3D shape of the dark matter halo. The halo of CDM models is very extended, at least 10 times the optical radius. The rotation curve cannot be easily probed at these large distances. Tidal tails produced during major mergers have however sizes that can exceed 100 kpc, reaching the outskirts of the dark matter halos: tails are thus a priori a convenient tool to probe the structure of cosmological halos. Numerical simulations have been used to study the effect of the size of the DM halo on the shape of tidal tails. Apparently contradictory results have been obtained, claiming or not a dependence with the halo mass, size, concentration or spin (Dubinski et al. 1996, Mihos et al. 1998, Dubinski et al. 1999, Springel & White 1999).
The shape of the DM halo, its triaxiality and presence of sub-halos might be probed by smaller, thiner tidal tails from minor mergers that wrap around galaxies. Those found around the Milky Way, such as the Sagittarius stream, are the target of numerous studies (e.g. Mayer et al. 2002, Helmi 2004, Peñarrubia et al. 2006, Varghese et al. 2011).
While no direct correlation between the size of the DM halo and the size of tidal tails has yet been established, the internal structure of tidal tails might be connected to the DM extent. Bournaud et al. (2003) argued that the massive condensations at the tip of tidal tails, associated with TDGs, cannot be formed if the halo of the parent galaxy is truncated. Duc et al. (2004) provided a toy model showing that in the case of a truncated halo, the tidal material is stretched along the tidal tails, preventing its collapse and the formation of massive sub-structures. When the halo is large enough, this stretching does no longer occur beyond a certain distance, and apparent massive condensations near the tip of the tail might form TDGs (see Figure 12). The observation of TDGs is thus consistent with the extended dark matter halos predicted by the CDM theory.
If large DM halos seem to be required to form TDGs and shape the inner structures of tidal tails, tails should themselves not contain large quantities of dark matter. Indeed the current picture of DM makes them collisionless particles distributed in a hot halo on which tidal forces have little impact. The tidal material originates from the disk, which is predicted to contain almost no DM. In practice, the DM content of tidal tails is difficult to probe. However in some special circumstances, it may be measured using the traditional method of rotation curves. Tidal dwarfs are gravitationally bound systems; their DM content may thus simply be derived determining their dynamical mass and subtracting it from the luminous one (consisting of HI, H2, stars and dust). This exercise has been carried out for a few systems (Bournaud et al. 2007, Duc et al. 2007, Belles et al 2012, in prep.). Even if the error bars are large, these measurements yield reliable dynamical to luminous mass ratios of 2-3. Assuming that the CDM theory is correct, one should conclude that TDGs (and thus the galactic disks) contain non-conventional dark matter, likely traditional baryonic matter which has not yet been detected by existing surveys. A possible candidate is molecular gas not accounted for by CO observations. The observations of dust in the far infrared by the Planck satellite supports the hypothesis of an unseen, dark, component in the gaseous disk of galaxies, which might contribute to the global budget of the missing baryons in the Local Universe (Planck Collaboration 2011). Alternatively, CDM might be wrong, as claimed by several groups who push for modified gravity. Modified Newtonian Dynamics (MOND) has retrieved the rotation curves of galaxies, including TDGs, without the need of a dark matter halo (Milgrom 2007, Gentile et al. 2007). Numerical simulations of galactic collisions in the MOND framework have been carried out: they also reproduce the long tidal tails made with classical Newtonian dynamics (Tiret & Combes 2007). The main difference is the absence of dynamical friction during the collision, which contributes to extend the time scale of the collision, and decrease the probability of a final coalescence.