Observations give some clues on the formation mechanism of tidal dwarfs. In the young TDGs observed so far, the atomic hydrogen makes the bulk of their mass. Therefore gas should play a key role. On the theoretical side, several scenarios have been proposed, supported by various types of numerical simulations:
The variety of proposed scenarios tells how much having ad-hoc initial conditions and all necessary ingredients in the simulations is important: if the dark matter halo is truncated in the numerical simulations (to lower their computational cost), scenario (4) will not work; scenarios (2)-(5) require proper treatment of the gaseous component, including feedback. Scenario (6) needs high resolution, so as to resolve Super Star Clusters. Fig. 2 presents one of such simulations fulfilling most of these criteria. The numerical model used a total of 36 million particles, including 12 million "sticky" particles for the gas component, and minimal grid cell size of 32 pc [6]. The production of star clusters with masses down to 105 M was directly resolved in these simulations. The mass spectrum of objects produced during the merger seems to be bimodal. Two distinct families arise: (a) compact SSCs, with masses less than 108 M which seem pressure supported and may be the progenitors of globular clusters; (b) extended objects with masses above 108 M which are usually supported by rotation. The latter have the properties of observed Tidal Dwarf Galaxies. Thus TDGs are not simply the high mass end of SSCs, a conclusion that was also reached from the analysis of HST images [16]. Furthermore, analyzing snapshots of the simulation for a period of one Gyr, we found no evidence that the latter evolve into the former, via merging. The TDG progenitors are visible soon after the first encounter, in the outskirts of the colliding galaxies, at a time when the tidal tails have not yet completely unfolded. They quickly collect all their building material. After about 100 Myr, their mass is stabilized. Rotational support appears as well very early on. Only a few massive objects are formed later on within the tidal tails. Note however that star-formation and gas feedback are not properly handled with the sticky particles used in these simulations. Investigations may now be carried using fully hydrodynamical simulations [26].
Figure 2. Formation of tidal dwarf galaxies in high resolution numerical simulation of a major merger [6]. Two snapshots are shown, resp. after the first encounter and the merger (Belles et al., in prep). |