5 Disruption, accretion and stripping of dwarf galaxies

What are the possible consequences, when dwarf galaxies of different types interact with the central galaxy, especially with respect to the formation of a cD halo and a rich GCS?

We make a distinction between two main cases:
(1) the infall of gas-poor dwarfs, for example dwarf ellipticals, where only the existing stellar component is involved in the interaction process, and
(2) the infall of gas-rich dwarfs, where the interaction of the gas has to be considered and might play an important role in the formation of new stellar populations.

A further sub-division of these cases is:
(a) the dwarf galaxy will be totally dissolved in the interaction process,
(b) only parts of the dwarf galaxy (for example gas and/or globular clusters) will be stripped during the passage through the central cluster region,
(c) the dwarf galaxy neither loses gas nor stars nor clusters to the cluster center, but might change its morphological shape because of tidal interactions (for example getting more compact or splitting into two).

In the next subsections we discuss the possible consequences for the different cases.

5.1 Gas-poor dwarfs

(1a): in this case the stellar population of the dwarf galaxy will be disrupted in tidal tails and the stellar light will be smeared out in the potential well of the cluster center. Most affected by this process are the faintest dEs (or dSphs, Thompson & Gregory 1993). In clusters with a low velocity dispersion or at the bottom of a local potential well in a rich cluster (Zabludoff et al. 1990), the light of several dissolved dwarfs may form an extended, diffuse cD halo. Existing GCs of the dwarfs will survive and contribute to the central GCS. In the Local Group, an example for this scenario may be the Sagittarius dSph which is dissolving into our Galaxy, adding 4 new GCs to the GCS of the Milky Way (Da Costa & Armandroff 1995). However, only few dwarf galaxies with a very rich GCS compared to their luminosity are known (Miller et al. 1998; Durrell et al. 1996). In Sect. 6 we estimate under which conditions the accretion of gas-poor dwarfs and their GCS can increase S_N of a central GCS.

Finally, the nuclei of dE,Ns can survive the dissolution of their parent galaxy and may appear as GCs (Zinnecker et al. 1988; Bassino et al. 1994). The nuclear magnitudes of all Virgo dE,Ns (Binggeli & Cameron 1991), for example, fall in the magnitude - surface brightness sequence is defined by the GCs (e.g. Binggeli 1994).

(1b): like in the case 1a the stripped stars and GCs will be distributed around the central galaxy. In this case the question arises on how large the number of stripped GCs is compared to the luminosity of the stripped stellar light. If GCs could be stripped from regions with a high local S_N, this would also increase S_N of the central GCS. According to model calculations by Muzzio et al. (1984, see also review by Muzzio 1987), the tidal accretion of GCs and stars can be an important process in the dynamical evolution of GCSs in galaxy clusters.

In some galaxies the GCS is more extended than the underlying stellar light, which has the consequence that the local S_N increases with galactocentric distance; for example NGC 4472 has a global S_N of 5.5 and a local S_N larger than 30 at 90 kpc (McLaughlin et al. 1994). Forbes et al. (1997) and Kissler-Patig et al. (1999) suggest that the stripping of the outermost GCs and stars from such a galaxy naturally increases the S_N of the central GCS. It would be interesting to investigate whether this is also true for the GCSs of dwarf galaxies.

Furthermore, it would be interesting to know how the tidal stripping process changes the shape of the remaining galaxy. Kroupa (1997) simulated the interaction of a spherical low-mass galaxy with a massive galactic halo and found that the model remnants share the properties of dwarf spheroidals. On the other hand, M 32 may be an example for a tidally compressed remnant, whose GCs have been stripped (e.g. Faber 1973; Cepa & Beckman 1988).

  

Table 4: Initial conditions for GCSs of dwarf galaxies in the Monte Carlo simulations. $N_{\rm gc}$ is the possible number of GCs in the magnitude bin. S_N gives the range of specific frequencies that can be achieved with the number of GCs

(1c): in this case the dwarf galaxy does not contribute to the formation of cD halo and central GCS. However, as in 1b one might speculate about the change of the morphological shape after a passage of the galaxy through the cluster center.

Note that, except in their nuclei, the metallicity of GCs in dEs as well as the metallicity of the bulk of their stars is very low ($-2.5 < {\rm [Fe/H]} < -1.0$ dex, see the review on Local Group dwarfs by Hodge 1994). Therefore, stripped GCs from these galaxies will only contribute to the metal-poor population of the central GCS. And accordingly, the cD halo should have quite a blue color.

5.2 Gas-rich dwarfs

(2a): for the stellar population and GCs see 1a. The infalling gas will experience the thermal pressure of the hot medium in the central galaxy. The densities can be high enough that star formation occurs and the formation of many dense and compact star clusters is possible (e.g. Ferland et al. 1994). As mentioned in Sect. 3.4, stripped gas that was not converted into stars may contribute to the intracluster X-ray gas in the cluster center (see Nath & Chiba 1995). The open questions here are, how many "new'' star clusters will survive the further cluster center evolution, and how large the number of surviving clusters is compared to the light of newly formed stars which contribute to the total light of the central galaxy and/or cD halo. In other words, it is unclear whether the formation of new GCs is so efficient that it can increase S_N of the central GCS.

Some constraints/estimations that can be made from observations of very young star clusters in merging galaxies and starburst galaxies are presented in Sect. 7.

(2b): the stripping of a gas-rich galaxy mainly affects the gas component that may form new stars and clusters as mentioned in case 2a. Nulsen (1982, see also Ferguson & Binggeli 1994) estimated a typical mass loss rate from infalling dwarfs of

$$\dot{M} = 7.4 \ 10^{-2}M_{\hbox{$\odot$}}\ {\rm yr}^{-1} n r_{\rm kpc}^2 \sigma_{\rm km s^{-1}}, \nonumber$$

where n is the gas density in the cluster, $r_{\rm kpc}$ the dwarf galaxy radius in kpc, and $\sigma_{\rm km s^{-1}}$ the velocity dispersion of the cluster. A stripping time scale for a typical dwarf irregular (dI), r = 4 kpc, gas mass $M_{\rm gas} = 10^8 M_{\hbox{$\odot$}}$, and $\sigma = 400$ km s^-1 (Fornax) is $t_{\rm S} = 0.5$ Gyr, when adopting $n = 8.2\ 10^8$ cm^-3 for the central gas density of the Fornax cluster (Ikebe et al. 1996). The fact that dEs are more concentrated towards cluster centers than dIs is interpreted as a result of this stripping scenario (dEs being the remnants of stripped dIs, e.g. Lin & Faber 1983; Kormendy 1985). Furthermore, non-nucleated dEs have a quite low GC S_N in contrast to dE,Ns, for which S_N increases with decreasing luminosity (Miller et al. 1998). One might speculate that not only gas but also GCs have been stripped, whereas the dE,Ns have a different evolution history.

(2c): the passage of a gas-rich dwarf through the intergalactic gas in a cluster might trigger star formation in the dwarf galaxy itself (Silk et al. 1987) and might enrich its GCS (see Sect. 7). Ferguson & Binggeli (1994) suggest that a galaxy falling into the cluster for the first time in the present epoch encounters such high densities that stars can form long before ram pressure becomes efficient. A further close passage then can result in the cases 2a or 2b.

In all these cases there are no restrictions for the metallicity of newly formed GCs and stars. Their metallicity depends on the gas enrichment history of the accreted or stripped galaxies themselves. Of course, a lower limit is the metallicity of the interstellar matter of the dwarf galaxy. As an estimation for this lower limit can serve the metallicity of the young populations in the Local Group dSphs and irregulars. It seems that the secondary stellar population has metallicities more metal-rich than ${\rm [Fe/H]} = -1.2$ dex, in some cases even up to solar values (e.g. Grebel 1997).