Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 327, 930-946 (1997)

Previous Section Next Section Title Page Table of Contents

1. Introduction

The centers of galaxy clusters are usually dominated by very massive ([FORMULA]) and extended ([FORMULA] kpc) galaxies, called brightest cluster members, or D or cD galaxies, whose particular physical properties require a distinct formation scenario. More detailed information on these objects is given in the reviews by Tonry (1987), Kormendy & Djorgovski (1989) and Schombert (1992). Four theories have been proposed so far to explain the properties of these central dominant galaxies.

The first theory is related to the presence of cooling flows in clusters of galaxies (Cowie & Binney 1977; Fabian & Nulsen 1977). If the central cluster density is high enough, intracluster gas can gradually condense and form stars at the bottom of the potential well. Andreon et al. (1992), however, show that colour gradients are small or absent, while McNamara & O'Connell (1992) find colour anomalies only in the inner 5-10% of the cooling radii estimated by X-ray observations. Furthermore, they find that the amplitudes of these colour anomalies imply star formation rates that account for at most a few percent of the material that is cooling and accreting on the central galaxy, if the initial mass function is the same as that of the solar neighbourhood.

The second theory involves tidal stripping. Cluster galaxies that pass near the center of the cluster may be stripped by the tidal forces arising from the cluster potential or the potential of the central galaxy itself. The stripped material eventually falls to the center of the potential well, where the giant galaxy resides, and may be responsible for the halo of cD galaxies. This theory was first proposed by Gallagher and Ostriker (1972) and later developed by Richstone (1975, 1976). It can explain the halos of cD galaxies, but it is unable to explain the differences between D galaxies, which are central dominant galaxies with no halo, and cD galaxies. Moreover, observations show that the velocity dispersion of the stars in cD halos is three times smaller than the velocity dispersion of galaxies in the cluster, and so this theory should work out how tidally stripped material is slowed down as it builds up a cD halo.

The third theory links the formation of the central galaxy to progressive mergings or captures of less massive galaxies by the central object of a cluster. This theory is known as "galactic cannibalism" and was first proposed by Ostriker & Tremaine (1975) and later developed by Ostriker & Hausman (1977). Merging might account for the formation of the central parts of first ranked galaxies with a de Vaucouleurs profile, since such a profile is often found in the simulations of galaxy mergers (Barnes & Hernquist 1992 and references therein). Photometric observations (Schombert 1987) discard the analytical approach of Ostriker and Hausman (1977), which is based on homology. The observation of multiple nuclei in central galaxies is often cited as evidence in favour of the merging theory. Nevertheless, the rates of mass increase which are obtained by analyses of samples of central galaxies with multiple nuclei (Lauer 1988; Merrifeld & Kent 1991; Blakslee & Tonry 1992) are more in agreement with a weak cannibalism than with a strong one. The observations of Thuan & Romanishin (1981), Morbey & Morris (1983) and Malumuth & Kirshner (1985) also give support to the theory of galactic cannibalism.

As a fourth alternative, Merritt (1983, 1984, 1985) suggests that the essential properties of cD galaxies are determined no later than cluster collapse. At later stages frequent merging between galaxies would be inhibited by the relatively high velocities between galaxies and the high fraction of the mass in a common background halo. Furthermore Merritt argues that truncation of galaxy halos during cluster collapse should make time scales for dynamical friction longer than a Hubble time and thus "turn off" subsequent evolution in the cluster.

In the eighties the dynamics of clusters of galaxies were explored by a number of studies which use different techniques and which include a variety of physical phenomena via numerical recipes. Such approaches have been criticized by García-Gómez, Athanassoula & Garijo (1996), who have compared a few of them to fully self-consistent simulations and have found them to be of very unequal quality.

More recently, with the advent of modern supercomputers, self-consistent simulations are possible. Funato et al. (1993) followed the evolution of 65536 particles with the special purpose GRAPE-3 machine. Mass was distributed between galaxies and a cluster background. The density profiles of both galaxies and background followed Plummer distributions with different scale sizes. In this simulation stripping was more important than merging for the evolution of the galaxies. In another simulation, however, Bode et al. (1994), using 40000 particles, found that merging was more important. In both works a central dominant object is formed as the result of the evolution of the systems. The differences in their results can be ascribed to the different density profiles and mass distributions selected to represent the galaxies (Bode et al. 1994). Bode et al. also looked for multiple nuclei in their simulations. In cases where the common halo has initially 50% of the total mass in the cluster they find that multiple nuclei are seen at least 20% of the time, with a maximum of  40% at  11 Gyrs, a number in agreement with what is expected from the projected surface density distributions. Bode et al. also showed that increasing the common halo mass slows the merging rate. For 90% of the mass in a common halo the merging time is longer than the Hubble time.

Our simulations are fully self-consistent, but, while most simulations concentrate on the dynamics of the cluster as a whole, our aim is to study the formation of the central dominant galaxy. Our simulations and their initial conditions are presented in Sect. 2 and their evolution is discussed in Sect. 3. In Sect. 4 we study the properties of the central object, both in three dimensions and projected, and compare them, whenever possible, with the observed properties of brightest cluster members. Finally we summarise our results in Sect. 5.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 1997

Online publication: April 6, 1998