El Boom de Opciones en un Gran Mercado Global

My thesis aims at characterizing the properties of galaxies in clusters at low and high red- shifts, as a method of identifying the physical processes which act on galaxies in clusters. Special attention is given to the Brightest Cluster Galaxies in the local cluster sample. The other cluster galaxies are characterized mainly by their stellar mass and their star formation history, rather than their morphology. This allows inferences to the physical processes which affect star formation in cluster galaxies.

A major technical focus of this thesis has been the compilation of a reliable sample of galaxy groups and clusters at low redshifts (z .0.1), to be used as a local comparison sample for cluster surveys at higher redshift, in particular for the ESO Distant Cluster Survey (EDisCS). The cluster sample must provide the local equivalent to the high quality deep observations being gathered for high-redshift cluster samples, while at the same time encompassing many more systems in order to reduce statistical uncertainties.

Such datasets have recently become possible with the increasing number of large sky survey projects. Arguably the one that is best suited for the construction of a local cluster sample with the desired properties is the Sloan Digital Sky Survey (SDSS). Various cluster catalogs have been compiled for the SDSS, and it was not the purpose of this thesis to provide yet an- other one. Rather, I have used an existing cluster sample, the C4 catalog, and have modified it to match the specifications.

An important improvement to the original C4 catalog has been the identification of the Brightest Cluster Galaxies (BCGs). As described in Sect. 2.1, BCGs are a particularly in- triguing class of galaxies. In Chapter 2, I ask “How special are Brightest Cluster Galaxies?” and investigate the systematic differences between BCGs and sample of non-BCGs of equal mass, redshift, and color. Between these matched samples, the difference lies in the location: the BCGs lie at the centers of galaxy clusters, the comparison galaxies not. This allows to single out systematic differences in galaxy properties due to the location, as well as the asso- ciated assembly history. Such a systematic comparison has only become with the sample of BCGs which I have compiled in this work.

There are two complementary methods to identify the physical processes which act on galaxies upon infall into the cluster: one is the fossil record in cluster galaxies, the other direct comparison of cluster samples at different cosmic epochs. In Chapters 3 and 4, I inves- tigate the fossil record, i.e. the galaxy populations in the local SDSS clusters. Chapter 3 aims to characterize the galaxy distribution within the clusters, i.e. the number density profile, and how the distribution depends upon galaxy color and luminosity. In Chapter 4 I make use of the Principal Component Analysis developed by Wild et al. (2007) to characterize the recent star formation history of galaxies in clusters and the field. Several studies have investigated the star formation rate in clusters, but this work sets itself apart by the large sample size, the

1.8 Aims and structure of this thesis

use of spectroscopic indicators for recent and current star formation and AGN activity, and characterizing the cluster galaxy population as function of galaxy stellar mass rather than luminosity.

Finally, in Chapter 5, I compare the population mix in the high-redshift EDisCS clusters to the low-redshift clusters. The EDisCS clusters have been selected optically rather than with X-rays, and are thus not biased towards more massive, relaxed systems as X-ray selected high-redshift cluster samples are. The distribution of cluster velocity dispersions suggests that the EDisCS clusters can indeed be regarded as progenitor systems of the SDSS cluster sample, making a direct comparison of the samples feasible. This comparison is performed via the same spectroscopic analysis used in Chapter 4. The compatibility in selection, mass range, and spectroscopic coverage make this a robust study of evolutionary trends of the galaxy populations in clusters.

2

How special are Brightest Cluster Galaxies?

In this chapter, I use the Sloan Digital Sky Survey to construct a sample of 625 brightest group and cluster galaxies (BCGs) together with control samples of non– BCGs matched in stellar mass, redshift, and color. I investigate how the systematic properties of BCGs depend on stellar mass and on their privileged location near the cluster center. The groups and clusters that I study are drawn from the C4 cata- logue of Miller et al., but I have developed improved algorithms for identifying the BCG and for measuring the cluster velocity dispersion. Since the SDSS photometric pipeline tends to underestimate the luminosities of large galaxies in dense environ- ments, I have developed a correction for this effect which can be readily applied to the published catalog data. I find that BCGs are larger and have higher velocity dis- persions than non-BCGs of the same stellar mass, which implies that BCGs contain a larger fraction of dark matter. In contrast to non–BCGs, the dynamical mass-to- light ratio of BCGs does not vary as a function of galaxy luminosity. Hence BCGs lie on a different fundamental plane than ordinary elliptical galaxies. BCGs also follow a steeper Faber–Jackson relation than non-BCGs, as suggested by models in which BCGs assemble via dissipationless mergers along preferentially radial orbits. I find tentative evidence that this steepening is stronger in more massive clusters. BCGs have similar mean stellar ages and metallicities to non-BCGs of the same mass, but they have somewhat higher α/Fe ratios, indicating that star formation may have occurred over a shorter timescale in the BCGs. Finally, I find that BCGs are more likely to host radio–loud active galactic nuclei than other galaxies of the same mass, but are less likely to host an optical AGN. The differences I find are more pronounced for the less massive BCGs, i.e. they are stronger at the galaxy group level.

This chapter has been published as von der Linden, A., et al. 2007, MNRAS, 379, 867 .

2.1 Introduction

The central galaxies in galaxy clusters seem to be special - in many cases, the differences are visually obvious, because central cluster galaxies often have extended envelopes (i.e. they are cD galaxies) and they are usually the brightest (and most massive) galaxies in their clusters. The termbrightest cluster galaxy (BCG) has thus become synonymous with the termcentral galaxy.

At first glance, it might seem evident that the location of the BCG at the bottom of the potential well of a cluster must be the cause for any property which distinguishes it from other (cluster) galaxies. However, BCGs are also the dominant population at the massive end of the galaxy luminosity function, and thus, their properties are influenced both by their large masses and by the cluster environment. It is very difficult to disentangle these two in- fluences, because it is difficult to find equally massive non-BCGs for comparison. Since most BCGs are early-type galaxies, their properties are often compared with the known scaling relations for elliptical galaxies. It has been claimed that BCGs lie on the same Fundamental Plane as other ellipticals (Oegerle & Hoessel 1991), but that they lie off its projections (e.g. the Faber–Jackson and Kormendy relations) in that they have lower velocity dispersions and larger radii than predicted by these relations (Thuan & Romanishin 1981; Hoessel et al. 1987; Schombert 1987; Oegerle & Hoessel 1991). More recently, it has been claimed that the slopes of the Faber–Jackson and Kormendy relations change as a function of galaxy luminosity for all elliptical galaxies (Lauer et al. 2006; Desroches et al. 2007). On the other hand, Brough et al. (2005) find that the surface brightness profiles (and thus the radii) of BCGs depend on the host cluster properties.

The formation mechanism of BCGs is also a subject of much debate. Early on, it was suggested that BCGs form when galaxies sink to the bottom of the potential well of a cluster and merge (termed galactic cannibalism; Ostriker & Tremaine 1975; White 1976). However, Merritt (1985) argued that tidal stripping would reduce the masses of cluster galaxies to the point where dynamical friction is too slow for this to be a viable mechanism. These analyses assumed that clusters are static entities. A further mechanism to form BCGs in situ in the cluster is star formation in cluster cooling flows. At the centers of clusters, gas reaches high enough density to cool and condense into the cluster core (Silk 1976; Fabian 1994). But while the mass deposition rates inferred from the X-ray luminosities of cooling flow clusters are of the order of several hundreds to >1000M⊙yr−1 (e.g. Allen et al. 1996),

observed star formation rates are at most _∼100M⊙yr−1 (Crawford et al. 1999). Recent X-

ray studies have furthermore demonstrated that the cluster gas does not cool below_∼2 keV (e.g. Peterson & Fabian 2006). Moreover, this scenario predicts that the stellar populations of BCGs should be young and blue, which is clearly inconsistent with observations.

These scenarios were proposed before hierarchical structure formation was fully established as the standard cosmological paradigm, and for simplicity they neglected many of the pro- cesses that take place when clusters assemble through mergers. Dubinski (1998) used N-body simulations to show that a dominant galaxy forms naturally by merging of massive galaxies when a cluster collapses along filaments. Recently, De Lucia & Blaizot (2007) investigated the formation of BCGs in the context of the Millenium Run simulation (Springel et al. 2005). In their model, the stars that make up BCGs today are formed in a number of galaxies at high redshifts, which subsequently merge to form larger systems (see Fig. 1.9). The final BCGs assemble rather late: by a redshift of z_∼0.5, on average about half of the final stellar mass