Model Settings: Size of the reference sample

In Chapter 1.2.4 and 2.6 the method of modeling color distributions by Monte-Carlo simula- tions was described, as well as the determination of the cumulative age distribution. With respect to the model parameters used previously it has been pointed out that the size of the simulated globular cluster sample was set with respect to the NGC 5846 sample, one of the most numerous globular cluster samples in our set of galaxies. From the stability tests (Section 2.4.2) we know that the sample size becomes important for theχ2-test and we therefore want to know whether for small samples the test results improve if the simulated color distributions agree in their size with the different observed globular cluster systems. For NGC 7192, NGC 3115 and M87 we simulated globular cluster with the blue/red (V −K) sub-population containing 21/40, 22/60 and 22/60 objects in NGC 7192, NGC 3115 and M87 respectively. The results of the comparison between the age distributions are shown in Figure 6.4. In the comparison with the corresponding plots in Figure 7.5 and 7.9 we do not find a significant improvement in the results. The age and size estimates for the small sample size are still much more uncertain than for NGC 5846 and NGC 4365. Thus we conclude that the minimum sample size, required for our semi-numerical method to describe the age structure in globular cluster systems, is mostly dependet on our choice of the applied SSP models in combination with the available broad band colors (see Chapter 3.5.2).

Fig. 6.4.— Results of theχ2- test if the globular cluster sample in the simulations is set similar to the observed system (previous runs used the significantly larger models than NGC 7192, NGC 3115 and M87). We exclude NGC 4478 from this test due to its extremely small sample (6 objects only). The modeling as well as the determination of the cumulative age distributions was done using the Bruzual & Charlot SSP models (2003).

Fig. 6.5.— Cumulative age distributions derived for simulated globular cluster systems using the Bruzual & Charlot SSP model isochrones (2000).

Fig. 6.6.— Cumulative age distributions derived for simulated globular cluster systems using the Bruzual & Charlot SSP model isochrones (2003).

Fig. 6.7.— Cumulative age distributions derived for simulated globular cluster systems using the Vazdekis SSP model isochrones (1999).

Fig. 6.8.— Cumulative age distributions derived for simulated globular cluster systems using the Maraston SSP model isochrones (2001).

References

Bruzual, G.A. & Charlot, S. 1993, ApJ, 405, 393B

Bruzual, G.A. 2000, private communication

Bruzual, G.A. & Charlot, S. 2003, MNRAS, 344, 1000

Maraston, C., Greggio, L. & Thomas, D. 2001, Ap&SS, 276, 893

Thomas, D.& Maraston, C. 2003, A&A, 401, 429

Thomas, D., Maraston, C. & Bender, R. 2003, MNRAS, 339, 897

7 Globular Cluster Systems

vs.

Globular Cluster

Systems

Abstract

The age structure of globular cluster systems in early-type galaxies provides important clues about their formation and evolution. The semi-numerical method of detecting globular cluster sub-populations and deriving age and size limits for them, as introduced in chapter 1.2.4 and 2.6, has been applied to various globular cluster systems. In this chapter the galaxy sample will be extended and the results will be compared. In this comparison we use the galaxy environment as external parameter, in order to investigate to which degree the evolution of the galaxy depends on where the galaxy is found, e.g. in the center or the fringe of a massive galaxy clusters, in smaller groups or in rather isolated environment. In a sample of eight early-type galaxies four have been found with globular cluster sub-populations of different age, three of them in smaller groups or in the outskirts of the galaxy cluster center.

7.1 Introduction

In the last decades an immense wealth on observational data on globular clusters and globu- lar cluster systems has become available and been mined fiercely. The tight link between globular cluster systems and their host galaxies on the one side, and the still ongoing discussion about ”the” formation scenario valid for early-type galaxies on the other side created an ever growing interest in this objects. Fact is, that various features of a given early-type galaxy are best explained by any of the commonly accepted formation scenarios, whereas others are in complete contradiction to it. The fundamental plane, for example, is defined by correlations between photometric, kinematic and chemical properties (e.g. Djorgovski & Davies 1987, Dressler et al. 1987, Djorgovski, Pahre & de Carvalho 1996) and has been proven valid for a large range in redshift/look back time. The conclusion seemed to be that early-type galaxies form a very homogeneous population, consisting of an old stellar population and being similar in their basic features (e.g. Bower, Lucey & Ellis 1992, Renzini & Ciotti 1993, Ellis et al. 1997). Already before the observational and theoretical tools to derive the ages in extragalactic stellar populations with high accuracy became available, evidence was found that these correlations may hide more complicated and variable galaxy structures. In Chapter and 5 galaxies are discussed, which show kinematically de-coupled cores or prominent dust lanes. Obviously, galaxies with such peculiar features are prime targets to study the age structure in their globular cluster systems. Previous attempts to derive the age of stellar pop- ulations, either photometrically or spectroscopically, have been hampered by the age-metallicity degeneracy in broad band colors (Worthey 1994), luminosity weighted contributions of different age sub-populations to the diffuse galaxy light or, from a more technical point of view, the large distance of early-type galaxies. The introduction of combined optical and near-infrared photometry, the discovery of globular clusters representing single stellar populations and 8-10 m class ground based or space based telescopes made it possible to overcome those obstacles. Consequently, the next step towards an understanding of the formation of early-type galaxies is to investigate their age structure and find systematic correlations to external parameters, such as the luminosity/mass or the galaxy environment. Since high signal to noise spectra for globular clusters are still very time consuming photometric studies are a powerful tool to search for age sub-populations in globular cluster systems and to extend the galaxy sample, i.e. to cover a range of luminosities and different galaxy environments (cluster/group galaxies, field galaxies).

This thesis introduces a semi-numerical approach to detect age sub-populations in globular cluster systems on the basis of combined optical and near-infrared photometry. After the method has been describes and first results presented in the previous chapters, we now complement the galaxy sample and compare, respectively discuss the results.