Data

For the purpose of this work, we use the dataset defined in the previous chapter. Briefly, we use a sample of member galaxies of a X-ray selected groups drawn from the ECDFS, CDFN, AEGIS and COSMOS X-ray survey. This group sample comprises 83 groups with masses ranging from 10^12.5 to 10^14.5 M⊙ and in the redshift range [0.15-1.1]. The group membership is obtained via dynamical analysis. The method and its reliability for deriving group members are explained in the previous chapter. Briefly, we use all galaxy groups with flag=1 & 2 (reliable identification) in four mentioned fields. Using dynamical velocity dispersion (σ) for those with more than 10 spectroscopic members and X-ray derived velocity dispersion for other ones, we choose member galaxies inside 2 × r₂₀₀ from X-ray center and 3σ in the line of sight direction from the group redshift. In order to follow the evolution of the position of group galaxies with respect to the MS, we divide the sample of group galaxy members in two redshift bins: one at low redshift (0.15<z<0.5) and one at high redshift (0.5<z<1.1) bins. In total we have 424 galaxies in low redshift bin and 511 galaxies in high redshift bin.

4.2.1 The local galaxy density

Most of the literature regarding the study of the role of the environment in the galaxy evolution define the environment through the local galaxy density field (e.g. Peng et al.

2010, 2012). Indeed, according to simulations, the galaxy density distribution should trace the mass density field with a bias that depends on the galaxy stellar mass, since massive

4.2 Data 93

systems tend to be more clustered while low mass objects are more uniformly distributed.

However, the accuracy with which the local galaxy density field can be reconstructed has to cope with the observational limits. Indeed, most of the density estimators, such as the nearest N^th neighbour or the density of galaxies within a fixed volume, require accurate spectroscopic redshifts of a rather complete sample of galaxies. However, any spectroscopic survey has to cope with a selection function and with an incompleteness level. In addition, any density estimator has to cope with projection and selection effects. All this makes quite complicated, to different extent, the proper definition of the galaxy density field. To show this aspect we use the mock catalogue of Kitzbichler & White (2007) described in the previous chapter. We use different flavours of the two major density estimators mentioned above to check if the reconstructed density field can properly trace the underlying mass distribution. Namely we check how tight is the correlation between the local galaxy density around each simulated galaxy and the mass of its parent dark matter halo. Left panel of Fig. 4.1 shows the correlation between the local galaxy density field and the parent halo mass for a sample of simulated galaxies with known actual 3D position. The correlation is rather tight and it shows a large dispersion only at very low parent halo masses. The right panel of the same figure shows, instead, the relation obtained by introducing the projection effects due to the use of ra, dec and redshift information and the selection effects due to a simulated spectroscopic selection function. For this latter plot we use the incomplete catalogue extracted from the Kitzbichler & White (2007) catalogue as explained in previous chapter. The correlation still exists but, according to a Spearman test, it is statistically much less significant than the one shown in the left panel of the same figure. On average, the lowest values of the local galaxy density correspond to very low halo masses. However, there is a large number of galaxies in low mass halos with rather high values of projected density. This is probably due to the fact that such low mass halos are in region of high density like filaments around more massive halos. The same result is obtained by using all the flavours of the density estimators mentioned above. Thus, our conclusion is that the local galaxy density field can not be considered a good definition of the environment, because due to several observational limits it does not trace the underlying matter field.

However, it can be used to isolate at least galaxies hosted by the lowest mass halos or the highest mass halos.

Given this conclusion, we adopt the following approach. The X-ray data used in this work allow us to create a galaxy group sample in the range of masses 10^12.5 −10^14.2 M⊙. Indeed, the X-ray deep surveys used here are not deep enough to let us observe lower mass groups. In order to identify the galaxy population of lower mass dark matter halos we use the properties of the galaxy density field described above. Namely, we identify the galaxy at the lowest values of the density distribution. According to the result of our simulation, this method should be able to identify galaxies of dark matter halos with mass below 10¹² M⊙. For this exercise we use, in particular, the galaxy density estimator that shows the lowest dispersion in the relation shown in the left panel of Fig. 4.1, since it can provide the best proxy for the underlying matter density field. The best proxy is provided by the galaxy number density of systems with a stellar mass above 10¹⁰ M⊙, measured within a cilinder with radius of 0.75 Mpc and length of 1000 km/s around each galaxy. This type

Figure 4.1: Left panel: The correlation between the local galaxy density field and the parent halo mass for a sample of simulated galaxies with known actual 3D position. Right panel: The same correlation with considering projection effects due to the use of ra, dec and redshift information and the selection effects due to a simulated spectroscopic selection function.

of density estimator takes advantage of the mass segregation observed in more massive halos (Scodeggio et al. 2009) and it is defined to sample a volume (0.75 Mpc and ±1000 km/s) quite similar to the one of a group/cluster sized halo. Indeed, ther200 radius varies from 0.3-0.5 Mpc for a group to 1-1.5 Mpc for a massive cluster and 1000 km/s is roughly twice the velocity dispersion of a group and quite similar to the velocity dispersion of a massive cluster. The galaxy density derived with this approach is further corrected for the spectroscopic incompleteness that would lead to an underestimation of the actual galaxy density. This correction is estimated with the same approach described in the previous chapter. Briefly, we estimate the ratio between the number of all galaxies and those with known spectroscopic redshift with mass above 10¹⁰ M⊙, in a cylinder along the line of sight of the considered galaxy with radius corresponding to 0.75 Mpc at the redshift of the considered source and with |zsource−zphot| < 10000 km/s, where zsource is the redshift of the central source and zphot is the photometric redshift of the surrounding galaxies. The limit of 10000 km/s is roughly 3 times the error of the photometric redshifts (Ilbert et al.

2010). Of course zphot is replaced by the spectroscopic redshift whenever this is available.

Fig. 4.2 shows the histogram of the density distribution obtained with our method for the whole galaxy population considered in this work (black histogram) and for the galaxy identified as groups spectroscopic members (red histogram). As confirmed by the simulations described above, this method is able to efficiently isolate galaxies that are not hosted by massive halos. We use this histogram to define the density cut for defining our

4.2 Data 95

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30

Normalized N(density)

density (Mpc⁻²) all galaxies group members

Figure 4.2: Density distribution around each galaxy with spectroscopic redshift in AEGIS, COSMOS, ECDFS and CDFN (in black) and group member galaxies in four mentioned fields (in red). The green dashed line at ρ=3 galaxies Mpc⁻² shows the threshold to separate group from field galaxies. 73 % of all field galaxies are found at densities below this limit and 90% of all group member galaxies are above that.

“field” galaxies sample, that is galaxies that should be hosted by DM halos of masses below 10¹² M⊙ according to our simulations. The green dashed line at ρ=3 galaxies Mpc⁻² in Fig. 4.2 shows the threshold to separate group from field galaxies. Indeed, 90% of all group member galaxies are above this limit. We do this exercise separately for the two redshift bins considered in our work. This leads to a sample of 4987 field galaxies in low redshift bin and 6063 field galaxies in high redshift bin.

Similarly to Ziparo et al. (2013) we define a third environmental class of galaxies identified by density values similar to the ones of group galaxy members but not associated to any X-ray extended emission observed in the X-ray surveys considered in this work.

These galaxies have density above the ρ=3 galaxies Mpc⁻² threshold and do not lie in the sky region defined by detected X-ray extended emissions. They likely belong to filaments, sheet like structures or to groups at lower mass with respect to the mass limit imposed by the CDFS, CDFN, AEGIS and COSMOS X-ray detection limits. We define this class of objects as “filament-like” galaxies. With this approach, we define a sample of 1246

“filament like” galaxies in low redshift bin and 2320 in high redshift bin. This additional class of objects will be used also to check whether the relative vicinity of galaxies can affect galaxy properties as suggested by Peng et al. (2010, 2012) or, instead, the membership to a massive halo is a key ingredient in the galaxy evolution.