Astron. Astrophys. 360, 861-870 (2000)

3. The definition of the galaxy systems in redshift space

Since the six cluster candidates are drawn from the EIS cluster catalogue, they obviously correspond to significant density enhancements in projection. Here we search for systems of galaxy redshifts that could be associated to the 2-d over-densities. In this way we assign a spectroscopic redshift to 4 of our cluster candidates. We also evaluate the probability that these systems correspond to a genuine three-dimensional cluster.

Olsen et al. (1999a) search for clusters in projection assuming a 2-d radial density profile with a cutoff radius h Mpc. This size is well fitted to the EFOSC2 field-of-view, corresponding to h Mpc², at (the average estimated redshift of our candidate clusters). Therefore, we search for cluster members in redshift-space within the whole EFOSC2 field.

Several refined algorithms for the definition of systems of galaxies in redshift space can be found in the literature (e.g. Katgert et al. 1996; Pisani 1993). However, with only a dozen galaxy redshifts per field, these sophisticated algorithms can not be applied. We choose to identify galaxy systems using a physical criterion based on the well established properties of nearby clusters of galaxies.

Within each EFOSC2 field, we identify any set of two or more galaxies contained within a given redshift range, . In order to define , we note that Abell-like clusters of galaxies have mean velocity dispersions  km s^-1 (Girardi et al. 1993). Since the line-of-sight velocity distributions of clusters are approximately gaussian in shape (Girardi et al. 1993), % of the cluster members have a velocity within . As a consequence, galaxies in a given cluster should be located within a redshift range , taking into account the cosmological factor (Danese et al. 1980).

We list in Table 3 the sets of galaxies we identify in redshift space. In Col. (1) we list the EIS cluster field identification, in Col. (2) the galaxy set identification, in Col. (3) the number of galaxies in the set, in Col. (4) the median redshift of the set, in Col. (5) the total redshift range covered by the galaxies within the galaxy set. In Col. (6) we list the probability of the galaxy set to correspond to a significant overdensity in redshift space, as estimated from resamplings of the Canada-France Redshift Survey data-base (CFRS; Lilly et al. 1995; Le Fèvre et al. 1995; Hammer et al. 1995; Crampton et al. 1995) - see Sect. 4. The real systems (probability ) are flagged in Col. (7).

In total we identify 19 galaxy sets along the line-of-sight of six EIS candidate clusters. It is interesting to detail the comparison of the estimated mean redshifts, 's, of these clusters (see Table 1), to the spectroscopic redshifts of the 19 sets (see Table 3). In this comparison, we take into account the uncertainties in the mean redshifts of the galaxy sets, and note that the matched-filter redshift estimates are at most accurate to within .

In the case of the candidate clusters EIS0533-2353, EIS0950-2154, and EIS0951-2047, we do not find any galaxy set close to the estimated cluster redshifts. According to the matched-filter algorithm, these three clusters are located at a higher redshift than any of the galaxy sets found in their fields. In each of the fields of EIS0955-2113 and of EIS0956-2009, there is one set of galaxies with mean redshift close to the estimated redshift ( vs. , and vs. , respectively). Finally, in the field of EIS0540-2418, there are two sets of galaxies with mean redshifts close to the cluster estimated redshift, .

The detection of galaxy concentrations close to the estimated redshifts of three EIS clusters supports the reliability of the matched-filter redshift estimates.

As far as the failed detections are concerned, EIS0533-2353 may have escaped detection because of its high (estimated) redshift, - we may simply have not been observing deep enough. Moreover, the line-of-sight to a single cluster can intercept several different galaxy sets. Katgert et al. (1996) estimate that 10% of nearby Abell clusters result from the superposition of two almost equally rich systems (and this fraction is probably higher for more distant clusters). The nearest of these systems has the highest chance of being detected. In this context, we note that the low-z set of galaxies, 3a, detected in the field of EIS0950-2154 is somewhat off-centered with respect to the nominal EIS cluster center (see Fig. 3). The same is not true for the sets with in the fields of EIS0540-2418, EIS0955-2113, and EIS0956-2009. This fact suggests that the set 3a does not correspond to the EIS cluster, but is a foreground group.

Da Costa et al. (2000) suggest that all the six EIS candidate clusters could be real clusters at redshift . Da Costa et al. base their suggestion on the detection of red-sequences of early-type galaxies in the colour-magnitude diagrams. They note that under-sampling of the redshift distribution of galaxies in the cluster fields may explain the lack of spectroscopic detections of some clusters.

© European Southern Observatory (ESO) 2000

Online publication: August 23, 2000
helpdesk.link@springer.de