CCD Imaging: First Data

 
Figure 1: A representative image of the field centred on Abell 22. The open circles represent all those galaxies in the field brighter than our spectroscopic limit of B=21.

 
Figure 2: Field-subtracted colour histograms of galaxies down to R = 20.0 for annuli at increasing radius from the centre of A1651.

The imaging component of our project is close to completion as a result of four observing runs at LCO over the period 1996 March to 1997 September. In that time, 17 clusters were imaged in both B and R, mostly in photometric conditions and in 1.1 arcsec -1.5 arcsec seeing. U-band data for a subset of clusters were also taken to better identify star-forming members. Exposures of 3 `field' regions which are known to be devoid of clusters (Dalton et al. 1997) were also obtained to compile a large statistical sample of field galaxies. Exposures taken in photometric conditions are indicated in Table 1. For the non-photometric cases, a series of short, follow-up exposures taken in photometric conditions were obtained to ensure zero-point continuity over each cluster.

The reduction of the data has been carried out using a combination of packages in IRAF and the automated image analysis package, SExtractor, of Bertin & Arnouts (1996). The former has been used to undertake the preliminary reduction steps: The COLBIAS task was used for bias subtraction. Images were then flat-fielded using ``super-flat'' frames created by median-combining an appropriate series of program images taken on a given night. The images were then aligned and coadded using the IMALIGN and IMCOMBINE routines, with cosmic ray events and chip imperfections eliminated in the process. The detection, photometry and star/galaxy separation of objects was then done automatically on the coadded frames using SExtractor. Object identification was conducted on the R-band images using a (above sky) detection threshold within 12 or more contiguous pixels. Total magnitudes (within an aperture of radius ) and fixed (4´´ ) aperture colours were then measured for all these objects using the R and B images. In cases where the seeing differed between the two bands, the images with the best seeing were smoothed to match those with the worst seeing before making the colour measurements. For star/galaxy separation, we used SExtractor's neural network-based algorithm.

Figure 1 shows a compilation of data from the 21 images taken of Abell 22. The open circles represent the location of the 907 galaxies detected within the field with , our limit for 2dF spectroscopy. The check on the uniformity of our photometry facilitated by the overlap of our CCD images, indicates that our magnitude and colour zero-points are stable to mag across the whole field. There is a clear overdensity of galaxies in the centre of the field which makes up the elongated cluster core. The asymmetric distribution seen here in the core of Abell 22 is typical of many of our clusters, suggesting that the traditional assumption of a spherically symmetric galaxy distribution is the exception rather than the rule.

Field-subtracted colour distributions of galaxies in the field of Abell 1651 are shown in Figure 2. The ``field'' is measured from the outskirts of the cluster itself to give a more accurate local field density subtraction. Here the data have been subdivided into annuli each 6arcmin in width, starting at the cluster centre and going out to a radius which encloses % of the cluster population. Importantly, the radius of the innermost annulus corresponds to R (Mpc at the redshift of the cluster), the standard Butcher-Oemler radius enclosing the inner 30% of the cluster population and used for measurements of blue galaxy fraction (Couch 1981, Butcher & Oemler 1984). We see there is a strong E/S0 sequence in this region of the cluster, giving rise to the dominant peak in the colour distribution at . Any population of blue galaxies, if present, would appear to be very small. There is also a conspicuous peak in the colour distribution at B - R of 2.25 within the annulus, suggestive perhaps of a second, more distant cluster within the two degree region of this field. Clearly spectroscopy is crucial for identifying blue cluster members and for detecting and eliminating contaminating objects (and members of other clusters!) within each field.

 


Figure 3: left-Abell 3112, right-Abell 3888. The grey scale is an R-band optical image of the cluster core taken at LCO. The overlayed contours represent the ROSAT HRI X-ray images which show the gas distribution within each of the clusters.

The distribution of galaxies relative to the hot X-ray gas lends insight into the complex dynamical and structural morphology of each cluster. We can easily compare these features by overlaying high resolution X-ray contours on optical images of our clusters. Figure 3 illustrates the variety in structural morphology of the central regions of clusters in our sample. The regular, concentric contours centred on the cD galaxy of Abell 3112 (left) show that it is a relaxed cluster. There is a strong radio source also coincident with the projected position of the cD, which, if it were an AGN, could have an effect on the seeming relaxedness of this cluster. However, the XBACs sample (E96) from which we drew our cluster subsample had removed all systems featuring prominent contaminating X-ray point sources, so this particular radio source does not affect the cluster X-ray map and we are indeed looking at a relaxed cluster. In cases where there is clear displacement of the cD galaxy from the X-ray centroid the X-ray centroid is taken as the centre of mass of the cluster. Abell 3888 (right) does not have a cD and is clearly not relaxed. There are several subclumps of galaxies and X-ray peaks and only one of the bright central galaxies coincides with a local X-ray maximum.

© Copyright Astronomical Society of Australia 1997