DEEP CCD PHOTOMETRY OF THE RICH GALAXY CLUSTER ABELL 1656: CHARACTERISTICS OF THE DWARF ELLIPTICAL GALAXY POPULATION IN THE CLUSTER CORE

Abstract

ABSTRACT I have developed a statistically rigorous and automated program code (DYNAMO) to implement the detection, photometry and classification of faint objects on digital images. Detected objects correspond to local maxima of the intensity distribution, whose peak-flux pixel intensity exceeds a user-defined threshold. For all detected objects, several intensity-weighted quantities are computed: average and centers, the radial moments r1 and r-2, the ellipticity epsilon and the concentration index cin. Then, Kron's 2r1 total magnitude, integrated colors derived from standard fixed aperture magnitudes, and measures of surface brightness (central and average) are then calculated for each object. The process of discrimination between galaxy and starlike objects is performed interactively, through parameter-space culling and color-magnitude discrimination. Statistical number correction with remote background fields can then be used to further reduce contamination in the resulting distributions. This program code was tested and characterized using artificial CCD images of star and galaxy fields: it is accurate, robust and versatile. I have analyzed deep R- and B-band KPNO 4-m CCD images of the central ~700 arcmin2 of the Coma cluster (Abell 1656, v ~= 7000 km/s, richness-class 2), using the new photometry software described above. I detected and measured total R magnitudes and (B-R) colors for a sample of 3741 objects in Coma cluster core and 1164 objects on a remote control field, complete to a limiting magnitude of R = 22.5 mag. The typical uncertainties are ±0.06 and ±0.12 mag in total magnitude and color respectively. The dwarf elliptical (dE) galaxies are confined to a tight sequence in the color range given by 0.7 ≤ (B-R) ≤ 1.9 mag; within this interval there are 2535 dE candidates in the cluster core, and 694 objects on the control field (2.57 X less area). Scaling the number of objects on the control field to the area of the program fields leaves a sample of about 750 candidate dwarf elliptical galaxies. I detected a significant color gradient in the radial distribution of dE galaxies, and calibrating a color-metallicity relationship based upon Galactic globular clusters, this color gradient corresponds to a metallicity gradient which goes as Z proportional to R-0.32 outwards from the cluster center at NGC 4874. As well, there is a strong color-luminosity correlation: for each one magnitude increase in R, the dEs become redder (in the mean) by 0.062 mag in (B-R). These observations are consistent with a model in which the intracluster gas exerted a confinement pressure (greatest near the cluster core), impeding the outflow of supernovae-driven metal-rich gas from the young dE galaxies. The spatial distribution of the faint dE galaxies is well fit by a standard King model with a central surface density of Sigma0 = 1.424 dEs arcmin-2 and a core radius of Rc = 18.73 arcmin (~= 0.44 Mpc). This core radius is significantly larger than Rc = 12.3 arcmin found for the brightest dEs and giant cluster galaxies, and it is consistent with tidal disruption of faint dEs in the dense cluster core. The composite luminosity function for the Coma cluster galaxies (completeness corrected and background subtracted) was modeled as the sum of a Gaussian distribution for the giant galaxies and a Schechter function for the dE galaxies. In this manner, weighted least-squares fits and subsequent numerical integration allowed us to solve for the relative contributions of the two populations. During this modeling and analysis, the dispersion of the Gaussian distribution was constrained to the observed range sigma = 1.0 ± 0.20 mag, while the faint-end slope of the Schechter function was constrained to alpha = -1.41 ± 0.05, calculated for this luminosity function from a separate analysis. Decomposing the galaxy luminosity function in this manner, I found that the early-type dwarf-to-giant ratio (EDGR) for the Coma cluster core is identical with that of the Virgo cluster. I proposed that the presence of substructure is an important factor in determining the cluster's EDGR, since during the merger of two or more richness-class 1 galaxy clusters, the total number of dwarf and giant galaxies will be conserved (to first order). Thus, this low EDGR result is consistent with the Coma cluster being formed as the merger of multiple, less-rich galaxy clusters.