CCD Observations of Clusters of Galaxies

Citation

Schneider, Donald P. (1982) CCD Observations of Clusters of Galaxies. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/tsbt-dz83. https://resolver.caltech.edu/CaltechETD:etd-09172008-091944

Abstract

A sample of 84 Abell clusters has been investigated to determine photometric and metric properties of brightest cluster galaxies as a function of cluster richness. The clusters are distributed fairly uniformly in Abell richness class. Seventy-five new measurements of cluster redshifts are presented (previously published redshifts were used for the nine other clusters). The selection criteria resulted in a strong redshift-richness correlation in the sample. Poor clusters (richness classes 0 and 1) have redshifts ~0.09, while the richest clusters have redshifts roughly as twice as large.

Direct imaging of the core of each cluster was affected using a CCD area photometer. The wavelength response of the CCD is so different from that of previous photometric devices that a new photometric system is required. The reductions employ photon based k-corrections (instead of energy based ones) and galactic absorption determined by the neutral hydrogen column density.

The luminosity of the brightest cluster galaxy (G1) within an aperture of 16 kpc radius is shown to be a good standard candle. Previous aperture measurements of 135 first-ranked cluster galaxies are placed on the CCD photometric system. Nearly 200 Gl's with redshifts less than 0.3 have been measured, and they effectively determine the luminosity of giant ellipticals at the present epoch. Their aperture luminosity dispersion is 0.34 mag, which can be reduced to slightly under 0.3 mag by removing richness and morphological trends. The richness correction is reasonably well established at 0.10 magnitudes per Abell richness class, with the rich clusters having brighter galaxies. The trend of luminosity with Bautz-Morgan type matches those of previous investigations, ~0.12 mag per subclass, with BM I clusters having brightest cluster galaxies which are 0.3 mag brighter than the average Gl.

Aperture magnitudes were also determined for the second and third ranked cluster galaxies (G2 and G3), which are defined as the next two brightest galaxies within 250 kpc of Gl. The limiting radius was adopted because of the relatively small area covered by the detector. The dispersion in the aperture luminosity for G2 is 0.55 mag; for G3 the dispersion is 0.65 mag. On average G2 is 0.8 mag and G3 1.3 mag fainter than Gl. The luminosity dependence of G2 and G3 on Abell richness class is roughly the same as that for Gl, except for the richest clusters where G2 and G3 are much brighter than expected. There is no significant BM-luminosity correlation for either G2 or G3 in this sample. The observed colors for the brightest three cluster galaxies indicate that no color evolution has taken place since redshifts of 0.25 (~four billion years).

The radius surface brightness profiles inside 16 kpc for Gl, G2, and G3 are fit fairly well by either a de Vaucouleurs model or a modified Hubble law. At 16 kpc the surface brightness for G1 falls off like a power law with an index of -1.6 to -1.8. For first-ranked galaxies the mean effective radius is 28 kpc and the mean core radius is 2.1 kpc. These scale lengths are three and five times the values for G2 and G3, respectively. The strong correlation of Gl's structure with its absolute magnitude and with cluster morphology are confirmed. The average aperture correction factor (α) for first-ranked cluster galaxies is 0.7; this reduces the sensitivity of the Hubble diagram to q_o by 35%.

Nearly half of the brightest cluster galaxies have multiple nuclei, roughly five times the number expected from projection effects. The multiple systems are, on average, ~0.13 magnitudes brighter than the single systems. An evolutionary correction to q_o of ~+1.5 is required if the multiple systems are interpreted as mergers induced by dynamical friction. The merger process, however, can be calibrated from the α-luminosity relations; this allows corrections to be applied to each galaxy individually.

There is a strong effective radius-surface brightness relation for brightest cluster galaxies. The surface brightness at the effective radius I(R_e), determines the effective radius (R_e) to ~25%. The observed effective radii (determined from fits to the inner 16 kpc) range from ~4 kpc to over 100 kpc. The outer regions (> 30 kpc) of galaxies with extended envelopes do not match the de Vaucouleurs profile found by fitting the inner regions. An angular diameter test based on the effective radii is impractical due to the large intrinsic scatter; a test using the surface brightness corrected effective radii conveys the same information as the standard redshift-magnitude test. The effective radius-surface brightness relation explains the small dispersion in the aperture magnitudes of Gl, and predicts that the total luminosities of brightest cluster galaxies grow as the 0.7 power of the scale length. If the mass-to-light ratio in ellipticals is constant, the luminosity-scale length correlation is incompatible with the Faber-Jackson relation. Application of the R_e-I(R_e) relation to the brightest galaxies in Virgo (NGC 4472 and NGC 4486) yields a null result (no infall) for the distortion of the local Hubble flow. Infall velocities of 250 km s^-1 are excluded at the 2σ level. The second and third ranked galaxies follow a similar R_e-I(R_e) relation. The exponent in the luminosity scale length relation for G2 is about 10% smaller than that for Gl. For G3 the luminosity increases as the square root of the scale length.

Data of sufficient quality to allow construction of luminosity functions were obtained for 60 of the 84 clusters. The limited size of the field required that the luminosity functions be determined inside a given metric radius (250 kpc). The observed luminosity functions were fit to Schechter functions using maximum likelihood techniques. The brightest cluster galaxy cannot be drawn from a universal luminosity function. It is impossible to reconcile the small total luminosity-richness correlation with the relatively large (0.6 mag) dispersion in their total luminosities. The first- ranked galaxies are also about one magnitude too bright to be drawn from a Schechter function. Excluding Gl from the luminosity function results in satisfactory Schechter function fits to the rest of the cluster members. The power-law slope at low luminosities is ~-1, but is not well determined. The observations find a mean M_* in close agreement with other investigations. The observed dispersion of M_* about the mean as a function of cluster richness is similar to that predicted from numerical simulations.

Cluster richness is defined as the total luminosity found by integrating over all luminosities the best fitting Schechter function determined from galaxies within 250 kpc of the brightest cluster galaxy. This definition correlates well but not perfectly with Abell richness class. Richness (actually central density) varies by nearly a factor of 40 from the poorest to the richest clusters in this sample,and in several poor clusters the brightest galaxy outshines the rest of the core. The total luminosity of G1 is weakly correlated with richness (at the same level as with Abell richness class). The luminosities of G2 and G3, however, exhibit a strong positive relationship with cluster richness.

Surface photometry of ~2000 cluster members indicates that they may form the basis for a very powerful angular diameter (or luminosity) test for the deceleration parameter, but uncertainties in the seeing corrections and object selection effects must first be resolved.

The evidence for dynamical evolution, while admittedly circumstantial, is nevertheless persuasive. The strong structure-luminosity relation and the frequency of multiple systems are strong arguments in favor of galactic cannibalism. A detailed spectroscopic and photometric study of a brightest cluster galaxy composed of nine nuclei, V Zw 311, indicates that dynamical friction can radically alter a galaxy in a time scale of only a billion years. The lack of strong luminosity-richness correlation is the most often advanced objection to the merger picture, but dynamical studies of rich systems are required before their capture rates can be calculated.

Thesis Files