Evolution in the structural properties of early-type Brightest Cluster Galaxies at small lookback time and dependence on the environment

TL;DR

This paper studies the evolution of early-type Brightest Cluster Galaxies (BCGs) over the last 3 billion years, revealing size and velocity dispersion changes, and suggests growth mainly through dry minor mergers rather than major mergers.

Contribution

It provides new insights into the size and velocity dispersion evolution of BCGs and supports the dry minor merger scenario for their growth, contrasting with previous major merger models.

Findings

BCGs are larger and have smaller velocity dispersions at lower redshifts.

Size evolution of BCGs follows (1+z)^{0.85(Mr+21)} at z~0.25.

BCGs are about 1 Gyr older and 0.5 mag more luminous than satellites.

Abstract

At the present time, SDSS early-type BCGs have larger Re than early-type galaxies of similar L, whether these other objects are in the field, or are satellites in clusters (Re ~ L for BCGs). At fixed M* and formation time, BCGs at lower z are larger and have smaller velocity dispersions, i.e. Re increases and sigma decreases with age. As a result, at z~0.25, corresponding to a lookback time of order 3 Gyrs, BCGs are smaller than their lower z counterparts by as much as 70% for the brightest BCGs: Re evolves as (1+z)^{0.85(Mr+21)}. Qualitatively similar but weaker evolution in the sizes is also seen in the bulk of the early-type population: at Mr<-22 Re evolves as (1+z)^{0.7(Mr+21)}, while at Mr>-22 the evolution is approximately (1+z)^{-0.7}, independent of Mr. The sigma-L correlation also evolves: (1+z)^{-0.2(Mr+21)} at Mr < -22 (as for the BCGs) and (1+z)^{0.2} for fainter galaxies. The Re- and sigma-M* correlations yield consistent results. These trends are most easily understood if early-type BCGs grew from many dry minor mergers rather than a few major mergers. Only in such a scenario can BCGs be the descendents of the superdense galaxies seen at z~2; major dry mergers, which increase the size in proportion to the mass, cannot bring these galaxies onto the BCG Re-M* relation at z~0. We also compared the ages and sizes of our early-type BCGs with satellites. BCGs are larger than satellites of similar L or M* at the same redshift. Although both satellites and BCGs trace the same weak age-L or age-M* relation, this can be understood by noting that BCGs are typically about 1 Gyr older than the satellites in their group, and they are about 0.5 mags more luminous. Finally, we find that the mean satellite L is approximately independent of BCG L, in agreement with recent predictions based on the L-dependence of clustering.