The galaxy population of the z = 1 cluster of galaxies MG2016+112

S. Toft 1 G. Soucail 0 J. Hjorth 1 0 Observatoire Midi-Pyrenees, Laboratoire d'Astrophysique , UMR 5572, 14 avenue E. Belin, F-31400 Toulouse, France 1 Astronomical Observatory, University of Copenhagen , Juliane Maries Vej 30, DK-2100 Copenhagen , Denmark a characteristic magnitude Ks = 18.90+00..4557 and faint-end slope = 0.60+00..3393, consistent with what is expected for a passively evolving population of galaxies formed at high redshift, zf > 2. From the total Ks-band flux of the cluster galaxies and a dynamical estimate of the total mass of the cluster, the rest-frame Ks-band mass-to-light ratio of the cluster is derived to be M /LKs = 27+6147 h50(M /LKs) , in agreement with the upper limit derived earlier from Chandra X-ray observations and the value derived locally in the Coma cluster. The cluster galaxies span a red sequence with a considerable scatter in the colour-magnitude diagrams, suggesting that they contain young stellar populations in addition to the old populations of main-sequence stars that dominate the Ks-band luminosity function. This is in agreement with spectroscopic observations which show that 5 out of the 6 galaxies in the field confirmed to be at the redshift of the lensing galaxy have emission lines. The projected spatial distribution of the cluster galaxies is filamentary-like rather than centrally concentrated around the lensing galaxy, and shows no apparent luminosity segregation. A handful of the cluster galaxies show evidence of merging/interaction. The results presented in this paper suggest that a young cluster of galaxies is assembling around MG2016+112. 1.1 Galaxy evolution in clusters The central parts of clusters of galaxies are dominated by a population of old elliptical galaxies which are observed to follow a tight red sequence in colourmagnitude diagrams, both in the local Universe (e.g. Bower, Lucey & Ellis 1992) and at intermediate (e.g. Stanford, Eisenhardt & Dickinson 1998) and high redshifts (Stanford et al. 1997; Rosati et al. 1999; van Dokkum et al. 2001). The small scatter of the relationship taken together with the observed Mg2 relationship [which suggests a close correlation between galaxy mass and metallicity: see Bender, Burstein & Faber (1993)] and the evolution of its slope and zero-point to the highest observed redshifts are well described by the monolithic burst formation scenario (Eggen, Lynden-Bell & Sandage 1962). In this scenario the majority of the stars in the galaxies are formed in a single starburst at high redshift (zf > 2), followed by passive evolution with little star formation activity. There is, however, accumulating evidence that the evolution of cluster galaxies is not as simple as suggested by the monolithic formation scenario. The increasing observed fractions of blue galaxies (Butcher & Oemler 1978; Rakos & Schombert 1995), merging galaxies (e.g. van Dokkum et al. 2000) and galaxies with emission lines and post-starburst features (Dressler et al. 1999) with redshift favour a hierarchical formation scenario in which massive cluster galaxies are built up from continuous merging of smaller galaxies (Kauffmann & Charlot 1998a). This picture is also supported by the decreasing number of S0 galaxies in clusters with increasing redshift (Dressler et al. 1997). The red sequence is expected to increase its scatter and eventually fall apart as the redshift approaches the formation redshift of the stars in the galaxies. In monolithic models the increasing scatter with redshift should not depend on magnitude, since all the galaxies form at the same time and have similar star formation histories, while in hierarchical models the scatter is predicted to be larger for the brighter galaxies since these form later than the fainter galaxies (Ferreras & Silk 2000). The observed tight red sequence in clusters, even at high redshifts, is consistent with the hierarchical formation scenario if the bulk of the stellar populations was formed at high redshift in the progenitors of the cluster galaxies and additional star formation in the subsequent merging process was suppressed. In support of this scenario, Hubble Space Telescope (HST) imaging of high-redshift cluster galaxies has revealed a population of luminous red mergers in high-redshift clusters (van Dokkum et al. 2000, 2001) which follow the same colourmagnitude relation as the cluster ellipticals but with a slightly larger scatter. Another potential way to distinguish between the two scenarios is by studying the evolution of the mass function of the cluster galaxies. Hierarchical galaxy formation models predict the mass assembly of galaxies to take place over a long time-scale (z 2) and to be quite decoupled from star formation (Kauffmann & Charlot 1998b). If this picture is true, a strong evolution of the stellar mass function with redshift is expected. This evolution is best studied in the Ks band because it is relatively unaffected by ongoing/recent star formation and hence directly measures the growth of galaxy mass (Kauffmann & Charlot 1998b). Locally, the cluster galaxy luminosity function has been studied in great detail at optical and near-infrared (NIR) wavelengths. Goto et al. (2002a) derive the composite u-, g-, r -, i - and zband luminosity function of 204 clusters of galaxies from the Sloan Digital Sky Survey Cut & Enhance Galaxy Cluster Catalog (Goto et al. 2002b) in the redshift range z = 0.020.25, and find that the faint-end slope of the luminosity function becomes flatter toward the redder wavebands, consistent with the hypothesis that the cluster luminosity function has two distinct underlying populations: a population of bright ellipticals with a Gaussian-like luminosity distribution that dominate the bright end, and a population of faint blue star-forming galaxies with a steep power-law-like luminosity distribution that dominate the faint end. de Propris et al. (1998) derive the NIR (H-band) luminosity function of the Coma cluster galaxies, and find results consistent with the above picture of a population of bright red galaxies and large population of faint blue dwarf galaxies. The bright (massive) end of the luminosity function is expected to steepen and shift to fainter magnitudes at high redshift, as the old bright cluster galaxies begin to break up into building blocks. In contrast to the bright cluster galaxies, the faint cluster galaxies seem to have a much greater diversity of star formation histories (Kodama & Bower 2001). The ButcherOemler effect at intermediate redshift is an example of this. If this progression of activity to lower-mass galaxies as the Universe ages is a consequence of the faint galaxies forming at later cosmic times than their massive counterparts, then distant clusters should have a luminosity function with a declining faint-end slope in contrast to the rising faint-end slope of local clusters (de Propris et al. 1998). The evolution of the K-band luminosity function in the redshift range z = 0.1 (...truncated)