A spectroscopic survey of dwarf galaxies in the Coma cluster: stellar populations, environment and downsizing

We investigate the stellar populations in a sample of 89 faint red galaxies in the Coma cluster, using high signal-to-noise ratio (S/N) spectroscopy from the 6.5-m MMT. Our sample is drawn from two 1 degrees fields, one centred on the cluster core and the other located 1 degrees to the south-west of the cluster centre. The target galaxies are mostly 2-4 mag fainter than M*; galaxies with these luminosities have been previously studied only using small samples, or at low S/N. For a comparison sample we use published high-S/N data for red-sequence galaxies in the Shapley supercluster. We use state-of-the-art stellar population models ( by R. Schiavon) to interpret the absorption-line indices and infer the single-burst-equivalent age and metallicity (Fe/H) for each galaxy, as well as the abundances of the light elements Mg, Ca, C and N. The ages of the Coma dwarfs span a wide range from <2 Gyr to >10 Gyr, with a strong environmental dependence. The oldest galaxies are found only in the core, while most of the galaxies in the outer south-west field have ages similar to 3 Gyr. The galaxies have a metallicity range -1.0 less than or similar to [Fe/H] less than or similar to 0.0, and follow the same age-metallicity-mass plane as high-mass galaxies, but with increased intrinsic scatter. The Mg/Fe abundance ratios are on average slightly supersolar, and span a range -0.1 less than or similar to [Mg/Fe] less than or similar to +0.4. The highest Mg enhancements are found only in the cluster core, while solar ratios predominate in the outskirts. We show that parametrized models with more complex star formation histories perform no better than single-burst models in reproducing the observed line indices. Assuming a star formation history dominated by a single burst, the number of dwarf galaxies on the red sequence in the Coma core has doubled since z approximate to 0.7. Assuming instead an abruptly truncated constant star formation rate, the equivalent redshift is z approximate to 0.4. These estimates bracket the red-sequence growth time-scales found by direct studies of distant clusters. In the south-west field, the red sequence was established only at z approximate to 0.2 for a burst-dominated star formation history (z approximate to 0.1 for the truncated case). Our observations confirm previous indications of very recently quenched star formation in this part of the cluster. Our results strongly support the scenario in which much of the cluster passive dwarf population (in this luminosity range) was generated by environment-driven transformation of infalling late-type galaxies.