Abstract

We analyse Spitzer Multiband Imaging Photometer (MIPS) 24-mu m observations, and Sloan Digital Sky Survey Data Release 7 optical broad-band photometry and spectra, to investigate the star formation (SF) properties of galaxies residing in the Coma supercluster region. We find that SF in dwarf galaxies is quenched only in the high-density environment at the centre of clusters and groups, but that passively evolving massive galaxies are found in all environments, indicating that massive galaxies can become passive via internal processes. The SF-density relation observed for the massive galaxies is weaker relative to the dwarfs, but both show a trend for the fraction of star-forming galaxies (f(SF)) declining to similar to 0 in the cluster cores. We find that active galactic nucleus activity is also suppressed among massive galaxies residing in the cluster cores. We present evidence for a strong dependence of the mechanism(s) responsible for quenching SF in dwarf galaxies on the cluster potential, resulting in two distinct evolutionary pathways. First, we find a significant increase (at the 3 Sigma level) in the mean equivalent width of H alpha emission among star-forming dwarf galaxies in the infall regions of the Coma cluster and the core of Abell 1367 with respect to the overall supercluster population, indicative of the infalling dwarf galaxies undergoing a starburst phase. We identify these starburst galaxies as the precursors of the post-starburst k + A galaxies. Extending the survey of k + A galaxies over the whole supercluster region, we find 11.4 per cent of all dwarf (z mag > 15) galaxies in the Coma cluster and 4.8 per cent in the Abell 1367 have post-starburst like spectra, while this fraction is just 2.1 per cent when averaged over the entire supercluster region (excluding the clusters). This points to a cluster-specific evolutionary process in which infalling dwarf galaxies undergo a starburst and subsequent rapid quenching due to their passage through the dense intracluster medium. In galaxy groups, the SF in infalling dwarf galaxies is instead slowly quenched due to the reduced efficiency of ram-pressure stripping. We show that in the central similar to 2 h-1(70) Mpc of the Coma cluster, the (24 - z) near-infrared/mid-infrared (MIR) colour of galaxies is correlated with their optical (g - r) colour and H alpha emission, separating all MIR-detected galaxies into two distinct classes of 'red' and 'blue'. By analysing the spatial and velocity distribution of galaxies detected at 24 mu m in Coma, we find that the (optically) red 24-mu m detected galaxies follow the general distribution of 'all' the spectroscopic members, but their (optically) blue counterparts (i) are almost completely absent in the central similar to 0.5 h-1(70) Mpc of Coma and (ii) have a remarkable peak in their velocity distribution, corresponding to the mean radial velocity of the galaxy group NGC 4839, suggesting that a significant fraction of the 'blue' MIR galaxies are currently on their first infall towards the cluster. The implications of adopting different SF-rate tracers for quantifying evolutionary trends like the Butcher-Oemler effect are also discussed.