LoCuSS: THE SLOW QUENCHING OF STAR FORMATION IN CLUSTER GALAXIES AND THE NEED FOR PRE-PROCESSING

Abstract

We present a study of the spatial distribution and kinematics of star-forming galaxies in 30 massive clusters at 0.15<z<0.30, combining wide-field Spitzer 24µm and GALEX NUV imaging with highly-complete spectroscopy of cluster members. The fraction ( f SF ) of star-forming cluster galaxies rises steadily with clustercentric radius, increasing fivefold by 2 r 200 , but remains well below field values even at 3 r 200 . This suppression of star formation at large radii cannot be reproduced by models in which star formation is quenched in infalling field galaxies only once they pass within r 200 of the cluster, but is consistent with some of them being first pre-processed within galaxy groups. Despite the increasing f SF -radius trend, the surface density of starforming galaxies actually declines steadily with radius, falling ∼15× from the core to 2 r 200 . This requires star-formation to survive within recently accreted spirals for 2-3 Gyr to build up the apparent over-density of star-forming galaxies within clusters. The velocity dispersion profile of the star-forming galaxy population shows a sharp peak of 1.44 σ ν at 0.3 r 500 , and is 10-35% higher than that of the inactive cluster members at all cluster-centric radii, while their velocity distribution shows a flat, top-hat profile within r 500 . All of these results are consistent with star-forming cluster galaxies being an infalling population, but one that must also survive ∼0.5-2 Gyr beyond passing within r 200 . By comparing the observed distribution of star-forming galaxies in the stacked caustic diagram with predictions from the Millennium simulation, we obtain a best-fit model in which SFRs decline exponentially on quenching time-scales of 1.73±0.25 Gyr upon accretion into the cluster.

LoCuSS: weak-lensing mass calibration of galaxy clusters

“…On the latter point, the COS-MOS photometry extends to 8µm, whereas we have observed half of the cluster sample discussed in this article with Spitzer/MIPS at 24µm and with Herschel/PACS and SPIRE at 100 − 500µm (Smith et al 2010), albeit only to a depth corresponding to a bolometric infrared luminosity of LIR ≃ 5 × 10 10 L⊙ at z ≃ 0.2. The spectroscopic completeness of follow-up observations with Hectospec (Fabricant et al 2005) is 96 per cent for objects detected with Spitzer down to 0.5mJy and and ∼ 80 per cent down to i = 20 (Mulroy et al 2014;Haines et al 2015). Clearly these data are not sensitive enough to obtain definitive estimates of the number and redshifts of dusty galaxies that satisfy our faint red optical background galaxy selection.…”

“…However they provide a useful sanity check based on bright galaxies. We select galaxies from the catalogues discussed by Haines et al (2015) at i < 20 that have (V − i) colours that would place them in our red background galaxy catalogues if they were faint enough. We find that ∼ 1 per cent of these bright optically red galaxies are LIRGs, and ∼ 1 per cent of the same bright optically red galaxies are cluster members, and there is no overlap between these two populations.…”

Evidence of pre-processing and a dependence on dynamical state for low-mass satellite galaxies

“…Bahé et al 2013). To address this, we also show pre-processing results for our 'strict' infall sample (2 < R < 3 R 200 ) which should be less susceptible to backsplash contamination, as many previous studies have shown that the majority of backsplashing galaxies are found within two virial radii (Mamon et al 2004;Mahajan et al 2011;Oman et al 2013;Haines et al 2015). If contamination from backsplash galaxies is low, this field excess should approximate the fraction of galaxies which have been pre-processed prior to infalling onto their present-day group.…”

“…Therefore the fact that our values for the fraction of pre-processed galaxies are consistently smaller than the quoted subhalo fractions is still consistent. Studying the star-forming fractions of cluster galaxies, Haines et al (2015) use a simple toy model in an attempt to reproduce the trend between cluster-centric radius and star-forming fraction. They find that in order to reproduce the observational trend, a 19 per cent decrease in the starforming fraction of cluster galaxies relative to the field is required on top of star formation quenching occuring within the virial radius.…”

“…In this work we find that the fraction of high-mass (the Haines et al sample consists of galaxy stellar masses > 2 × 10 10.2 M ) pre-processed galaxies for high-mass clusters is at most 7 ± 2 per cent. Therefore, this work is only able to account for a portion of the amount of pre-processing required by the Haines et al (2015) model, although a more complete comparison would require samples matched in halo mass and galaxy stellar mass.…”