We present the deepest far-IR observations obtained with Herschel and examine the 3-500um SEDs of galaxies at 0<z<2.5, supplemented by a local reference sample from IRAS, ISO, Spitzer and AKARI data. We find that the ratio of total IR luminosity to r
est-frame 8um luminosity, IR8 (=Lir/L8), follows a Gaussian distribution centered on IR8=4 and defines an IR main sequence (MS). A minority population (<20 %) of outliers producing a tail skewed toward higher values of IR8 consist of starbursts (SB) with compact projected star formation densities. IR8 can be used to separate galaxies with normal and extended modes of star formation from compact SBs with high-IR8, high projected IR surface brightness (>3x10^10 Lsun kpc^-2) and a high specific SFR (i.e., SBs). The rest-frame, UV-2700A size of these distant SBs is typically half that of MS galaxies, supporting the correlation between star formation density and SB activity that is measured for the local sample. Locally, (U)LIRGs are systematically in the SB mode, whereas most distant (U)LIRGs form stars in the normal MS mode. This confusion between two modes of star formation is the cause of the so-called mid-IR excess population of galaxies found at z>1.5 by previous studies. MS galaxies have strong PAH emission line features, a broad far-IR bump resulting from a combination of dust temperatures (Tdust~15-50 K), and an effective Tdust~31 K, as derived from the peak wavelength of their IR SED. Galaxies in the SB regime instead exhibit weak PAH EW and a sharper far-IR bump with an effective Tdust~40 K. Finally, we present evidence that the mid-to-far IR emission of X-ray AGNs is predominantly produced by star formation and that candidate dusty AGNs with a power-law emission in the mid-IR systematically occur in compact, dusty SBs. After correcting for the effect of SBs on IR8, we identify new candidates for extremely obscured AGNs.
The Herschel Space Observatory enables us to accurately measure the bolometric output of starburst galaxies and active galactic nuclei (AGN) by directly sampling the peak of their far-infrared (IR) emission. Here we examine whether the spectral energ
y distribution (SED) and dust temperature of galaxies have strongly evolved since z~2.5. We use Herschel deep extragalactic surveys from 100 to 500um to compute total IR luminosities in galaxies down to the faintest levels, using PACS and SPIRE in the GOODS-North field (PEP and HerMES key programs). We show that measurements in the SPIRE bands can be used below the statistical confusion limit if information at higher spatial resolution is used to identify isolated galaxies whose flux is not boosted by bright neighbors. Below z~1.5, mid-IR extrapolations are correct for star-forming galaxies with a dispersion of only 40% (0.15dex), therefore similar to z~0 galaxies. This narrow distribution is puzzling when considering the range of physical processes that could have affected the SED of these galaxies. Extrapolations from only one of the 160um, 250um or 350um bands alone tend to overestimate the total IR luminosity. This may be explained by the lack of far-IR constraints around and above ~150um (rest-frame) on local templates. We also note that the dust temperature of luminous IR galaxies around z~1 is mildly colder by 10-15% than their local analogs and up to 20% for ULIRGs at z~1.6. Above z=1.5, distant galaxies are found to exhibit a substantially larger mid- over far-IR ratio, which could either result from stronger broad emission lines or warm dust continuum heated by a hidden AGN. Two thirds of the AGNs identified in the field with a measured redshift exhibit the same behavior as purely star-forming galaxies. Hence a large fraction of AGNs harbor star formation at very high SFR and in conditions similar to purely star-forming galaxies.
We discuss observational evidence that quasars play a key role in the formation of galaxies starting from the detailed study of the quasar HE0450-2958 and extending the discussion to a series of converging evidence that radio jets may trigger galaxy
formation. The direct detection with VISIR at the ESO-VLT of the 7 kpc distant companion galaxy of HE0450-2958 allows us to spatially separate the sites of quasar and star formation activity in this composite system made of two ultra-luminous infrared galaxies (ULIRGs). No host galaxy has yet been detected for this quasar, but the companion galaxy stellar mass would bring HE0450-2958 in the local M(BH)-M(stellar bulge) relation if it were to merge with the QSO. This is bound to happen because of their close distance (7 kpc) and small relative velocity (~60-200 km/s). We conclude that we may be witnessing the building of the M(BH)-M(stellar bulge) relation, or at least of a major event in that process. The star formation rate (~340 Msun/yr), age (40-200 Myr) and stellar mass ([5-6]x10^10 Msun) are consistent with jet-induced formation of the companion galaxy. We suggest that HE0450-2958 may be fueled in fresh material by cold gas accretion from intergalactic filaments. We map the projected galaxy density surrounding the QSO as a potential tracer of intergalactic filaments and discuss a putative detection. Comparison to other systems suggests that inside-out formation of quasar host galaxies and jet-induced galaxy formation may be a common process. Two tests are proposed for this new paradigm: (1) the detection of offset molecular gas or dust emission with respect to the position of distant QSOs, (2) the delayed formation of host galaxies as a result of QSO activity, hence the two step building of the M(BH)/M(stellar bulge) ratio.
