Herschel Unveils Galaxy Evolution – HUGE

A decade ago, it was thought that the global star formation history of galaxies at cosmological scales was relatively well constrained by the optical and UV data. Soon after, 15 microns observations with ISOCAM, onboard the Infrared Space Observatory (ISO), and sub-millimeter SCUBA results from the ground, brought a revolution in the field by revealing that Luminous and Ultra-luminous infrared galaxies were much more common in the past than they are now. These galaxies are forming stars at a high pace, and most of the energy radiated by the forming stars and also by active galactic nuclei, when present, is absorbed by dust and re-radiated in the IR. It became possible to re-establish the global history of star formation at z'2, and to identify the sources from which the bulk of the Cosmic Infrared Background originates. The next IR Observatory, Spitzer, combined with multiwavelength observations on specific areas of the sky, and with the UV satellite Galex, brought a second revolution in the field: the abundance of active galactic nuclei in distant massive galaxies had been largely underestimated, even when the deepest X-ray surveys were used; galaxies behave in what appears as anti-hierarchical with massive galaxies formed first and then stopping abruptly to form stars, but environment effects and AGNs thought to quench star formation, either accelerate or are coeval with star formation in galaxies. At each of these stages of understanding of galaxy evolution, members of the team in this ANR proposal played a leading role at the world level. But mostly it was shown by us and others that the far infrared luminosity, hence the star formation rate (SFR), of distant galaxies (z>1) cannot be safely extrapolated from data at other wavelengths: the correction for extinction of the optical-UV regime saturates at high SFR; z~2 SCUBA galaxies exhibit unique spectral energy distributions in the IR with e.g. colder dust temperatures than extrapolated from local galaxies; the total IR luminosity of z'1.5 galaxies extrapolated from the mid IR is overestimated by increasingly large factors with increasing luminosity, partly due to the presence of Compton Thick AGNs. Thus, many of the key questions on galaxy formation/evolution, such as the global star formation rate or the frequency of AGNs in galaxy cores, and their link with star formation/quenching, remain open. When the Herschel satellite, to be launched on April 12, 2009, will produce the deepest images of the sky in the far infrared, it will become possible, for the first time, to address these unsolved questions, by measuring the bolometric luminosity of distant galaxies, as produced not only by young and massive stars, but also by accreting super massive black holes. An international team (PI David Elbaz; this is the only Herschel Open Time Key Program with a French PI) has obtained considerable Herschel time to carry out crucial observations in a region of the sky where data at other wavelengths is already available. The French part of this team, author of this ANR proposal, will profit from the know how of the colleagues, but also suffer from the competition with the rest of the team, especially considering that the US part of the team will be well funded to carry out this project. To enable the French team to extract the key observational information from Herschel and to use/design the modeling tools necessary for the interpretation, it is mandatory that, in France too, young researchers (postdocs from the ANR and PhD students from other funding sources). The main scientific objectives of this ANR project rely on the GOODS-Herschel Open Time Key Program that we are leading: (A) To resolve most of the cosmic SFR density up to z~4, thanks to the ~2000 galaxies that will be detected in GOODS-Herschel in the unexplored regimes of normal galaxies up to z~1, luminous and ultra-luminous infrared galaxies up to z~2 and 4 respectively. This will be done by bridging IR and UV selected galaxies down to the level where both SFR agree up to z~1.5 and potentially up to z~4. Hence the need to use/design models dealing with UV-optical-IR emission of galaxies. (B) To identify the buried Compton Thick AGNs responsible for the still unresolved 30% fraction of the cosmic X-ray background (CXB), which peaks at 30 keV, hence above the 10 keV limit of XMM-Newton and Chandra, and study their link with star formation or its quenching in distant massive galaxies. This will be done by searching for mid over far infrared excess galaxies using and extending a technique that we previously validated on a statistical scale (stacking). (C) To identify the sources making the bulk of the extragalactic infrared background, second in intensity after the CMB and which contains the majority ofthe energy radiated by star formation over the Hubble time.