Studying Galaxy Evolution with Spitzer and Herschel CUP Conference Series, Vol. **VOLUME**, 2006 V. Charmandaris, D. Rigopoulou, N. Kylafis

Evidence for enhanced star formation during structure formation at z ∼ 1 David Elbaz1 , Emanuele Daddi1 , Damien Le Borgne1 ,Mark Dickinson2 , Dave M. Alexander3 , Ranga-Ram Chary4 , Jean-Luc Starck1 , William Nielsen Brandt5 , Manfred Kitzblicher6 , Emily MacDonald2 , Mario Nonino7 , Paola Popesso8 , Dan Stern9 , Eros Vanzella7 1 CEA–Saclay,

AIM-UMR (7158), DAPNIA/Service d’Astrophysique, France 2 National Optical Astronomy Observatory, Tucson, USA 3 Department of Physics, Durham University, Durham, England 4 Spitzer Science Center, California Institute of Technology, USA 5 Department of Astronomy and Astrophysics, The Pennsylvania State University, USA 6 Max-Planck Institute of Astrophysics, Garching, Germany 7 INAF - Osservatorio Astronomico di Trieste, Trieste, Italy 8 European Southern Observatory, Garching, Germany 9 Jet Propulsion Laboratory, California Institute of Technology, USA Abstract. It is known since decades that most massive galaxies in the local Universe are spheroids that inhabit preferentially high density environments, while star formation mainly takes place in field disks. Well established correlations exist locally between galaxy density, colour, morphology and star formation. However, the physical processes responsible for these relations and the epoch at which they were forged remain unclear. Here we show that the star formation-density relation is reversed at a redshift of 1, i.e. when the universe was less than half of its present age, such that the typical level of activity increases with local density. We present evidence that this reversal is connected to the formation of large-scale structures at the galaxy group scale. This is exemplified by a remarkable structure at z=1.007-1.025, containing X-ray traced galaxy concentrations, that will eventually merge into a Virgo-like cluster. These findings were made possible thanks to the ultra-deep 24 µm Great Observatories Origin Deep Survey (GOODS) with Spitzer able to measure star formation rates down to 3 M yr−1 at z ∼ 1.

The morphologies and stellar populations of galaxies are the fossil memory of the complex physical processes that led to their formation. This is exemplified by the global properties of the two broad classes of galaxies in the local Universe. Red, early-type galaxies (those with prominent stellar bulges) have old stars, exhibit little to no star formation, and contain small gas reservoirs, suggesting that they formed their stars early in the distant Universe. By comparison, blue, latetype galaxies (typically, disk-dominated galaxies with prominent spiral arms) are gas rich and harbour on-going star formation. Intriguingly, red galaxies are preferentially more massive and are located in denser environments than blue galaxies, suggesting that the environment has a large affect on the star-formation history of galaxies (Dressler 1980, Kauffmann et al. 2004). The origin of these 1

2

David Elbaz

fundamental correlations between basic galaxy properties is not yet understood. Two opposite interpretations of the poor activity of galaxies found in dense environments in the local Universe have been invoked: star formation was either quenched or triggered in the past by the local galaxy density. An approach to directly probe the origin of the activity-density relation is to map the spatial distribution of distant galaxies and determine their local environment, star formation rate (SFR) and stellar mass. We can perform these measurements reliably at z∼1 (i.e. 7.7 billion years ago) in the Great Observatories Origins Deep Survey (GOODS, Dickinson & Giavalisco 2001). GOODS consists of two ultra-deep multi-wavelength fields spanning 10x16 comoving Mpc at z=1 (for a cosmology of H0 = 70 km s-1 Mpc−1 , Λ= 0.7, Ωm =0.3), each wide enough to adequately sample the characteristic 1 Mpc scale over which galaxy activity is expected to be affected (Blanton 2006). The GOODS fields are widely separated, located in the two hemispheres (individually termed GOODS-N and GOODS-S), to guard against the affects of cosmic variance, and have been the subject of some of the largest, deepest, and most complete spectroscopic campaigns to date. Spectroscopic redshifts have been identifed for 60 and 50 % of the GOODS-N and S galaxies respectively down to zAB =23, i.e. rest-frame BAB = -20.4. The spectroscopic completeness of the GOODS survey provides a highly representative sampling of galaxies as a function of stellar mass, colour and SFR within the range considered in this analysis. A crucial strength of the GOODS fields is their ultradeep coverage at 24 µm with the MIPS instrument onboard the Spitzer Space Telescope (Chary et al. 2004). Heavily obscured star-formation can be detected down to faint limits (∼2-3 M .yr−1 ) using the tight correlation between the rest-frame 12 µm and 8-1000 µm luminosities observed at z∼0 (Chary & Elbaz 2001) which has been shown to remain valid at z∼1 (Appleton et al. 2004). Accounting for the contribution of the unobscured UV luminosity to the SFR using Eq.5 of Daddi et al. (2004) for a Salpeter IMF, allows us to extend the census of star-formation activity down to even fainter limits (∼0.5 M .yr−1 ). We have investigated the dependence of galaxy activity (SFR) with local environmental density in the range z=0.8 to 1.2 (Figure 1). At z∼1, we find a reversal of the local activity-density relation: both fields, GOODS-N (filled circles) and GOODS-S (small filled squares, blue), exhibit a consistent increase of the average SFR of galaxies by a factor 3±0.5 and 5±1.5 respectively with increasing density (Σ=0.4 to ∼5 gal.Mpc−2 ). The probability that the observed increase is obtained by chance is 3 and 7x10−5 for GOODS-N and S respectively. We have verified that the reversal is robust and not due to biases in the spectroscopic sample selection (in magnitude, colour, SFR and stellar mass). This finding implies that star formation is affected by environment at z∼1. This trend is opposite to that which is observed for local galaxies from the 2 degree Field survey (Lewis et al. 2002) and Sloan Digital Sky Survey (Brinchmann et al. 2004), using the same technique and selection (Figure 1). High density environments statistically contain galaxies with larger stellar masses. Our finding implies that these objects were formed with a major role of vigorous in-situ star formation, as opposed to simple mass assembly by merging. A model prediction for galaxies at z=1±0.2 is shown for comparison, from the semi-analytical treatment of baryons in the Millennium numerical simulation (filled green squares, Kitzbichler & White 2006). While the average SFR of the model is consistent with observations at low densities, the model

Evidence for enhanced star formation during structure formation at z ∼ 1

Figure 1. Environmental dependance of SFR at z∼1 and 0. The mean SFR of galaxies is plotted as a function of projected galaxy density, Σ, in boxes of 1.5x1.5 in the Ra,Dec directions (1.5 comoving-Mpc on a side) and ∆v=±1500 km s−1 (the typical width of redshift peaks in GOODS). ∆z is adjusted to keep the same comoving volume of 97 comoving Mpc3 per box as a function of redshift. SFRs in GOODS computed from mid IR and UV uncorrected.

predicts a decrease of SFR with density, instead of the observed increase. This behavior is mostly due to the implementation in the model of the quenching of star formation by AGN feedback required to avoid too much recent star formation in massive galaxies (e.g., Schawinski et al. 2006). At redshifts larger than z∼2, the model predicts a rise of SFR with density similar to the observed z∼1 Universe. This suggests that star formation in massive galaxies is taking place, or has been stopped, too early in this realization of the model. Finally, we also find evidence for a turnover in the increase of the SFR with density at the highest densities, suggesting that there is a typical scale above which the SFR of galaxies is turning down. This is consistent with existing evidence that at z≥1 gravitationally relaxed galaxy clusters (thus at environmental densities larger than probed in GOODS) are dominated by passive ellipticals (Rosati et al. 1999). More generally, this implies a shift of the star formation activity with cosmic time towards lower density environments. This opens a third dimension in the manifestation of the downsizing effect, which states that the characteristic stellar mass of star forming galaxies increases with redshift (Cowie et al. 1996): this increase is linked with large-scale structure formation.

3

4

David Elbaz

A possible explanation for the observed positive correlation of the activity of galaxies with local density at z∼1 is the role of major mergers of galaxies. However, we find that only a third of the luminous infrared galaxies (LIRGs, with L(8-1000 µm)≥ 1011 M yr−1 ) at z ∼ 1 exhibit the morphological signature of a major merger (mass ratio lower than 4), whereas nearly 100 % of them are major mergers at z ∼ 0. The non dominant role of mergers, also argued by Bell et al. (2005) and Hogg et al. (2005), suggests that other mechanisms might play a dominant role such as minor mergers or gas infall associated with bar formation/destruction (Combes et al. 2004). A possible triggering mechanism might also be related to AGN activity (e.g., Silk 2005). More generally, one may wonder whether the reason for galaxies to exhibit higher SFR in denser environments is due to nature or nurture. A purely internal triggering of star formation might in fact also be the cause of the activity-density relation, since denser environments harbor more massive galaxies. However, if the reason for the enhanced activity in dense environments is internal, this raises the question of the origin of the gas fueling this star formation activity at z ∼ 1. References Dressler, A., 1980, AJ 236, 351 Kauffmann, G., White, S.D.M., Heckman, T.M. et al., 2004, MNRAS 353, 713 Dickinson, M., Giavalisco, M., 2001, ESO/USM Workshop ”The Mass of Galaxies at Low and High Redshift”, eds. R. Bender and A. Renzini (astro-ph/020213) Blanton, M.R., 2006, ApJ 648, 268 Chary, R., Casertano, S., Dickinson, M. E., et al., 2004, ApJS 154, 80 Chary, R.R., Elbaz, D., 2001, ApJ 556, 562 Appleton, P.N., Fadda, D.T., Marleau, F.R. et al., 2004, ApJS 154, 147 Daddi, E., Cimatti, A., Renzini, A. et al., 2004, ApJ 617, 746 Lewis, I., Balogh, M., De Propris, R., et al., 2002, MNRAS 334, 673 Brinchmann, J., Charlot, S., White, S.D.M., et al., 2004, MNRAS 351, 1151 Kitzbichler, M.G., White, S.D.M., MNRAS (submitted, astro-ph/0609636) Schawinski, K., Khochfar, S., Kaviraj, S., et al., 2006, Nature 442, 888 Rosati, P., Stanford, S.A., Eisenhardt, P.R. et al., 1999, AJ 118, 76 Cowie, L.L., Songaila, A., Hu, E.M., Cohen, J.G., 1996, AJ 112, 839 Ponman, T.J., Cannon, D.B., Navarro, J.F., The thermal imprint of galaxy formation on X-ray clusters, Nature 397, 135 (1999) Elbaz, D., Cesarsky, C.J., A fossil record of galaxy encounters, Science 300, 270 (2003) Bell, E.F., Papovich, C., Wolf, C. et al., 2005, ApJ 625, 23 Hogg, D.W., 2005, in ”The Fabulous Destiny of Galaxies: Bridging Past and Present”, 2005 June 20-24, Marseille (astro-ph/0512029) Combes, F., in The evolution of starbursts, AIP conference proceedings 43, 783 (2005, astro-ph/0410410) Silk, J., 2005, MNRAS 364, 1337