Jerusalem 2004 Hans-Walter Rix - MPIA The Evolution of the High-z Galaxy Populations

Jerusalem 2004 Hans-Walter Rix - MPIA The Evolution of the Cosmic Star Formation Rate When did it start / ramp up? When were half of the stars formed? Are we at the end of star-formation? Estimating the “cosmic star-formation rate” –Estimating the SFRs in individual objects –Are all relevant sources included Faint Obscured

Jerusalem 2004 Hans-Walter Rix - MPIA First ( ) estimates of the cosmic star-formation history Warning: historical plots. Do not use for research!! after Madau et al 1996  =1!! Lilly et al 1995 HDF (too “empty”) Steidel et al The impact of LSS!

Jerusalem 2004 Hans-Walter Rix - MPIA A current “UV-based” version Giavalisco et al 2003 (GOODS)  =0.3  =0.7 H 0 =70 from measured UV density after dust correction

Jerusalem 2004 Hans-Walter Rix - MPIA And from the perspective of sub-mm (thermal dust emission) Barger et al 2000 thermal IR observed in sources Completeness corrected x 11 (undetected sources)  =1 Dust corrected UV (Giavalisco et al 03)

Jerusalem 2004 Hans-Walter Rix - MPIA The Evolution of Intense Starbursts SFR from ULIRGs has dropped by > 100 since ~3! NB: many high-z QSO’s also show enormous thermal dust emission  phases of intense SFR Detected sources are likely to have luminosities ~ 5x Arp 220 SFR from ULIRGS now

Jerusalem 2004 Hans-Walter Rix - MPIA Cosmic Star-Formation History: Upshot has risen by 5-7 x from z~0 to z~1.2 UV based estimates (after correction by 10) and sub-mm/thermal IR (after different correction of 10) give consistent estimates 2<z<5 1.5 approximately constant (~2) (Note: there is some evidence for a drop beyond 5) ULIRG (>few 100M sun /yr) mode of star- formation has dropped by >100 since z~3

Jerusalem 2004 Hans-Walter Rix - MPIA Brief, but important aside Galaxies with low SFR exist at 1<z<3.5 they can be found in IR-selected samples They seem to make up ½ of the stellar mass (see below) at z~2-3 FIRES: van Dokkum et al 2003 with Spectra

Jerusalem 2004 Hans-Walter Rix - MPIA The Build-Up of the Stellar Mass Step 1 The mean color of galaxies as function of redshift (Rudnick et al 2003): –Optical colors   M/L B as long as star-formation history is not too “bursty” (Bell and de Jong 2001) –Individual galaxies may have bursts, but an ensemble of galaxies at a given epoch (say  z~0.5) should not have their bursts all at the same time –  Look at the mean color of the galaxies as a function of

Jerusalem 2004 Hans-Walter Rix - MPIA Mean Color of the Galaxy Population as a Function of Redshift On average, galaxies were much bluer in the past was 10 x lower at z~3 K-band selected sample ( rest >5000A) x10 in M/L! SFR ~ e -t/7Gyr since z~4-5  redshift

Jerusalem 2004 Hans-Walter Rix - MPIA The Build-Up of Stellar Mass Rudnick et al 2003 Take IR-selected sample Multiply j V with V to get ½ of stars since z~1.4 50% between 2>z>1 10% before z~3 Now Big Bang Integrated SFR estimates (e.g. Giavalisco 03)

Jerusalem 2004 Hans-Walter Rix - MPIA Galaxy Clustering and its Evolution Generic prediction of hierarchical CDM models: –More massive halos (=more luminous galaxies?) are more strongly clustered –(Luminous) Galaxies at early epochs are increasingly “biased”  their clustering remains high Present epoch: redder galaxies are more clustered than blue ones –Has that always been true?

Jerusalem 2004 Hans-Walter Rix - MPIA Example: Red Galaxies in the Hubble Deep Field

Jerusalem 2004 Hans-Walter Rix - MPIA SUBARU Deep Field Ouchi et al 2002

Jerusalem 2004 Hans-Walter Rix - MPIA Ly-break galaxies are observed to be clustered (e.g. Giavalisco et al 1998; Ouchi et al 2003) Z~4 galaxies in the Subaru deep field

Jerusalem 2004 Hans-Walter Rix - MPIA S.A.M prediction Kauffmann et al 99 Baugh et al 99 Correlation Length Bias Ouchi et al 2003

Jerusalem 2004 Hans-Walter Rix - MPIA How are galaxies and the IGM (=Ly-a forest) correlated? Search for galaxies near the line of sight to a distant QSO Is there an overabundance of neutral gas in the vicinity of galaxies at z~3? Yes  Galaxies lived in large-scale overdense regions  more HI Distance LBG to Ly-a Absorption ??

Jerusalem 2004 Hans-Walter Rix - MPIA Internal Structure of Early Galaxies (Astro-)folklore: –High-z galaxies are compact –There are no big disks at high-z Basis of that lore: –Observations (see below) –Theoretical expectation: Halos that collapse early are denser Spin parameter  is universal Size of galaxies reflects angular momentum of halo  baryons have less ang. Mom.  small galaxies

Jerusalem 2004 Hans-Walter Rix - MPIA Are there large (disk?) galaxies at high redshift? (Labbe et al 2003; see also Lowenthal et al 1997) M81 At the present, “normal” disk galaxies look completely different in the UV than in the optical Z spec =2.9 “peculiar”, or star- forming ring seen in the UV Older / redder bulge / bar?

Jerusalem 2004 Hans-Walter Rix - MPIA Several examples in the HDF South Can’t prove that these objects are disks, but they sure look like local examples However, they are likely to be single entities that are not grossly disturbed

Jerusalem 2004 Hans-Walter Rix - MPIA Size Evolution from GOODS (Ferguson et al 2003)  UV size at a given UV luminosity was smaller in the past Increasing redshift  Apparent size [“] Rest-frame 1500A Size Constant size

Jerusalem 2004 Hans-Walter Rix - MPIA Can one map the evolution of size at a given (stellar) mass? (FIRES; Trujillo et al 2003) What is the empirical null hypothesis? Size  LuminositySize  Mass (stellar) SDSS z~0 Shen et al 2003

Jerusalem 2004 Hans-Walter Rix - MPIA What if the SDSS Galaxies were observed at 1<z<3 in FIRES? Simulated distribution of sizes, assuming SDSS size distribution holds high-z same stellar mass as observed in FIRES all observational effects

Jerusalem 2004 Hans-Walter Rix - MPIA What if the SDSS Galaxies were observed at 1<z<3 in FIRES? Actually observed in FIRES Note: rest-frame V-band sizes at all z’s

Jerusalem 2004 Hans-Walter Rix - MPIA Describing (the lack of) size evolution with redshift Let’s assume that the SDSS distribution holds in its functional form at all redshifts, but that r e (M * )~(1+z)  Fit a from the FIRES data… At a given stellar mass, galaxies have the same size at all (observed) redshifts! R e (z | L v ) R e (z | M * )

Jerusalem 2004 Hans-Walter Rix - MPIA Summary Different approaches to dM * /dt(z) and M * (z) give (surprisingly?) consistent answers, at the ~2 x level. –We live at the end of star/galaxy formation –The epoch where most, say 80%, of the stars form has been identified: 3>z>0.5 –No clear evidence for the “onset” of SF, yet. Galaxies at high-z are observed to be highly clustered –They are “rarer” objects  bias Galaxy sizes do not show the expected imprint of “formation epoch” –R e (M * )~ const. with z