Observing the Assembly of Galaxies HansWalter Rix MaxPlanckInstitute

Observing the Assembly of Galaxies Hans-Walter Rix Max-Planck-Institute for Astronomy Heidelberg HWR Princeton, 2005

Overview I. The Build-Up of the Stellar Mass in Galaxies II. The Formation and Evolution of Massive Galaxies Thursday May 5, 2: 00 PM III. The Evolution of (Internal) Galaxy Structure Wednesday May 11, 2: 00 PM IV. Archeo-Cosmology in the Local Group Friday, May 13, 2: 00 PM HWR Princeton, 2005

I. The Build-Up of Stellar Mass 1. Casting the problem into specific questions 2. Diagnostic Tools 3. A brief survey of surveys 4. Estimating the star-formation rate = f(z) 5. Estimating the stellar mass density = f(z) 6. Results HWR Princeton, 2005

1. Re-phrasing “the build-up of stellar mass” • What is <SFR(z)> and <r*(z) >? • What epoch encloses the formation of most stars? • How to best measure <SFR (z) > and <r*(z) > ? • How much important are mergers in triggering SF and in setting the present-day mass function? • What are the expectations from models? HWR Princeton, 2005

2. Diagnostic tools for star-formation rates and stellar masses • Star formation rate estimates are based on UV luminosity produced by hot, massive, short-lived stars – Observe the UV – Observe Ha – Observe absorbed UV flux, reradiated by dust in thermal IR ! LIR(re-radiated) >> LUV(escaped) ! – Mtot estimate is based on stars >10 Mo, which are small fraction of Mtot HWR Princeton, 2005 Kroupa 2002

Starlight and Re-processed Starlight Single-age, dust-free stellar population Devriend et al 2000 HWR Princeton, 2005

ft shi Red ground Herschel (2007) Spitzer SED of an ageing stellar population of solar metalicity with dust HWR Princeton, 2005

f(24 mm) vs Lbol Papovich and Bell 2003 Given that Spitzer can only observe well at 24 mm, what are the bolometric corrections? HWR Princeton, 2005

Mass measurements in cosmologically distant galaxies • Dynamics: – OK to z~1, but quite expensive. – Very limited spatial resolution conceptually problematic – Currently not feasible for most galaxies z>1. 5 • Clustering: – Measures halo mass, not stellar mass • M* = L x (M/L)* with M/L from SEDs HWR Princeton, 2005

Stellar Masses from Spectral Energy Distributions Near-degeneracy of age, metallicity and dust Source of despair or opportunity? tstars = [Gyrs] B K Bell and de Jong 2001 Optical/near-IR spectra of galaxies are a nearly 1 D sequence HWR Princeton, 2005

• Mapping one or few integrated galaxy colors to – age – dust extinction – metallicity is poor! • Mapping (optical -- across 4000 A break) color to M/L should be robust! HWR Princeton, 2005

M/L from Colors? Compare to Mdyn! Van der Wel, Franx, can Dokkum and Rix, 2004 at z~1 HWR Princeton, 2005

Look-back Galaxy Surveys: Desiderata • • Select SFR surveys by SFR, and mass surveys by stellar mass – SFR: assure most of the intense star-burst are not missing due to dust – Stellar mass: select galaxies lobs > (1+z) 4000 A break Number of galaxies as a function of – – Epoch redshift (few %) Luminosity/stellar mass – Color/stellar age 1, 000 – 10, 000 galaxies • Measure galaxy sizes/internal structure ~0. 3” resolution • Either Nfield >> 1 or qfield > 2 xcorrelation length ~10’ HWR Princeton, 2005

A Survey Name Nfield Field size HST imaging # of bands Depth Nredshift HDFs/UDF 3 2. 5’ + 7 R=29 700 GOODS 2 12’ + 10 i=27. 5 400 3000 (5%) FIRES 2 5’ + 10 KAB=26 600 (5%) COMBO-17 GEMS 3 30’ + 22 R=24 30, 000 (1%) MUNICS 3 30’ - 7 K=19. 5 20. 000 (5%) GDDS/LCIRS 2 30’ - 7 H=21. 5 500(2000) SUBARU Rest. UV Steidel et al Rest. UV HWR Princeton, 2005

HWR Princeton, 2005

COMBO-17 Wolf, Meisenheimer, Rix et al. 01/03 Heidelberg, Oxford, Potsdam, Edinburgh • • • 3 fields @ 30’x 30’ 17 filters to mr~23. 6 ~10. 000 redshifts (1. 5%)+ SEDs per field Z Wavelength HWR Princeton, 2005 [nm] MB

Comparison of COMBO-17 with VIMOS Spectra (data from Le Fevre et al 2004) HWR Princeton, 2005

A quick Tour through Redshift Space GEMS(CDFS) HWR Princeton, 2005 Abell 901 S 11 (random)

HWR Princeton, 2005

Stellar Masses from the COMBO-17 Survey Borch, Rix, Meisenheimer et al 2005 • Stellar masses to z~1 can be estimated for 10. 000 s of galaxies • Flux limit (R-band) is VERY different from mass limits. 0. 65<z<0. 75 HWR Princeton, 2005

FIRES: Faint. Infra-Red-Extragalactic-Survey ultra-deep VLT survey *HDF-south 100 hours in JHK FWHM=0. 45” *MS 1054: 5 xlarger area 25 hours in JHK per pointing Franx, Rix, Rudnick, Labbe, van Dokkum, Foerster-Schreiber, Trujillo, Moorwood, et al. 2001 -2005 Selecting and studying galaxies z>2 in their rest-frame optical bands HWR Princeton, 2005

Not a Ly-break!! Just a red SED HWR Princeton, 2005

What kind of galaxies are found in such a search? • Galaxies without many (really) young stars won’t be found by their Ly-break or their sub-mm dust emission. • Ditto for galaxies with significant dust extinction that are not powerful enough for a sub-mm detection. • Remember: both UV searches (dust) and sub-mm searches (fainter galaxies) have ~10 corrections to get total SFR HWR Princeton, 2005

SED fits for DRGs Near-IR selected UV selected HWR Princeton, 2005 Förster-Schreiber, Franx, Rix et al; FIRES

Improving Mass, SFR and Av Estimates at z~2. 5 through IRAC (3. 6 mm-8 mm) data Labbe, Franx, Rix et al 2005 Förster-Schreiber, Rix et al 2005; FIRES HWR Princeton, 2005

Comparing dynamical (? ) with SED masses HWR Princeton, 2005 Van Dokkum, Franx, Rix, et al. 2004

Results I: Cosmic Star-Formation Rate HWR Princeton, 2005

SFR’s from thermal-IR flux 0<z<1 Zheng, Rix, Rieke, Bell et al 2004 Stacking galaxy classes (z, L) from COMBO -17 and measuring the 24 mm flux HWR Princeton, 2005

SFR’s from thermal-IR flux 0<z<1 Zheng, Rix, Rieke, Bell et al 2004 • Through stacking, (single source) confusion limit LIR/LSpitzer’s UV = f(SFR) @ all z, Lopt can be beat by >10 to <10 m. Jy • IR flux dominates in all galaxies (to 3% of L*) to z~1. 2; – large majority of UV photons absorbed. Local relation • Mean LIR/LUV drops with galaxy luminosity faint galaxies contribute hardly to SF integral • “Correction” seems to be a function of (absolute) SFR only – Insensitive to stellar luminosity, redshift HWR Princeton, 2005

State of Affairs: Star-fomration rate Borch, Rix, Meisenheimer et al 2005 HWR Princeton, 2005

Why the drop of the SFR since z~1? or In what type of galaxies did stars form back then? HWR Princeton, 2005

Whence the UV flux at z~0. 7? j 280 (z~0. 7) ~ 4 x j 280 nm (now) [not necessarily true in massive, old systems] Explore “morphology” of galaxies that give rise to these photons Subjective – use 6 eyes [Morphologies from GEMS, see Thursday] HWR Princeton, 2005 UV luminous “blue” Pick f(2800 A) as a proxy for young stars (t<tdyn) UV-to-optical flux (M 280 nm – V) Wolf, Bell, Rix et al 2004 0. 65<z<0. 75

UV-light contribution by • At MV>-19 and z~0. 75 – ½ the flux comes from seemingly normal spirals – 20% from visibly interacting systems Galaxy type at z~0. 75 Normal spirals ar uli ec /P ng HWR Princeton, 2005 cti • drop in (major) merger rate not cause of SFR drop a er Int • only minority of UV flux from manifestly interacting systems at z~0. 75

Results II: Evolution of the Stellar Mass Density with Redshift HWR Princeton, 2005

The Evolution of the Stellar Mass Function over the Last 7 Gyrs COMBO-17 survey; 30, 000 galaxies Mean stellar mass Build-Up Present-day stellar mass function HWR Princeton, 2005 Borch, Meisenheimer, Rix, Bell et al 2005, COMBO-17

Where is the stellar mass at z=2 -3. 5? DRGs (“distant red galaxies”) vs Ly-Break Galaxies Distant red galaxies likely dominate the mass budget HWR Princeton, 2005

<r*(z)>: State of Affairs Borch, Meisenheimer, Rix et al 2005 HWR Princeton, 2005

…half the mass since z~1. 5… Borch et al 2005 HWR Princeton, 2005

Putting it together Borch, Meisenheimer, Rix et al 2005 HWR Princeton, 2005

Summary • Waning SFR not a consequence of waning major mergers – Waning cold gas supply • SED-based stellar mass estimates now available for 1000’s of galaxies to z~3 – Need to observe at least to lrest>4000 A – Available testing against dynamics OK • “Distant red galaxies”, between Ly-break and sub-mm galaxies, may contain the bulk of stellar mass 2<z<3. 5 – Found through near-IR surveys – Quite frequent objects with SFR x t. SFR ~1010 -11 M • <r*(z) > can be traced from z~3. 5 to 0 – enclosing ~90% of all stars formed • Integral over SFR estimate agrees with <r*(z) > to < 2 – Assuming diet-Salpeter IMF (e. g. Kroupa 2002) – Leaves not much room for overlooked SFR HWR Princeton, 2005

Where do we go from here? • Role of merging in the build-up of the galaxy mass function is observationally barely constrained • Comprehensive linkeage of SED-based and dynamical masses • Beat field-to-field variations at z>2 • Relate stellar masses at different z to halo masses – Lensing, clustering HWR Princeton, 2005

HWR Princeton, 2005

HWR Princeton, 2005

HWR Princeton, 2005

HWR Princeton, 2005

HWR Princeton, 2005

HWR Princeton, 2005

HWR Princeton, 2005

HWR Princeton, 2005

HWR Princeton, 2005

Improving Mass, SFR and Av Estimates at z~2. 5 through IRAC (3. 6 mm-8 mm) data Labbe, Franx, Rix et al 2005 Förster-Schreiber, Rix et al 2005; FIRES HWR Princeton, 2005

SED Fitting of FIRES Galaxies HWR Princeton, 2005

Where is the stellar masses at z=2 -3. 5 DRGs (“distant red galaxies”) vs Ly-Break Galaxies HWR Princeton, 2005 Förster-Schreiber, Franx, Rix et al 2005