Starbursts from z3 to 7 10 Daniel Schaerer

Starbursts from z~3 to 7 -10 Daniel Schaerer (Geneva Observatory, OMP Toulouse) • Stellar populations: ages, star formation histories, masses, (IMF, metallicities, …) • Reddening: amount (attenuation law, …) • Star. Formation. Rate and SFR density in z ~ 3 to 7 -10 galaxies (LBG and Lyman-α emitters)

Outline • Observables & methods – brief comments • Lyman Break Galaxies (LBG) at z~3 – Stellar populations, SF histories, reddening. . . • LBG at z ~4 – 6 – Comparing their properties with z~3 – …including UDF • Lyman-α emitters (LAE) • Distant galaxies seen through the Gravitational Telescope – Properties of two lensed z ~6 -7 galaxies – Searches for z ~ 7 – 10 galaxies and first results Not discussed: ERO, sub-mm galaxies, red galaxies at z>3, …

Observables & methods Main observables: • Detailed spectroscopy -- rarely available at z>~3 age, SF history, IMF (from line fits) cf. talk Leitherer attenuation (from UV slope or Balmer decrement) cf. poster Noll+ abundances (eg. from R 23, or from UV lines) cf. talks Rix, Mehlert, de Mello kinematics, masses cf. talk Erb • Narrowband flux, « poor » spectroscopy z, emission line flux, EW • Broad-band flux SED, zphot – Detailed SED fits… – Beta-slope attenuation • Average SED(z), luminosity functions, … this review

Observables & methods – SED degeneracies 1) age – reddening degeneracy (UV restframe): UV slope depends on age, SF history, reddening (+law) Can be « broken » in some special cases! E. g. : - Presence of emission line (EL) ongoing SF (young burst or continuous SF) - Flat/rising slope + EL Strong extinction - Very steep (blue) slope young + no/little extinction Or by adding: * restframe optical data (near-IR, SPITZER) * UV lines (cf. Leitherer talk) 2) A priori UV slope NOT metallicity indicator! For « normal » metallicities (Z>~1/50 Zsun) little dependent on Z. Very metal-poor populations: FLATTER slope due to nebular continuum ! Only possible if statistical correlations hold (e. g. Heckman et al. 1998)

LBG at z~3 Large samples with spectroscopic redshift (~1000, e. g. Shapley et al. 2003) Imaging: mostly optical (UV restframe), some also with near-IR Sawicki & Yee (1998, Ap. J, 115, 1329): * 17 spectroscopically confirmed LBG in HDF with photometry in seven bands (UBVIJHK) UV-optical (restframe) coverage to break the UV age-degeneracy * Assume: Calzetti attenuation law, Salpeter IMF, Bruzual & Charlot (BC) synthesis models, variable metallicity, * SED chi 2 fitting free parameters: age, reddening , SF history (burst/SFR=const) Also: SFR, stellar mass estimates (from best-fit model, not standard conversion factors)

LBG at z~3 Sawicki & Yee (1998, Ap. J, 115, 1329): * Extinction: non-zero, median E(B-V)~0. 28 [A_V~1, factor ~16 at 1600 Ang] * Age (of dominant population): young (<~0. 2 Gyr) From sample spanning z~2 to 3. 5, i. e. ~1 Gyr episodic SF, not extended and continuous * SFR: median ~59 Msun/yr (h 100 -2) * Stellar mass: from burst or SFR=const models median ~109 Msun

LBG at z~3 Papovich et al. (2001, Ap. J, 559, 620): * 33 LBG in HDF-N with UBVIJHK * Assume: Calzetti attenuation law, Bruzual & Charlot (BC) synthesis models, variable metallicity, Madau Ly-forest attenuation * SED chi 2 fitting free parameters: age, reddening , SF history (exp. declining), IMF Similar results as Sawicki & Yee (1998) See also study of Shapley et al. (2001) But: none of these studies includes information from presence of Lyman-α

Age – e-folding time prob. distribution Age – dust attenuation probability distribution Age – stellar mass probability distribution Main results: * mass estimates * typical ages: ~30 Myr to 1 Gyr now confirmed by SPITZER/IRAC mid-IR obs. (Barmby et al. 2004) * no young AND dust free object * From Δt of sample and relative absence of quiescent objects recurrent SF Papovich et al. (2001)

LBG at z > 3: from 3 to 4 0 • No detailed SED analysis of galaxies with z >>3 (cf. Schaerer & Pello 2004: 3 lensed galaxies with z>~6) • • UV luminosity density @ z=4 ~ identical to z=3: Papovich et al. (2004) -- GOODS Ouchi et al. (2004) -- SUBARU Deep Fields global average SFR ~const if same reddening Reddening ~constant between z~3 and 4 (from UV colors: i-z’; Ouchi et al. 2004) But: Papovich et al. (2004): possible « blueing » (decrease of mean age, reddening, decrease of M/L) Ouchi et al. (2004) 1 2 3 4 5

LBG at z > 3: from 3 to 5 Ouchi et al. (2004) -- SUBARU Deep Fields (BVRiz’): ~2600 LBG at z~4 -5 Possible decrease of UV luminosity density from z~3 to 5 – also Iwata et al. (2003) – HDF-N/SUBARU (Viz’) Lehnert & Bremer (2003) – Riz + spectroscopy z~5: basically no information on reddening since i band affected by Lyman-α forest but: colors consistent with little reddening (L&B 03+) • Important uncertainty at z>~5: integration of LF ! No difference in observed LF from z ~3 to 5 (Ouchi et al. ) 0 1 2 3 4 5 z~5 spectroscopic follow-ups: * Iwata et al. poster * Douglas, Bremer+ (talk, poster) after before LF integration

LBG at z ~ 5 (cont. ) Samples & follow-ups: 1) Iwata et al. (2003) – HDF-N/SUBARU (Viz’) Deep spectroscopy of 17 objects with FOCAS/SUBARU: Ando et al. (2004) 8 confirmed at z~4. 5 – 5. 2 7 with no or weak Ly-α emission, but relatively strong IS abs. lines In contrast with z~3 LBG ! Due to selection of brightest objects (>L*) ? 2) Lehnert & Bremer (2003) – Riz + spectroscopy cf. poster+talk Bremer+ Bremer et al. (2004) – Viz (Chandra Deep Field S) 6 of 13 galaxies confirmed by Ly-α emission (flux ~(0. 2 -2. 5)*e-17) + break « high » EW(Ly-α) indicative of young age and/or ongoing SF X-ray non-detections: SB or < weak AGN UV luminosity density from these and brighter objects insufficient to maintain ionisation Sources of reionisation fainter than MAB(1700 Ang) > - 21 3) Ouchi et al. (2004) -- SUBARU Deep Fields (BVRiz’) no spectroscopic follow-up yet ? !

i-dropouts: LBG at z~6 HST -- UDF: (ACS i’z’+ NICMOS JH) * Bunker et al. (2004): 54 candidates i’-dropouts over 11 arcmin 2 part of UDF (ACS only). Determine LF, SFR density, … * Stanway et al. (2004): ~27 candidates with i’z’JH (z-J) color ~flat spectral slope – SB! possibly bluer than LBGs at lower z (lower reddening? , young populations? ) z_phot extinction z_phot Stanway et al.

i-dropouts: LBG at z~6 HDF-N (i’z’/ACS + JH/NICMOS) and RDCS field (i’z’/ACS + JKs/ISAAC): Bouwens et al. (2003 , 2004) * 11+1 objects with optical + near-IR, total 21+2 candidates with z>~6 * (z-J) color ~flat spectral slope – SB * H or K: large uncertainties no information on stellar populations * detailed derivations of UV luminosity density (SFR density) using different methods and accounting for surface brightness dimming => SFRD ~14 x Stanway small decrease (39± 21 %) of SFRD from z ~3 to 6

Going beyond z ~ 6 -6. 5 … Requires: * HST: UDF -- ACS+NICMOS (JH): some z dropouts with blue J-H ? ! cf. talk Thompson, poster Bouwens et al. * HST: ACS grism spectroscopy (Ly-α break up to z ~7; cf. Rhoads et al. 04) * deep JH AND K photometry (J: 7 -10 dropout) present: combined with gravitational lensing !! Future: 30 m tel. , JWST * other selection technique: emission line search (Lyman-α emitters (LAE)) with narrowband filters, tunable filters, « blind » searches (long-slit spectroscopy, IFU)… in optical or near-IR So far: successfully applied to from z~2 -4 to 6. 58

Lyman-α emitters (LAE) • Numerous LAE searches: see e. g. reviews by Spinrad (2003), Taniguchi et al. (2003) • Currently used to trace SFR(z) out to z~6. 6 (also clustering properties…) But: Lyman-α gives only lower limit on SFR, since affected by several « destruction » processes (dust, ISM geometry + kinematics) • Most LAE are detected in very few (1!) or no broad-band filter Little known about their properties, stellar populations, nature, relation with LBG … Taniguchi et al. (2004)

Lyman-α emitters (LAE) at z~ 4. 5 – 5. 7 LALA survey (4 m Kitt Peak): BVRIz’+ 2 narrowband at z=4. 5, 5. 7 * 157 z=4. 5 LAE candidates (Malhotra & Rhoads 2002) * 18 z=5. 7 LAE candidates (Rhoads & Malhotra & 2001) High median EW(Ly-a) ! AGN ? NO ! Wang et al. (2004) Very-metal poor objects or Pop III ? Extreme/”massive” IMFs ? Nature puzzling ! BUT: Many metal free objects at z < 6 expected ? ? Z=1/50 - 2 Zsun High EW real? Uncertainties in EW from NB ? Z=0 Z=10 -7 Z=10 -5 Schaerer (2003)

Lyman-α emitters (LAE) at z~ 4. 5 – 5. 7 * Keck Follow-up spectroscopy of z=5. 7 candidates: 3 of 4 confirmed (Rhoads et al. 2002) * no other UV lines detected (but deep enough? ) (Wang et al. , Dawson et al. 2004) * overall SED ? Nature of LALA sources puzzling ! Similar programs: " Hu et al. (2004): SUBARU deep imaging, similar selection criteria 26 z=5. 7 candidates, 19 confirmed (DEIMOS/Keck) less than 25% have EW(Ly-a) > 240 Ang! Difference due to deeper imaging !? " Ajiki et al. (2002, 2004): several LAE, none with EW(Ly-a) > 200 Ang!

z >~6 -7 galaxies seen through the “Gravitational Telescope” Abell 370 HCM 6 A, z=6. 56 Hu et al. 2002, Ap. J, 568, L 75 • NB excess • asymetric emission line (Lya) • no secondary image • magnification 4. 5 (1. 6 mag)

Abell 370 HCM 6 A, z=6. 56 Main results from spectral fitting: * Good fits with burst models: age ~ 100 -200 Myr, ~no extinction -- BUT no Lyα emission expected then !! * Good fits with SFR=const + non negligible extinction (AV~1. ) SFR ~ 80 -300 Msun/yr (cf. Hu et al. : 9 Msun/yr) Also mass, luminosity estimate Observed/predicted Lyα flux ~ 9 -66% (Hu+Haiman 2002: ~1/5) No indication on age, metallicity Prediction: IRAC/SPITZER (non detection) could confirm strong extinction Schaerer & Pelló (2004) SPITZER/IRAC sensitivity

Abell 2218 KESR, z undetermined (~6 -7) Observations: VIZJHK (HST: WFPC 2, ACS, NICMOS), spectroscopic UL on continuum flux (9000 -9300 Ang), no emission line – Kneib et al. (2004) Main results from spectral fitting: * Photometric redshift well behaved z~5. 8— 6. 8 * Age: 5 -90 Myr (up to 200 Myr) * Best fits: generally little / no extinction * Absence of Lyα NOT SURPRISING ! Intrinsic: too old population Emission present but destroyed (…) * Quite strong degeneracies in age, SF histories, extinction law ! Object detected with IRAC/SPITZER ! (Egami et al. 2004) Above results compatible with SPITZER observations Schaerer & Pelló (2004)

Searching for z ~7 -10 galaxies with the “Gravitational Telescope” * Target clusters: " “lensing” galaxy clusters with well-defined mass models " existing deep optical imaging (ground/HST) typically 1— 3 mag amplification * Ultra-deep NIR (JHK) exposures in cluster core Prime targets: z ~ 7 to 10 galaxies Abell 1689 - ACS / HST

Search for lensed distant/primeval/Pop. III galaxies at z > 7 Step 1) Ultra-deep JHK (ISAAC/VLT) + existing optical imaging (HST, …): Traditional drop-out technique + blue rest-frame UV spectrum photometric redshift estimate + selection of starbursts JH Step 2) Follow-up near-IR « high-res » spectroscopy (ISAAC): emission line (Ly-a, He. II? ) search redshift + other properties !? Pelló, Schaerer (2001 -2003), Barton et al. (2004) H-K

J-H Optical dropouts H-K Z~9 critical line

First spectroscopic confirmation of a possible z=10 lensed source (Abell 1835 -IR 1916) Pelló, Schaerer, Richard, Le Borgne, Kneib, 2004, A&A 415, L 19

First spectroscopic confirmation of a possible z=10 lensed source (Abell 1835 -IR 1916) Redshift indicators: 1) photometric-z ~ 9 -11 2) emission line with z=10. 0 if Ly-α 3) galaxy on top of critical line for z>~9 * Intrinsic flux: >~ 28. 5 – 29. mag. AB in H and Ks * Magnification factor ~25 -100 (3. 5 to 5 magnitudes) * Ly-alpha line flux ~ (4. 1± 0. 5)x 10 -18 erg/s/cm 2 Derived properties: (z=10. 460 Myr after Big Bang) * SFR(UV) ~ 2 -3 Msun yr-1, SFR(Ly-a) >~ 0. 03 -0. 09 Difference due to: loss of Ly-a photons (ISM geometry, …) partial IGM transmission as source z >> 6! * UV slope no extinction & young population * Mass estimate (Salpeter IMF 1 -100 Msun): M* ~ (9 -50). 106 Msun (young bursts or const SFR) heavier than massive GC, typical for super star cluster …properties as expected for young z~10 proto-galaxy…

News from Abell 1835 -IR 1916 * H-band spectroscopy: low z (~2. 5) solution excluded * additional photometry (Z, SZ bands): non-detections - compatible with A&A results Schaerer et al. (2004) -- also other attempts to detect IR 1916 in optical * re-analysis of ISAAC spectroscopy (Weatherley et al. , astro-ph): non-standard technique not suited to complex observational setup * GEMINI/NIRI H-band imaging (Bremer et al. 2004): not detected in H spurious? transient source ? Probability = ? ! z=2. 52 ed d u Metal-poor HII galaxy (SBS 0335 -052) l xc !! AVe=3. 6 Wto verify… easy NO upcoming HST ACS+NICMOS observations of Abell 1835 & AC 114 fields other z ~7 -9 candidates (Richard et al. + Pello et al. 2004)

A 1835: 2 observing runs: 4 priority targets 1 confirmed/ 1 no-detected/ 1 low-z/ 1 tb confirmed 2 secondary targets 1 confirmed high-z / 1 low-z A 1835 -2582 (Richard et al. 2003): z=1. 67, [OIII]5007, 4959, H_beta detected In Summary: 3 spectroscopic "confirmation" runs/ 2 clusters 6 priority 1 targets: 1 confirmed high-z A 1835 -775: z=1. 888 double line 10760 A /10765. 5 A; likely [OII]3727 2 to be confirmed 2 no-detected 1 low-z Efficiency ~30 50% 2 secondary targets: 1 confirmed high-z A 1835 -1055: z=7. 89 if Ly ; unlikely [OII]3727 z=1. 9; 1 low-z or z= 1. 16 if [OIII]5007 (no [OIII]4959) or z= 1. 2238 if H_beta (no [OIII]5007) A 1835 -1143: faint line to be confirmed Data reduction ongoing A 1835 -1736: no detection within the J band A 1835 -1916: z=10. 0 Ly emitter Also: AC 114: 1 observing run/ 2 nights/ 2 1 rst priority candidates

Summary • • LBGs at z~3: – Relatively young populations (<~0. 3 Gyr) – Recurrent/episodic SF! – Moderate extinction (E(B-V) ~0. 3) – Masses (~5 -10 x 109 Msun, SFR ~60 Msun/yr), also metallicities … Reddening in LBGs: similar between z ~3 and 4. No clear indication for z > 4. • Tendency towards bluer colors (less reddening, younger pops ? ) at z~5 to 6. SFR density (from LBGs and LAE): ~ constant between z ~3 and 4. Possible decrease at z>~5 -- Important uncertainty: low end of LF • LALA sources (high EW(Ly-α)): Nature puzzling! • Two lensed z~6 -7 galaxies: 1 « high » extinction, 1 negligible. SPITZER/IRAC observations: age up to 200 -400 Myr. Search for z~7 -10 galaxies with VLT + Gravitational Telescope: Quite efficient! z=10 object !? , Other z~7 -8. 5 confirmed + candidates •