1. Massive galaxies at z ~ 2 from K20+ Preamble: (Cimatti, Daddi, Fontana, Arimoto, GOODS, GRAPES, COSMOS ) Looking at z~2 massive galaxies gives us more leverage than looking at z~1 ✰ as their predicted abundance and properties (e.g. Passive vs. starbursting) are most critically dependent on theoretical model *assumptions, *algorithms and *parameters → allow to narrow down our model shopping list. A Renzini, STScI, September 27, 2004 Fontana et al. 2004

2. The Local Mass Function (from the SDSS) et al. (2003) In the local universe: Among the MMGs the old passively evolving, early- type galaxies outnumber the starforming galaxies by more than a factor ~10

3. The High-z Tail in the K20 Sample K 1.6: ~ 0.0 predicted by SAMs 32 observed in the the K20 sample (apparent agreement with PLE models) (old) SAM PLE Cimatti et al. 2002

4. The Nature of the K 1.6 galaxies 1. The Star-Forming Galaxies Daddi et al. (2004)

5. The K<20, z≈2 starforming galaxies Coadded spectra of 5 best S/N galaxies (de Mello et al. 2004). More stronglined than LBGs both for ISM and photospheric absorptions ➔ >~solar metallicity which along with: ● M * >10 11 M ⊙ ● SFR > ~ 100 M ⊙ /yr ● Strong clustering ➔ Likely progenitors of local ellipticals and big bulges

6. 2. The Passively Evolving Galaxies → z F > 2.5-3 Coadded VLT spectra ACS/GOODS images Cimatti et al. (Nature, July 8, 2004) (The Highest redshift ellipticals)

7. The Mg UV Feature: a key to hi-z passive galaxies First used by Dunlop/Spinrad et al '96; Cimatti et al '04; McCarthy et al. '04)

8. Passively Evolving galaxies (cont.) In red LBDS 53w091 (z=1.55) Blue=average of the 4 gals. F2V F5V F2V SDSS =0.5 z=1 old EROs Z=1 SF ERO 0.5 Gyr 1 Gyr 3Gyr

9. The 4 galaxies in real color GOODS ACS BViz images

10. Old galaxies at high redshift In the K20/GOODS field (32 □ ) Passive galaxies with R-K>6, z>1.5, K<20: 4 objects with z spec 3 objects with z phot 7 objects in total In the currently “best” semi-analytic model (Somerville 2004): Mock catalog for a whole GOODS field (160 □): Only one object with the same characteristics. 35 expected scaling from the K20

11. The ACS/GOODS Morphologies 32 K20 galaxies at z>1.4 with deep z-band ACS imaging S=Starforming ➔ mergers, starbursts P=Passively evolving ➔ Elliptical and Bulge-dominated galaxies

12. Some Inferences (so far) ● Massive (M*>10 11 M ⊙ ) galaxies appear to be in place at z ~ 2 in much greater number than predicted by most CDM Simulations Models with strong SN/AGN feedback do better (Nagamine et al. 2001; Granato et al. 2004) [Anti-hierarchical galaxy assembly!] While at z=0 most “most massive galaxies” are passively evolving, old ellipticals, by z~2 passive and active star-forming galaxies coexist in nearly equal number ☹ Limitations of the K20 survey: ✈ relatively small area (prone to cosmic variance) ✈ relatively shallow (K<20)

13. Post-K20: go wider, go deeper, focus on z~2 Daddi et al. (2004) The BzK criterion for selecting BOTH starforming (reddening independent!) and passively evolving galaxies at 1.4<z<2.5 Calibrated on the K20, GOODS, and GDDS datasets. ⇦ K20/GOODS data only shown here.

14. Why the BzK criterion works Daddi et al. (2004) BC03 models with various ages, SF histories, and reddening: Nice agreement with the K20 empirical findings SSPs Cont. SF SFR ~ e(-t/ τ) z<1.4

15. BzK- vs. LBG UGR-selected Galaxies A preliminary comparison: K20 BzK sample vs UGR sample (Steidel et al. 2004) K20 BzK Objects (K<20) > 0.3 = ~ 200 M ⊙ /yr ~1 Object/□' ~1/3 of the BzK+UGR Star Formation Rate @ z=2 Misses objects with K>20 Misses Objects wit /y ~ 1 Object/□' ~1/3 of the BzK+UGR Star Formation Rate @ z = 2 Misses objects with K>20 UGR Objects (R<25) E(B-V) < 0.3 =~50 M ⊙ /yr ~9 objects/□' ~2/3 of the BzK+UGR Star Formation Rate @ z = 2 Misses highly reddened objects ~ 9 Objects/□' ~2/3 of the BzK+UGR Star Formation Rate @ z = 2 Misses highly reddened objects / che-1 drwx------ 31 alvio alvio 4096 Feb 20 12:15. [alvio@nb004358 alvio]$

16. Going deeper: the GMASS project VLT/FORS2 ultradeep spectroscopy (~30 hours integration per mask) over the UDF/GOODS-South field. PI A. Cimatti, Co-Is as usual Scheduled for September-November 2004. (Also called the K21 project) The selected targets on the BzK diagram All targets with: K>20 (average 21.2) (4.5) AB <23.5 (FromSpitzer/GOODS) Perhaps the first Spitzer-selected sample.

17. Going Deeper: the UDF+GRAPES * Sample * Malhotra et al. 2004 Daddi et al. 2004 (in prep) Selecting candidate passively evolving galaxies at z>1.4 from the UDF field (Bz) & GOODS (K) with the BzK plot. ACS/GRISM spectra from the GRAPES project. GOODS K20 UDF Targets down to K=21.1

18. ACS Cutouts & ACS/GRISM Spectra The Mg UV Feature as a function of SSP age and for Continous SF + E(B-V)=1.2 ← Redshifts identified by the Mg UV Feature agree with photo-z's

19. ACS&NICMOS Morphologies n~1.0 n=2.9 n= 4.7 n=9.0 n=4.2 n=4.3 merger? N=8.2 Sersic index in z

20. Passive or Active? Object #4950 @ z=1.55

21. Masses and Mass Densities Typical (stellar) masses of these galaxies are ~ 0.5-2 10 11 M ⊙ The 6 passively evolving galaxies at 1.4 ~10 11 M ⊙ correspond to 20-40% of the number density of such galaxies at z=0. Assuming r o = 10 Mpc for their correlation length the 1 range becomes 10-80% (!) COSMIC VARIANCE dominates BIGGER FIELDS ARE NEEDED!

22. Going Wider: BzK-selected galaxies Over ~ 1000 □ ESO/SUBARU Collaboration: Cimatti, Daddi, Renzini, + Arimoto, Ikuta, Kong, Broadhurst, Pozzetti et al. Onodera, et al. K-band (<20.2) from NTT, Bz bands from SuprimeCam Objects observed with VIMOS, February '04 ❍ Blue grism, R=200 Red grism, R=600 Reductions in progress

23. Work in Progress Typical VIMOS/Blue grism spectra of BzK-selected galaxies (R=200) Z = 1.822 B=23.6 Z=2.360 B=24.4 Z=1.565 B=24.3 Z=2.200 B=24.3

24. GOODS, COSMOS,..... GOODS: 160 + 160 □ ' COSMOS: 2 □ ° ACS Equatorial Field Goal: ~90,000 redshifts w/ VLT/VIMOS GOODS-South: To be proposed... + Magellan/IMACS ~6000 redshifts from the VLT + whoever may like to join with other ~2000 released, the rest next fall telescopes COSMOS: Mapping galaxy evolution as a function of redshift and LSS Context by 2006- 07.

25. COSMOS Targets BzK+UGR selected Passive, z>1.4 Starforming, 1.4<z<2.5 Stars Galaxies, z<1.4 Red, Blue, Black: (K)<0.2 2000; 20,000; 100,000 (BzK-selected galaxies) Yellow, Cyan: (K)>0.2 ~50,000 (UGR-selected, à la Steidel) Yellow+Blue: VLT/VIMOS/LR-blue Black: VLT/VIMOS/LR-Red Red: Magellan/IMACS Data: Bz (SUBARU), K (NOAO, IfA), U (CFHT) from B. Mobasher catalog

26. COSMOS Targets UGR BzK Bottom line: all selection criteria are +/- biased: the BzK+UGR selection is the most un-biased criterion we were able to invent. COSMOS PIs: ACS (Scoville), SUBARU (Taniguchi) CFHT (LeFevere) VLT (Lilly) Magellan (McCarthy) XMM (Hasinger) Galex (Rich) VLA (Schinnerer).................

27. SUMMARY Different CDM Models tuned to match some z=0 observable diverge widely by z=2. By z=2 the massive galaxy mix is ~50-50 passive and starbursting Current data favor an early build up of galactic spheroids, with the more massive ones completing their star formation ahead of the less massive ones (seemingly “anti-hierarchical”, “downsizing”) Fully explored fields (< 100 □) are still dominated by cosmic variance. But..................

28. 2005-2007 (~full) Mapping of Galaxy Evolution up to z = 6.7 GOODS, IDDS, GDDS, GMASS, GEMS, COMBO17, UDF, CDFS, SDF, DEEP2, VVDS, COSMOS, SWIRE, GALEX, DEIMOS, VIMOS, IMAX, MOIRCS, GMOS, ACS, SuPrime, CHANDRA, XMM, SPITZER, SCUBA, VLA, FORS2, ISAAC, NICMOS,............................................................. END

30. Ellipticals in clusters and field look much the same Bernardi et al. (2003): ~9000 ellipticals in the SDSS sample. No appreciable trend of the stellar population content with local density. “Field” ellipticals less than ~1 Gyr younger than cluster ellipticals.

31. Searching for the high-z Precursors of the MMGs at z=0 By and large, this is ~equivalent to searching for the precursors of elliptical galaxies. @ z=1: Passively evolving “EROs” (R-K>5) in (almost) enough number Beyond z=1: ■Up to which redshift passively evolving galaxies can be found? ■Given that the bulk of stars in ellipticals formed at z>~3: where are the ELLIPTICALS IN FORMATION?

32. Searching for High-z Massive Galaxies (the K20 Project, and Beyond) 1999: in absence of a better criterion, pick near-IR bright galaxies as best proxies to MMGs: i.e. Select for VLT spectroscopy K 20 galaxies ➔ the “K20 project” PI: A. Cimatti, Co-Is: E. Daddi, A. Fontana, L. Pozzetti, A. Renzini, G. Zamorani, T. Broadhurst, M. Mignoli, et al. Deep VLT Spectroscopy of 546 K 20 objects, over two fields (52 □) ● Now 92% complete ( 95 % over the subfield included in the GOODS field) 1999: in absence of a better criterion, pick near-IR bright galaxies i.e. Select for VLT spectroscopy K 20 galaxies ➔ the “K20 project” PI: A. Cimatti, Co-Is: E. Daddi, A. Fontana, L. Pozzetti, A. Renzini, G. Zamorani, T. Broadhurst, M. Mignoli, et al.

33. The Redshift Distribution of the K<20 Sample of Galaxies Pure Luminosity Evolution (PLE) Models apparently do better than CDM (<2002) Semianalytic Models (SAM) (Cimatti et al. 2002) PLE SAM

34. The Stellar Masses of K20 Galaxies Stellar masses derived from either color (R-K) (MA) or SED (UBVRIzJHK) fits (BF) and the K magnitude. ● Old Passive (Early type) ❍ Early + Emission line Photo-z only X Star forming

35. The Evolution of the Galaxy Mass Function Fontana et al. (2004) □ BF Masses ∇ MA Masses ❍ z spec only Color code: Bad fit Poor fit Fair fit Good fit Menci et al '02, '04 Nagamine et al '01 Cole et al. '00; Granato et al. '04 (Salpeter IMF) (Gould IMF) Somerville et al. '04a, '04b (Kennicutt IMF)

36. Drory et al. (2004) The K20 vs the Munics Mass Function up to z=1.2 (Very nice agreement) 1 □ o ; K<19.5; mostly photo-z's

37. The Build Up of Stellar Mass Through Cosmic Time ⇧ is my preferred value from the fossil evidence (mass and formation redshift of local spheroids), i.e. >30% of the stellar mass done by z=3 (Renzini 1998, 1999) ⇧ From Fontana et al. (2004)

38. Baryon-to-star conversion as a function of galaxy mass PLE Available baryons from CDM simulation assuming cosmic share ( b / m ) Two Main Points: There are enough massive DM halos to account for the massive galaxies we see (No fundamental failure of the DM paradigm). Strong mass dependence of the baryon-to-star conversion efficiency (many interesting ramifications... ).