1. Star Forming Proto-Elliptical Galaxies @ z>2 ? N.Arimoto (NAOJ) Subaru/Sup-Cam C.Ikuta (NAOJ) X.Kong (NAOJ) M.Onodera (Tokyo) K.Ohta (Kyoto) N.Tamura (Durham) VLT/VMOS A.Renzini (ESO) E.Daddi (ESO) A.Cimatti (Arcetri) T.Broadhurst (Hebrew)

2. Why universe @ z ～ 2? 1) A large fraction of the stars in the present- day universe formed ! 2) Bright QSO activity reached its peak ! 3) Rapidly star forming galaxies of compact and disordered morphologies became the normal Hubble sequence galaxies of the z<1 universe ! © Toho Co., Ltd.

3. Formation of Massive Galaxies Massive galaxies are the product of rather recent hierarchical merging of pre-existing disk galaxies taking place largely at z<1.5 with moderate SFRs (eg, cole et al. 2000). Mass Function Evolution (Baugh et al. 2002) Fully assembled massive galaxies with Ms>1E11Mo at z>2 are extremely rare.

4. Formation of Massive Galaxies Alternatively, massive galaxies formed at much higher redshifts (z>2-3), through a short and intense epoch of star formation, followed by passive evolution. Ages of Virgo Ellipticals Y.Yamada et al. (2004)

5. Observational Test Search for high-z Massive Systems Deep Spectroscopic Surveys of Field Galaxies selected in the K-band (Broadhurst et al. 1992) K20 Project (Daddi, Renzini, Cimatti et al.)

6. A New Population of near-IR bright, z ～ 2 Galaxies K>20 HST/ACS F435W, F850LP & K-band (VLT+ISAAC) A sample of 9 galaxies at 1.7<z<2.23 with bright K-band magnitudes 18.7<K<20 has recently been discovered (Daddi et al. 2003, astro-ph/0308456).

7. Why New Population? Because 1)they lie in the redshift-desert at z ～ 1.5-2 where no strong spectral features, 2) they are not LBGs, because UV continuum slope is too red, 3) they are not EROs, because R-K colours are often too blue. The existence of NIR Bright z ～ 2 Galaxies clearly outlines a major failure of semianalytical CDM modeling of high-z massive galaxies.

8. Nature of NIR Bright Galaxies 1)actively star forming galaxies (>100-500Mo/yr) 2)irregular light distribution, clumpy SF regions, high detected asymmetries, ongoing mergers? 3) low light concentration similar to local starbursts and ULIRGs 4) old underlyings, half light radii ～ 6kpc (large)

9. Proto-Ellipticals @ z ～ 2? If high SFRs > 100 Mo/yr are really present, it is clear that in a few 100 Myr a full M* galaxy will be assembled. However, high SFRs prevent a fair mass estimate from the photometry. The galaxies appear significantly clustered, similar to EROs and local massive ellipticals. Are these galaxies massive early-type galaxies in the act of their formation?

10. Subaru/Sup-Cam Observation K20 52 arcmin^2 FIRES 20 arcmin^2 GOOD-South 160 arcmin^2 Subaru/Suprime-Cam Observations March 2-4, 2003, 0. ” 5-0. ” 7 seeing 1) BRIz ’ +JK 900 arcmin^2 Ks(AB)=22.6(5σ) 400 arcmin^2 z ’ (AB)=26.1, B(AB)=27.7(5σ) 2) BRIz ’ + K 900 arcmin^2 (Daddi-F) Ks(AB)=21.3(5σ) 500 arcmin^2 z ’ (AB)=25.4, B(AB)=27.0(5σ)

11. VLT/VMOS Observations New populations are easily identified photometrically. Targets: NewPops EROs PEs(z>1.5) FIRES(J-K>2.3) Field 1 464 1200: 82 253 Daddi-F 146 400 46 ? We aim at determining redshifts, space densities, and clustering properties of a statistically significant populations of massive starburst galaxies at z ～ 2, such as NewPops, EROs, PEs, & FIRES, in total ～ 1000. VLT/VMOS 4 nights in January 2004

12. Clustering of z ～ 2 Galaxies New Pops & EROs (Daddi-F) New Pops & FIRES (Field-1)

13. Pilot Exploratory Programs in view of Subaru/FMOS A new population of galaxies appear to have high star formation rate (>100Mo/yr), irregular and possibly merging-like morphologies, large masses, and strong redshift clustering, suggesting that they are massive early-type galaxies in the act of major assembly episodes. 1) VLT/ISAAC NIR Spectroscopy 2.5n 2) Subaru/CISCO/OHS NIR Spectroscopy 4n

14. Crucial Questions 1) How massive are they? 2) How severe is the apparent conflict with the ΛCDM hierarchical model predictions? 4) What is their SFR? 5) Are they assembling by merging? 3) Are they early-type galaxies in formation?

15. Subaru/FMOS Observation Targeted Lines: [O II]λ3727, [O III]λ4959, 5007, Hα(λ6563 Å ), [N II]λλ6548, 6584 [S II]λ6717, and possibly Hβ(λ4861 Å ), in the near-IR domain for New Pops and other Galaxies (EROs, FIRES, LBGs) with 1.5<z<2.5. Key Words: Mass, SFR, Metallicity, Merging, AGN

16. SFRs and Reddening 1)Estimate reddening and SFRs from the observed emission line luminosities. The comparison of UV-band and Hα, [O II], [O III], and possibly Hβ based SFRs will allow an independent and more reliable estimate of the reddenning, thus corrected SFRs for our galaxies. If high SFRs (>100Mo/yr) will be confirmed, it is clear that in a few 100 Myr a full M* galaxy will be assembled.

17. Metallicity 2) Estimate metallicity of z ～ 2 starburst galaxies from R23 by using [O II], [O III] vs Hβ (or Hα by assuming intrinsic ratio of Hβ/Hα constant) as metallicity indicator. If these z ～ 2 starbursts are progenitors of z ～ 1 EROs and local ellipticals, their metallicity should be high, at least near-solar or more. Metallicity is a good indicator of stellar mass.

18. Dynamical Masses 3) Derive stringent limits on the dynamical masses of our galaxies, by measuring the intrinsic width of the Hα emission line, deriving in this way the velocity dispersion. Expected masses of our targets are order of one magnitude larger than the observed z ～２ LBGs (Erb et al. 2003). We will obtain Mdyn/L(K) to compare the Mdyn with stellar mass derived from UV-Submm SEDs.

19. Formation Scenario 5) Compare the masses with the predictions of different models of massive galaxy formation. Baugh et al (2000) predicted ～６ massive galaxies of Mstar=1E11Mo for 1.7<z<2.3 for our total field of 900 arcmin^2. Even a detection of single galaxy with mass larger than 1E11Mo would allow us to set stringent observational constraints on theoretical models of galaxy formation.

20. Stellar Populations 4) Derive stellar ages and metallicities by measuring line indices such as Hβ, Mgb, CN, G-band, Fe4383, Fe5270, Fe5335, Fe5406, Ca4227, from stacked spectra. In a current cosmology Ho=70km/s/Mpc, Ωm=0.3, Ω λ =0.7, the universe at z=2.3 is 2.8 Gyr old.

21. Merging? 6) Investigate the presence of merging systems with a help of HST/ACS images. Quite often a few distinct blobs are detected in UV. The secure detection of large velocity separations would give the first direct demonstration of assembling systems at z=2.

22. AGNs 7) Use [S II]λ6717/Hα and [N II]λ6548, 6584/Hα to check for the presence of AGNs. [N II]/Hα line ratio for abundance and ionization degree. We have XMM images of the fields. Our preliminary analysis shows that there are lots of New Pops detected in X-rays!