Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.arcetri.astro.it/science/k20/papers/og_final.pdf Äàòà èçìåíåíèÿ: Tue Nov 30 19:23:24 2004 Äàòà èíäåêñèðîâàíèÿ: Fri Feb 28 07:24:52 2014 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: redshift survey

Old Galaxies in the Young Universe
A. Cimatti1 , E. Daddi2 , A. Renzini2 , P. Cassata3 , E. Vanzella3 , L. Pozzetti4 , S. Cristiani5 , A. Fontana6 , G. Ro dighiero3 , M. Mignoli4 , G. Zamorani4
1

INAF - Osservatorio Astrofisico di Arcetri, Largo E. Fermi 5, I-50125, Firenze, Italy
2

Europ ean Southern Observatory, Karl-Schwarzschild-Str. 2, D-85748, Garching, Germany

3

Dipartimento di Astronomia, Universit` di Padova, Vicolo dell'Osservatorio, 2, I-35122 a Padova, Italy
4

INAF - Osservatorio Astronomico di Bologna, via Ranzani 1, I-40127, Bologna, Italy
5

INAF - Osservatorio Astronomico di Trieste, Via Tiep olo 11, I-34131 Trieste, Italy

6

INAF - Osservatorio Astronomico di Roma, via dell'Osservatorio 2, Montep orzio, Italy

In the lo cal universe, over half of all stars are found in massive spheroidal galaxies1 characterized by old stellar p opulations2,3 and absent or weak star formation. In current galaxy formation scenarios, such early-typ e galaxies app ear rather late as the culmination of a hierarchical merging pro cess. However, observations have not yet established how and when such systems formed2,3 , and if their seemingly sudden app earance when the universe was ab out half its present age (z 1) results from a real evolutionary effect or from the observational difficulty of identifying them at earlier ep o chs. Here we rep ort the sp ectroscopic and morphological identification of four old, fully assembled massive (> 1011 solar masses) spheroidal galaxies at 1.6 < z < 1.9, the farthest such ob jects currently known. The existence of such systems when the universe was only 1/4 its present age, shows that the build-up of massive early-typ e galaxies was much faster in the early universe than so far exp ected from theoretical simulations4 . In the CDM scenario5 , galaxies are thought to build-up their present-day mass through a continuous assembly driven by the hierarchical merging of dark matter halos, with the most massive galaxies b eing the last to form. However, the formation and evolution of massive spheroidal early-typ e galaxies is still an op en question. Recent results indicate that early-typ e galaxies are found up to z 1 with a numb er density comparable to that of lo cal luminous E/S0 galaxies6,7 , suggesting a slow evolution of their stellar mass density from z 1 to the present ep o ch. The critical question is whether these galaxies do exist in substantial numb er8,9 at earlier ep o chs, or if they were assembled later10,11 as favored by most renditions of the hierarchical galaxy formation scenario4 . The problem is complicated also by the difficulty of identifying such galaxies due to their faintness and, for z > 1.3, the lack of strong sp ectral 1


features in optical sp ectra, placing them among the most difficult targets even for the largest optical telescop es. For example, while star-forming galaxies are now routinely found up to z 6.612 , the most distant sp ectroscopically confirmed old spheroid is still a radio­selected ob ject at z = 1.552 discovered almost a decade ago 13,14 . One way of addressing the critical question of massive galaxy formation is to search for the farthest and oldest galaxies with masses comparable to the most massive galaxies in the present-day universe (1011-12 M ), and to use them as the "fossil" tracers of the most remote events of galaxy formation. As the rest-frame optical ­ near-infrared luminosity traces the galaxy mass15 , the Ks -band ( 2.2 µm in the observer frame) allows a fair selection of galaxies according to their mass up to z 2. Following this approach, we recently conducted the K20 survey16 with the Very Large Telescop e (VLT) of the Europ ean Southern Observatory (ESO). Deep optical sp ectroscopy was obtained for a sample of 546 ob jects with Ks < 20 (Vega photometric scale) and extracted from an area of 52 arcmin2 , including 32 arcmin2 within the GOODS­South field 17 (hereafter the GOODS/K20 field). The sp ectroscopic redshift (zspec ) completeness of the K20 survey is 92%, while the available multi-band photometry (B V RI z J H Ks ) allowed us to derive the sp ectral energy distribution (SED) and photometric redshift (zphot ) of each galaxy. The K20 survey sp ectroscopy was complemented with the ESO/GOODS public sp ectroscopy (Supplementary Table 1). The available sp ectra within the GOODS/K20 field were then used to search for old, massive galaxies at z > 1.5. We sp ectroscopically identified four galaxies with 18 Ks 19 and 1.6 zspec 1.9 which have rest-frame mid-UV sp ectra with shap es and continuum breaks compatible with b eing dominated by old stars and R - Ks 6 (the colour exp ected at z > 1.5 for old passively evolving galaxies due to the combination of old stellar p opulations and k-correction effects9 ). The Supplementary Table 1 lists the main galaxy information. The sp ectrum of each individual ob ject allows a fairly precise determination of the redshift based on absorption features and on the overall sp ectral shap e (Fig. 1). The co-added average sp ectrum of the four galaxies (Fig. 2­3) shows a near-UV continuum shap e, breaks and absorption lines that are intermediate b etween those of a F2 V and a F5 V star18 , and typical of ab out 1-2 Gyr old synthetic stellar p opulations19,20 . It is also very similar to the average sp ectrum of z 1 old Extremely Red Ob jects7 (EROs), and slightly bluer than that of the z 0.5 SDSS red luminous galaxies21 and of the z = 1.55 old galaxy LBDS 53w09113 . However, it is different in shap e and slop e from the average sp ectrum of z 1 dusty star-forming EROs7 . The multi-band photometric SED of each galaxy was successfully fitted without the need for dust extinction, and using a library of simple stellar p opulation (SSP) mo dels19 with a wide range of ages, Z = Z and Salp eter IMF. This pro cedure yielded b est-fitting ages of 1.0-1.7 Gyr, the mass-to-light ratios and hence the stellar mass of each galaxy, which results in the range of 1­3â1011 h-2 M . H0 = 70 km s-1 Mp c-1 70 (with h70 H0 /70), m = 0.3 and = 0.7 are adopted. In addition to sp ectroscopy, the nature of these galaxies was investigated with the fundamental complement of Hubble Space Telescop e+ ACS (Advanced Camera for Surveys) imaging from the GOODS public Treasury Program17 . The analysis of the

2


ACS high-resolution images reveals that the surface brightness distribution of these galaxies is typical of elliptical/early-typ e galaxies (Fig. 4). Besides pushing to z 1.9 the identification of the highest redshift elliptical galaxy, these ob jects are very relevant to understand the evolution of galaxies in general for three main reasons: their old age, their high mass, and their substantial numb er density. Indeed, an average age of ab out 1-2 Gyr (Z = Z ) at < z > 1.7 implies that the onset of the star formation o ccurred not later than at z 2.5 - 3.4 (z 2 - 2.5 for Z = 2.5Z ). These are strict lower limits b ecause they follow from assuming instantaneous bursts, whereas a more realistic, prolonged star formation activity would push the bulk of their star formation to an earlier cosmic ep o ch. As an illustrative example, the photometric SED of ID 646 (z = 1.903) can b e repro duced (without dust) with either a 1 Gyr old instantaneous burst o ccurred at z 2.7, or with a 2 Gyr old stellar p opulation with a star formation rate declining with exp(-t/ ) ( = 0.3 Gyr). In the latter case, the star formation onset would b e pushed to z 4 and half of the stars would b e formed at z 3.6. In addition, with stellar masses M > 1011 M , these systems would rank among the most massive galaxies in the present-day universe, suggesting that they were fully assembled already at this early ep o ch. Finally, their numb er density is considerably high. Within the comoving volume relative to 32 arcmin2 and 1.5 < z < 1.9 (40,000 h-3 Mp c3 ), the comoving density of 70 3 such galaxies is ab out 10-4 h70 Mp c-3 , corresp onding to a stellar mass density of ab out 2 â 107 h3 M Mp c-3 , i.e. ab out 10% of the lo cal (z = 0) value22 for masses greater 70 than 1011 M . This mass density is comparable to that of star-forming M > 1011 M galaxies at z 2 23 , suggesting that while the most massive galaxies in the lo cal universe are now old ob jects with no or weak star formation, by z 2 passive and active star-forming massive galaxies co exist in nearly equal numb er. Although more successful than previous mo dels, the most recent realizations of semi-analytic hierarchical merging simulations still severely underpredict the density of such old galaxies: just one old galaxy with Ks < 20, R - Ks > 6, and z > 1.5 is present in the mo ck catalog4 for the whole five times wider GOODS/CDFS area. As exp ected for early-typ e galaxies9,24 , the three galaxies at z 1.61 may trace the underlying large scale structure. In this case, our estimated numb er density may b e somewhat biased toward a high value. On the other hand, the numb er of such galaxies in our sample is likely to b e a lower limit due to the sp ectroscopic redshift incompleteness. There are indeed up to three more candidate old galaxies in the GOODS/K20 sample with 18.5 Ks 19.5, 1.5 zphot 2.0, 5.6 R - Ks 6.8 and compact HST morphology. Thus, in the GOODS/K20 sample the fraction of old galaxies among the whole z > 1.5 galaxy p opulation is 15±8% (sp ectroscopic redshifts only), or up to 25±11% if also all the 3 additional candidates are counted. It is generally thought that the so-called "redshift desert" (i.e. around 1.4 < z < 2.5) represents the cosmic ep o ch when most star formation activity and galaxy mass assembly to ok place25 . Our results show that, in addition to actively star forming galaxies26 , also a substantial numb er of "fossil" systems already p opulate this redshift range, and hence remain undetected in surveys biased towards star-forming systems.

3


The luminous star-forming galaxies found at z > 2 in sub-mm27 and near-infrared surveys may represent the progenitors of these old and massive systems.

23,28

1. Fukugita, M., Hogan, C.J., Peebles, P.J.E. The Cosmic Baryon Budget. Astrophys. J. 503, 518-530 (1998). 2. Renzini, A. Origin of Bulges. In "The formation of galactic bulges", ed. C.M. Carollo, H.C. Ferguson, R.F.G. Wyse, Cambridge University Press, p.9-26 (1999). 3. Peebles, P.J.E. When did the Large Elliptical Galaxies Form? In "A New Era in Cosmology", ASP Conference Pro ceedings, Vol. 283. ed. N. Metcalfe and T. Shanks, Astronomical So ciety of the Pacific, 2002., p.351-361 (2002). 4. Somerville, R.S. et al. The Redshift Distribution of Near-Infrared-Selected Galaxies in the Great Observatories Origins Deep Survey as a Test of Galaxy Formation Scenarios. Astrophys. J., 600, L135-139 (2004). 5. Freedman, W.L. & Turner, M.S. Collo quium: Measuring and understanding the universe. Reviews of Mo dern Physics, 75, 1433-1447 (2003). 6. Im, M. et al. The DEEP Groth Strip Survey. X. Numb er Density and Luminosity Function of Field E/S0 Galaxies at z < 1. Astrophys. J. 571, 136-171 (2002). 7. Cimatti, A. et al. The K20 survey. I. Disentangling old and dusty star-forming galaxies in the ERO p opulation. Astron. Astrophys. 381, L68-73 (2002). 8. Benitez, N. et al. Detection of Evolved High-Redshift Galaxies in Deep NICMOS/VLT Images. Astrophys. J. 515, L65-69 (1999). 9. Daddi, E. et al. Detection of strong clustering of extremely red ob jects: implications for the density of z > 1 ellipticals. Astron. Astrophys. 361, 535-549 (2000). 10. Zepf, S.E. Formation of elliptical galaxies at mo derate redshifts. Nature 390, 377-380 (1997). 11. Ro dighiero G., Franceschini A., Fasano G. Deep Hubble Space Telescop e imaging surveys and the formation of spheroidal galaxies. Mon. Not. R. Astron. So c. 324, 491-497 (2001). 12. Taniguchi, Y. et al. Lyman Emitters b eyond Redshift 5: The Dawn of Galaxy Formation. Journal of the Korean Astronomical So ciety 36, no.3, 123-144 (2003). 13. Dunlop, J.S. et al. A 3.5-Gyr-old galaxy at redshift 1.55. Nature, 381, 581-584 (1996). 14. Spinrad, H., Dey, A., Stern, D., Dunlop, J., Peaco ck, J., Jimenez, R., Windhorst, R. LBDS 53W091: an Old, Red Galaxy at z=1.552. Astrophys. J., 484, 581-601 (1997). 15. Gavazzi, G., Pierini, D., Boselli, A., The phenomenology of disk galaxies. Astron. Astrophys. 312, 397-408 (1996). 16. Cimatti, A. et al. The K20 survey. I I I. Photometric and sp ectroscopic prop erties of the sample. Astron. Astrophys. 392, 395-406 (2002). 17. Giavalisco, M. et al., The Great Observatories Origins Deep Survey: Initial Results from Optical and Near-Infrared Imaging. Astrophys. J. 600, L93-98 (2004). 18. Pickles, A.J. A Stellar Sp ectral Flux Library: 1150-25000 ° PASP 110, 863-878 A. (1998). 19. Bruzual, G. & Charlot, S. Stellar p opulation synthesis at the resolution of 2003. Mon. Not. R. Astron. So c. 344, 1000-1028 (2003). 4


20. Jimenez, R. et al. Synthetic stellar p opulations: single stellar p opulations, stellar interior mo dels and primordial proto-galaxies, Mon. Not. R. Astron. So c. 349, 240254 (2004). 21. Eisenstein, D.J. et al. Average Sp ectra of Massive Galaxies in the Sloan Digital Sky Survey. Astrophys. J. 585, 694-713 (2003). 22. Cole, S. et al. The 2dF galaxy redshift survey: near-infrared galaxy luminosity functions. Mon. Not. R. Astron. So c., 326, 255-273 (2001). 23. Daddi, E. et al., Near-Infrared Bright Galaxies at z 2. Entering the Spheroid Formation Ep o ch ? Astrophys. J., 600, L127-131 (2004). 24. Davis, M., Geller, M.J. Galaxy Correlations as a Function of Morphological Typ e. Astrophys. J., 208, 13-19 (1976). 25. Dickinson, M., Pap ovich, C., Ferguson, H.C., Budavari, T. The Evolution of the Global Stellar Mass Density at 0 < z < 3. Astrophys. J., 587, 25-40 (2003). 26. Steidel, C.C. et al. A Survey of Star-Forming Galaxies in the z=1.4-2.5 `Redshift Desert': Overview. Astrophys. J. 604, 534-550 (2004). 27. Genzel, R., Baker, A.J., Tacconi, L.J., Lutz, D., Cox, P.; Guilloteau, S., Omont, A. Spatially Resolved Millimeter Interferometry of SMM J02399-0136: A Very Massive Galaxy at z = 2.8. Astrophys. J. 584, 633-642 (2003). 28. Franx, M. et al. A Significant Population of Red, Near-Infrared-selected HighRedshift Galaxies. Astrophys. J., 587, L79-L83 (2003). 29. Pignatelli, E. & Fasano, G. GASPHOT: A To ol for Automated Surface Photometry of Galaxies. Astrophys. Sp. Sci. 269, 657-658 (1999). 30. Peng, C.Y., Ho, L.C., Imp ey, C.D., Rix, H.-W., Detailed Structural Decomp osition of Galaxy Images. Astron. J., 124, 266-293 (2002). Corresp ondence and requests for material should b e sent to Andrea Cimatti (cimatti@arcetri.astro.it). This work is based on observations made at the Europ ean Southern Observatory, Paranal, Chile, and with the NASA/ESA Hubble Space Telescop e obtained at the Space Telescop e Science Institute, which is op erated by the Asso ciation of Universities for Research in Astronomy (AURA). We acknowledge Rachel Somerville for information on the GOODS/CDFS mo ck catalog and the referees for useful and constructive comments.

5


Figure 1 The individual and average spectra of the detected galaxies. From b ottom to top: the individual sp ectra smo othed to a 16 ° b oxcar (26 ° for ID 237) and the average A A sp ectrum of the four old galaxies (zaverage = 1.68). The red line is the sp ectrum of the old galaxy LBDS 53w091 (z = 1.55) used to search for sp ectra with a similar continuum shap e. Weak features in individual sp ectra (e.g. MgI I2800 and the 2640 ° continuum break, B2640) b ecome clearly visible in the average sp ectrum. A The ob ject ID 235 has also a weak [OI I]3727 emission (not shown here). The sp ectra were obtained with ESO VLT+FORS2, grisms 200I (R(1 ) 400) (ID 237) and 300I (R(1 ) 600) (IDs 235,270,646), 1.0 wide slit and 1 seeing conditions. The integrations times were 3 hours for ID 237, 7.8 hours for IDs 235 and 270, and 15.8 hours for ID 646. For ID 646, the ESO/GOODS public sp ectrum was co-added to our K20 sp ectrum (see Supplementary Tab. 1). "Dithering" of the targets along the slits was applied to remove efficiently the CCD fringing pattern and the strong OH sky lines in the red. The data reduction was done with the IRAF software package (see16 ). The sp ectrophotometric calibration of all sp ectra was achieved and verified by observing several standard stars. The average sp ectrum, corresp onding to 34.4 hours integration time, was obtained by co-adding the individual sp ectra convolved to the same resolution, scaled to the same arbitrary flux (i.e. with each sp ectrum having the same weight in the co-addition), and assigning wavelength­dep endent weights which take into account the noise in the individual sp ectra due to the OH emission sky lines. Figure 2 The detailed average spectrum of the detected galaxies. A zo om on the average sp ectrum (blue) compared with the synthetic sp ectrum19 of a 1.1 Gyr old simple stellar p opulation (SSP) with solar metallicity (Z = Z ) and Salp eter IMF (red). The observed average sp ectrum was compared to a library of synthetic SSP template sp ectra19,20 with a range of ages of 0.1-3.0 Gyr with a step of 0.1 Gyr, and with assumed metallicities Z =0.4â, 1.0â, and 2.5âZ . The b est fit age for each set of synthetic templates was derived through a 2 minimization over the rest-frame wavelength range 2300­3400 ° The rms as a function of wavelength used in the 2 pro cedure A. was estimated from the average sp ectrum computing a running mean rms with a step of 1 ° and a b ox size of 20 ° corresp onding to ab out three times the resolution of the A A, observed average sp ectrum. The median signal-to-noise ratio is 20 p er resolution element in the 2300­3400 ° range. The wavelength ranges including the strongest A real features (i.e. absorptions and continuum breaks) were not used in the estimate of the rms. The resulting reduced 2 is of the order of unity for the b est fit mo dels. In the case of solar metallicity, the ranges of ages acceptable at 95% confidence level . are 1.0+0.5 Gyr and 1.4+0.5 Gyr for SSP mo dels of19 and20 resp ectively (see also Fig. -0.1 -0 4 3, top panel). Ages 50% younger or older are also acceptable for Z = 2.5Z or A A Z = 0.4Z resp ectively. The 2640 ° and 2900 ° continuum break13 amplitudes measured on the average sp ectrum are B2640=1.8±0.1 and B2900=1.2±0.1. These values are consistent with the ones exp ected in SSP mo dels19-20 for ages around 1­1.5 Gyr and solar metallicity. For instance, the SSP mo del sp ectrum shown here has B2640=1.84 and B2900=1.27. 6


Figure 3 The comparison between the average spectrum and a set of spectral templates. The average sp ectrum (blue) compared to a set of template sp ectra. From b ottom: F2 V (green) and F5 V (red) stellar sp ectra18 with Z = Z , the comp osite sp ectrum (red) of 726 luminous red galaxies at 0.47 < z < 0.55 selected from the SDSS21 (available only for > 2600 ° the average sp ectra of z 1 old (red) and dusty star-forming A), 7 (green) EROs , SSP synthetic sp ectra19 (Z = Z , Salp eter IMF) with ages of 0.5 Gyr (magenta), 1.1 Gyr (green) and 3.0 Gyr (red). Figure 4 The morphological properties of the detected galaxies. Images of the four galaxies taken with the Hubble Space Telescop e +ACS through the F850LP filter (from GOODS data17 ) which samples the rest-frame 3000-3500 ° for 1.6 < z < 2. The imA ages are in logarithmic grey­scale and their size is 2 â 2 , corresp onding to 17 â 17 kp c for the average redshift z = 1.7 and the adopted cosmology. At a visual insp ection, the galaxies show rather compact morphologies with most of the flux coming from the central regions. A fit of their surface brightness profiles was p erformed with a "Sersic law" ( r 1/n ) convolved with the average p oint spread function extracted from the stars in the ACS field and using the GASPHOT29 and GALFIT30 software packages. Ob jects ID 237 and ID 646 have profiles with acceptable values of n in the range of 4 < n < 6, i.e., typical of elliptical galaxies, ob ject ID 270 is b etter repro duced by a flatter profile (1 < n < 2), whereas a more ambiguos result is found for the ob ject showing some evidence of irregularities in the morphology (ID 235, 1 < n < 3). These latter ob jects may b e bulge-dominated spirals but no bulge/disk decomp osition was attempted. Ground-based near-infrared images taken under 0.5 seeing conditions with the ESO VLT+ISAAC through the Ks filter (rest-frame 6000-8000 ° show A) very compact morphologies, but no surface brightness fitting was done.

7


646 (1.903)

237 (1.615)

235 (1.610)

270 (1.605)

Figure 1:

8


Figure 2:

9


Figure 3:

10


Figure 4:

11