Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.arcetri.astro.it/~sperello/astroph0607325.pdf Äàòà èçìåíåíèÿ: Wed Sep 12 19:17:03 2007 Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 03:19:50 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: http www.stsci.edu science starburst

Mass and Environment Drive the Evolution of Early­Typ e Galaxies
Sperello di Serego Alighieri INAF ­ Osservatorio Astrofisico di Arcetri, Largo E. Fermi 5, 50125 Firenze, Italy Barbara Lanzoni INAF ­ Osservatorio Astronomico di Bologna, Via Ranzani 1, 40127 Bologna, Italy
New information has recently b ecome available on the fundamental plane for various samples of early-typ e galaxies with redshift up to 1.3, b oth in clusters and in the field. This information is reviewed and clues are derived on the evolution of spheroids over the last two thirds of the Universe lifetime as a function of galaxy mass and environment, in comparison with the predictions of the hierarchical models of galaxy formation. Using the Universe as a time-machine and interpreting changes in M/L ratio as age differences, we see that the age increases with galaxy mass in all environments, cluster galaxies with any mass are older than field galaxies with the same mass, and the age difference b etween cluster and field galaxies increases with mass. The first two results confirm those obtained with other methods, and are reproduced by the most recent incarnation of the hierarchical models, while the third result is new and app ears in contrast with the predictions of these models.

arXiv:astro-ph/0607325v1 14 Jul 2006

1

Intro duction

Early­typ e galaxies (ETG) contain most of the visible mass in the Universe 1 and are thought to reside in the highest density p eaks of the underlying dark matter distribution. Therefore, understanding their evolution is crucial for understanding the evolution of galaxies and structures in general. ETG have the very interesting prop erty that in the 3D space of their main parameters (the effective radius Re , the central velocity disp ersion , and the average surface luminosity 2 within Re , I e = L/2 Re ) they concentrate on a plane, therefore called the fundamental plane 2,3 ). This implies a striking regularity in their structure and stellar p opulation 4 , which calls (FP for a uniform process of formation and evolution. Studies of the FP at high redshift allow for a check on the p ersistence of this regularity back in time and offer the p ossibility to derive more stringent constraints on the formation history of ETG, using the Universe as a time-machine. We have therefore made a uniform comparison of the b est data available on the FP at z 1, the highest redshift for which these data are currently available, and, by interpreting the differences in M/L ratio directly derived from the FP parameters as age differences, we study how the ages of ETG dep end on b oth galaxy mass and environment. We assume a flat Universe with m = 0.3, = 0.7, and H0 = 70 km s-1 Mp c-1 , and we use magnitudes based on the Vega system.


2

The data on the FP at z 1

In selecting the data on the FP at z 1 we pay particular attention to the completeness of the samples to faint optical magnitudes, b ecause the galaxies with corresp ondingly small masses are those likely to show "downsizing" effects, i.e. the later and/or longer lasting formation of smaller mass galaxies 5 , and differences with environment. In fact it has b een shown that massive galaxies in the field b ehave similarly to massive galaxies in the clusters 6 . We also consistently use data in the rest­frame B band. In the field the only ETG galaxy sample complete to faint magnitudes at z 1, for which FP studies have b een made, is the one derived from the K20 survey 6 , which is complete down to MB = -20.0, and covers the redshift range 0.88 < z < 1.3 with 15 ETG over a total sky solid angle of 52 arcmin2 . We have used also the sample of Treu et al.7 , which, although not complete, contains 24 ETG in the same redshift range over 160 arcmin2 in the GOODS-N area, some of which are as faint as the faintest galaxies of di Serego Alighieri et al.6 . We have also examined the data presented by van der Wel et al.8 , who have studied the FP on a sample of 27 field ETG in the range 0.6 < z < 1.15 in two Southern fields, but eventually were forced not to include them in our analysis by the imp ossibility of understanding how they have selected the 13 primary ETG plotted in their figures from the 27 galaxies in their sample. Until very recently there was no sample of cluster ETG at high redshift with FP information and complete to faint magnitudes 9 . This is probably due to an observational selection effect, since on the clusters the slits of a single mask in a MOS sp ectrograph are easily filled with bright galaxies, and none is left for the fainter ones. We were very lucky that shortly b efore this Conference a FP study 10 of two clusters at z = 0.835 and z = 0.892, complete down to MB = -19.8, has b een published, and, although the data for the individual galaxies are not yet published, Inger JÜrgensen has kindly made them available to us for the present analysis. We have uniformly analized these three sets of data 6,7,10 and we show in Fig. 1 the corresp onding edge­on FP, compared to a recent analysis of the Coma cluster ETG 10 , where galaxies with M < 1010.3 M and emission­line galaxies have b een excluded. It is evident from Fig. 1 that the evolution of the FP consists b oth of a vertical shift and of a rotation, in the sense that the FP at z 1 app ears steep er that the local one, b oth in the field and in the clusters. As has b een shown b efore 6 , the change in the FP tilt is a necessary manifestation of "downsizing". Notwithstanding this evolution, the FP has a remarkably small scatter also at z 1. 3 Deriving ages from the FP parameters

It is customary and straightforward to interpret the evolution of the FP as changes in the M/L ratio. In fact the FP parameters can b e used to estimate the dynamical mass of the galaxies. For instance, from the Virial theorem and assuming R1/4 homology, the mass is given by 11 : M = 5Re 2 /G. (1)

Then the M/LB ratio can b e obtained for each high redshift galaxy and compared to that of local galaxies with the same mass, as shown in Fig. 2. Evidently the difference in the M/L ratio b etween a local ETG and one at z 1 with the same mass is larger for smaller mass galaxies, b oth in the field and in the clusters, and seems smaller in the cluster than in the field. However the redshift of the two clusters is smaller that the average redshift of the field galaxies: therefore a deep er analysis is necessary to ensure that the differences with environment are not due just to redshift differences. Usually this is achieved by taking for each high redshift ETG the logarythmic difference in M/L ratio (log(M/LB )) from that of a local galaxy with the same mass and comparing it to the straight line fit to the


Figure 1: Th at z 1 from squares), for at z = 0.835 to the Coma

e edge­on Fundamental Plane for local ETG in the Coma Cluster (empty triangles), for field ETG the K20 survey b oth for the CDFS field (filled black circles) and for the Q0055 field (filled black field ETG at z 1 in the GOODS area (filled black triangles), and for the ETG in two clusters (red filled squares) and at z = 0.892 (red filled triangles). The dashed line is the b est fit plane cluster galaxies. Compared to the local one, the FP at high redshift is offset and rotated in all environments.


Figure 2: The M/L ratio in the B-band as a function of the galaxy mass for the ETG samples given in Figure 1 (same symb ols). The dotted line is a fit to the Coma ETG. The changes in M/LB from high redshift to today decrease with galaxy mass in all environments and are larger in the field than in the clusters.

log(M/LB ) of ETG with M > 1011 M in clusters with redshift ranging from 0 to 1.3 9 . However this analysis is unsatisfactory, b ecause it is not clear why the massive cluster galaxies should b e taken as a reference, and b ecause it excludes the lower mass cluster galaxies from the analysis. We have therefore devised a different analysis with the aim of studying how the formation history of ETG dep ends on galaxy mass and environment. We interpret the changes in M/L ration as differences in age. Maraston 12 has develop ed simple stellar p opulation synthesis models that allow one to estimate the M/L ratio as a function of age (Fig. 3). By fitting this relation, e.g. for solar metallicity, nd for a Kroupa 13 we can therefore get the age (at z = 0) of each galaxy from the observed log(M/LB ) for our assumed cosmology. The results are shown in Figure 4, compared to the median ages derived from a semianalytic model of hierarchical galaxy evolution 14 . ETG age increases with mass (downsizing) b oth in the field and in the clusters. Field galaxies are younger than cluster galaxies with the same mass. The age difference b etween cluster and field galaxies with the same mass increases with mass. The first two results have b een obtained b efore by other means and are reproduced by the most recent incarnation of the hierachical models. The third result is new and app ears to go in the opp osite way as predicted by the hierarchical models (see fig. 1 of De Lucia et al. 14 ). Note that after the the metallicity for each synthesis model approp also analysed selection our conclusions are not Conference we have refined our estimate of galaxy ages, by evaluating ETG from the velocity disp ersion and then direclty using the p opulation riate for the derived metallicity 12 without any fitting involved. We have effects. The resulting ages are very similar to those presented here and affected 15 .


Figure 3: The relationship b etween age and M/L ratio obtained by Maraston (2005) for a simple stellar p opulation with solar metallicity. The dashed line shows the linear fit that we have used.

Acknowledgments We would like to thank Inger JÜrgensen for allowing us to use her still unpublished data, and Claudia Maraston and Alvio Renzini for useful suggestions and discussions. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] Renzini, A., ARA&A, in press (2006, also astro-ph/0603479) Djorgovski, S. & Davis, M., ApJ 313, 59 (1987) Dressler, A. et al., ApJ 313, 42 (1987) Renzini, A. & Ciotti, L., ApJ 416, L49 (1993) Cowie, L.L., Songaila, A., Hu, E.M., & Cohen, J.G., AJ 112, 839 (1996) di Serego Alighieri, S. et al., A&A 442, 125 (2005) Treu, T. et al., ApJ 633, 174 (2005) van der Wel, A. et al., ApJ 631, 145 (2005) van Dokkum, P. & Stanford, S.A., ApJ 585, 78 (2003) JÜrgensen, I. et al., ApJLett 639, L9 (2006) Michard, R., A&A 91, 122 (1980) Maraston, C., MNRAS 362, 799 (2005) Kroupa, P., MNRAS 322, 231 (2001) De Lucia, G., Springel, V., White, S.D.M., Croton, D. & Kauffmann, G., MNRAS 366, 499 (2006) [15] di Serego Alighieri, S., Lanzoni, B. & JÜrgensen, I., ApJLett, in press (2006, also astroph/0607231)


Figure 4: The formation ep och of the ETG from the samples given in Figure 1 (same symb ols), evaluated as explained in the text. The continuous line shows the median model ages obtained by De Lucia et al. (2006) from a semianalytic model of hierarchical galaxy formation, while the dashed lines are their upp er and lower quartiles. More massive ETG form earlier in all environments, and cluster ETG are older than those in the field at all masses, while their age difference app ears to increase with galaxy mass. We stress that, although the absolute ages that we derive are somewhat model dep endent and reach a few absurdly high values, the trends of age differences b etween high redshift and local ETG, and b etween galaxies with different masses and in different environments are much more robust.