Ages and metallicities of faint red galaxies in the Shapley Supercluster
aa r X i v : . [ a s t r o - ph ] D ec Proceedings Title IAU SymposiumProceedings IAU Symposium No. 245, 2007M. Bureau, E. Athanassoula, and B. Barbuy, eds. c (cid:13) Ages and metallicities of faint red galaxies inthe Shapley Supercluster
Russell J. Smith , John R. Lucey and Michael J. Hudson Department of Physics, University of Durham, United Kingdom Department of Physics and Asronomy, University of Waterloo, Canada
Abstract.
We present results on the stellar populations of 232 quiescent galaxies in the ShapleySupercluster, based on spectroscopy from the AAOmega spectrograph at the AAT. The keycharacteristic of this survey is its coverage of many low-luminosity objects ( σ ∼
50 km s − ), withhigh signal-to-noise ( ∼
45 ˚A − ). Balmer-line age estimates are recovered with ∼
25% precisioneven for the faintest sample members. We summarize the observations and absorption line data,and present correlations of derived ages and metallicities with mass and luminosity. We highlightthe strong correlation between age and α -element abundance ratio, and the anti-correlation ofage and metallicity at fixed mass, which is shown to extend into the low-luminosity regime.
1. Introduction
The evolution of galaxy populations, subsequent to the peak of cosmic star formationat z ∼
2, has been driven by a progressive extinguishing of activity as galaxies lose orexhaust their gas. Recent surveys of galaxies at redshifts up to z ∼ today’s Red Sequence galaxies. In par-ticular, comparison of observed metal and Balmer line strengths against model predic-tions can yield estimates of stellar population “age”, weighted towards the most recentstar-formation episode. Until recently, high signal-to-noise (S/N) absorption-line datawere limited to fairly small samples of galaxies, and generally to giant elliptical galaxies(e.g. Kuntschner et al. 2001). Samples based on much larger spectroscopic surveys (e.g.Nelan et al. 2005; Bernardi et al. 2006; Graves et al. 2007) generally do not have suf-ficient S/N to measure ages for individual galaxies, except at the highest luminosities.Nonetheless, these surveys do recover an average trend of younger ages for galaxies withlower σ , which is the expected signature of downsizing in the low-redshift fossil record.Downsizing implies that today’s faint Red Sequence galaxies are of crucial importanceto understanding the evolution of galaxy populations at z <
1. Some early progressin measuring galaxy ages in the low-luminosity regime was made by Mobasher et al.(2001) and Caldwell, Rose & Concannon (2003). Here, we present new results from a1 Smith, Lucey & Hudsoncomprehensive spectroscopic survey of galaxies in the Shapley Supercluster ( z = 0 .
2. AAOmega observations and the index- σ relations In April 2006, we obtained spectra for a total of 416 galaxies in the Shapley Super-cluster, forming a magnitude-limited (
R <
18) sample from the photometric catalogue ofthe NOAO Fundamental Plane Survey (NFPS, Smith et al 2004). The sample covers thecentral 40 ×
40 arcmin region in each of the three clusters Abell 3556, Abell 3558 andAbell 3562. Two AAOmega fibre configurations were employed, with ∼ σ <
100 km s − ) is ∼
45 per ˚A, (at rest-frame4400–5400 ˚A). For galaxies with σ >
100 km s − , the median S/N is ∼
90 per ˚A.Redshifts and velocity dispersions were measured using standard cross-correlationmethods. Emission line equivalent widths were measured after removing a best-fitingstellar continuum model. The Lick absorption line indices (Trager et al. 1998), weremeasured on the blue-arm spectra, and converted to the Lick system using resolutiontransformations based on model SSP spectra, to avoid smoothing the observed spectra.Our stellar population analysis is restricted to supercluster members using the mea-sured redshift, and to quiescent galaxies defined by their H α emission, EW(H α ) < . ∼
30 galaxies have mea-sured velocity dispersions consistent with zero and are shown at an arbitrary σ at theleft of the figure. Index − σ slopes are fitted to the ∼
200 galaxies with measured σ .The slopes of the index − σ relations can be used to infer the scaling relations followedby the stellar population parameters (e.g. Nelan et al. 2005; Smith et al. 2006). Forthe Thomas et al. (2003) models, the derived parameters are age, total metallicity, [Z/H]and α -element abundance ratio, [ α /Fe], estimated by comparison to predictions for simplestellar populations (SSPs). Fitting for the slopes of nine index − σ relations, in comparisonto the these models, we recover: Age ∝ σ . ± . , Z / H ∝ σ . ± . , and α/ Fe ∝ σ . ± . . The quoted errors include a (dominant) contribution from sytematic errors,estimated using a large set of different indices in the fitting process.The age– σ scaling relation is consistent with that obtained by Nelan et al. (2005) fromsimilar analysis of the NFPS, and implies that even if the most massive red galaxiesformed at very high redshift, much of the faint red population became quiescent only atrecent epochs ( z < . quantitatively with the observed evolution in theRed Sequence luminosity function (Stott et al. 2007; De Lucia et al. 2007).
3. The Age–Metallicity–Mass relations
To move beyond the average scaling relations, we use stellar population models todetermine the values of age, [Z/H] and [ α /Fe] which reproduce a non-degenerate set ofobserved indices, for each galaxy individually. The results shown here are derived from ge and metallicities of faint red galaxies log ( s ) H g F − − log ( s ) M gb5177 log ( s ) F e5015 Figure 1.
The index- σ relations for selected indices. Objects to the left of the dashed line havevelocity dispersions consistent with zero, and are not used in fitting the relations. Open symbolsrepresent these and also other galaxies excluded from the fit by an iterative outlier rejection.The dotted line at σ = 100 km s − shows the typical low-mass limit of many other studies. log ( s ) l og ( A ge ) . . . . log ( L R L sun ) [ Z / H ] − . . . log(Age) [ a F e ] − . . . . . Figure 2.
Left: Ages versus velocity dispersion. Centre: total metallicity versus luminosity.Right: [ α /Fe] versus age. The SSP-equivalent parameters were derived by comparison to modelsof Thomas et al. (2003). Galaxies with unresolved velocity dispersions are shown by open symbolsand placed at arbitrary position in the left panel. fitting to the models of Thomas et al. (2003), using indices H γ F , Fe5015 and Mgb5177.For the faintest galaxies in the sample ( L R ∼ . L ⊙ ), the median formal errors are 25%in age, 0.08 dex in Z/H and α /Fe. Some key correlations of the derived parameters areshown in Figure 2. Intriguingly, we find that although age is correlated more tightly with σ than with luminosity, metallicity is better correlated with luminosity. The abundanceratio [ α /Fe] is well-correlated with σ and also follows a tight age–[ α /Fe] relation. Thelatter suggests that “younger” galaxies formed stars over more extended periods, broadlyconsistent with a quenching scenario rather than formation in a single rapid burst.Figure 3 shows the relationship between age and metallicity for three intervals in σ .As expected from the previous figure, both age and [Z/H] increase with increasing σ , onaverage. However, at fixed σ , they are anti-correlated: galaxies which are younger thanaverage for their mass are also more metal-rich than average. The sample populates aplane (the “Z-plane”) described by [Z / H] = 0 . ± .
04 log σ − . ± .
04 log Age − . ± . , with scatter ∼ Figure 3.
The Age–Metallicity–Mass relations. The black points and red ellipses show theestimated and 1 σ confidence intervals for galaxies in each of three intervals in velocity dispersion.The rest of the sample is shown in grey/yellow for reference. The arrow indicates the directionin which the median age and metallicity move with increasing mass. The heavy black line (samein all panels) indicates the direction of the age–metallicity degeneracy: movement parallel tothis line generates no change in galaxy broadband colours. ness of the tight colour–magnitude sequence is partly a consequence of this “conspiracy”.Note that although the error ellipses are oriented parallel to the intrinsic dispersion, themeasurement errors can account for only ∼
20% of the observed variance along the plane.The existence and slope of the age–metallicity anti-correlation at fixed mass is crucialto understanding the star-formation history of the galaxies prior to quenching. In Smith,Lucey & Hudson (2007c, in preparation), we describe a simple self-enrichment model forthe pre-cursors of today’s Red Sequence galaxies, and show that the rate of metallicitygrowth required to fit the Z-plane agrees quantitatively with the evolution observed inthe metallicity–mass relation for star-forming field galaxies (e.g. from Erb et al. 2006).
Acknowledgements
RJS is supported under the STFC rolling grant PP/C501568/1 “Extragalactic Astron-omy and Cosmology at Durham 2005–2010”.
References
Bell, E.F., et al. 2004,
ApJ , 608, 752Bernardi, M., Nichol, R.C., Sheth, R.K., Miller, C.J., & Brinkmann, J. 2006, AJ , 131, 1288Caldwell, N., Rose, J.A., & Concannon, K.D. 2003, AJ , 125, 2891De Lucia, G., et al. 2007, MNRAS , 374, 809Erb, D.K., Shapley, A.E., Pettini, M., et al. 2006,
ApJ , 644, 813Faber, S.M., et al. 2007,
ApJ , 665, 265Graves, G.J., Faber, S.M., Schiavon, R.P., & Yan, R. 2007,
ApJ , in press, arXiv:0707.1523Kuntschner, K., Lucey, J.R., Smith, R.J., et al. 2001,
MNRAS , 323, 615Mobasher, B., et al. 2001,
ApJS , 137, 279Nelan, J.E., Smith, R.J., Hudson, M.J., et al. 2005,
ApJ , 632, 137Smith, R.J., et al. 2004, AJ , 128, 1558Smith, R.J., Hudson, M.J., Lucey, J.R., et al. 2006, MNRAS , 369, 1419Smith, R.J., Lucey, J.R., & Hudson, M.J. 2007a,
MNRAS , 381, 1035Stott, J.P., Smail, I., Edge, A.C., et al. 2007,
ApJ , 661, 95Thomas, D., Maraston, C. & Bender 2003 2003,