SPHEROIDAL PANCHROMATIC INVESTIGATION IN

DIFFERENT ENVIRONMENTAL REGIONS(SPIDER)

F. La Barbera(1); R.R. de Carvalho(2); I.G. de La Rosa(3); P.A. Lopes(4); I. Ferreras(5); R.R. Gal(6); H.V. Capelato(2)

(1) INAF-OAC; (2) INPE-DAS, Sao José dos Campos, Brazil; (3) IAC, Tenerife, Spain; (4) OV/UFRJ, Rio de Janeiro, Brazil; (5) University College London; (6) Institute for Astronomy, Hawaii, USA

Early-type galaxies (ETGs)

Spheroidal systems contribute with a significant fraction of the total stellar mass in the local universe

number density

stellar mass function

fraction of dry mergers

stellar population content(age, metallicity, .....)

size evolution

Environment might play a crucial role

Constraining their observed properties may help understand better the hierarchical scenario of galaxy formation and evolution

Constraining the formation/evolution (of ETGs)

Paleontology of galactic properties Studying the galaxy’s progenitors

Large samplesBetter data

Small differences Smaller samples Less data

Large(r) differences

Early-type galaxies (ETGs)

Low redshift High redshiftReference point

The SDSS+UKIDSS dataset

SDSS-DR7 (u=22.0,g=22.2, r=22.2, i=21.3, z=20.5)

UKIDSS-Large Area Survey (LAS; Y=20.5, J=20, H=18.8, K=18.4)

THE SPIDER

SDSS+UKIDSS photometric system (overall throughput curves)

The SDSS+UKIDSS dataset

griz+YJHK: advantages

NIR data describe the old, quiescent stellar population in galaxies, hence following more closely the mass distribution

NIR bands are much less sensitive to metallicity (through line-blanketing), and less sensitive to age than optical data

Using grizYJHK bands, we minimize the age-metallicity degeneracy

true value

true value

GOALS

Establishing the waveband dependence of the scaling relations of ETGs (Faber-Jackson, Kormendy relations, Fundamental Plane)

Measuring the variation of stellar population properties INSIDE galaxies and along the mass sequence of ETGs

…….. as a function of the environment - characterized by either local density (potential) or global properties (halo mass)

Systematic comparison of different approaches/techniques to measure stellar population properties (e.g. diagnostic diagrams of spectral indices, spectral fitting, PCA spectral analysis, SED fitting of the grizYJHK photometry)

Sample selection and

galaxy parameters

z≥0.05; minimizing the aperture bias (Gomez et al. 2003, ApJ 584)

z≤0.095; where Mr matches the SDSS spectroscopic completeness limit (r*~17.8)

We select a volume-limited sample of early-type galaxies from SDSS-DR6 as in Miller et al. (2003, ApJ 586), and Sorrentino et al. (2006, A&A 460)

spectroscopy available

Mr<-20 (~ separation between ordinary and bright spheroids; Capaccioli et al. 1992, MNRAS 259)

The ETG’s sample

Velocity dispersion available, with 70≤σ0≤420 km s-1 and zWarn=0

early-type galaxies

eclass<0, FracDevr>0.8 (as in Bernardi et al. 2003, AL 125,

1849)

The ETG’s sample

The optical sample of 39,993 ETGs is matched to UKIDSS-LAS DR4, resulting into an sample of 5,080 optical+NIR ETGs

Galaxy parameters - 2DPHOT

SDSS r band

Re, <μ>e, Sersic index n, b/a, disky/boxy parameter (a4)

g r i z Y J H K

low 2 (<1.5)

high 2 (<1.5)

Examples of 2D fitting from g through K

Galaxy parameters - 2DPHOT

The SDSS spectra are re-analyzed with the software STARLIGHT (Cid Fernandes et sl. 2005, MNRAS 358)

Example of STARLIGHT spectral fitting

STARLIGHT provides the linear combination of SSP models, that broadened with a given σ0, best matches the observed spectrum

The SSP models are those of the new high resolution (2.3Å FWHM) α-enhanced MILES library.

Galaxy parameters - spectra

Residual spectrum revealing the Hβ nebular emission.

Regions possibly contaminated by nebular emission are masked out in the fit

The environment

We use the largest group/Cluster catalogue generated from SDSS at low redshift (z<0.1; 5162 groups at z>0.05) using a 3D FoF algorithm as in Berlind et al. 2006, ApJSS 167

The sample of ETGs is also matched to the catalogue of Compact Groups (CGs) from McConnachie et al. 2009, MNRAS 395. 400 Compact Groups have at least one ETG member.

25 Fossil Groups from the largest homogenous sample available at low redshift (La Barbera et al. 2009, AJ 137)

The environment

Cluster propertiesFor each group of the FoF catalogue, we re-select the group members (from SDSS-DR7), and re-estimate velocity dispersion (σcl), physical radius (R200), and mass (M200) as in Lopes

et al. 2009, MNRAS 392. For each group,

we re-estimate the central redshift and velocity limits by the gap-technique (Katgert et

al. 1996, A&A 310; Adami et al. 1998, A&A 331; Olsen et al. 2005, A&A 435);

the quantities σcl, R200, M200 are measured using the virial analysis as in Girardi et al.

1998, ApJ 505; Biviano et al. 2006, A&A 456 ; Popesso et al. 2007, A&A 464.

member galaxies are identified by the shifting gapper technique (Fadda et al. 1996,

ApJ 473);

Group members (points); interlopers (circles)

we flag those groups with substructures detected (at the 5% level) by the (3D) Δ test (Dressler & Shectman 1988, AJ 95)

Cluster properties

Distribution of ETGs with respect to local galaxy density. The sample covers three orders of magnitude in ΣN.

For a given group, local galaxy density is estimated by using only the (projected) distribution of its member galaxies.

We define the local density, ΣN, as N/(πdN

2), where N is the square root of the number of group members.

This ensures that dN scales with cluster mass.

Cluster properties

Distribution of ETGs with respect to the mass of the parent cluster where they reside. The sample covers almost two orders of magnitude in M200.

Example of a rich cluster in the updated FoF catalogue. Velocity dispersion is 1010 km s-1, while cluster mass is 2.7×1015 Msun. Red squares mark the objects with spectra available from SDSS-DR7 in the cluster field (not necessarily cluster members).

HIGH DENSITY-HIGH MASS

INTERMEDIATE MASS

Example of a group in the updated FoF catalogue. Velocity dispersion is 520 km s-1, while cluster mass is 5×1014 Msun. Red squares mark the objects with spectra available from SDSS-DR7 in the cluster field.

HIGH DENSITY-LOW MASS

Examples of (the 400) Compact groups (CGs) from the catalogue of McConnachie et al. 2009,

MNRAS 395. The CGs are defined according to the original Hickson criteria. 70% of these groups have a counterpart in the FoF catalogue.

SDSS SDSS SDSS

UKIDSS-LAS UKIDSS-LAS UKIDSS-LAS

LOW DENSITY-HIGH MASS

X-ray emission map from the RASS for the same FG. The LX is typical for an entire galaxy group/cluster.

Example of Fossil Group (FG) from La Barbera et al. 2009,

AJ 137. The optical light is dominated by a bright elliptical galaxy in the center, with a large magnitude gap (≥2mag) between first and second rank galaxies.

Some results for the entire sample

Internal color gradients

Mean internal color gradient, g-X, between g band and one of the other wavebands (from r through K).

From La Barbera & de Carvalho 2009, ApJL 699, 76

The trend implies both a negative metallicity gradient (higher metallicity towards the center), AND a small but significantly positive age gradient (younger stars towards the center)

Evolving back in time the age gradient, we find that this might explain (some of) the compactness of ETGs at high z.

Kormendy relations

MAG.LIM

g r i z

Y J H K

We find that the slope, β, of the KR depends significantly on the waveband, with larger slope values at longer wavelengths.

Is this variation important to analyze the evolution in size of ETGs at high z?

The Fundamental Plane (FP) from g through K

The FP and its waveband dependence

log Re = a log σ

0 + b <µ>

e + c optical wavebands

a~1.2 b~0.3

virial theorem+homology+M/L=const. a=2 b=0.4expected values

WHY THE WAVEBAND DEPENDENCE ?

Constraining the origin of the TILT: a change of stellar population vs. galaxy mass is expected to be wavelength dependent, while other effects (homology breaking, dark-matter content variations, ….) are not

Studying the FP at high redshifts, where observations are done in different wavebands

The FP and its waveband dependence

Jorgensen et al .96 1.08 ± 0.08 0.34 ± 0.02 41 U

Dressler et al 87 1.32 ± 0.05 0.33 ± 0.02 40 B

De Carvalho and Djorgovski 92 1.25 ± 0.07 0.32 ± 0.01 55 B

Busarello et al. 97 1.11 ± 0.20 0.36 ± 0.04 40 B

Graham 98 1.10 ± 0.14 0.22 ± 0.04 25 B

Prugniel and Simien 94 1.42 ± 0.05 0.35 ± 0.01 102 B

Saglia et al. 93 1.05 0.35 15 B

Graham and Colless 96 1.33 ± 0.10 0.32 ± 0.04 26 V

Guzman et al. 93 1.13 0.31 37 V

Djorgovski and Davies 87 1.39 ± 0.14 0.36 ± 0.036 260 r

Hudson et al. 97 1.38 ± 0.04 0.33 ± 0.01 325 R

Jorgensen et al. 96 1.24 ± 0.07 0.328 ± 0.008 226 r

Gibbons et al. 01 1.37 ± 0.05 0.336 ± 0.001 400 R

Colless et al. 01 1.22 ± 0.09 0.33 ± 0.009 255 R

D’Onofrio et al. 08 1.36 ± 0.02 0.325 ± 0.003 1579 V

Bernardi et al .03 1.45 ± 0.06 0.296 ± 0.004 9000 g

Bernardi et al 03 1.49 ± 0.06 0.300 ± 0.004 9000 r

Bernardi et al. 03 1.52 ± 0.05 0.312 ± 0.004 9000 i

Bernardi et al. 03 1.51 ± 0.05 0.308 ± 0.004 9000 z

Hyde & Bern. 09 1.40->1.47 0.305->0.329 50000 g->z

Scodeggio et al.98 1.51 ± 0.09 0.32 ± 0.01 29 H

Zibetti et al. 02 1.38 ± 0.1 0.35 ± 0.03 135 H

Pahre et al. 95 1.29 ± 0.08 0.284 ± 0.024 12 K

Pahre et al.98 1.53 ± 0.08 0.32 ± 0.01 251 K

Mobasher et al. 99 1.36 ± 0.26 0.30 ± 0.02 48 K

total 475 gals

OPTICAL FPs SDSS FP

NIR FPs

a~1.05-52

a~1.29-53

a~1.45-52a b a

a

b

b

La Barbera et al. 08: a=1.42±0.05 (r band); a=1.53±0.04 (K band) 1430 gals

The FP from g through K

Edge-on projection of the FP in the grizYJHK wavebands. The best-fits (orthogonal fit) are shown by the dashed lines.

Slopes of the FP from g through K for the entire sample (all environments). Ellipses denote 1σ confidence contours.

The FP from g through K

For the orthogonal fit, we find that the coefficient “b” is independent of the waveband, consistent with previous studies. The “a” changes by only 15% from g through K..

No variation is found for the log σ0 fit (MIST algorithm, La Barbera+2000).

We model the tilt as a variation of age, (log t), and metallicity, (log Z), between more and less massive ETGs per decade in mass, and a variation of M/L with M which is NOT due to stellar populations (but, for instance, to either a change of dark matter content with mass or to non-homology). We indicate the fraction of the tilt, in the NIR, which is not due to (log t), and (log Z), as f.

The NIR tilt of the FP is NOT due to stellar populations (f=0), with more massive galaxies being more metal rich and having the SAME age as less massive systems.

The FP from g through K

We fit this model to the grizYJHK values of the FP slopes (as in La Barbera+08).

The FP from g through K in different environments

The FP and the environment

de La Rosa et al. 2001 (AJ 122) found no difference between the FP of ETGs in CGs and those in other environments.

Bernardi et al. 2006 (AJ 131, 1288) found that the FP relation is very similar between ETGs in high and low density regions, but with a small significant offset in the zero-point. This offset is consistent with a pure age difference of 1Gyr.

D’Onofrio et al.2008 found that the FP coefficients are strongly correlated with the environment (cluster-centric distance and local density)

Zepf & Whitmore 1993 (ApJ 418) found that the FP of ETGs in CGs differs from that of field galaxies, with ETGs in CGs having lower σo.

Jorgensen et al. 1996 found that the FP coefficients are consistent for samples of galaxies residing in different nearby clusters (although with small sample sizes)

The r-band FP –LOCAL DENSITY

We bin the sample of ETGs in groups with respect to local density (Nbin=50; Ngal=362).

Variation of the FP slopes as a function of local density.

We find that the slope “a” is constant; while “b” increases with ΣN. However, we have not accounted for the fact that ETGs in different bins might have different distributions in the space of FP parameters

After correcting for that, we find that the variation of “b” tends to become less significant (from 4 to 2σ). These results seem to be in disagreement with those of D’Onofrio et al. 2008, but…...

Is the GLOBAL environment (halo mass)producing such discrepancy ?

The FP for Compact Groups

Edge-on projection of the FP for ETGs in the entire sample (black) and the CG catalogues.

The slopes for the three samples are very similar.

Zepf & Whitmore ‘93

Coefficient “c” and scatter of the FP. The variation is consistent with a pure age difference of 1Gyr.

First results of the XMM to confirm the nature of FGs from La Barbera et al. 2009 (P.I. M. Paolillo)

Although the sample is small, we can tentatively measure the scatter and offset of the FP for FGs.

The scatter is significantly smaller (at 4σ) than that of the entire sample, as expected for a population of isolated, passive galaxies.

CONTAMINATION ?

The FP for Fossil Groups

The grizYJHK FP – LOCAL DENSITY

Variation of the FP slopes from g through K in bins of local density. “a” smoothly increases with the waveband, while “b” is constant.

The amount of variation in “a” seems to decrease from high to low density.

“a” vs. the effective filter wavelength for the field and high density samples. The difference is significant at ~3σ.

The grizYJHK FP – LOCAL DENSITY

Field

Variation of the FP offset, “c”, from g through K, as a function of local density. “c” smoothly decreases towards higher ΣN for ALL wavebands.

Coefficient “c”, in the r-band, as a function of local density. The variation is consistent with a pure age difference of 1Gyr, consistent with Bernardi et al. 2006 (AJ 131). .

BC03, SSP, Z=Zsun, t=10Gyr

r-band

What is the role of metallicity?

Conclusions

The variation of the FP relation from g through K implies significant differences in the mass sequence of (bright) ETGs between low and high density environments.

Can we reconcile these results into a consistent picture for the mass assembling of ETGs in the different environments ?