AKARI Observation of the North Ecliptic Pole (NEP) Supercluster at z = 0.087: mid-infrared view of transition galaxies
Jongwan Ko, Myungshin Im, Hyung Mok Lee, Myung Gyoon Lee, Seong Jin Kim, Hyunjin Shim, Yiseul Jeon, Ho Seong Hwang, Christopher N. A. Willmer, Matthew A. Malkan, Casey Papovich, Benjamin J. Weiner, Hideo Matsuhara, Shinki Oyabu, Toshinobu Takagi
aa r X i v : . [ a s t r o - ph . C O ] N ov AKARI
Observation of the North Ecliptic Pole (NEP) Supercluster at z =0.087: mid-infrared view of transition galaxies
Jongwan Ko , , , , Myungshin Im , , Hyung Mok Lee , Myung Gyoon Lee , Seong Jin Kim ,Hyunjin Shim , Yiseul Jeon , , Ho Seong Hwang , Christopher N. A. Willmer , Matthew A.Malkan , Casey Papovich , Benjamin J. Weiner , Hideo Matsuhara , Shinki Oyabu , ToshinobuTakagi [email protected] ABSTRACT
We present the mid-infrared (MIR) properties of galaxies within a supercluster inthe North Ecliptic Pole region at z ∼ ) IR survey and the CLusters of galaxiesEVoLution studies (CLEVL) mission program. We show that near-IR (3 µ m)–mid-IR (11 µ m) color can be used as an indicator of the specific star formation rate andthe presence of intermediate age stellar populations. From the MIR observations, wefind that red-sequence galaxies consist not only of passively evolving red early-typegalaxies, but also of 1) “weak-SFG” (disk-dominated star-forming galaxies which havestar formation rates lower by ∼ × than blue-cloud galaxies), and 2) “intermediate-MXG” (bulge-dominated galaxies showing stronger MIR dust emission than normal redearly-type galaxies). Those two populations can be a set of transition galaxies fromblue, star-forming, late-type galaxies evolving into red, quiescent, early-type ones. Wefind that the weak-SFG are predominant at intermediate masses (10 M ⊙ < M ∗ < Yonsei University Observatory, Yonsei University, Seoul 120-749, Korea Korea Astronomy and Space Science Institute, Daejeon 305-348, Korea Astronomy Program, Department of Physics & Astronomy, FPRD, Seoul National University, Seoul 151-742,Korea Center of the Exploration of the Origin of the Universe (CEOU), Seoul National University, Seoul, Korea Spitzer Science Center, California Institute of Technology, MS 220-6, Pasadena, CA 91125 Service d’Astrophysique, CEA Saclay, F-91191 Gif-sur-Yvette, France Steward Observatory, University of Arizona, 933 N. Cherry Avenue, Tucson, AZ 85721, USA Department of Physics and Astronomy, University of California at Los Angeles, CA 90095, USA George P. and Cynthia W. Mitchell Institute for Fundamental Physics and Astronomy, Department of Physics,Texas A&M University, College Station, TX 77843, USA Institute of Space and Astronautical Science, Japan Aerospace Exploration Agency, Kanagawa 229-8510, Japan Graduate School of Science, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, Aichi 464-8602, Japan . M ⊙ ) and are typically found in local densities similar to the outskirts of galaxyclusters. As much as 40% of the supercluster member galaxies in this mass range can beclassified as weak-SFGs, but their proportion decreases to <
10% at larger masses ( M ∗ > . M ⊙ ) at any galaxy density. The fraction of the intermediate-MXG among red-sequence galaxies at 10 M ⊙ < M ∗ < M ⊙ also decreases as the density and massincrease. In particular, ∼
42% of the red-sequence galaxies with early-type morphologiesare classified as intermediate-MXG at intermediate densities. These results suggestthat the star formation activity is strongly dependent on the stellar mass, but that themorphological transformation is mainly controlled by the environment.
Subject headings: surveys: galaxies — galaxies: clusters and groups: — galaxies: evo-lution — galaxies: stellar content — infrared: galaxies
1. INTRODUCTION
One of the leading factors that can strongly influence galaxy evolution is the environment.Observationally, it has been known that the environment plays an important role in shaping galaxyproperties (see Blanton & Moustakas 2009 for a review). The morphology-density relation (MDR)was first described by Dressler (1980), who found a strong correlation between the morphologicaltype fraction and local galaxy surface density, where, for increasing local density, the fraction ofelliptical galaxies increases, while the spiral fraction decreases. Since then, a number of studieshave reported that galaxy properties such as colors, and star formation activity (SFA) are alsostrongly dependent on the local density (e.g., Park & Hwang 2009). MDR was also found at z ∼ × M ⊙ : lower-mass galaxies have young stellar populations and low concentrationindices typical of disk systems. Baldry et al. (2006) also found that the color-mass relations donot depend strongly on environment, while the fraction of red galaxies depends both on mass andenvironment. They also found that models with internally driven feedback mechanisms can explain 3 –the observed properties better. In a different study, Bundy et al. (2006) investigated the mass-dependent evolution of galaxies for 0.4 < z < ∼ M ∗ ∼ M ⊙ ) to the most massive galaxies ( M ∗ ∼ M ⊙ ). Inaddition to the MDR and the CDR mentioned earlier, the number of SF galaxies (blue galaxies)in clusters increases toward higher redshift (Butcher-Oemler effect; Butcher & Oemler 1984), andthe number of S0 galaxies seems to decline rapidly with redshift, with a corresponding increasein the blue, spiral galaxies (e.g., Dressler et al. 1997; Poggianti et al. 2001). At higher redshift,the star formation rates (SFRs) in high-density environments seem to be steadily increasing (Elbazet al. 2007). The lack of SF galaxies in cluster centers, and the possible transformation of bluegalaxies into red, S0 galaxies prompted the suggestion of many physical mechanisms to carry outthis transformation, whose cause is either gravitational (tidal interaction), or hydrodynamic (e.g.,ram-pressure stripping of interstellar gas; removal of halo reservoir halting the gas supply andquenching star formation; see Boselli & Gavazzi 2006 and Park & Hwang 2009).Many researchers have used optical colors or spectral features as a proxy of SFA, to studythe galaxy properties in cluster environments. Stellar masses of galaxies are usually obtained byfitting the spectral energy distributions (SED) measured by UV and optical photometry. Largedatasets of galaxies, such as the SDSS, have been used to understand the galaxy evolution incluster environments (e.g., Blanton & Moustakas 2009; Park & Hwang 2009). However, an obviousdisadvantage of this approach using the UV/optical light is dust obscuration. The interstellarmedium in galaxies has dust, which absorbs the UV light from stars (mostly coming from young,hot stars). This leads to significant extinction of the UV and optical light, complicating theinterpretation of the SFA. Fortunately, the absorbed UV/optical light is re-emitted in the infrared,where one can obtain an unobscured view of the SFA. Earlier studies of the SFA in clusters were 4 –made with ISO (e.g., Boselli et al. 1998; Biviano et al. 2004). More recently, Spitzer has revealedthe SFA in clusters in the IR, finding a few clusters with exceptionally high SFRs, and the rapidevolution of the SFA as a function of redshift (Bai et al. 2009). Nevertheless, the study of the SFAin clusters, as a function of environment, has been limited.In cluster environments, early-type galaxies follow a tight color-magnitude relation (red-sequence),indicating their stellar population is homogeneously old and passively evolving (e.g., Bower et al.1992; Kodama & Arimoto 1997). However, early-type galaxies do not contain homogeneous stel-lar populations when we examine them at different wavelengths, particularly in the IR. PreviousIR observations showed that there are some early-type galaxies with excess far-IR (FIR) emission(Knapp et al. 1989), and mid-IR (MIR) emission (Knapp et al. 1992; Xilouris et al. 2004).Recently, Clemens et al. (2009), using Spitzer-IRS peakup images (16 µ m), found that about32% of the early-type galaxies in the Coma cluster have excess flux over photospheric emission inthe MIR. Bressan et al. (2006) also detected MIR emission in early-type galaxies with Spitzer,showing a wide emission feature around 10 µ m and another broad feature near 18 µ m. Unusualpolycyclic aromatic hydrocarbons (PAHs) are also detected in the NIR/MIR spectra of nearbyearly-type galaxies (Kaneda et al. 2005, 2008; Lee et al. 2010; Vega et al. 2010; Panuzzo et al.2011). This suggests that they are associated with intermediate-age stellar populations, formed ina post-starburst phase.The IR emission from early-type galaxies is attributed either to the Rayleigh-Jeans tail of thestellar photosphere, or to circumstellar dust around evolved stars in the Asymptotic Giant Branch(AGB). Theoretical works show that the MIR-excess emission of AGB stars is well-correlated withstellar age (Piovan et al. 2003), and several studies suggest that MIR-excess is a good age indicator(Temi et al. 2005; Ko et al. 2009; Shim et al. 2011). Indeed, the MIR-excess emission can beuseful in tracing the past SFA, because other mean stellar age indicators cannot discriminate theemission of these stars.In summary, earlier studies find that IR data are useful and perhaps critical in some cases to(i) obtain an unobscured view of the SFA and (ii) trace recent SFA in early-type or red galaxies.However, the previous studies of IR properties of cluster galaxies have been mostly limited to thederivation of global properties of clusters, or examining a limited number of individual galaxies. Inaddition, there are few studies focusing on the galaxy environment on much larger scales (such asacross a supercluster). Therefore, it is necessary to study the effects of environment and/or masson the IR properties of galaxies in clusters and superclusters.Batuski & Burns (1985) first discovered the large-scale structure in the North Ecliptic Pole(NEP) region as an association of six clusters of galaxies, using a percolation analysis of clustersin the Abell’s (1958) catalog. Subsequently, Burg et al. (1992), using early observations with theROSAT satellite, reported five X-ray clusters and groups at 0.08 < z < ◦ of the NEP.With the deepest exposure of the ROSAT All-Sky Survey, Mullis et al. (2001) found an extendedlarge-scale structure in the NEP region at 0.07 < z < − × h − ergs s − (0.5 − − × h − ergs s − (0.5 − ) as part of the AKARI NEP-Widesurvey (ANWS, see Figs.1 and 3). The main advantage of AKARI against previous IR satellites isthe continuous wavelength coverage from the near- to mid-IR (2 to 24 µ m), especially at 11 and 15 µ m, which allows detecting the MIR-excess dust emission from circumstellar matter around AGBstars. Therefore, AKARI observations are well-suited for the study of MIR-excess from AGB stars,as well as obscured SFA. The ANWS covers intermediate- and low-density regions at 0.07 < z < < z < µ m flux traces not only the mean stellar age and the specificSFR (SSFR) of SF galaxies, but also the presence of intermediate age stellar populations, detectingeven tiny amounts of past star formation in early-type galaxies.Throughout this paper, we use H = 70 km s − Mpc − , Ω M = 0.3 and Ω Λ = 0.7. In thiscosmology, an angular scale of 1 arcsec at the distance of the NEP supercluster corresponds to1.629 kpc. All magnitudes are given in the AB system.
2. THE DATA2.1. IR & Optical Imaging
The ANWS was carried out using all available filters of the InfraRed Camera (IRC). For eachof the AKARI cameras, there are three associated channels : NIR (N2, N3, N4), MIR-S (S7,S9W, S11), and MIR-L (L15, L18W, and L24), all with a field-of-view covering 10 ′ × ′ . The 6 –numbers next to each letter represent the central wavelengths in µ m, and the W’s for 9 and 18 µ m represent the wider bandwidths. The ANWS was completed with 446 pointed observationscovering a large area of ∼ towards the NEP. Each pointing was done with the ‘IRC03’Astronomical Observation Template (AOT; see AKARI Observer’s Manual version 1.2 ), with2 dithered pointings per filter. For detailed descriptions of the survey strategy, its observationalproperties and the reduction of the AKARI IR images we refer the reader to Matsuhara et al.(2006) and Lee et al. (2009a).The optical survey covers the ANWS field centered at α = 18 h m s , δ = +66 ◦ ′ ′′ . Thecentral 2 deg were covered by the CFHT Megacam u ∗ , g ′ , r ′ , i ′ , and z ′ filters (Hwang et al. 2007).The remaining area, (which includes a small overlap with the CFHT Megacam data) is coveredwith the SNUCAM (Im et al. 2010) on the 1.5m telescope at Maidanak Observatory in Uzbekistanusing the Bessell B , R , and I filters (Jeon et al. 2010). We convert the CFHT Megacam (of ANWS)and SDSS (of A2255) photometry into the photometric system of Maidanak (i.e. Bessell B and R ) using best-fit spectral energy distribution (SED) model colors. In the following analysis thefinal uncertainties combine (in quadrature) the original photometric errors with the uncertaintiesderived from the SED fits which are typically < in the J and H bands, respectively. Figure 1 shows the coverage of eachsurvey, and the mean depth and the FWHM of each band are summarized in Table 1. We didnot use the AKARI L24 data in this study, due to its insufficient sensitivity. The photometry hasbeen corrected for foreground Galactic extinction using the Schlegel et al. (1998) dust maps andCardelli Milky Way extinction curve (Cardelli et al. 1989), assuming R V = 3.1.The object detection and photometry was done with SExtractor (Bertin & Arnouts 1996)on the coadded images of each individual band. We consider sources as real detections if theyhave more than five contiguous pixels above 3 × the rms fluctuations of the sky. However it wasnecessary to match the AKARI objects with optical counterparts (FWHM of 0.8 ′′ – 1.4 ′′ ) due tothe low resolution (FWHM of 5.5 ′′ – 6.6 ′′ ) of AKARI images. Thus, sources that SExtractor cannotseparate properly because of blending with neighbors in the AKARI IR images are excluded. Thephotometry was done using SExtractor in a single-band mode, and we used MAG-AUTO for thetotal magnitudes. To check the MAG-AUTO values we performed large-aperture photometry forseveral isolated galaxies in the final image. These showed that the difference between MAG-AUTOand MAG-APER were smaller than the typical measurement errors (NIR: ≤ ≤ R band or CFHT r ′ band and the 11 ′′ ( ∼ × FWHM) diameteraperture flux. Figure 2 shows the relation between the second-order moments for isolated galaxies R band image and the magnitude difference between MAG-AUTO and MAG-APER for N S
11. To determine the aperture correction factors, we used the best-fit (reddashed lines) for sample galaxies that are not contaminated by nearby sources.Fig. 1.— The imaging coverage of the ANWS. The center coordinates of the ANWS are at ( α =18 h m s , δ = +66 ◦ ′ ′′ ). North is up and east is to the left. The AKARI IR (2–24 µ m) data(5.4 deg ), CFHT Megacam u ∗ g ′ r ′ i ′ z ′ data (2 deg ), Maidanak BRI data (4.9 deg ), KPNO JH data (5 deg ) are represented by the red solid circle, blue dashed squares, the grey shaded region,and green squares, respectively. The spectroscopic follow-up of galaxies in the ANWS field used MMT/Hectospec and WIYN/Hydra.Based on the optical and IR fluxes, we selected objects with power-law SEDs as AGN candidates(N2–N4 > >
0, see Lee et al. 2007) and S11-detected objects with 15 µ m flux brighterthan 250 µ Jy as SF galaxy candidates. Supercluster member candidates (red-sequence galaxies)were selected using the NIR color-magnitude diagram (-0.7 < N3–N4 < -0.4), and a brightnessrequirement in the N2 band (N2 < R or r ′ band images to exclude stars. In summary,we selected as spectroscopic targets galaxies with a wide range of IR fluxes to study their IR prop-erties in a variety of local density environments. However, because of signal to noise limitations,identifying the correct redshifts of faint absorption-line galaxies is often very difficult, so that less 8 –Fig. 2.— Second-order moments (from SExtractor) in the Maidanak R band image versus mag-nitude difference between MAG-AUTO and MAG-APER with 11 ′′ diameter for N3 ( lef t ) and S11( right ). To determine aperture correction factors, we used the best-fit (red dashed lines) for galax-ies which are not contaminated by nearby objects. Black dashed and dotted lines show the meanvalue and the standard deviation of all sample galaxies in N3 ( lef t ) and S11 ( right ), respectively.massive galaxies with absorption line spectra are under-represented in the sample. We took intoaccount this incompleteness in our analysis.Fig. 3.— Follow-up spectroscopic observations of the NEP-Wide survey. The five cyan circlesindicate MMT/Hectospec fields, and the ten yellow circles indicate WIYN/Hydra fields. Therepresentation of the AKARI and optical surveys are the same as in Fig.1. 9 – Hectospec is a multiobject, moderate-dispersion spectrograph, covering a 1 ◦ diameter field ofview at the f/5 focus of the 6.5m MMT (Fabricant et al. 2005). It comprises 300 fibers of 1. ′′ ≤
22 in the region, but that were un-detected by AKARI. In addition to galaxies, this catalog also contains candidate F stars, selectedfrom the photometry, which are used to flux-calibrate the spectra. The assignment of objects tofibers is done using XFITFIBS , which takes into account the number of configurations (5 in thecase of the ANWS), the object priorities and number of sky positions. Typically 250 fibers persetup were assigned to NEP objects, 40 to random sky positions and up to 6 fibers placed on thecandidate F stars. The spectroscopic reduction used the HSRED package of
IDL scripts writtenby R. Cool , which is based on the IDL pipeline developed for the reduction of SDSS spectra.
HSRED does the standard reduction by correcting for bias, flatfields, illumination (if twilight flatswere taken), performing wavelength calibration (from HeNeAr lamps), sky-subtraction and extract-ing one-dimensional spectra. The flux-calibration is done by combining the 1-dimensional F-starspectra with the multiband photometry (to obtain the spectro-photometric zero-point) and Kuruczstellar models (to rectify the spectra). Redshifts for the wavelength- and flux-calibrated spectraare obtained from the cross-correlation with a series of galaxy, QSO and stellar template spectra.All Hectospec redshifts were individually validated and assigned a quality code ranging from 1 to4, as used for the DEEP2 survey (e.g., Newman et al. in preparation; Willmer et al. 2006). Onlyqualities of 4 or 3 are used in the analyses, meaning that the probability of the redshift beingcorrect is greater than 95% and 90% respectively.
We also obtained optical spectra with the Hydra multiobject spectrograph on the WIYN 3.5mtelescope at Kitt Peak National Observatory. We used 98 red fibers of 2 ′′ diameter feeding thebench spectrograph with a 316 lines mm − grating, yielding a dispersion 2.64 of ˚A pixel − . Thewavelength range is 4500–9000 ˚A, but the spectrum quality is low beyond 8000 ˚A due to the strong
10 –Table 1. The imaging dataset in the ANWS
CFHT Maidanak KPNO AKARICovering area 2 deg Band u ∗ g’ r’ i’ z’ B R I J H N2 N3 N4 S7 S9W S11 L15 L18WDepth [AB mag.] 26.0 26.1 25.6 24.7 23.7 23.2 22.0 21.2 20.5 19.3 20.9 21.1 21.1 19.5 19.3 19.0 18.6 18.8FWHM [arcsec] 1.13 1.05 0.93 0.84 0.79 1.4 1.2 1.1 1.4 1.3 5.5 6.0 6.0 5.9 6.6 5.9 6.2 6.2(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18)Note. − Col. (1)-(5): CFHT Megacam u ∗ , g’, r’, i’, and z’ from Hwang et al. (2007). Col. (6)-(8): Maidanak B, R, and I from Jeon et al. (2010).Col. (9)-(10): KPNO Flamingo J and H from Jeon et al. (2011, in preparation). Col. (11)-(18): AKARI N2, N3, N4, S7, S9W, S11, L15, andL18W from Lee et al. (2009). Table 2. MMT Hectospec fields
Field R.A. (J2000) DEC. (J2000) t exp (minutes) Observation Date (UTC)nep-hecto-1 17 54 50.66 +66 38 47.48 100 2008 May 03nep-hecto-2 18 04 29.63 +65 53 32.20 80 2008 June 02nep-hecto-3 17 55 09.27 +65 49 28.81 80 2008 September 03nep-hecto-4 17 59 29.51 +67 16 03.83 80 2008 November 17nep-hecto-5 18 07 39.24 +66 38 56.78 80 2008 November 20
11 –sky emission lines. The field of view of Hydra is approximately 1 ◦ and we observed ten fields overthe ANWS field (Fig. 3).Target assignment in each configuration was done using the WHYDRA software. For eachconfiguration, 10–15 fibers were assigned to blank sky positions, and 3–6 fibers were assigned tospectrophotometry standard stars. Excluding broken fibers, fibers assigned to blank sky positions,and fibers assigned to standard stars, we obtained spectra of 60–70 targets in each configuration.In Figure 3, the locations of the Hydra configurations are shown by ten yellow circles, and details ofthe observations are summarized in Table 3. Depending on the observing conditions, the exposuretime for each field varied from 3 ×
20 minutes to 5 ×
20 minutes.We used IRAF to reduce the spectra. First, we performed the pre-processing which includesthe corrections for overscan, bias, dark and flat, and trimming the image. A flatfield image wascreated by averaging dome flats taken before and after the observations. We removed cosmic raysusing L.A.Cosmic (van Dokkum 2001). After the pre-processing, we extracted one-dimensionalspectra using the Hydra reduction package DOHYDRA (F. Valdes 1995) . We extracted one-dimensional spectra from all apertures and did wavelength calibration with a Cu–Ar comparisonlamp. The master sky spectrum produced by coadding sky spectra of blank skies, was subtractedwith DOHYDRA. Finally, the extracted one-dimensional spectral images were combined usingIRAF task scombine to improve the final signal-to-noise ratio. Redshifts were determined by identifying high signal-to-noise emission lines and/or multipleabsorption lines. Three individuals (J. Ko, M. Im, and H. Shim) independently determined redshiftsfor all objects. Then, each of them flagged objects according to their spectral features. We flagged Guide to the HYDRA Reduction Task DOHYDRA, available at http://iraf.net/irafdocs/dohydra.pdf
Table 3. WIYN Hydra fields
Field R.A. (J2000) DEC. (J2000) t exp (minutes) Observation Date (UTC)NEP00 18 00 10.031 +66 34 00.00 80 2008 June 27NEP01 17 56 29.855 +66 21 00.00 80 2008 June 27NEP02 17 54 09.796 +65 54 00.00 100 2008 June 27NEP03 17 51 51.806 +66 16 00.00 80 2008 June 28NEP04 17 52 58.195 +66 58 00.00 80 2008 June 28NEP05 17 54 51.441 +67 10 00.00 60 2008 June 29NEP06 17 59 28.904 +67 16 00.00 60 2008 June 30NEP07 18 05 22.322 +67 10 00.00 60 2008 June 30NEP08 18 09 25.763 +66 59 00.00 60 2008 June 30NEP09 18 08 18.026 +66 20 00.00 60 2008 June 30
12 –the objects with at least two distinct spectral features as those with a secure redshift. All individualsgenerally agree on the secure redshifts, but faint or distant galaxies can be ambiguous due to theweak line features. These objects were not used in the analyses.To verify our redshift determination, we ran
XCSAO in the
RV SAO package, which computesradial velocities by cross-correlating spectra against templates of known redshift (Kurtz et al.1992). In this test, we used 61 supercluster member galaxies flagged as having secure redshiftswith Hectospec galaxy templates, and found that the difference of radial velocities for all samplesis 119 ±
123 km s − , consistent with no difference in radial velocities measured from both methodsand telescopes.We were able to successfully determine secure redshifts for 1026 and 400 of 1195 Hectospecobjects and 600 Hydra objects, respectively. Figure 4 shows a sample of WIYN spectra with linesused to determine redshifts, and the redshift distribution of all objects with secure redshifts in theANWS is shown in Figure 5. To this sample we added 241 galaxies for which redshifts are availablein the NASA Extragalactic Database (NED). An additional 18 redshifts come from long slit spectraobtained by Matthew Malkan using the Kast spectrograph at the Lick Observatory 3-m telescope.Fig. 4.— Sample of WIYN/Hydra spectra with sky (grey), absorption (red), and emission (blue)lines used for redshift determination.In Figure 6, we plot the spectroscopic completeness as a function of the observed N B − R ) color for extended sources. Here we use N B − R ) color because we willadopt the absolute N B − R ) color to separate redsequence galaxies from blue cloud galaxies. The vertical line represents our N3 magnitude cut (N3 <
19) used in this study. 13 –Fig. 5.— Redshift distribution of galaxies with secure redshifts in the ANWS. The black line showsthe redshift distribution of the total spectroscopic sample.Fig. 6.— Spectroscopic completeness as a function of the N3 apparent magnitude (top panel) andobserved ( B − R ) color and N3 apparent magnitude (bottom panel) of galaxies in the ANWS.Vertical dashed lines indicate the N3 magnitude cut used in this study. 14 – The sample of NEP supercluster members used in this paper consists of the spectroscopicsample of galaxies at 0.07 < z < < z < J and H ), and GALEX UVdata where available. The matching radius was chosen to be 5.0 arcseconds. All matched objectswere then visually validated by examining postage stamp images for all bands.The A2255 spectroscopic and photometric data are drawn from Shim et al. (2011). We used313 objects covered by the GALEX UV data. The redshift identification is nearly complete forgalaxies with r < < z < µ Jy in the N3, N4, S7, S11, L15, and L24 bands,respectively (Shim et al. 2011). The depth of the A2255 data is comparable to that of the ANWSdata within the measurement errors.Fig. 7.— The redshift distribution of galaxies in A2255 ( lef t ) and the ANWS ( right ). The verticaldotted lines indicate the redshift range of the NEP supercluster (0.07 < z <
3. PHYSICAL PROPERTIES
In this section, we describe the method used to derive the galaxy stellar masses, SFR, andlocal galaxy density.
We derive the stellar mass, the mean stellar age, the reddening parameter, and the SFR foreach object through SED fitting, which uses a standard χ minimization procedure with varioustemplates. We used two different SED libraries – the Bruzual & Charlot (2003; BC03) spectralsynthesis models to estimate the stellar mass and age, and the IR templates of SF galaxies fromChary & Elbaz (2001; CE01) to determine the IR luminosity. When performing the fits for SFgalaxies, we only use the IR data in the fits, as the empirical IR templates do not include thediverse range of ages, metallicities and star formation histories that are modeled by the BC03library. Figure 8 shows examples of the observed UV-optical-MIR SEDs of galaxies in the ANWS,and the best-fit SEDs from each model are overplotted. 16 –Fig. 8.— Examples of the SEDs of galaxies in the ANWS. Filled circles indicate observed datapoints from GALEX UV to AKARI MIR. All objects are detected in S11, except in cases whereobjects were not detected in the AKARI L15 and L18 bands (Id: 1881, 1266, 1882, 1216, 65, 1264,and 1332), in which cases no points are shown. Overplotted lines represent the best-fit SEDs: theblue solid lines indicate SEDs from Bruzual & Charlot (2003) and the red solid lines indicate infraredgalaxy templates of Chary & Elbaz (2001). Each SED fit lists the object id on top-left, stellar age(years), stellar mass ( M ⊙ ), and total L IR ( L ⊙ ) in unit of dex, from left to right, respectively.Each row shows a set of four objects that are representative of our classification scheme (describedin Table 4): weak-MXG, intermediate-MXG, weak-SFG, dusty-SFG, and blue-SFG, from top tobottom. 17 –Stellar masses were estimated using stellar population synthesis models (e.g., Bell et al. 2003;Ilbert et al. 2010). During the SED-fitting process, the redshift was kept fixed, and all bandsfrom the NUV to NIR (NUV – N4) were used to fit a purely stellar SED. The SED templateswere generated with BC03 models assuming a Chabrier (2003) initial mass function (IMF) and anexponentially declining star formation history SFR ∝ e − t/τ ( τ between 0.1 Gyr to 30 Gyr). TheSEDs were generated for a grid of 44 ages (0.1 Gyr to 13.5 Gyr) and three different metallicities(0.02, 0.008, and 0.004 Z ⊙ ), and dust extinction was added using the formula of Calzetti et al.(2000) for E B − V between 0 to 0.5. Figure 9 shows the stellar masses computed with BC03 versusthe N N N = -19 corresponds to a stellar mass of ∼ log M ∗ = 9.1 M ⊙ , the stellar mass range of our sample is roughly log M ∗ = 9.1 – 11.5 M ⊙ .The histogram of the N M ∗ /M ⊙ = –0.524 × M N – 0.810), and the two dashed linesrepresent the standard deviation of residuals to this fit ( σ = 0.255).To derive the L IR for each galaxy, we use the IR templates of CE01. Templates are shifted toeach galaxy’s redshift and then matched to all the available bands longer than 7 µ m (i.e. the AKARIS7, S9W, S11, L15, and L18W bands). We first subtract the stellar contribution (i.e. the best-fitBC03 template to the UV–NIR data) from the observed IR data, and then fit them with CE01templates to estimate the total IR luminosity at 8-1000 µ m ( L IR ). We convert L IR into SFR using 18 –Fig. 10.— Upper : Histogram of the N3 absolute magnitude of galaxies in A2255 ( lef t ) and inthe ANWS ( right ). The top axis shows stellar masses calculated from the linear relation in Fig.9. Vertical lines are our N3 magnitude cut.
Lower : Distribution of derived total IR luminosity L IR (8–1000) for S11 detected galaxies in A2255 ( lef t ) and the ANWS ( right ). The Top axiscorresponds to SFR, using the Kennicutt (1998) relation.Kennicutt (1998) relation: SFR ( M ⊙ yr − ) = 1.72 × − L IR /L ⊙ . The lower panel of Figure 10shows the distribution of derived total IR luminosity L IR for S11 detected galaxies. We only findtwo Luminous IR Galaxies (LIRG; log L IR /L ⊙ >
11) candidates in the ANWS field. Conservatively,we estimate a SFR limit of 0.1 M ⊙ yr − in A2255 and 0.2 M ⊙ yr − in the ANWS. There is a likelycontribution of AGN to the IR luminosity. However, the contamination in our sample is very lowif we assume many AGNs have power-law IR SEDs (e.g., Lee et al. 2007). Among the superclustermember galaxies in the ANWS, there is only one X-ray point source (detected by ROSAT), whichis also a LIRG. Thus, AGN contamination is expected to be negligible in our results. 19 – Although our dataset does not cover the entire NEP supercluster, it covers a wide range ofenvironments, including a rich galaxy cluster (A2255), three galaxy groups, and low density regionsin the supercluster outskirts.To characterize the galaxy properties as a function of the environment, we adopt a local surfacedensity estimator Σ, which is the surface number density of galaxies within a projected distance of0.5 Mpc and within a relative velocity of 1000 km s − for each galaxy in the sample. The velocitycut is adopted to exclude foreground and background galaxies. Our method of measuring galaxyenvironment (i.e. the fixed aperture environment measure) is found to be the best estimate of galaxyenvironmental density when the virial radius of the host halo is difficult to measure according toHaas et al. (2011), so we chose this as our density parameter for discussion. As a cautionarymeasure, we compared several other environment indicators against this measure, using galaxiesin A2255. Two other indicators are tested, one is the surface number density of galaxies within aprojected distance of 1.0 Mpc and within a relative velocity of 1000 km s − (Σ . Mpc ), and anotheris the surface number density of galaxies within to the 5th-nearest neighbor and within a relativevelocity of 1000 km s − (Σ th − nearest ). In the right panel of Figure 11, we show the correlationbetween the local density adopted in this work (Σ . Mpc ) and other environmental parameters.The Spearman rank correlation coefficient is also printed, indicating other choice of environmentalparameters correlate well with our choice. Indeed, we performed the analysis on our main resultsusing these different environment parameters, and find that the results are not affected by a choiceof the environmental parameter. To account for the spectroscopic incompleteness of the NEP-Widesurvey, the number density is computed by weighing each galaxy by the inverse of completenesscorresponding to its N B − R ) color in Figure 6. In the left panel ofFigure 11 shows the distribution of the local surface density for galaxies in A2255 (dotted line) andin the ANWS (dashed line). The peak density in the ANWS area corresponds to the lowest densityregion of A2255, i.e. the infall region of A2255. We refer to this as an intermediate local density inthe following analysis. The spatial distribution of all supercluster member galaxies in A2255 andin the ANWS field is overplotted on their number density map in Figure 12. 20 –Fig. 11.— Left : Distribution of the local density Σ for galaxies in A2255 (dotted line) and in theANWS (dashed line).
Right : Compasiron between the local density Σ . Mpc adopted in this work,which is the number density within a projected distance of 0.5 Mpc, and Σ . Mpc within a projecteddistance of 1.0 Mpc ( upper ) and Σ th − nearest within a projected distance to 5th-nearest neighbor( lower ). The neighbors were identified among all members (black-solid lines) and massive (log M ∗ /M ⊙ >
10) ones (red-dotted lines) within ∆ v < − . The Spearman rank correlationcoefficient r s between local densities are printed in the bottom-right of each panel.Fig. 12.— The spatial distribution of galaxies in A2255 ( lef t ) and the ANWS ( right ), on thesmoothed galaxy number density map. Red or blue circles represent galaxies with redshifts aboveor below z=0.087, respectively. The symbol size is proportional to the redshift deviation from theNEP supercluster mean redshift (z=0.087). 21 –
4. GALAXY CLASSIFICATIONS: MIR View of Galaxies4.1. Motivation
The optical color-magnitude relation (CMR) is commonly used to separate red-sequence (here-after red) galaxies from blue-cloud (hereafter blue) galaxies. Blue optical colors can mean thatgalaxies are actively star-forming. While the presence of an AGN can make a galaxy blue (e.g.,Obric et al. 2006; Choi et al. 2009), SF is in general the cause of blue galaxy colors. The red colorsof galaxies can be due to a dominant population of passively evolving old stars (even though somegalaxies with predominantly old stars have low-level SF, e.g., Trager et al. 2000). Galaxies canalso be red in the optical/NUV if their light is extinguished by interstellar dust. Thus we need todifferentiate such causes of red galaxy colors.We show the CM diagram of A2255 galaxies (the left side of Fig. 13) compared to galaxies inthe ANWS (the right side of Fig. 13) in three colors: B − R (panels a, d), N U V − R (panels b,e), and N − S
11 (panels c, f). The absolute N M N absolute magnitude,we apply only the luminosity distance of each galaxy (i.e. no correction for peculiar motions orbandpass shift are included). In the following analysis, we only consider galaxies brighter than M N = -19 (corresponding N ≈
19) due to the relatively shallow S
11 detection limit. Specifically, thiscut allows us to construct an unbiased sample of galaxies with N − S > B − R optical CMR for galaxies brighter than M N = − B − R versus M N ,shown as a dashed line, by rejecting outliers iteratively based on the bi-weight calculation. Thestandard deviation of residuals to the fit is 0.07 mag ( σ ), implying a tight optical red-sequence.The horizontal solid line indicates the color cut we adopted to separate red galaxies (redward of thesolid line) from blue galaxies (blueward of the solid line). The CMR was moved to a bluer color by∆( B − R ) = 0 . ∼ σ ) to define the color cut. The same color criterion is applied to the galaxiesin the ANWS.Fig. 13(b,e) shows the N U V − R CM diagram for the same sample as in panels (a,d). The NUVCM diagram shows a scatter an order-of-magnitude larger than that in the optical data, indicatingthat a number of red galaxies have been forming stars (e.g., Yi et al. 2005). Since the NUV fluxis much more sensitive to young stellar populations than the optical flux, it can be a good tracerof recent star-formation (within ∼ N − S
11 colors for red galaxies in theMIR CM diagram (Fig. 13(c,f)), suggests that these galaxies present a variety of MIR properties.We focus in particular on the MIR. The choice of the S
11 band has several advantages overother MIR bands. First, the S
11 flux correlates with the Polycyclic Aromatic Hydrocarbons (PAHs)emission features at 11.3 and 12.7 µ m, which may be related to current star formation. Becauseof the PAH features, MIR emission (especially around 11 µ m) correlates well with the total IR 22 –luminosity (Spinoglio et al. 1995), which can be converted into SFR (Chary & Elbaz 2001). Second,the MIR emission may also contain a contribution from the envelopes of evolved stars, showingbroad silicate emission features around 10 µ m (e.g., Bressan et al. 2006) or/and unusual PAHfestures (e.g., Vega et al. 2010). In particular, this MIR emission from dust surrounding AGBstars is also sensitive to stellar ages, because it declines with time (e.g., Piovan et al. 2003; Temiet al. 2005). Thus the MIR emission may be tracing not only the current SF, but also past SFA.For the model SEDs in Figure 13(c), we used Single Stellar Population (SSP) models including thedust emission from circumstellar dust around AGB stars (Piovan et al. 2003; hereafter P03). Ashas been recognized, the optical CMR can be well-described with a single age model assuming ametallicity gradient (the dashed line in Fig. 13(a); e.g., Kodama et al. 1997). The same model fitsthe N − S
11 versus N N − S
11 colors (see Ko et al. 2009), whichsuggests either the existence of younger stellar populations or of some other mechanism.We divide the red galaxies into several subsamples with different SFA by using the N − S M ∗ ;SSFR) of SF galaxies, and also of the luminosity-weighted mean stellar age in passively evolvingobjects. 23 –Fig. 13.— Top : the B − R vs. absolute N M N = −
19 in A2255 ( lef t ) and in the ANWS ( right ). The CMR is shown by the dashedline and the horizontal solid line indicates the color cut adopted to classify red galaxies (redwardof the solid line). The CMR in the ANWS is normalized to the red sequence of A2255. The colordeviation from the CMR is defined as ∆( B − R ) = 0 . ∼ σ ), where σ is the standard deviationof residuals to CMR fit. The cross in the upper left corner indicates the typical errors. Middle : the
N U V − R vs. absolute N lef t ) and in the ANWS ( right ). The GALEX
Nearby Galaxy Survey
N U V detection limits are shown for undetected galaxies (arrows).
Bottom : the N − S
11 vs. absolute N lef t ) and in the ANWS( right ). The dotted lines indicate the CMR calculated from the P03 AGB model SSPs, assuming ametallicity sequence at three different stellar ages (1, 5 , and 12 Gyr), respectively. The horizontalsolid line represents the P03 model SSPs without AGB dust. The AKARI S
11 detection limitsare given for undetected samples in S
11 (arrows). The vertical line indicates the magnitude cut( N ≈
19 corresponding to M N = -19 at the distance of the NEP supercluster). 24 – Using the SWIRE (
Spitzer
Wide-area InfraRed Extragalactic survey) templates of Polletaet al. (2007), we find that early-type galaxies (e.g., ellipticals, S0, and Sa) have N − S <
0, while SF, late-type galaxies have N − S > µ m. These are generated with the GRASIL code (Silva et al. 1998) includingdusty envelopes of AGB stars following the prescription by Bressan et al. (1998). In particular,in the MIR spectral region between 5 and 12 µ m, the spiral and starburst templates encompassthe variety of observed MIR spectra (Polleta et al. 2007). Also shown as a comparison are SSPtemplates of P03 incorporating AGB dust with a metallicity (Z=0.02) and three stellar ages (2,5, and 13 Gyr). The P03 model SEDs show somewhat redder N − S
11 colors than the GRASILmodel due to difference in how dust emission from evolved stars is predicted, but they are broadlyin agreement that N − S
11 color as the mean stellar age decreases. For comparison, we also plotSSPs from different libraries (BC03, CB07, and Ma05) with a Salpeter (1955) IMF and a fixedmetallicity (Z=0.02), but without the inclusion of the MIR dust emission prescription. CB07 is anew version of BC03 including the new stellar evolution prescription of Marigo & Girardi (2007)for the thermally-pulsing (TP) AGB evolution, and Ma05 is also including the TP-AGB phaseof stellar evolution (Maraston 2005) differently from previous models. However, these two SSPs(CB07 and Ma05) do not include the circumstelalr dust emissions. These three models withoutAGB dust cannot produce the MIR-excess colors of N − S > − N − S
11) to divide galaxies into MIR-red galaxies ( N − S >
0) and MIR-blue galaxies( N − S < N − S > − N − S
11 color versus SSFR derived from the SED fits. This reveals thatthe N − S
11 color correlates well with SSFR, indicating that the N − S
11 color probes differentlevels of SFA. Specifically, we find that our MIR color cut ( N − S
11 = 0) is comparable to log(SSFR) ∼ -10.7, i.e. galaxies with log (SSFR) < -10.7 (which corresponds to a SFR of 0.2 M ⊙ yr − at a stellar mass of 10 M ⊙ ) can be considered as passively evolving galaxies. This SSFR cut wasadopted by Gallazzi et al. (2009) to separate SF galaxies from quiescent ones. However, in thecase of MIR-blue galaxies, the derived SSFR may not have any physical significance as a measureof the current SFA, but may indicate a wide dispersion of mean stellar ages among those passivegalaxies if the P03 model SEDs are adopted.Now we further divide optically red galaxies into four classes depending on their N − S
11 25 –Fig. 14.— The rest-frame N − S
11 vs.
N U V − R ( lef t ) and N − S
11 vs. B − R ( right )color-color distribution of galaxies detected in all four bands, compared with SWIRE templates ofPolletta et al. (2007) including 3 ellipticals (2, 5, 13 Gyr), 7 spirals (S0, Sa, Sb, Sc, Sd, Sdm, Spi4),and 6 starbursts (M82, Arp220, N6090, N6240, I20551, I22491). Also plotted are SSP templatesfrom different libraries (BC03, CB07, Ma05, and Piovan03) with a Salpeter (1995) IMF and a solarmetallicity (Z=0.02) for the same ages as SWIRE ellipticals. Templates with N − S > N − S <
0. The arrow shows the mean value of areddening vector of E(B-V) = 0.1 using the Calzetti et al. (2001) extinction law, and the cross inthe bottom right corner indicates the typical errors.colors. We made the morphological classification of galaxies in the ANWS field sample using theCFHT r ′ band images of 0. ′′
187 pixel − : early types are bulge-dominated with good symmetry (0and 1 in Figs 16-20), while late types are disk-dominated with asymmetric internal structure (2, 3,and 4 in Fig. 16-20). The dominant morphological type and its fraction are shown in Table 4.First, we classify optically red galaxies with N − S < − N − S
11 colorvaries depending on the model used. They may have a small amount of MIR-excess at 11 µ m, butthat can be well understood within the framework of passively evolving galaxies with AGB dust.Figure 16 shows the CFHT r ′ -band images of these galaxies. As the figure shows, their morphologiesare predominately early-type. Therefore, these are passively evolving early-type galaxies.The second class comprises optically red galaxies with − < N − S < ≤ > -10.7, but are very near the cut in Fig. 15). If left 26 –Fig. 15.— SSFRs derived from the SED fits and N − S
11 colors of galaxies detected in the N S
11 bands. The filled red and blue circles are optically red and blue galaxies classified by theoptical CMR (Fig. 13). The open circle indicates massive galaxies (log M ∗ /M ⊙ > − MIR color ( N − S
11 = 0) and SSFR (log(SSFR) = − N − S > < -10 ( ∼ > -10 ( ∼ ′ -band images of weak-SFG and dusty-SFG, respectively,which suggest that they have disk-like morphology. Interestingly, more than half of dusty-SFG areedge-on disks, indicating that half of those are viewed at higher disk inclination. Thus, we canspeculate that the dusty-SFG are strong SF late-type galaxies, but optically reddened. On theother hand, weak-SFG have relatively lower SSFR than the blue-SFG, thus we can expect that 27 –their current/recent star formation is insufficient to change their optical color. In other words, inthe weak-SFG, the dominant stellar populations are generally old (no hot young stars, i.e. theaverage stellar age is greater than 1 Gyr) although their MIR-red colors ( N − S >
0) indicatethat they have recently formed some stars ( < Galaxy type Optical color IR color log(SSFR [yr − ]) Morphology fraction Comments(1) (2) (3) (4) (5)weak-MXG red N3-S11 < -1 -11.2 early type ( > < N3-S11 < > > ∼ -10.0) late type ( > > ∼ -9.0) late type ( > > ∼ -8.9) late type ( > − Col. (1): Classified galaxy type. Col. (2): First, the optical color-magnitude relation (CMR) is used to separate red from blue galaxies. Col. (3): Second, the red galaxies are subdivided by NIR − MIR ( N − S
11) color. Col. (4): The meanvalues of specific star formation rate (SSFR) in units of dex. For SF populations, the range of SSFR is enclosed in brackets.Col. (5): The dominant morphology from visual classification, and the fractions of early types: 0, 1 and late types: 2, 3, 4 (seethe caption in Fig. 16).
Fig. 16.— Optical (CFHT r ′ band) images (60 ′′ × ′′ ) of the most massive galaxies with opticallyred and N − S < -1 (weak-MXG). Each image lists the object id, morphology, absolute N3 mag-nitude, and N − S
11 color, on the top-left, top-right, bottom-right, and bottom-left, respectively.The morphology is classified by eye: 0 = bulge-dominated, 1 = bulge-dominated with disk feature,2 = disk-dominated with symmetric spiral, 3 = disk-dominated with asymmetric spiral, 4 = others(merging system or unclassified). 29 –Fig. 17.— Same as Fig. 16, but for the galaxies with optically red colors and 0 < N − S < -1(intermediate-MXG).Fig. 18.— Same as Fig. 16, but for the galaxies with optically red colors and N − S > < -10; weak-SFG).Fig. 19.— Same as Fig. 16, but for the galaxies with optically red colors and N − S > > -10; dusty-SFG). 30 –Fig. 20.— Same as Fig. 16, but for the galaxies with optically blue colors and N − S > The weak-SFG in our sample show low levels of SFA (on average ∼ M ∗ /M ⊙ = [9.5, 11]). However, their optical red colors aremostly due to their underlying old stellar populations, and not an effect of dust reddening. In otherwords, they have older stellar populations and smaller SFRs (insufficient to change the optical color)compared to blue-SFG. Furthermore, they have mostly disk-dominated morphologies. Therefore,in terms of SFA in galaxy evolution, our weak-SFG are possible candidates for the transitionpopulation in which star formation is suppressed, compared to blue-SFG, and their red colors aredue to old stellar populations. These populations are very similar to ‘anemic spirals’ discoveredby van den Bergh (1976) which are thought to be in the transition from blue, SF field spirals tored, non-SF cluster S0 galaxies. They are also generally at a similar evolutionary stage to thetransition populations (red spirals or red SFs) selected using different criteria in other works usingSDSS galaxies (e.g., Bamford et al. 2009; Masters et al. 2010) and galaxies in the Abell 901/902supercluster at redshift ∼ The intermediate-MXG in our sample are defined as bluer in N − S
11 colors (indicating lowerSSFRs) than weak-SFG, but redder (indicating younger ages) than weak-MXG. This implies thattheir MIR emission arises from very weak SFA (on average ∼ ∼ B − V ( ∼ − µ m region in nearby early-type galaxies arises from intermediate-age carbon stars, andthey can be formed by rejuvenation episodes within the last few Gyr at the 1% level of the massof the galaxy. 32 –Our intermediate-MXG seem to be closely related to the UV-excess galaxies in Yi et al. (2005).They found that roughly 15% of nearby early-type galaxies have excess NUV emission indicatingrecent ( < −
2% of the total stellar mass. Around 1 Gyr aftera single-burst of star formation, massive stars (O, B, and A stars) have expired, and recent starformation indicators (e.g., NUV flux and H β line index) are no longer good tracers of the starformation history. However, the MIR-excess emission over stellar light can trace star formationover a much longer period, since low to intermediate mass (1 − ⊙ ) stars evolve to the AGBphase, and their circumstellar dust emission is strong in the MIR. Frogel et al. (1990) found that thecontribution of AGB stars to the bolometric luminosity peaks at more than 40% at ages from 1.1 to3.3 Gyr, but rapidly falls to less than 5% at 10 Gyr. In other words, our intermediate-MXG seem tobe a population of descendants of objects showing recent SFA within ∼
5. ENVIRONMENTAL AND MASS DEPENDENCE OF GALAXIES INTRANSITION PHASE
We now investigate how the properties of galaxies are influenced by their stellar mass and localenvironment. We focus on the intermediate-MXG and the weak-SFG as objects that might be inthe evolutionary transition from blue, star-forming late-types to red, passive early-types.If the properties of galaxies are affected by environmental mechanisms, such as hydrodynamicor gravitational processes (e.g., Boselli & Gavazzi 2006 and Park & Hwang 2009 for a review),we expect to see an increase of transition populations in specific environments. For example,when blue spiral galaxies infall from low- to high-density regions, (although it depends on therelevant time-scale) both global halo properties and/or the local galaxy density can produce agradual/sharp decline of star formation and morphological transformation by external factors, suchas tidal interactions and gas stripping. As a result, SF galaxies in dense regions show lower SFR,and the fraction of early-type galaxies increases with increasing density or toward the center of thegalaxy cluster.On the other hand, if the effects of stellar mass on the galaxy properties are stronger thanthose of environment, then we would not expect any significant changes in the distribution oftransition galaxies over a range of local density at fixed stellar masses. In this case, the stellar massis the primary parameter governing the suppression of star formation and the transformation ofmorphology, and thus transition populations would only show a trend with their stellar mass.In an attempt to analyze the dependence of the properties on the stellar mass and the envi-ronment independently, we plot the change in the relative fraction of different types of galaxies asa function of the stellar mass at low, intermediate, and high density regions (Fig. 21), and as a 33 –function of the local density at different mass bins (Fig. 22). Here, Σ is used for the local density,as described in Section 3.2. The type fraction ( f t ) is defined as the ratio between the number N t of each galaxy-type to the N total total number galaxies in fixed mass and density bins. We include S N total . The uncertainties of the fractions are estimated bycalculating the variance in the likelihood of the fraction (e.g., De Propris et al. 2004). Assumingthat the fraction has the form of the likelihood function L ∝ f N t t (1 − f t ) N total and its maximum is N t N total , the variance of the fraction is σ ( f t ) = ( d lnLdf t ) − when the likelihood function has a Gaussian form. Note that our estimate of the standard de-viation might be low due to the distribution of our small samples (i.e. the likelihood function maynot be Gaussian). In order to avoid the lower detection rate of galaxies with N − S < N S M ∗ /M ⊙ = [10, 11]. We include S −
23 and table 5, our results can be summarizedas follows. • We find that the weak-SFGs are mostly dominant at mass bin log M ∗ /M ⊙ = [10, 10.5] andat density bin log Σ = [-0.5, 0.5] (Fig. 21(d) and Fig. 22(e)). In this mass range (Fig. 22(e)),more than 20% of the population in all density bins are the weak-SFG, and, in particular,the weak-SFG are the dominant population in the density bin of log Σ = [-0.5, 0.5] ( ∼ M ∗ /M ⊙ >
10) onthe number density maps of all member galaxies. The weak-SFG tend to avoid the centralregion of A2255 (the two weak-SFG close to the center have redshifts of 0.07379 and 0.07387,placing them at the edge of the velocity distribution of A2255, so that their location is likely aprojection effect). However, when galaxies in both more-massive (log M ∗ /M ⊙ = [10.5, 11.0];Fig. 22(f)) and less-massive (log M ∗ /M ⊙ = [9.5, 10.0]; Fig 22(d)) bins are considered, theweak-SFG are not as significant, contributing with less than 20% of the population in anydensity bins, and their environmental trend is weaker. Note that, although the weak- andintermediate-MXG are not detected in the lowest mass bin because of the S
11 detection limit,our result matches well the trend of red spirals of Wolf et al. (2009; see their Fig. 15). • At log M ∗ /M ⊙ >
10, the overall fraction of SF galaxies (blue-, dusty-, and weak-SFG)decreases gradually by ∼ ∼ ∼
14% in low-, intermediate-, and high-densitybins respectively (see Fig. 23(c)). The overall fraction of red galaxies increases increasinglocal density (see Fig. 23(d)). This is basically a confirmation of the SFR–density relation 34 –Fig. 21.—
Upper : SSFR as a function of stellar mass in three density bins for our N S
11 (see Fig. 13), and the solidlines in each panel show the detection limit of 0.2 M ⊙ yr − for the ANWS sample. Lower : Thefraction of each galaxy-type as a function of stellar mass. Type fractions are defined as the ratiobetween the number galaxies in each type class and the total number of galaxies at a fixed mass.The errors are calculated assuming that the fractions are the maximum values of the likelihoodfunction (see text for details). The total number galaxies includes objects that are undetected in S
11. Mass bins with less than three galaxies are excluded.(e.g., Lewis et al. 2002; Kauffmann et al. 2004; Weinmann et al. 2006; Hwang et al. 2010)and CDR (e.g., Pimbblet et al. 2002; Blanton et al. 2005; Cucciati et al. 2006). This trendindicates that massive galaxies are mostly red in all environments, and comprise most of thegalaxies in the highest density regions. However, if optically red galaxies are divided into fourdifferent sub-populations, then the interpretation needs to be done carefully. If we focus onthe weak-MXG (Fig. 23(c)), there is a strong environmental dependence, there being a verysmall fraction of weak-MXGs in the lowest density bin. However, the fraction of red galaxiespresents a high value ( ∼ Upper : SSFR as a function of galaxy local density in three stellar mass bins for our N Lower : The fraction of each galaxy type as a function of local density.proportion of SF galaxies (weak- and dusty-SFG) that are red. • The relative fraction of weak-SFGs among SF galaxies (weak-, dusty-, and blue-SFG), as afunction of local density at fixed mass, allows us to study the importance of mass/environmentin SF quenching. In Figure 25(a), at all local densities, the relative weak-SFG fraction ishigher at large stellar mass (log M ∗ /M ⊙ >
10) than at low stellar mass (log M ∗ /M ⊙ < ∼ twice as high at the highest density bin, compared withthe lowest density bin when only massive galaxies are considered. However, for less-massivegalaxies, the fraction does not change significantly with the local density. This suggests thatthe weak-SFG are likely to be more massive than strong SF galaxies (dusty- and blue-SFG)and comprise a significant fraction of all massive SF galaxies in high-density environments. • We find roughly 10% intermediate-MXG among all massive (log M ∗ /M ⊙ >
10; Fig. 23(c))galaxies in all density bins. However, the relative fraction of these versus the weak-MXGdecreases as the density increases, by ∼ ∼ ∼ M ∗ /M ⊙ = 36 –Fig. 23.— Upper : SSFR as a function of galaxy local density at stellar mass bin log M ∗ /M ⊙ =[10, 11] for our N B − R vs. M N CMR (see Fig.13 for details).
Lower : The fraction of each galaxy type as a function of local density. Undetectedsamples in S
11 are counted as the weak-MXG ( lef t ) and red ( right ) due to their low SSFR.[10, 10.5], [10.5, 11]), the relative fraction of the intermediate-MXG is higher in the lower-mass bin. It is thus possible to infer that the intermediate-MXG are less-massive and tendto be located in the outer parts of the cluster. This is largely consistent with our studyof the galaxies in A2218 (Ko et al. 2009). Although the size of our mass-limited sample ofintermediate-MXG is too small to determine the mass-dependence for this population, we cansee the environmental-dependence for massive galaxies. In Figure 24, the intermediate-MXGare likely to be located in the outskirts of the cluster and near group centers (corresponding 37 –Fig. 24.— Spatial distribution of each galaxy type in Fig. 23, on the smoothed galaxy numberdensity maps for all supercluster member galaxies in A2255 ( lef t ) and in the ANWS ( right ). Inthe right panel, the three “+” signs indicate the center of X-ray detected groups with the meanredshifts given (Henry et al. 2006). North is up, and east is to the left.to the lowest density bin in Fig. 25(b)). Therefore, an environmental action is necessarilyrequired to explain the properties of this population. It is expected that the intermediate-MXG evolve into weak-MXG. Furthermore, SSFRs (approximately MIR-weighted mean stel-lar ages) of the intermediate-MXG are much smaller than those of weak-SFG, indicating thatthe weak-SFG are at an earlier evolutionary stage than the intermediate-MXG. Also, themain difference between both transition populations is the morphology (disk-dominated forweak-SFG and bulge-dominated for intermediate-MXG).Table 5. The fraction of each galaxy-type in Figure 23(c). log M ∗ /M ⊙ = [10, 11]Galaxy type low-density (%) intermediate-density (%) high-density (%)(-0.5 < log Σ < < log Σ < < log Σ <
38 –Fig. 25.—
Left : The fraction of weak-SFG among SF galaxies (weak-, dusty-, and blue-SFG) isplotted as a function of local density in three stellar mass bins: log M ∗ /M ⊙ = [9.5, 10], [10, 10.5],[10.5, 11]. Right : The relative fraction of the intermediate-MXG versus the weak-MXG is plottedas a function of local density in two stellar mass bins: log M ∗ /M ⊙ = [10, 10.5], [10.5, 11]. Becauseof the cut in S
11, less-massive (log M ∗ /M ⊙ <
10) MIR-blue ( N − S <
0) galaxies are notdetected.
6. DISCUSSION
Recent studies of red galaxies defined by optical CMR cuts indicate that they contain severalpopulations at different evolutionary stages (e.g., Lee et al. 2008; Ko et al. 2009; Cortese & Hughes2009; Wolf et al. 2009; Gallazzi et al. 2009; Tran et al. 2009; Bamford et al. 2009; Bundy et al.2010; Masters et al. 2010; Salim & Rich 2010). For example, red, early-type galaxies are found tohave a wide range of MIR-excess of non-stellar origin, suggesting that some of these experiencedrecent star formation episodes. In addition, some red-sequence galaxies are found to have UV excesssuggesting weak SFA.We expect that there are two different phases (star formation quenching and morphologychanging), when a blue, star-forming, late-type galaxy turns into a red, quiescent, early-type galaxy.Recent studies revealed that star formation quenching (optical color change) is not always accompa-nied by morphological change (e.g., Blanton et al. 2005, S´anchez et al. 2007, Bamford et al. 2009;Wolf et al. 2009). In other words, the time scale of transition from blue to red and of morphologicalchange from late-type to early-type is different, and seems to be a function of stellar mass and localgalaxy density. Observationally, the existence of red spirals (i.e. weak-SFG) and blue early-typegalaxies, and their preference for specific masses and local densities supports this idea.To trace mass- and environment-dependence of changes in color and morphology, we focusedon two different categories of transition galaxies (intermediate-MXG and weak-SFG). Specifically,their NIR-MIR color ( N − S
11) is a good tracer of SSFR, and the SSFR of SF galaxies is notsensitive to their mass and environment (e.g., Peng et al. 2010). 39 –
Our result that the weak-SFG are mainly dominant at intermediate mass (log M ∗ /M ⊙ = [10,10.5]) in the cluster outskirts agrees with previous studies: both Wolf et al. (2009) and Masterset al. (2010) show that red spirals are predominant at intermediate local density (infall regions ofclusters), and at the higher mass end ( > M ⊙ ). Quantitatively, we find that the fraction ofweak-SFG among SF galaxies (blue-, dusty-, and weak-SFG) range from 71% at the cluster core to36% in the outskirts of the cluster, at log M ∗ /M ⊙ = [10, 11]. This suggests that the suppressionof star formation progresses rapidly in high-density environments, so that the weak-SFG shiftedinto quiescence earlier than those in the lower density environments, at a fixed mass ( > M ⊙ ).This suggests the acceleration of “downsizing” in overdense regions (e.g., Bundy et al. 2006). Ifwe assume the “downsizing” scenario, then galaxies with the same mass and SSFR should havesimilar star formation histories. However, the much higher weak-SFG fraction among SF galaxiesand much lower weak-SFG fraction among red galaxies in high-density environments compared tolow-density environments suggests that the star formation of their progenitor galaxies (blue-SFG)could be suppressed efficiently because of the high galaxy number density and/or much longerinteraction with the cluster environment. Thus massive weak-SFG in high-density regions could bealready replaced by intermediate-MXG and/or weak-MXG.In contrast, at the highest masses (log M ∗ /M ⊙ = [10.5, 11]), the fraction of weak-SFG is > M ∗ /M ⊙ < <
10% of all galaxies).This can be interpreted as a mass-dependent star formation history where massive galaxies aremuch older, became passive earlier than less massive galaxies, and have been undergoing a muchhigher frequency of mergers, so that their morphologies have already transformed to early types.At lower masses (log M ∗ /M ⊙ = [9.5, 10]), even considering our S
11 detection limit, the fractionof weak-SFG among SF galaxies appears to decrease sharply at all density bins, compared to largermasses (see the left panel of Fig. 25). This is consistent with the results of Wolf et al. (2009) wherethey suggest that the star formation of low-mass galaxies in clusters is suppressed quickly andthe morphological change happens simultaneously, hence red spirals are very rare (see also Boselliet al. 2008). Therefore, we confirm the previous results that low-mass blue-SFG infalling fromoutskirts of clusters have experienced halo and disk gas stripping via some environmental effects,and almost simultaneously spiral structures have disrupted into early-type morphology, in contrastto higher-mass galaxies whose spiral structures persist much longer. It is supposed that low-massgalaxies with depleted gas disks are susceptible to morphological change through minor mergerevents (Hopkins et al. 2009). Furthermore, Masters et al. (2010) also found that the fraction ofred spirals with smaller masses ( < M ⊙ ) is very low, in contrast to the larger masses.From the behavior of the weak-SFG, star formation quenching is affected both by the stellarmass and environment. However, the environmental dependence works differently in each mass bin. 40 – Another proposed transition population is the intermediate-MXG, which has been alreadyclassified in our analysis of A2218 and A2255 (Ko et al. 2009; Shim et al. 2011). These galaxies areoptically red and have early-type morphologies, but show broad emission in the MIR (e.g., Bressan2006). These suggest that the MIR-weighted mean stellar ages of these galaxies are younger thanthose of the weak-MXG. They also show a wide range of MIR-excess emission, suggesting a varietyof star formation histories among red, early-type galaxies.In the right panel of Figure 25, although we can only explore massive intermediate-MXG due toour S
11 detection limit, we find that these galaxies are relatively low-mass systems among massive( > M ⊙ ) red, early-type galaxies, and are likely to be located in the outer parts of the cluster.This is consistent with the results in A2218.At larger masses ( > M ⊙ ), about 51% of the galaxies show quenched or decreased starformation (i.e. weak- and intermediate-MXG, and weak-SFG) and around 24% have early-typemorphology (i.e. weak-and intermediate-MXG) in the outskirts of clusters; when the cluster core isconsidered, these proportions increase to 96% and 86% respectively. Furthermore, with increasinglocal density, the relative fraction of these transition galaxies versus the weak-MXG decreasessharply, while versus the weak-SFG it increases. These trends can be explained if the morphologicaltransformation starts at the outskirts and the process is mostly completed at high density, and theintermediate-MXG are a set of products of morphological transition between the weak-SFG and theweak-MXG at all cluster environments. From the behavior of the intermediate-MXG, morphologicaltransformation of galaxies can be explained by the environmental effects.The behavior of our transition populations (weak-SFG and intermediate-MXG) suggests apossible scenario of evolutionary history of galaxies, from star-forming, late-type galaxies to non-star-forming, early-type galaxies. On this issue, based on the findings above, we can speculatethat as the gas supply decreases and the SFRs continue to decline, blue-SFG may naturally changeto weak-SFG with gas-poor disks, mainly governed by their mass (massive galaxies evolve fasterthan less-massive ones in optical color change), and that this transition is accelerated in the high-density environment. Then a large fraction of massive weak-SFG proceed slowly through severalenvironmental processes, particularly starting at the outskirts of clusters, while for less-massivegalaxies this happens faster. Finally they transform into the weak-MXG through the intermediate-MXG. S´anchez et al. (2007) suggested a two-step scenario in which star formation is quenchedfirst, and morphological transformation follows on longer timescale, from the analysis of A2218.S´anchez-Bl´azquez et al. (2009) also suggested that the timescale of morphological transformationof the galaxies entering the red-sequence is different from that of star formation quenching. 41 –
7. SUMMARY AND CONCLUSIONS
We have investigated the MIR properties of optical red-sequence galaxies within a superclusterin the NEP region at redshift ∼ ) and AKARI IR (2 − µ m)observations of A2255, in conjunction with NUV-optical SEDs and optical spectroscopy. AKARI11 µ m flux traces not only the amount of recent SFA, but also the presence of intermediate agestellar populations (i.e. past SFA). Therefore, the NIR − MIR ( N − S
11) color can be a goodindicator of SSFRs, whereby we can identify dusty SF galaxies (dusty-SFG) and transition galaxiesamong red-sequence galaxies.We find that ∼
22% of the massive ( > M ⊙ ) red-sequence galaxies are dusty-SFG in low-density regions. Their SSFRs are comparable to those of blue-SFG and their environmental trendis similar to SF populations (blue- and weak-SFG). Thus their optical red colors are due to dustextinction or/and higher disk inclination.We also find that red-sequence galaxies, excepting dusty-SFG, consist not only of passivelyevolving galaxies, but also of weak-SFG (disk-dominated SF galaxies which have SSFR lower thanblue-SFG), and intermediate-MXG (bulge-dominated galaxies showing broad non-stellar MIR emis-sion compared to weak-MXG). These two populations may represent transition galaxies from blue,star-forming, late-type galaxies evolving into red, quiescent, early-type ones. In this study, we havefocused on properties of these transition galaxies, and on how the fraction of transition galaxiesdepends on the stellar mass and the local density. Our main conclusions are summarized as follows. • The weak-SFG are found to be similar to the red spirals of the Galaxy Zoo (Masters et al.2010) and optically passive spirals (Wolf et al. 2009) in the A901/2 cluster. Consistent withprevious studies, at the same mass range (log M ∗ /M ⊙ = [10, 11]), they show a lower level ofSFA (on average ∼ • The intermediate-MXG show little recent ( < • In the evolution of a galaxy, the weak-SFG could be candidates for the transition stagebetween blue-SFG and intermediate-MXG, where the star formation is quenched, while theintermediate-MXG are likely to be placed in an intermediate stage between blue/weak-SFGand weak-MXG, where the morphology is transformed into early types. • The transition population is the most abundant at intermediate local densities (outskirtsof clusters), suggesting that most of the action takes place at intermediate densities. Therelative fraction of weak-SFG versus intermediate-MXG increases as local density decreases.This indicates that the star formation quenching is ongoing at the outskirts, and the process 42 –is mostly completed at high density. The morphologies of the intermediate-MXG are mostlyearly-type while the weak-SFG are late-type, meaning that the quenching of star formationoccurs earlier than the morphological transformation. • The fraction of the weak-SFG to SF galaxies shows a different environmental dependence fordifferent stellar masses. For low-mass galaxies, there are no strong environmental effects. Thisindicates that SF quenching occurs rapidly in low-mass galaxies, and thus they have alreadyevolved into intermediate-MXG. However, our shallow detection limit at S
11 does not allowconfirming the intermediate-MXG at low mass ( < M ⊙ ), except that intermediate-MXGare relatively low-mass systems among massive ( > M ⊙ ) galaxies, and likely to be locatedin the outer parts of the cluster.This work is based on observations with AKARI , a JAXA project with the participation ofESA. This work was supported by the Korea Science and Engineering Foundation (KOSEF) grantNo. 2009-0063616, funded by the Korea government (MEST). JK was supported by ‘KASI − YonseiJoint Research for the Frontiers of Astronomy and Space Science’ program (2011) funded by KoreaAstronomy and Space Science Institute. We thank L. Piovan for providing his SED model, andStephane Charlot and Gustavo Bruzual for kindly sending us the new version of their CB07 model.HSH acknowledges the support of the Centre National d’Etudes Spatiales (CNES). This work ispartly based on observations obtained with the MMT, a joint facility operated by the SmithsonianAstrophysical Observatory and the University of Arizona, and with the telescopes at the Kitt PeakNational Observatory. This research has made use of the NASA/IPAC Extragalactic Database(NED) which is operated by the Jet Propulsion Laboratory, California Institute of Technology,under contract with the National Aeronautics and Space Administration.
REFERENCES
Abell, G. O. 1958, ApJS, 3, 211Adami, C., Ilbert, O., Pell´o, R., et al. 2008, A&A, 491, 681Bai, L., Rieke, G. H., Rieke, M. J., Christlein, D., & Zabludoff, A. I. 2009, ApJ, 693, 1840Baldry, I. K., Balogh, M. L., Bower, R. G., et al. 2006, MNRAS, 373, 469Balogh, M., Eke, V., Miller, C., et al. 2004, MNRAS, 348, 1355Bamford, S. P., Nichol, R. C.; Baldry, I. K., et al. 2009, MNRAS, 393, 1324Batuski, D. J., & Burns, J. O. 1985, AJ, 90, 1413Bekki, K., Couch, W. J., & Shioya, Y. 2002, ApJ, 577, 651 43 –Bell, E. F., McIntosh, D. H., Katz, N., & Weinberg, M. D. 2003, ApJS, 149, 289Bertin, E., & Arnouts, S. 1996, A&AS, 117, 393Biviano, A., Metcalfe, L., McBreen, B., et al. 2004, A&A, 425, 33Blanton, M. R., Eisenstein, D., Hogg, D. W., Schlegel, D. J., & Brinkmann, J. 2005, ApJ, 629, 143Blanton, M. R., & Moustakas, J. 2009, A&A Rev., 47, 159Boselli, A., Lequeux, J., Sauvage, M., et al. 1998, A&A, 335, 53Boselli, A., & Gavazzi, G. 2006, PASP, 118, 517Boselli, A., Boissier, S., Cortese, L., & Gavazzi, G. 2008, ApJ, 674, 742Bower, R. G., Lucey, J. R., & Ellis, R. S. 1992, MNRAS, 254, 601Bressan, A., Granato, G. L., & Silva, L. 1998, A&A, 332, 135Bressan, A., Panuzzo, P., Buson, L., et al. 2006, ApJ, 639, L55Bruzual, G., & Charlot, S. 2003, MNRAS, 344, 1000Bundy, K., Ellis, R. S., Conselice, C. J., et al. 2006, ApJ, 651, 120Bundy, K., Scarlata, C., Carollo, C. M., et al. 2010, ApJ, 719, 1969Burg, R., Giacconi, R., Huchra, J., et al. 1992, A&A, 259, L9Butcher, H., & Oemler, A. Jr. 1984, ApJ, 285, 426Calzetti, D., Armus, L.; Bohlin, R. C., et al. 2000, ApJ, 533, 682Cardelli, J. A., Clayton, G. C., & Mathis, J. S. 1989, ApJ, 345, 245Chabrier, G. 2003, PASP, 115, 763Chary, R., & Elbaz, D. 2001, ApJ, 556, 562Clemens, M. S., Bressan, A., Panuzzo, P., et al. 2009, MNRAS, 392, 982Choi, Y.-Y., Woo, J.-H., & Park, C. 2009, ApJ, 699, 1679Choi, Y., Goto, T., & Yoon, S.-J. 2009, MNRAS, 395, 637Cortese, L., & Hughes, T. M. 2009, MNRAS, 400, 1225Cowie, L. L., Songaila, A., Hu, E. M., & Cohen, J. G. 1996, AJ, 112, 839Cucciati, O., Iovino, A., Marinoni, C., et al. 2006, A&A, 458,39 44 –De Propris, R., Colless, M., Peacock, J. A., et al. 2004, MNRAS, 351, 125Dressler, A. 1980, ApJ, 236, 351Dressler, A., Oemler, A. Jr., Couch, W. J., et al. 1997, ApJ, 490, 577Elbaz, D., Daddi, E., Le Borgne, D., et al. 2007, A&A, 468, 33Fabricant, D., Fata, R., Roll, J., et al. 2005, PASP, 117, 1411Frogel, J. A., Mould, J., & Blanco, V. M. 1990, ApJ, 352, 96Gallazzi, A., Bell, E. F., Wolf, C., et al. 2009, ApJ, 690, 1883Haas, M. R., Schaye, J., & Jeeson-Daniel, A. 2011, MNRAS, accepted, preprint (astro-ph/1103.0547)Henry, J. P., Mullis, C. R., Voges, W., et al. 2006, ApJS, 162, 304Hopkins, P. F., Somerville, R. S., Cox, T. J., et al. 2009, MNRAS, 397, 802Hwang, H. S., & Park, C. 2009, ApJ, 700, 791Hwang, H. S., Elbaz, D., Lee, J. C., et al. 2010, A&A, 522, 33Hwang, N., Lee, M. G., Lee, H. M., et al. 2007, ApJS, 172, 583Ilbert, O., Capak, P., Salvato, M., et al. 2009, ApJ, 690, 1236Ilbert, O., Salvato, M., Le Floc’h, E., et al. 2010, ApJ, 709, 644Im, M., Lee, H. M., Lee, M. G., et al. 2008, ASPC, 399, 382Im, M., Ko, J., Cho, Y., et al. 2010, JKAS, 43, 75Jeon, Y., Im, M., Ibrahimov, M., et al. 2010, ApJS, 190, 166Kauffmann, G., Heckman, T. M., White, S. D. M., et al. 2003, MNRAS, 341, 54Kauffmann, G., White, S. D. M., Heckman, T. M., et al. 2004, MNRAS, 353, 713Kaneda, H., Onaka, T., & Sakon, I. 2005, ApJ, 632, L83Kaneda, H., Onaka, T., Sakon, I., et al. 2008, ApJ, 684, 270Kennicutt, R. C. 1998, ARA&A, 36, 189Knapp, G. R., Guhathakurta, P., Kim, D.-W., & Jura, M. A. 1989, ApJS, 70, 329Knapp, G. R., Gunn, J. E., & Wynn-Williams, C. G. 1992, ApJ, 399, 76 45 –Ko, J., Im, M., Lee, H. M., et al. 2009, ApJ, 695, L198Kodama, T., & Arimoto, N. 1997, A&A, 320, 41Koopmann, R. A., & Kenney, J. D. P. 2004, ApJ, 613, 866Kurtz, M. J., Mink, D. J., Wyatt, W. F., et al. 1992, ASPC, 25, 432Lee, H. M., Im, M., Wada, T., et al. 2007, PASJ, 59S, 529Lee, H. M., Kim, S. J., Im, M., et al. 2009a, PASJ, 61, 375Lee, M. G., Lee, H. M., Im, M., et al. 2009b, ASPC, 418, 241Lee, J. H., Lee, M. G., Park, C., & Choi, Y.-Y. 2008, MNRAS, 389, 1791Lee, J. H., Hwang, H. S., Lee, M. G., Lee, J. C., & Matsuhara, H. 2010, ApJ, 719, 1946Lewis, I., Balogh, M., De Propris, R., et al. 2002, MNRAS, 334, 673Li, I. H., Glazebrook, K., Gilbank, D., et al. 2011, MNRAS, 411, 1869Lin, C. C., & Shu, F. H. 1964, ApJ, 140, 646Lintott, C. J., Schawinski, K., Slosar, A., et al. 2008, MNRAS, 389, 1179Maraston, C. 2005, MNRAS, 362, 799Marigo, P., & Girardi, L. 2007, A&A, 469, 239Masters, K. L., Mosleh, M., Romer, A. K., et al. 2010, MNRAS, 405, 783Mullis, C. R., Henry, J. P., Gioia, I. M., et al. 2001, ApJ, 553, L115Murakami, H., Baba, H., Barthel, P., et al. 2007, PASJ, 59, S369Obri´c, M., Ivezi´c, ˇZ.; Best, P. N., et al. 2006, MNRAS, 370, 1677Panuzzo, P., Rampazzo, R., Bressan, A., et al. 2011, A&A, 528, 10Park, C., & Hwang, H. S. 2009, ApJ, 699, 1595Peng, Y., Lilly, S. J., Kovaˇc, K., et al. 2010, ApJ, 721, 193Pimbblet, K. A., Smail, I., Kodama, T., et al. 2002, MNRAS, 331, 333Piovan, L., Tantalo, R., & Chiosi, C. 2003, A&A, 408, 559Poggianti, B. M, Bridges, T. J., Carter, D., et al. 2001, ApJ, 563, 118Polletta, M., Tajer, M., Maraschi, L., et al. 2007, ApJ, 663, 81 46 –Postman, M., Franx, M., Cross, N. J. G., et al. 2005, ApJ, 623, 721Salim, S., & Rich, R. M. 2010, ApJ, 714, L290Salpeter, E. E. 1955, ApJ, 121, 161S´anchez-Bl´azquez, P., Jablonka, P., Noll, S., et al. 2009, A&A, 499, 47S´anchez, S. F., Cardiel, N., Verheijen, M. A. W., Pedraz, S., & Covone, G. 2007, MNRAS, 376, 125Schlegel, D. J., Finkbeiner, D. P., & Davis, M. 1998, ApJ, 500, 525Shim, H., Im, M., Lee, H. M., et al. 2011, ApJ, 727, 14Silva, L., Granato, G. L., Bressan, A., & Danese, L. 1998, ApJ, 509, 103Sobral, D., Best, P. N., Smail, I., et al. 2011, MNRAS, 411, 675Spinoglio, L., Malkan, M. A., Rush, B., Carrasco, L., & Recillas-Cruz, E. 1995, ApJ, 453, 616Temi, P., Mathews, W. G., & Brighenti, F. 2005, ApJ, 622, 235Tran, K.-V. H., Saintonge, A., Moustakas, J., et al. 2009, ApJ, 705, 809Trager, S. C., Faber, S. M., Worthey, G., & Gonz´alez, J. J. 2000, AJ, 120, 165van Dokkum, P. G. 2001, PASP, 113, 1420van den Bergh, S. 1976, ApJ, 206, 883Vega, O., Bressan, A., Panuzzo, P., et al. 2010, ApJ, 721, 1090Weinmann, S. M., van den Bosch, F. C., Yang, X., & Mo, H. J. 2006, MNRAS, 366, 2Willmer, C. N. A., Faber, S. M., Koo, D. C., et al. 2006, ApJ, 647, 853Wolf, C., Arag´on-Salamanca, A., Balogh, M., et al. 2009, MNRAS, 393, 1302Xilouris, E. M., Madden, S. C., Galliano, F., Vigroux, L., & Sauvage, M. 2004, A&A, 416, 41Yi, S. K., Yoon, S.-J., Kaviraj, S., et al. 2005, ApJ, 619, L111York, D. G., Adelman, J., Anderson, J. E. Jr., et al. 2000, AJ, 120, 1579